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 for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3694 (unchanged sound) to 10.0 (maximum effect).
3695 To inverse filtering use negative value.
3698 Enable clipping. By default is enabled.
3701 @subsection Commands
3703 This filter supports the all above options as @ref{commands}.
3706 Apply a DC shift to the audio.
3708 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3709 in the recording chain) from the audio. The effect of a DC offset is reduced
3710 headroom and hence volume. The @ref{astats} filter can be used to determine if
3711 a signal has a DC offset.
3715 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3719 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3720 used to prevent clipping.
3725 Apply de-essing to the audio samples.
3729 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3733 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3737 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3741 Set the output mode.
3743 It accepts the following values:
3746 Pass input unchanged.
3749 Pass ess filtered out.
3754 Default value is @var{o}.
3760 Measure audio dynamic range.
3762 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3763 is found in transition material. And anything less that 8 have very poor dynamics
3764 and is very compressed.
3766 The filter accepts the following options:
3770 Set window length in seconds used to split audio into segments of equal length.
3771 Default is 3 seconds.
3775 Dynamic Audio Normalizer.
3777 This filter applies a certain amount of gain to the input audio in order
3778 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3779 contrast to more "simple" normalization algorithms, the Dynamic Audio
3780 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3781 This allows for applying extra gain to the "quiet" sections of the audio
3782 while avoiding distortions or clipping the "loud" sections. In other words:
3783 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3784 sections, in the sense that the volume of each section is brought to the
3785 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3786 this goal *without* applying "dynamic range compressing". It will retain 100%
3787 of the dynamic range *within* each section of the audio file.
3791 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3792 Default is 500 milliseconds.
3793 The Dynamic Audio Normalizer processes the input audio in small chunks,
3794 referred to as frames. This is required, because a peak magnitude has no
3795 meaning for just a single sample value. Instead, we need to determine the
3796 peak magnitude for a contiguous sequence of sample values. While a "standard"
3797 normalizer would simply use the peak magnitude of the complete file, the
3798 Dynamic Audio Normalizer determines the peak magnitude individually for each
3799 frame. The length of a frame is specified in milliseconds. By default, the
3800 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3801 been found to give good results with most files.
3802 Note that the exact frame length, in number of samples, will be determined
3803 automatically, based on the sampling rate of the individual input audio file.
3806 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3807 number. Default is 31.
3808 Probably the most important parameter of the Dynamic Audio Normalizer is the
3809 @code{window size} of the Gaussian smoothing filter. The filter's window size
3810 is specified in frames, centered around the current frame. For the sake of
3811 simplicity, this must be an odd number. Consequently, the default value of 31
3812 takes into account the current frame, as well as the 15 preceding frames and
3813 the 15 subsequent frames. Using a larger window results in a stronger
3814 smoothing effect and thus in less gain variation, i.e. slower gain
3815 adaptation. Conversely, using a smaller window results in a weaker smoothing
3816 effect and thus in more gain variation, i.e. faster gain adaptation.
3817 In other words, the more you increase this value, the more the Dynamic Audio
3818 Normalizer will behave like a "traditional" normalization filter. On the
3819 contrary, the more you decrease this value, the more the Dynamic Audio
3820 Normalizer will behave like a dynamic range compressor.
3823 Set the target peak value. This specifies the highest permissible magnitude
3824 level for the normalized audio input. This filter will try to approach the
3825 target peak magnitude as closely as possible, but at the same time it also
3826 makes sure that the normalized signal will never exceed the peak magnitude.
3827 A frame's maximum local gain factor is imposed directly by the target peak
3828 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3829 It is not recommended to go above this value.
3832 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3833 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3834 factor for each input frame, i.e. the maximum gain factor that does not
3835 result in clipping or distortion. The maximum gain factor is determined by
3836 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3837 additionally bounds the frame's maximum gain factor by a predetermined
3838 (global) maximum gain factor. This is done in order to avoid excessive gain
3839 factors in "silent" or almost silent frames. By default, the maximum gain
3840 factor is 10.0, For most inputs the default value should be sufficient and
3841 it usually is not recommended to increase this value. Though, for input
3842 with an extremely low overall volume level, it may be necessary to allow even
3843 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3844 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3845 Instead, a "sigmoid" threshold function will be applied. This way, the
3846 gain factors will smoothly approach the threshold value, but never exceed that
3850 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3851 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3852 This means that the maximum local gain factor for each frame is defined
3853 (only) by the frame's highest magnitude sample. This way, the samples can
3854 be amplified as much as possible without exceeding the maximum signal
3855 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3856 Normalizer can also take into account the frame's root mean square,
3857 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3858 determine the power of a time-varying signal. It is therefore considered
3859 that the RMS is a better approximation of the "perceived loudness" than
3860 just looking at the signal's peak magnitude. Consequently, by adjusting all
3861 frames to a constant RMS value, a uniform "perceived loudness" can be
3862 established. If a target RMS value has been specified, a frame's local gain
3863 factor is defined as the factor that would result in exactly that RMS value.
3864 Note, however, that the maximum local gain factor is still restricted by the
3865 frame's highest magnitude sample, in order to prevent clipping.
3868 Enable channels coupling. By default is enabled.
3869 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3870 amount. This means the same gain factor will be applied to all channels, i.e.
3871 the maximum possible gain factor is determined by the "loudest" channel.
3872 However, in some recordings, it may happen that the volume of the different
3873 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3874 In this case, this option can be used to disable the channel coupling. This way,
3875 the gain factor will be determined independently for each channel, depending
3876 only on the individual channel's highest magnitude sample. This allows for
3877 harmonizing the volume of the different channels.
3880 Enable DC bias correction. By default is disabled.
3881 An audio signal (in the time domain) is a sequence of sample values.
3882 In the Dynamic Audio Normalizer these sample values are represented in the
3883 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3884 audio signal, or "waveform", should be centered around the zero point.
3885 That means if we calculate the mean value of all samples in a file, or in a
3886 single frame, then the result should be 0.0 or at least very close to that
3887 value. If, however, there is a significant deviation of the mean value from
3888 0.0, in either positive or negative direction, this is referred to as a
3889 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3890 Audio Normalizer provides optional DC bias correction.
3891 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3892 the mean value, or "DC correction" offset, of each input frame and subtract
3893 that value from all of the frame's sample values which ensures those samples
3894 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3895 boundaries, the DC correction offset values will be interpolated smoothly
3896 between neighbouring frames.
3898 @item altboundary, b
3899 Enable alternative boundary mode. By default is disabled.
3900 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3901 around each frame. This includes the preceding frames as well as the
3902 subsequent frames. However, for the "boundary" frames, located at the very
3903 beginning and at the very end of the audio file, not all neighbouring
3904 frames are available. In particular, for the first few frames in the audio
3905 file, the preceding frames are not known. And, similarly, for the last few
3906 frames in the audio file, the subsequent frames are not known. Thus, the
3907 question arises which gain factors should be assumed for the missing frames
3908 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3909 to deal with this situation. The default boundary mode assumes a gain factor
3910 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3911 "fade out" at the beginning and at the end of the input, respectively.
3914 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3915 By default, the Dynamic Audio Normalizer does not apply "traditional"
3916 compression. This means that signal peaks will not be pruned and thus the
3917 full dynamic range will be retained within each local neighbourhood. However,
3918 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3919 normalization algorithm with a more "traditional" compression.
3920 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3921 (thresholding) function. If (and only if) the compression feature is enabled,
3922 all input frames will be processed by a soft knee thresholding function prior
3923 to the actual normalization process. Put simply, the thresholding function is
3924 going to prune all samples whose magnitude exceeds a certain threshold value.
3925 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3926 value. Instead, the threshold value will be adjusted for each individual
3928 In general, smaller parameters result in stronger compression, and vice versa.
3929 Values below 3.0 are not recommended, because audible distortion may appear.
3932 Set the target threshold value. This specifies the lowest permissible
3933 magnitude level for the audio input which will be normalized.
3934 If input frame volume is above this value frame will be normalized.
3935 Otherwise frame may not be normalized at all. The default value is set
3936 to 0, which means all input frames will be normalized.
3937 This option is mostly useful if digital noise is not wanted to be amplified.
3940 @subsection Commands
3942 This filter supports the all above options as @ref{commands}.
3946 Make audio easier to listen to on headphones.
3948 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3949 so that when listened to on headphones the stereo image is moved from
3950 inside your head (standard for headphones) to outside and in front of
3951 the listener (standard for speakers).
3957 Apply a two-pole peaking equalisation (EQ) filter. With this
3958 filter, the signal-level at and around a selected frequency can
3959 be increased or decreased, whilst (unlike bandpass and bandreject
3960 filters) that at all other frequencies is unchanged.
3962 In order to produce complex equalisation curves, this filter can
3963 be given several times, each with a different central frequency.
3965 The filter accepts the following options:
3969 Set the filter's central frequency in Hz.
3972 Set method to specify band-width of filter.
3987 Specify the band-width of a filter in width_type units.
3990 Set the required gain or attenuation in dB.
3991 Beware of clipping when using a positive gain.
3994 How much to use filtered signal in output. Default is 1.
3995 Range is between 0 and 1.
3998 Specify which channels to filter, by default all available are filtered.
4001 Normalize biquad coefficients, by default is disabled.
4002 Enabling it will normalize magnitude response at DC to 0dB.
4005 Set transform type of IIR filter.
4014 Set precison of filtering.
4017 Pick automatic sample format depending on surround filters.
4019 Always use signed 16-bit.
4021 Always use signed 32-bit.
4023 Always use float 32-bit.
4025 Always use float 64-bit.
4029 @subsection Examples
4032 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4034 equalizer=f=1000:t=h:width=200:g=-10
4038 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4040 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4044 @subsection Commands
4046 This filter supports the following commands:
4049 Change equalizer frequency.
4050 Syntax for the command is : "@var{frequency}"
4053 Change equalizer width_type.
4054 Syntax for the command is : "@var{width_type}"
4057 Change equalizer width.
4058 Syntax for the command is : "@var{width}"
4061 Change equalizer gain.
4062 Syntax for the command is : "@var{gain}"
4065 Change equalizer mix.
4066 Syntax for the command is : "@var{mix}"
4069 @section extrastereo
4071 Linearly increases the difference between left and right channels which
4072 adds some sort of "live" effect to playback.
4074 The filter accepts the following options:
4078 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4079 (average of both channels), with 1.0 sound will be unchanged, with
4080 -1.0 left and right channels will be swapped.
4083 Enable clipping. By default is enabled.
4086 @subsection Commands
4088 This filter supports the all above options as @ref{commands}.
4090 @section firequalizer
4091 Apply FIR Equalization using arbitrary frequency response.
4093 The filter accepts the following option:
4097 Set gain curve equation (in dB). The expression can contain variables:
4100 the evaluated frequency
4104 channel number, set to 0 when multichannels evaluation is disabled
4106 channel id, see libavutil/channel_layout.h, set to the first channel id when
4107 multichannels evaluation is disabled
4111 channel_layout, see libavutil/channel_layout.h
4116 @item gain_interpolate(f)
4117 interpolate gain on frequency f based on gain_entry
4118 @item cubic_interpolate(f)
4119 same as gain_interpolate, but smoother
4121 This option is also available as command. Default is @code{gain_interpolate(f)}.
4124 Set gain entry for gain_interpolate function. The expression can
4128 store gain entry at frequency f with value g
4130 This option is also available as command.
4133 Set filter delay in seconds. Higher value means more accurate.
4134 Default is @code{0.01}.
4137 Set filter accuracy in Hz. Lower value means more accurate.
4138 Default is @code{5}.
4141 Set window function. Acceptable values are:
4144 rectangular window, useful when gain curve is already smooth
4146 hann window (default)
4152 3-terms continuous 1st derivative nuttall window
4154 minimum 3-terms discontinuous nuttall window
4156 4-terms continuous 1st derivative nuttall window
4158 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4160 blackman-harris window
4166 If enabled, use fixed number of audio samples. This improves speed when
4167 filtering with large delay. Default is disabled.
4170 Enable multichannels evaluation on gain. Default is disabled.
4173 Enable zero phase mode by subtracting timestamp to compensate delay.
4174 Default is disabled.
4177 Set scale used by gain. Acceptable values are:
4180 linear frequency, linear gain
4182 linear frequency, logarithmic (in dB) gain (default)
4184 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4186 logarithmic frequency, logarithmic gain
4190 Set file for dumping, suitable for gnuplot.
4193 Set scale for dumpfile. Acceptable values are same with scale option.
4197 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4198 Default is disabled.
4201 Enable minimum phase impulse response. Default is disabled.
4204 @subsection Examples
4209 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4212 lowpass at 1000 Hz with gain_entry:
4214 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4217 custom equalization:
4219 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4222 higher delay with zero phase to compensate delay:
4224 firequalizer=delay=0.1:fixed=on:zero_phase=on
4227 lowpass on left channel, highpass on right channel:
4229 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4230 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4235 Apply a flanging effect to the audio.
4237 The filter accepts the following options:
4241 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4244 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4247 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4251 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4252 Default value is 71.
4255 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4258 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4259 Default value is @var{sinusoidal}.
4262 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4263 Default value is 25.
4266 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4267 Default is @var{linear}.
4271 Apply Haas effect to audio.
4273 Note that this makes most sense to apply on mono signals.
4274 With this filter applied to mono signals it give some directionality and
4275 stretches its stereo image.
4277 The filter accepts the following options:
4281 Set input level. By default is @var{1}, or 0dB
4284 Set output level. By default is @var{1}, or 0dB.
4287 Set gain applied to side part of signal. By default is @var{1}.
4290 Set kind of middle source. Can be one of the following:
4300 Pick middle part signal of stereo image.
4303 Pick side part signal of stereo image.
4307 Change middle phase. By default is disabled.
4310 Set left channel delay. By default is @var{2.05} milliseconds.
4313 Set left channel balance. By default is @var{-1}.
4316 Set left channel gain. By default is @var{1}.
4319 Change left phase. By default is disabled.
4322 Set right channel delay. By defaults is @var{2.12} milliseconds.
4325 Set right channel balance. By default is @var{1}.
4328 Set right channel gain. By default is @var{1}.
4331 Change right phase. By default is enabled.
4336 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4337 embedded HDCD codes is expanded into a 20-bit PCM stream.
4339 The filter supports the Peak Extend and Low-level Gain Adjustment features
4340 of HDCD, and detects the Transient Filter flag.
4343 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4346 When using the filter with wav, note the default encoding for wav is 16-bit,
4347 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4348 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4350 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4351 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4354 The filter accepts the following options:
4357 @item disable_autoconvert
4358 Disable any automatic format conversion or resampling in the filter graph.
4360 @item process_stereo
4361 Process the stereo channels together. If target_gain does not match between
4362 channels, consider it invalid and use the last valid target_gain.
4365 Set the code detect timer period in ms.
4368 Always extend peaks above -3dBFS even if PE isn't signaled.
4371 Replace audio with a solid tone and adjust the amplitude to signal some
4372 specific aspect of the decoding process. The output file can be loaded in
4373 an audio editor alongside the original to aid analysis.
4375 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4382 Gain adjustment level at each sample
4384 Samples where peak extend occurs
4386 Samples where the code detect timer is active
4388 Samples where the target gain does not match between channels
4394 Apply head-related transfer functions (HRTFs) to create virtual
4395 loudspeakers around the user for binaural listening via headphones.
4396 The HRIRs are provided via additional streams, for each channel
4397 one stereo input stream is needed.
4399 The filter accepts the following options:
4403 Set mapping of input streams for convolution.
4404 The argument is a '|'-separated list of channel names in order as they
4405 are given as additional stream inputs for filter.
4406 This also specify number of input streams. Number of input streams
4407 must be not less than number of channels in first stream plus one.
4410 Set gain applied to audio. Value is in dB. Default is 0.
4413 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4414 processing audio in time domain which is slow.
4415 @var{freq} is processing audio in frequency domain which is fast.
4416 Default is @var{freq}.
4419 Set custom gain for LFE channels. Value is in dB. Default is 0.
4422 Set size of frame in number of samples which will be processed at once.
4423 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4426 Set format of hrir stream.
4427 Default value is @var{stereo}. Alternative value is @var{multich}.
4428 If value is set to @var{stereo}, number of additional streams should
4429 be greater or equal to number of input channels in first input stream.
4430 Also each additional stream should have stereo number of channels.
4431 If value is set to @var{multich}, number of additional streams should
4432 be exactly one. Also number of input channels of additional stream
4433 should be equal or greater than twice number of channels of first input
4437 @subsection Examples
4441 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4442 each amovie filter use stereo file with IR coefficients as input.
4443 The files give coefficients for each position of virtual loudspeaker:
4446 -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"
4451 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4452 but now in @var{multich} @var{hrir} format.
4454 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"
4461 Apply a high-pass filter with 3dB point frequency.
4462 The filter can be either single-pole, or double-pole (the default).
4463 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4465 The filter accepts the following options:
4469 Set frequency in Hz. Default is 3000.
4472 Set number of poles. Default is 2.
4475 Set method to specify band-width of filter.
4490 Specify the band-width of a filter in width_type units.
4491 Applies only to double-pole filter.
4492 The default is 0.707q and gives a Butterworth response.
4495 How much to use filtered signal in output. Default is 1.
4496 Range is between 0 and 1.
4499 Specify which channels to filter, by default all available are filtered.
4502 Normalize biquad coefficients, by default is disabled.
4503 Enabling it will normalize magnitude response at DC to 0dB.
4506 Set transform type of IIR filter.
4515 Set precison of filtering.
4518 Pick automatic sample format depending on surround filters.
4520 Always use signed 16-bit.
4522 Always use signed 32-bit.
4524 Always use float 32-bit.
4526 Always use float 64-bit.
4530 @subsection Commands
4532 This filter supports the following commands:
4535 Change highpass frequency.
4536 Syntax for the command is : "@var{frequency}"
4539 Change highpass width_type.
4540 Syntax for the command is : "@var{width_type}"
4543 Change highpass width.
4544 Syntax for the command is : "@var{width}"
4547 Change highpass mix.
4548 Syntax for the command is : "@var{mix}"
4553 Join multiple input streams into one multi-channel stream.
4555 It accepts the following parameters:
4559 The number of input streams. It defaults to 2.
4561 @item channel_layout
4562 The desired output channel layout. It defaults to stereo.
4565 Map channels from inputs to output. The argument is a '|'-separated list of
4566 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4567 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4568 can be either the name of the input channel (e.g. FL for front left) or its
4569 index in the specified input stream. @var{out_channel} is the name of the output
4573 The filter will attempt to guess the mappings when they are not specified
4574 explicitly. It does so by first trying to find an unused matching input channel
4575 and if that fails it picks the first unused input channel.
4577 Join 3 inputs (with properly set channel layouts):
4579 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4582 Build a 5.1 output from 6 single-channel streams:
4584 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4585 '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'
4591 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4593 To enable compilation of this filter you need to configure FFmpeg with
4594 @code{--enable-ladspa}.
4598 Specifies the name of LADSPA plugin library to load. If the environment
4599 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4600 each one of the directories specified by the colon separated list in
4601 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4602 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4603 @file{/usr/lib/ladspa/}.
4606 Specifies the plugin within the library. Some libraries contain only
4607 one plugin, but others contain many of them. If this is not set filter
4608 will list all available plugins within the specified library.
4611 Set the '|' separated list of controls which are zero or more floating point
4612 values that determine the behavior of the loaded plugin (for example delay,
4614 Controls need to be defined using the following syntax:
4615 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 Alternatively they can be also defined using the following syntax:
4618 @var{value0}|@var{value1}|@var{value2}|..., where
4619 @var{valuei} is the value set on the @var{i}-th control.
4620 If @option{controls} is set to @code{help}, all available controls and
4621 their valid ranges are printed.
4623 @item sample_rate, s
4624 Specify the sample rate, default to 44100. Only used if plugin have
4628 Set the number of samples per channel per each output frame, default
4629 is 1024. Only used if plugin have zero inputs.
4632 Set the minimum duration of the sourced audio. See
4633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4634 for the accepted syntax.
4635 Note that the resulting duration may be greater than the specified duration,
4636 as the generated audio is always cut at the end of a complete frame.
4637 If not specified, or the expressed duration is negative, the audio is
4638 supposed to be generated forever.
4639 Only used if plugin have zero inputs.
4642 Enable latency compensation, by default is disabled.
4643 Only used if plugin have inputs.
4646 @subsection Examples
4650 List all available plugins within amp (LADSPA example plugin) library:
4656 List all available controls and their valid ranges for @code{vcf_notch}
4657 plugin from @code{VCF} library:
4659 ladspa=f=vcf:p=vcf_notch:c=help
4663 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4666 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4670 Add reverberation to the audio using TAP-plugins
4671 (Tom's Audio Processing plugins):
4673 ladspa=file=tap_reverb:tap_reverb
4677 Generate white noise, with 0.2 amplitude:
4679 ladspa=file=cmt:noise_source_white:c=c0=.2
4683 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4684 @code{C* Audio Plugin Suite} (CAPS) library:
4686 ladspa=file=caps:Click:c=c1=20'
4690 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4692 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4696 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4697 @code{SWH Plugins} collection:
4699 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4703 Attenuate low frequencies using Multiband EQ from Steve Harris
4704 @code{SWH Plugins} collection:
4706 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4710 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4713 ladspa=caps:Narrower
4717 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4719 ladspa=caps:White:.2
4723 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4725 ladspa=caps:Fractal:c=c1=1
4729 Dynamic volume normalization using @code{VLevel} plugin:
4731 ladspa=vlevel-ladspa:vlevel_mono
4735 @subsection Commands
4737 This filter supports the following commands:
4740 Modify the @var{N}-th control value.
4742 If the specified value is not valid, it is ignored and prior one is kept.
4747 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4748 Support for both single pass (livestreams, files) and double pass (files) modes.
4749 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4750 detect true peaks, the audio stream will be upsampled to 192 kHz.
4751 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4753 The filter accepts the following options:
4757 Set integrated loudness target.
4758 Range is -70.0 - -5.0. Default value is -24.0.
4761 Set loudness range target.
4762 Range is 1.0 - 20.0. Default value is 7.0.
4765 Set maximum true peak.
4766 Range is -9.0 - +0.0. Default value is -2.0.
4768 @item measured_I, measured_i
4769 Measured IL of input file.
4770 Range is -99.0 - +0.0.
4772 @item measured_LRA, measured_lra
4773 Measured LRA of input file.
4774 Range is 0.0 - 99.0.
4776 @item measured_TP, measured_tp
4777 Measured true peak of input file.
4778 Range is -99.0 - +99.0.
4780 @item measured_thresh
4781 Measured threshold of input file.
4782 Range is -99.0 - +0.0.
4785 Set offset gain. Gain is applied before the true-peak limiter.
4786 Range is -99.0 - +99.0. Default is +0.0.
4789 Normalize by linearly scaling the source audio.
4790 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4791 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4792 be lower than source LRA and the change in integrated loudness shouldn't
4793 result in a true peak which exceeds the target TP. If any of these
4794 conditions aren't met, normalization mode will revert to @var{dynamic}.
4795 Options are @code{true} or @code{false}. Default is @code{true}.
4798 Treat mono input files as "dual-mono". If a mono file is intended for playback
4799 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4800 If set to @code{true}, this option will compensate for this effect.
4801 Multi-channel input files are not affected by this option.
4802 Options are true or false. Default is false.
4805 Set print format for stats. Options are summary, json, or none.
4806 Default value is none.
4811 Apply a low-pass filter with 3dB point frequency.
4812 The filter can be either single-pole or double-pole (the default).
4813 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4815 The filter accepts the following options:
4819 Set frequency in Hz. Default is 500.
4822 Set number of poles. Default is 2.
4825 Set method to specify band-width of filter.
4840 Specify the band-width of a filter in width_type units.
4841 Applies only to double-pole filter.
4842 The default is 0.707q and gives a Butterworth response.
4845 How much to use filtered signal in output. Default is 1.
4846 Range is between 0 and 1.
4849 Specify which channels to filter, by default all available are filtered.
4852 Normalize biquad coefficients, by default is disabled.
4853 Enabling it will normalize magnitude response at DC to 0dB.
4856 Set transform type of IIR filter.
4865 Set precison of filtering.
4868 Pick automatic sample format depending on surround filters.
4870 Always use signed 16-bit.
4872 Always use signed 32-bit.
4874 Always use float 32-bit.
4876 Always use float 64-bit.
4880 @subsection Examples
4883 Lowpass only LFE channel, it LFE is not present it does nothing:
4889 @subsection Commands
4891 This filter supports the following commands:
4894 Change lowpass frequency.
4895 Syntax for the command is : "@var{frequency}"
4898 Change lowpass width_type.
4899 Syntax for the command is : "@var{width_type}"
4902 Change lowpass width.
4903 Syntax for the command is : "@var{width}"
4907 Syntax for the command is : "@var{mix}"
4912 Load a LV2 (LADSPA Version 2) plugin.
4914 To enable compilation of this filter you need to configure FFmpeg with
4915 @code{--enable-lv2}.
4919 Specifies the plugin URI. You may need to escape ':'.
4922 Set the '|' separated list of controls which are zero or more floating point
4923 values that determine the behavior of the loaded plugin (for example delay,
4925 If @option{controls} is set to @code{help}, all available controls and
4926 their valid ranges are printed.
4928 @item sample_rate, s
4929 Specify the sample rate, default to 44100. Only used if plugin have
4933 Set the number of samples per channel per each output frame, default
4934 is 1024. Only used if plugin have zero inputs.
4937 Set the minimum duration of the sourced audio. See
4938 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4939 for the accepted syntax.
4940 Note that the resulting duration may be greater than the specified duration,
4941 as the generated audio is always cut at the end of a complete frame.
4942 If not specified, or the expressed duration is negative, the audio is
4943 supposed to be generated forever.
4944 Only used if plugin have zero inputs.
4947 @subsection Examples
4951 Apply bass enhancer plugin from Calf:
4953 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4957 Apply vinyl plugin from Calf:
4959 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4963 Apply bit crusher plugin from ArtyFX:
4965 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4970 Multiband Compress or expand the audio's dynamic range.
4972 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4973 This is akin to the crossover of a loudspeaker, and results in flat frequency
4974 response when absent compander action.
4976 It accepts the following parameters:
4980 This option syntax is:
4981 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4982 For explanation of each item refer to compand filter documentation.
4988 Mix channels with specific gain levels. The filter accepts the output
4989 channel layout followed by a set of channels definitions.
4991 This filter is also designed to efficiently remap the channels of an audio
4994 The filter accepts parameters of the form:
4995 "@var{l}|@var{outdef}|@var{outdef}|..."
4999 output channel layout or number of channels
5002 output channel specification, of the form:
5003 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5006 output channel to define, either a channel name (FL, FR, etc.) or a channel
5007 number (c0, c1, etc.)
5010 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5013 input channel to use, see out_name for details; it is not possible to mix
5014 named and numbered input channels
5017 If the `=' in a channel specification is replaced by `<', then the gains for
5018 that specification will be renormalized so that the total is 1, thus
5019 avoiding clipping noise.
5021 @subsection Mixing examples
5023 For example, if you want to down-mix from stereo to mono, but with a bigger
5024 factor for the left channel:
5026 pan=1c|c0=0.9*c0+0.1*c1
5029 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5030 7-channels surround:
5032 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5035 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5036 that should be preferred (see "-ac" option) unless you have very specific
5039 @subsection Remapping examples
5041 The channel remapping will be effective if, and only if:
5044 @item gain coefficients are zeroes or ones,
5045 @item only one input per channel output,
5048 If all these conditions are satisfied, the filter will notify the user ("Pure
5049 channel mapping detected"), and use an optimized and lossless method to do the
5052 For example, if you have a 5.1 source and want a stereo audio stream by
5053 dropping the extra channels:
5055 pan="stereo| c0=FL | c1=FR"
5058 Given the same source, you can also switch front left and front right channels
5059 and keep the input channel layout:
5061 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5064 If the input is a stereo audio stream, you can mute the front left channel (and
5065 still keep the stereo channel layout) with:
5070 Still with a stereo audio stream input, you can copy the right channel in both
5071 front left and right:
5073 pan="stereo| c0=FR | c1=FR"
5078 ReplayGain scanner filter. This filter takes an audio stream as an input and
5079 outputs it unchanged.
5080 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5084 Convert the audio sample format, sample rate and channel layout. It is
5085 not meant to be used directly.
5088 Apply time-stretching and pitch-shifting with librubberband.
5090 To enable compilation of this filter, you need to configure FFmpeg with
5091 @code{--enable-librubberband}.
5093 The filter accepts the following options:
5097 Set tempo scale factor.
5100 Set pitch scale factor.
5103 Set transients detector.
5104 Possible values are:
5113 Possible values are:
5122 Possible values are:
5129 Set processing window size.
5130 Possible values are:
5139 Possible values are:
5146 Enable formant preservation when shift pitching.
5147 Possible values are:
5155 Possible values are:
5164 Possible values are:
5171 @subsection Commands
5173 This filter supports the following commands:
5176 Change filter tempo scale factor.
5177 Syntax for the command is : "@var{tempo}"
5180 Change filter pitch scale factor.
5181 Syntax for the command is : "@var{pitch}"
5184 @section sidechaincompress
5186 This filter acts like normal compressor but has the ability to compress
5187 detected signal using second input signal.
5188 It needs two input streams and returns one output stream.
5189 First input stream will be processed depending on second stream signal.
5190 The filtered signal then can be filtered with other filters in later stages of
5191 processing. See @ref{pan} and @ref{amerge} filter.
5193 The filter accepts the following options:
5197 Set input gain. Default is 1. Range is between 0.015625 and 64.
5200 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5201 Default is @code{downward}.
5204 If a signal of second stream raises above this level it will affect the gain
5205 reduction of first stream.
5206 By default is 0.125. Range is between 0.00097563 and 1.
5209 Set a ratio about which the signal is reduced. 1:2 means that if the level
5210 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5211 Default is 2. Range is between 1 and 20.
5214 Amount of milliseconds the signal has to rise above the threshold before gain
5215 reduction starts. Default is 20. Range is between 0.01 and 2000.
5218 Amount of milliseconds the signal has to fall below the threshold before
5219 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5222 Set the amount by how much signal will be amplified after processing.
5223 Default is 1. Range is from 1 to 64.
5226 Curve the sharp knee around the threshold to enter gain reduction more softly.
5227 Default is 2.82843. Range is between 1 and 8.
5230 Choose if the @code{average} level between all channels of side-chain stream
5231 or the louder(@code{maximum}) channel of side-chain stream affects the
5232 reduction. Default is @code{average}.
5235 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5236 of @code{rms}. Default is @code{rms} which is mainly smoother.
5239 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5242 How much to use compressed signal in output. Default is 1.
5243 Range is between 0 and 1.
5246 @subsection Commands
5248 This filter supports the all above options as @ref{commands}.
5250 @subsection Examples
5254 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5255 depending on the signal of 2nd input and later compressed signal to be
5256 merged with 2nd input:
5258 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5262 @section sidechaingate
5264 A sidechain gate acts like a normal (wideband) gate but has the ability to
5265 filter the detected signal before sending it to the gain reduction stage.
5266 Normally a gate uses the full range signal to detect a level above the
5268 For example: If you cut all lower frequencies from your sidechain signal
5269 the gate will decrease the volume of your track only if not enough highs
5270 appear. With this technique you are able to reduce the resonation of a
5271 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5273 It needs two input streams and returns one output stream.
5274 First input stream will be processed depending on second stream signal.
5276 The filter accepts the following options:
5280 Set input level before filtering.
5281 Default is 1. Allowed range is from 0.015625 to 64.
5284 Set the mode of operation. Can be @code{upward} or @code{downward}.
5285 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5286 will be amplified, expanding dynamic range in upward direction.
5287 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5290 Set the level of gain reduction when the signal is below the threshold.
5291 Default is 0.06125. Allowed range is from 0 to 1.
5292 Setting this to 0 disables reduction and then filter behaves like expander.
5295 If a signal rises above this level the gain reduction is released.
5296 Default is 0.125. Allowed range is from 0 to 1.
5299 Set a ratio about which the signal is reduced.
5300 Default is 2. Allowed range is from 1 to 9000.
5303 Amount of milliseconds the signal has to rise above the threshold before gain
5305 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5308 Amount of milliseconds the signal has to fall below the threshold before the
5309 reduction is increased again. Default is 250 milliseconds.
5310 Allowed range is from 0.01 to 9000.
5313 Set amount of amplification of signal after processing.
5314 Default is 1. Allowed range is from 1 to 64.
5317 Curve the sharp knee around the threshold to enter gain reduction more softly.
5318 Default is 2.828427125. Allowed range is from 1 to 8.
5321 Choose if exact signal should be taken for detection or an RMS like one.
5322 Default is rms. Can be peak or rms.
5325 Choose if the average level between all channels or the louder channel affects
5327 Default is average. Can be average or maximum.
5330 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5333 @subsection Commands
5335 This filter supports the all above options as @ref{commands}.
5337 @section silencedetect
5339 Detect silence in an audio stream.
5341 This filter logs a message when it detects that the input audio volume is less
5342 or equal to a noise tolerance value for a duration greater or equal to the
5343 minimum detected noise duration.
5345 The printed times and duration are expressed in seconds. The
5346 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5347 is set on the first frame whose timestamp equals or exceeds the detection
5348 duration and it contains the timestamp of the first frame of the silence.
5350 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5351 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5352 keys are set on the first frame after the silence. If @option{mono} is
5353 enabled, and each channel is evaluated separately, the @code{.X}
5354 suffixed keys are used, and @code{X} corresponds to the channel number.
5356 The filter accepts the following options:
5360 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5361 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5364 Set silence duration until notification (default is 2 seconds). See
5365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5366 for the accepted syntax.
5369 Process each channel separately, instead of combined. By default is disabled.
5372 @subsection Examples
5376 Detect 5 seconds of silence with -50dB noise tolerance:
5378 silencedetect=n=-50dB:d=5
5382 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5383 tolerance in @file{silence.mp3}:
5385 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5389 @section silenceremove
5391 Remove silence from the beginning, middle or end of the audio.
5393 The filter accepts the following options:
5397 This value is used to indicate if audio should be trimmed at beginning of
5398 the audio. A value of zero indicates no silence should be trimmed from the
5399 beginning. When specifying a non-zero value, it trims audio up until it
5400 finds non-silence. Normally, when trimming silence from beginning of audio
5401 the @var{start_periods} will be @code{1} but it can be increased to higher
5402 values to trim all audio up to specific count of non-silence periods.
5403 Default value is @code{0}.
5405 @item start_duration
5406 Specify the amount of time that non-silence must be detected before it stops
5407 trimming audio. By increasing the duration, bursts of noises can be treated
5408 as silence and trimmed off. Default value is @code{0}.
5410 @item start_threshold
5411 This indicates what sample value should be treated as silence. For digital
5412 audio, a value of @code{0} may be fine but for audio recorded from analog,
5413 you may wish to increase the value to account for background noise.
5414 Can be specified in dB (in case "dB" is appended to the specified value)
5415 or amplitude ratio. Default value is @code{0}.
5418 Specify max duration of silence at beginning that will be kept after
5419 trimming. Default is 0, which is equal to trimming all samples detected
5423 Specify mode of detection of silence end in start of multi-channel audio.
5424 Can be @var{any} or @var{all}. Default is @var{any}.
5425 With @var{any}, any sample that is detected as non-silence will cause
5426 stopped trimming of silence.
5427 With @var{all}, only if all channels are detected as non-silence will cause
5428 stopped trimming of silence.
5431 Set the count for trimming silence from the end of audio.
5432 To remove silence from the middle of a file, specify a @var{stop_periods}
5433 that is negative. This value is then treated as a positive value and is
5434 used to indicate the effect should restart processing as specified by
5435 @var{start_periods}, making it suitable for removing periods of silence
5436 in the middle of the audio.
5437 Default value is @code{0}.
5440 Specify a duration of silence that must exist before audio is not copied any
5441 more. By specifying a higher duration, silence that is wanted can be left in
5443 Default value is @code{0}.
5445 @item stop_threshold
5446 This is the same as @option{start_threshold} but for trimming silence from
5448 Can be specified in dB (in case "dB" is appended to the specified value)
5449 or amplitude ratio. Default value is @code{0}.
5452 Specify max duration of silence at end that will be kept after
5453 trimming. Default is 0, which is equal to trimming all samples detected
5457 Specify mode of detection of silence start in end of multi-channel audio.
5458 Can be @var{any} or @var{all}. Default is @var{any}.
5459 With @var{any}, any sample that is detected as non-silence will cause
5460 stopped trimming of silence.
5461 With @var{all}, only if all channels are detected as non-silence will cause
5462 stopped trimming of silence.
5465 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5466 and works better with digital silence which is exactly 0.
5467 Default value is @code{rms}.
5470 Set duration in number of seconds used to calculate size of window in number
5471 of samples for detecting silence.
5472 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5475 @subsection Examples
5479 The following example shows how this filter can be used to start a recording
5480 that does not contain the delay at the start which usually occurs between
5481 pressing the record button and the start of the performance:
5483 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5487 Trim all silence encountered from beginning to end where there is more than 1
5488 second of silence in audio:
5490 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5494 Trim all digital silence samples, using peak detection, from beginning to end
5495 where there is more than 0 samples of digital silence in audio and digital
5496 silence is detected in all channels at same positions in stream:
5498 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5504 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5505 loudspeakers around the user for binaural listening via headphones (audio
5506 formats up to 9 channels supported).
5507 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5508 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5509 Austrian Academy of Sciences.
5511 To enable compilation of this filter you need to configure FFmpeg with
5512 @code{--enable-libmysofa}.
5514 The filter accepts the following options:
5518 Set the SOFA file used for rendering.
5521 Set gain applied to audio. Value is in dB. Default is 0.
5524 Set rotation of virtual loudspeakers in deg. Default is 0.
5527 Set elevation of virtual speakers in deg. Default is 0.
5530 Set distance in meters between loudspeakers and the listener with near-field
5531 HRTFs. Default is 1.
5534 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5535 processing audio in time domain which is slow.
5536 @var{freq} is processing audio in frequency domain which is fast.
5537 Default is @var{freq}.
5540 Set custom positions of virtual loudspeakers. Syntax for this option is:
5541 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5542 Each virtual loudspeaker is described with short channel name following with
5543 azimuth and elevation in degrees.
5544 Each virtual loudspeaker description is separated by '|'.
5545 For example to override front left and front right channel positions use:
5546 'speakers=FL 45 15|FR 345 15'.
5547 Descriptions with unrecognised channel names are ignored.
5550 Set custom gain for LFE channels. Value is in dB. Default is 0.
5553 Set custom frame size in number of samples. Default is 1024.
5554 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5555 is set to @var{freq}.
5558 Should all IRs be normalized upon importing SOFA file.
5559 By default is enabled.
5562 Should nearest IRs be interpolated with neighbor IRs if exact position
5563 does not match. By default is disabled.
5566 Minphase all IRs upon loading of SOFA file. By default is disabled.
5569 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5572 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5575 @subsection Examples
5579 Using ClubFritz6 sofa file:
5581 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5585 Using ClubFritz12 sofa file and bigger radius with small rotation:
5587 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5591 Similar as above but with custom speaker positions for front left, front right, back left and back right
5592 and also with custom gain:
5594 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5601 This filter expands or compresses each half-cycle of audio samples
5602 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5603 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5605 The filter accepts the following options:
5609 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5610 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5613 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5614 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5615 would be such that local peak value reaches target peak value but never to surpass it and that
5616 ratio between new and previous peak value does not surpass this option value.
5618 @item compression, c
5619 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5620 This option controls maximum local half-cycle of samples compression. This option is used
5621 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5622 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5623 that peak's half-cycle will be compressed by current compression factor.
5626 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5627 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5628 Any half-cycle samples with their local peak value below or same as this option value will be
5629 compressed by current compression factor, otherwise, if greater than threshold value they will be
5630 expanded with expansion factor so that it could reach peak target value but never surpass it.
5633 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5634 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5635 each new half-cycle until it reaches @option{expansion} value.
5636 Setting this options too high may lead to distortions.
5639 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5640 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5641 each new half-cycle until it reaches @option{compression} value.
5644 Specify which channels to filter, by default all available channels are filtered.
5647 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5648 option. When enabled any half-cycle of samples with their local peak value below or same as
5649 @option{threshold} option will be expanded otherwise it will be compressed.
5652 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5653 When disabled each filtered channel gain calculation is independent, otherwise when this option
5654 is enabled the minimum of all possible gains for each filtered channel is used.
5657 @subsection Commands
5659 This filter supports the all above options as @ref{commands}.
5661 @section stereotools
5663 This filter has some handy utilities to manage stereo signals, for converting
5664 M/S stereo recordings to L/R signal while having control over the parameters
5665 or spreading the stereo image of master track.
5667 The filter accepts the following options:
5671 Set input level before filtering for both channels. Defaults is 1.
5672 Allowed range is from 0.015625 to 64.
5675 Set output level after filtering for both channels. Defaults is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set input balance between both channels. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set output balance between both channels. Default is 0.
5684 Allowed range is from -1 to 1.
5687 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5688 clipping. Disabled by default.
5691 Mute the left channel. Disabled by default.
5694 Mute the right channel. Disabled by default.
5697 Change the phase of the left channel. Disabled by default.
5700 Change the phase of the right channel. Disabled by default.
5703 Set stereo mode. Available values are:
5707 Left/Right to Left/Right, this is default.
5710 Left/Right to Mid/Side.
5713 Mid/Side to Left/Right.
5716 Left/Right to Left/Left.
5719 Left/Right to Right/Right.
5722 Left/Right to Left + Right.
5725 Left/Right to Right/Left.
5728 Mid/Side to Left/Left.
5731 Mid/Side to Right/Right.
5734 Mid/Side to Right/Left.
5737 Left/Right to Left - Right.
5741 Set level of side signal. Default is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set balance of side signal. Default is 0.
5746 Allowed range is from -1 to 1.
5749 Set level of the middle signal. Default is 1.
5750 Allowed range is from 0.015625 to 64.
5753 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5756 Set stereo base between mono and inversed channels. Default is 0.
5757 Allowed range is from -1 to 1.
5760 Set delay in milliseconds how much to delay left from right channel and
5761 vice versa. Default is 0. Allowed range is from -20 to 20.
5764 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5767 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5769 @item bmode_in, bmode_out
5770 Set balance mode for balance_in/balance_out option.
5772 Can be one of the following:
5776 Classic balance mode. Attenuate one channel at time.
5777 Gain is raised up to 1.
5780 Similar as classic mode above but gain is raised up to 2.
5783 Equal power distribution, from -6dB to +6dB range.
5787 @subsection Commands
5789 This filter supports the all above options as @ref{commands}.
5791 @subsection Examples
5795 Apply karaoke like effect:
5797 stereotools=mlev=0.015625
5801 Convert M/S signal to L/R:
5803 "stereotools=mode=ms>lr"
5807 @section stereowiden
5809 This filter enhance the stereo effect by suppressing signal common to both
5810 channels and by delaying the signal of left into right and vice versa,
5811 thereby widening the stereo effect.
5813 The filter accepts the following options:
5817 Time in milliseconds of the delay of left signal into right and vice versa.
5818 Default is 20 milliseconds.
5821 Amount of gain in delayed signal into right and vice versa. Gives a delay
5822 effect of left signal in right output and vice versa which gives widening
5823 effect. Default is 0.3.
5826 Cross feed of left into right with inverted phase. This helps in suppressing
5827 the mono. If the value is 1 it will cancel all the signal common to both
5828 channels. Default is 0.3.
5831 Set level of input signal of original channel. Default is 0.8.
5834 @subsection Commands
5836 This filter supports the all above options except @code{delay} as @ref{commands}.
5838 @section superequalizer
5839 Apply 18 band equalizer.
5841 The filter accepts the following options:
5848 Set 131Hz band gain.
5850 Set 185Hz band gain.
5852 Set 262Hz band gain.
5854 Set 370Hz band gain.
5856 Set 523Hz band gain.
5858 Set 740Hz band gain.
5860 Set 1047Hz band gain.
5862 Set 1480Hz band gain.
5864 Set 2093Hz band gain.
5866 Set 2960Hz band gain.
5868 Set 4186Hz band gain.
5870 Set 5920Hz band gain.
5872 Set 8372Hz band gain.
5874 Set 11840Hz band gain.
5876 Set 16744Hz band gain.
5878 Set 20000Hz band gain.
5882 Apply audio surround upmix filter.
5884 This filter allows to produce multichannel output from audio stream.
5886 The filter accepts the following options:
5890 Set output channel layout. By default, this is @var{5.1}.
5892 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5893 for the required syntax.
5896 Set input channel layout. By default, this is @var{stereo}.
5898 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5899 for the required syntax.
5902 Set input volume level. By default, this is @var{1}.
5905 Set output volume level. By default, this is @var{1}.
5908 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5911 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5914 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5917 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5918 In @var{add} mode, LFE channel is created from input audio and added to output.
5919 In @var{sub} mode, LFE channel is created from input audio and added to output but
5920 also all non-LFE output channels are subtracted with output LFE channel.
5923 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5924 Default is @var{90}.
5927 Set front center input volume. By default, this is @var{1}.
5930 Set front center output volume. By default, this is @var{1}.
5933 Set front left input volume. By default, this is @var{1}.
5936 Set front left output volume. By default, this is @var{1}.
5939 Set front right input volume. By default, this is @var{1}.
5942 Set front right output volume. By default, this is @var{1}.
5945 Set side left input volume. By default, this is @var{1}.
5948 Set side left output volume. By default, this is @var{1}.
5951 Set side right input volume. By default, this is @var{1}.
5954 Set side right output volume. By default, this is @var{1}.
5957 Set back left input volume. By default, this is @var{1}.
5960 Set back left output volume. By default, this is @var{1}.
5963 Set back right input volume. By default, this is @var{1}.
5966 Set back right output volume. By default, this is @var{1}.
5969 Set back center input volume. By default, this is @var{1}.
5972 Set back center output volume. By default, this is @var{1}.
5975 Set LFE input volume. By default, this is @var{1}.
5978 Set LFE output volume. By default, this is @var{1}.
5981 Set spread usage of stereo image across X axis for all channels.
5984 Set spread usage of stereo image across Y axis for all channels.
5986 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5987 Set spread usage of stereo image across X axis for each channel.
5989 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5990 Set spread usage of stereo image across Y axis for each channel.
5993 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5996 Set window function.
5998 It accepts the following values:
6021 Default is @code{hann}.
6024 Set window overlap. If set to 1, the recommended overlap for selected
6025 window function will be picked. Default is @code{0.5}.
6028 @section treble, highshelf
6030 Boost or cut treble (upper) frequencies of the audio using a two-pole
6031 shelving filter with a response similar to that of a standard
6032 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6034 The filter accepts the following options:
6038 Give the gain at whichever is the lower of ~22 kHz and the
6039 Nyquist frequency. Its useful range is about -20 (for a large cut)
6040 to +20 (for a large boost). Beware of clipping when using a positive gain.
6043 Set the filter's central frequency and so can be used
6044 to extend or reduce the frequency range to be boosted or cut.
6045 The default value is @code{3000} Hz.
6048 Set method to specify band-width of filter.
6063 Determine how steep is the filter's shelf transition.
6066 Set number of poles. Default is 2.
6069 How much to use filtered signal in output. Default is 1.
6070 Range is between 0 and 1.
6073 Specify which channels to filter, by default all available are filtered.
6076 Normalize biquad coefficients, by default is disabled.
6077 Enabling it will normalize magnitude response at DC to 0dB.
6080 Set transform type of IIR filter.
6089 Set precison of filtering.
6092 Pick automatic sample format depending on surround filters.
6094 Always use signed 16-bit.
6096 Always use signed 32-bit.
6098 Always use float 32-bit.
6100 Always use float 64-bit.
6104 @subsection Commands
6106 This filter supports the following commands:
6109 Change treble frequency.
6110 Syntax for the command is : "@var{frequency}"
6113 Change treble width_type.
6114 Syntax for the command is : "@var{width_type}"
6117 Change treble width.
6118 Syntax for the command is : "@var{width}"
6122 Syntax for the command is : "@var{gain}"
6126 Syntax for the command is : "@var{mix}"
6131 Sinusoidal amplitude modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6138 (20 Hz or lower) will result in a tremolo effect.
6139 This filter may also be used as a ring modulator by specifying
6140 a modulation frequency higher than 20 Hz.
6141 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6144 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6145 Default value is 0.5.
6150 Sinusoidal phase modulation.
6152 The filter accepts the following options:
6156 Modulation frequency in Hertz.
6157 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6160 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6161 Default value is 0.5.
6166 Adjust the input audio volume.
6168 It accepts the following parameters:
6172 Set audio volume expression.
6174 Output values are clipped to the maximum value.
6176 The output audio volume is given by the relation:
6178 @var{output_volume} = @var{volume} * @var{input_volume}
6181 The default value for @var{volume} is "1.0".
6184 This parameter represents the mathematical precision.
6186 It determines which input sample formats will be allowed, which affects the
6187 precision of the volume scaling.
6191 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6193 32-bit floating-point; this limits input sample format to FLT. (default)
6195 64-bit floating-point; this limits input sample format to DBL.
6199 Choose the behaviour on encountering ReplayGain side data in input frames.
6203 Remove ReplayGain side data, ignoring its contents (the default).
6206 Ignore ReplayGain side data, but leave it in the frame.
6209 Prefer the track gain, if present.
6212 Prefer the album gain, if present.
6215 @item replaygain_preamp
6216 Pre-amplification gain in dB to apply to the selected replaygain gain.
6218 Default value for @var{replaygain_preamp} is 0.0.
6220 @item replaygain_noclip
6221 Prevent clipping by limiting the gain applied.
6223 Default value for @var{replaygain_noclip} is 1.
6226 Set when the volume expression is evaluated.
6228 It accepts the following values:
6231 only evaluate expression once during the filter initialization, or
6232 when the @samp{volume} command is sent
6235 evaluate expression for each incoming frame
6238 Default value is @samp{once}.
6241 The volume expression can contain the following parameters.
6245 frame number (starting at zero)
6248 @item nb_consumed_samples
6249 number of samples consumed by the filter
6251 number of samples in the current frame
6253 original frame position in the file
6259 PTS at start of stream
6261 time at start of stream
6267 last set volume value
6270 Note that when @option{eval} is set to @samp{once} only the
6271 @var{sample_rate} and @var{tb} variables are available, all other
6272 variables will evaluate to NAN.
6274 @subsection Commands
6276 This filter supports the following commands:
6279 Modify the volume expression.
6280 The command accepts the same syntax of the corresponding option.
6282 If the specified expression is not valid, it is kept at its current
6286 @subsection Examples
6290 Halve the input audio volume:
6294 volume=volume=-6.0206dB
6297 In all the above example the named key for @option{volume} can be
6298 omitted, for example like in:
6304 Increase input audio power by 6 decibels using fixed-point precision:
6306 volume=volume=6dB:precision=fixed
6310 Fade volume after time 10 with an annihilation period of 5 seconds:
6312 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6316 @section volumedetect
6318 Detect the volume of the input video.
6320 The filter has no parameters. The input is not modified. Statistics about
6321 the volume will be printed in the log when the input stream end is reached.
6323 In particular it will show the mean volume (root mean square), maximum
6324 volume (on a per-sample basis), and the beginning of a histogram of the
6325 registered volume values (from the maximum value to a cumulated 1/1000 of
6328 All volumes are in decibels relative to the maximum PCM value.
6330 @subsection Examples
6332 Here is an excerpt of the output:
6334 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6342 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6348 The mean square energy is approximately -27 dB, or 10^-2.7.
6350 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6352 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6355 In other words, raising the volume by +4 dB does not cause any clipping,
6356 raising it by +5 dB causes clipping for 6 samples, etc.
6358 @c man end AUDIO FILTERS
6360 @chapter Audio Sources
6361 @c man begin AUDIO SOURCES
6363 Below is a description of the currently available audio sources.
6367 Buffer audio frames, and make them available to the filter chain.
6369 This source is mainly intended for a programmatic use, in particular
6370 through the interface defined in @file{libavfilter/buffersrc.h}.
6372 It accepts the following parameters:
6376 The timebase which will be used for timestamps of submitted frames. It must be
6377 either a floating-point number or in @var{numerator}/@var{denominator} form.
6380 The sample rate of the incoming audio buffers.
6383 The sample format of the incoming audio buffers.
6384 Either a sample format name or its corresponding integer representation from
6385 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6387 @item channel_layout
6388 The channel layout of the incoming audio buffers.
6389 Either a channel layout name from channel_layout_map in
6390 @file{libavutil/channel_layout.c} or its corresponding integer representation
6391 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6394 The number of channels of the incoming audio buffers.
6395 If both @var{channels} and @var{channel_layout} are specified, then they
6400 @subsection Examples
6403 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6406 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6407 Since the sample format with name "s16p" corresponds to the number
6408 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6411 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6416 Generate an audio signal specified by an expression.
6418 This source accepts in input one or more expressions (one for each
6419 channel), which are evaluated and used to generate a corresponding
6422 This source accepts the following options:
6426 Set the '|'-separated expressions list for each separate channel. In case the
6427 @option{channel_layout} option is not specified, the selected channel layout
6428 depends on the number of provided expressions. Otherwise the last
6429 specified expression is applied to the remaining output channels.
6431 @item channel_layout, c
6432 Set the channel layout. The number of channels in the specified layout
6433 must be equal to the number of specified expressions.
6436 Set the minimum duration of the sourced audio. See
6437 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6438 for the accepted syntax.
6439 Note that the resulting duration may be greater than the specified
6440 duration, as the generated audio is always cut at the end of a
6443 If not specified, or the expressed duration is negative, the audio is
6444 supposed to be generated forever.
6447 Set the number of samples per channel per each output frame,
6450 @item sample_rate, s
6451 Specify the sample rate, default to 44100.
6454 Each expression in @var{exprs} can contain the following constants:
6458 number of the evaluated sample, starting from 0
6461 time of the evaluated sample expressed in seconds, starting from 0
6468 @subsection Examples
6478 Generate a sin signal with frequency of 440 Hz, set sample rate to
6481 aevalsrc="sin(440*2*PI*t):s=8000"
6485 Generate a two channels signal, specify the channel layout (Front
6486 Center + Back Center) explicitly:
6488 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6492 Generate white noise:
6494 aevalsrc="-2+random(0)"
6498 Generate an amplitude modulated signal:
6500 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6504 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6506 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6513 Generate a FIR coefficients using frequency sampling method.
6515 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6517 The filter accepts the following options:
6521 Set number of filter coefficents in output audio stream.
6522 Default value is 1025.
6525 Set frequency points from where magnitude and phase are set.
6526 This must be in non decreasing order, and first element must be 0, while last element
6527 must be 1. Elements are separated by white spaces.
6530 Set magnitude value for every frequency point set by @option{frequency}.
6531 Number of values must be same as number of frequency points.
6532 Values are separated by white spaces.
6535 Set phase value for every frequency point set by @option{frequency}.
6536 Number of values must be same as number of frequency points.
6537 Values are separated by white spaces.
6539 @item sample_rate, r
6540 Set sample rate, default is 44100.
6543 Set number of samples per each frame. Default is 1024.
6546 Set window function. Default is blackman.
6551 The null audio source, return unprocessed audio frames. It is mainly useful
6552 as a template and to be employed in analysis / debugging tools, or as
6553 the source for filters which ignore the input data (for example the sox
6556 This source accepts the following options:
6560 @item channel_layout, cl
6562 Specifies the channel layout, and can be either an integer or a string
6563 representing a channel layout. The default value of @var{channel_layout}
6566 Check the channel_layout_map definition in
6567 @file{libavutil/channel_layout.c} for the mapping between strings and
6568 channel layout values.
6570 @item sample_rate, r
6571 Specifies the sample rate, and defaults to 44100.
6574 Set the number of samples per requested frames.
6577 Set the duration of the sourced audio. See
6578 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6579 for the accepted syntax.
6581 If not specified, or the expressed duration is negative, the audio is
6582 supposed to be generated forever.
6585 @subsection Examples
6589 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6591 anullsrc=r=48000:cl=4
6595 Do the same operation with a more obvious syntax:
6597 anullsrc=r=48000:cl=mono
6601 All the parameters need to be explicitly defined.
6605 Synthesize a voice utterance using the libflite library.
6607 To enable compilation of this filter you need to configure FFmpeg with
6608 @code{--enable-libflite}.
6610 Note that versions of the flite library prior to 2.0 are not thread-safe.
6612 The filter accepts the following options:
6617 If set to 1, list the names of the available voices and exit
6618 immediately. Default value is 0.
6621 Set the maximum number of samples per frame. Default value is 512.
6624 Set the filename containing the text to speak.
6627 Set the text to speak.
6630 Set the voice to use for the speech synthesis. Default value is
6631 @code{kal}. See also the @var{list_voices} option.
6634 @subsection Examples
6638 Read from file @file{speech.txt}, and synthesize the text using the
6639 standard flite voice:
6641 flite=textfile=speech.txt
6645 Read the specified text selecting the @code{slt} voice:
6647 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6651 Input text to ffmpeg:
6653 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6657 Make @file{ffplay} speak the specified text, using @code{flite} and
6658 the @code{lavfi} device:
6660 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6664 For more information about libflite, check:
6665 @url{http://www.festvox.org/flite/}
6669 Generate a noise audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Specify the sample rate. Default value is 48000 Hz.
6678 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6682 Specify the duration of the generated audio stream. Not specifying this option
6683 results in noise with an infinite length.
6685 @item color, colour, c
6686 Specify the color of noise. Available noise colors are white, pink, brown,
6687 blue, violet and velvet. Default color is white.
6690 Specify a value used to seed the PRNG.
6693 Set the number of samples per each output frame, default is 1024.
6696 @subsection Examples
6701 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6703 anoisesrc=d=60:c=pink:r=44100:a=0.5
6709 Generate odd-tap Hilbert transform FIR coefficients.
6711 The resulting stream can be used with @ref{afir} filter for phase-shifting
6712 the signal by 90 degrees.
6714 This is used in many matrix coding schemes and for analytic signal generation.
6715 The process is often written as a multiplication by i (or j), the imaginary unit.
6717 The filter accepts the following options:
6721 @item sample_rate, s
6722 Set sample rate, default is 44100.
6725 Set length of FIR filter, default is 22051.
6728 Set number of samples per each frame.
6731 Set window function to be used when generating FIR coefficients.
6736 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6738 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6740 The filter accepts the following options:
6743 @item sample_rate, r
6744 Set sample rate, default is 44100.
6747 Set number of samples per each frame. Default is 1024.
6750 Set high-pass frequency. Default is 0.
6753 Set low-pass frequency. Default is 0.
6754 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6755 is higher than 0 then filter will create band-pass filter coefficients,
6756 otherwise band-reject filter coefficients.
6759 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6762 Set Kaiser window beta.
6765 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6768 Enable rounding, by default is disabled.
6771 Set number of taps for high-pass filter.
6774 Set number of taps for low-pass filter.
6779 Generate an audio signal made of a sine wave with amplitude 1/8.
6781 The audio signal is bit-exact.
6783 The filter accepts the following options:
6788 Set the carrier frequency. Default is 440 Hz.
6790 @item beep_factor, b
6791 Enable a periodic beep every second with frequency @var{beep_factor} times
6792 the carrier frequency. Default is 0, meaning the beep is disabled.
6794 @item sample_rate, r
6795 Specify the sample rate, default is 44100.
6798 Specify the duration of the generated audio stream.
6800 @item samples_per_frame
6801 Set the number of samples per output frame.
6803 The expression can contain the following constants:
6807 The (sequential) number of the output audio frame, starting from 0.
6810 The PTS (Presentation TimeStamp) of the output audio frame,
6811 expressed in @var{TB} units.
6814 The PTS of the output audio frame, expressed in seconds.
6817 The timebase of the output audio frames.
6820 Default is @code{1024}.
6823 @subsection Examples
6828 Generate a simple 440 Hz sine wave:
6834 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6838 sine=frequency=220:beep_factor=4:duration=5
6842 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6845 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6849 @c man end AUDIO SOURCES
6851 @chapter Audio Sinks
6852 @c man begin AUDIO SINKS
6854 Below is a description of the currently available audio sinks.
6856 @section abuffersink
6858 Buffer audio frames, and make them available to the end of filter chain.
6860 This sink is mainly intended for programmatic use, in particular
6861 through the interface defined in @file{libavfilter/buffersink.h}
6862 or the options system.
6864 It accepts a pointer to an AVABufferSinkContext structure, which
6865 defines the incoming buffers' formats, to be passed as the opaque
6866 parameter to @code{avfilter_init_filter} for initialization.
6869 Null audio sink; do absolutely nothing with the input audio. It is
6870 mainly useful as a template and for use in analysis / debugging
6873 @c man end AUDIO SINKS
6875 @chapter Video Filters
6876 @c man begin VIDEO FILTERS
6878 When you configure your FFmpeg build, you can disable any of the
6879 existing filters using @code{--disable-filters}.
6880 The configure output will show the video filters included in your
6883 Below is a description of the currently available video filters.
6887 Mark a region of interest in a video frame.
6889 The frame data is passed through unchanged, but metadata is attached
6890 to the frame indicating regions of interest which can affect the
6891 behaviour of later encoding. Multiple regions can be marked by
6892 applying the filter multiple times.
6896 Region distance in pixels from the left edge of the frame.
6898 Region distance in pixels from the top edge of the frame.
6900 Region width in pixels.
6902 Region height in pixels.
6904 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6905 and may contain the following variables:
6908 Width of the input frame.
6910 Height of the input frame.
6914 Quantisation offset to apply within the region.
6916 This must be a real value in the range -1 to +1. A value of zero
6917 indicates no quality change. A negative value asks for better quality
6918 (less quantisation), while a positive value asks for worse quality
6919 (greater quantisation).
6921 The range is calibrated so that the extreme values indicate the
6922 largest possible offset - if the rest of the frame is encoded with the
6923 worst possible quality, an offset of -1 indicates that this region
6924 should be encoded with the best possible quality anyway. Intermediate
6925 values are then interpolated in some codec-dependent way.
6927 For example, in 10-bit H.264 the quantisation parameter varies between
6928 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6929 this region should be encoded with a QP around one-tenth of the full
6930 range better than the rest of the frame. So, if most of the frame
6931 were to be encoded with a QP of around 30, this region would get a QP
6932 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6933 An extreme value of -1 would indicate that this region should be
6934 encoded with the best possible quality regardless of the treatment of
6935 the rest of the frame - that is, should be encoded at a QP of -12.
6937 If set to true, remove any existing regions of interest marked on the
6938 frame before adding the new one.
6941 @subsection Examples
6945 Mark the centre quarter of the frame as interesting.
6947 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6950 Mark the 100-pixel-wide region on the left edge of the frame as very
6951 uninteresting (to be encoded at much lower quality than the rest of
6954 addroi=0:0:100:ih:+1/5
6958 @section alphaextract
6960 Extract the alpha component from the input as a grayscale video. This
6961 is especially useful with the @var{alphamerge} filter.
6965 Add or replace the alpha component of the primary input with the
6966 grayscale value of a second input. This is intended for use with
6967 @var{alphaextract} to allow the transmission or storage of frame
6968 sequences that have alpha in a format that doesn't support an alpha
6971 For example, to reconstruct full frames from a normal YUV-encoded video
6972 and a separate video created with @var{alphaextract}, you might use:
6974 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6979 Amplify differences between current pixel and pixels of adjacent frames in
6980 same pixel location.
6982 This filter accepts the following options:
6986 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6987 For example radius of 3 will instruct filter to calculate average of 7 frames.
6990 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6993 Set threshold for difference amplification. Any difference greater or equal to
6994 this value will not alter source pixel. Default is 10.
6995 Allowed range is from 0 to 65535.
6998 Set tolerance for difference amplification. Any difference lower to
6999 this value will not alter source pixel. Default is 0.
7000 Allowed range is from 0 to 65535.
7003 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7004 This option controls maximum possible value that will decrease source pixel value.
7007 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7008 This option controls maximum possible value that will increase source pixel value.
7011 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7014 @subsection Commands
7016 This filter supports the following @ref{commands} that corresponds to option of same name:
7028 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7029 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7030 Substation Alpha) subtitles files.
7032 This filter accepts the following option in addition to the common options from
7033 the @ref{subtitles} filter:
7037 Set the shaping engine
7039 Available values are:
7042 The default libass shaping engine, which is the best available.
7044 Fast, font-agnostic shaper that can do only substitutions
7046 Slower shaper using OpenType for substitutions and positioning
7049 The default is @code{auto}.
7053 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7055 The filter accepts the following options:
7059 Set threshold A for 1st plane. Default is 0.02.
7060 Valid range is 0 to 0.3.
7063 Set threshold B for 1st plane. Default is 0.04.
7064 Valid range is 0 to 5.
7067 Set threshold A for 2nd plane. Default is 0.02.
7068 Valid range is 0 to 0.3.
7071 Set threshold B for 2nd plane. Default is 0.04.
7072 Valid range is 0 to 5.
7075 Set threshold A for 3rd plane. Default is 0.02.
7076 Valid range is 0 to 0.3.
7079 Set threshold B for 3rd plane. Default is 0.04.
7080 Valid range is 0 to 5.
7082 Threshold A is designed to react on abrupt changes in the input signal and
7083 threshold B is designed to react on continuous changes in the input signal.
7086 Set number of frames filter will use for averaging. Default is 9. Must be odd
7087 number in range [5, 129].
7090 Set what planes of frame filter will use for averaging. Default is all.
7093 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7094 Alternatively can be set to @code{s} serial.
7096 Parallel can be faster then serial, while other way around is never true.
7097 Parallel will abort early on first change being greater then thresholds, while serial
7098 will continue processing other side of frames if they are equal or below thresholds.
7103 Set sigma for 1st plane, 2nd plane or 3rd plane. Default is 32767.
7104 Valid range is from 0 to 32767.
7105 This options controls weight for each pixel in radius defined by size.
7106 Default value means every pixel have same weight.
7107 Setting this option to 0 effectively disables filtering.
7110 @subsection Commands
7111 This filter supports same @ref{commands} as options except option @code{s}.
7112 The command accepts the same syntax of the corresponding option.
7116 Apply average blur filter.
7118 The filter accepts the following options:
7122 Set horizontal radius size.
7125 Set which planes to filter. By default all planes are filtered.
7128 Set vertical radius size, if zero it will be same as @code{sizeX}.
7129 Default is @code{0}.
7132 @subsection Commands
7133 This filter supports same commands as options.
7134 The command accepts the same syntax of the corresponding option.
7136 If the specified expression is not valid, it is kept at its current
7141 Compute the bounding box for the non-black pixels in the input frame
7144 This filter computes the bounding box containing all the pixels with a
7145 luminance value greater than the minimum allowed value.
7146 The parameters describing the bounding box are printed on the filter
7149 The filter accepts the following option:
7153 Set the minimal luminance value. Default is @code{16}.
7156 @subsection Commands
7158 This filter supports the all above options as @ref{commands}.
7161 Apply bilateral filter, spatial smoothing while preserving edges.
7163 The filter accepts the following options:
7166 Set sigma of gaussian function to calculate spatial weight.
7167 Allowed range is 0 to 512. Default is 0.1.
7170 Set sigma of gaussian function to calculate range weight.
7171 Allowed range is 0 to 1. Default is 0.1.
7174 Set planes to filter. Default is first only.
7177 @subsection Commands
7179 This filter supports the all above options as @ref{commands}.
7181 @section bitplanenoise
7183 Show and measure bit plane noise.
7185 The filter accepts the following options:
7189 Set which plane to analyze. Default is @code{1}.
7192 Filter out noisy pixels from @code{bitplane} set above.
7193 Default is disabled.
7196 @section blackdetect
7198 Detect video intervals that are (almost) completely black. Can be
7199 useful to detect chapter transitions, commercials, or invalid
7202 The filter outputs its detection analysis to both the log as well as
7203 frame metadata. If a black segment of at least the specified minimum
7204 duration is found, a line with the start and end timestamps as well
7205 as duration is printed to the log with level @code{info}. In addition,
7206 a log line with level @code{debug} is printed per frame showing the
7207 black amount detected for that frame.
7209 The filter also attaches metadata to the first frame of a black
7210 segment with key @code{lavfi.black_start} and to the first frame
7211 after the black segment ends with key @code{lavfi.black_end}. The
7212 value is the frame's timestamp. This metadata is added regardless
7213 of the minimum duration specified.
7215 The filter accepts the following options:
7218 @item black_min_duration, d
7219 Set the minimum detected black duration expressed in seconds. It must
7220 be a non-negative floating point number.
7222 Default value is 2.0.
7224 @item picture_black_ratio_th, pic_th
7225 Set the threshold for considering a picture "black".
7226 Express the minimum value for the ratio:
7228 @var{nb_black_pixels} / @var{nb_pixels}
7231 for which a picture is considered black.
7232 Default value is 0.98.
7234 @item pixel_black_th, pix_th
7235 Set the threshold for considering a pixel "black".
7237 The threshold expresses the maximum pixel luminance value for which a
7238 pixel is considered "black". The provided value is scaled according to
7239 the following equation:
7241 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7244 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7245 the input video format, the range is [0-255] for YUV full-range
7246 formats and [16-235] for YUV non full-range formats.
7248 Default value is 0.10.
7251 The following example sets the maximum pixel threshold to the minimum
7252 value, and detects only black intervals of 2 or more seconds:
7254 blackdetect=d=2:pix_th=0.00
7259 Detect frames that are (almost) completely black. Can be useful to
7260 detect chapter transitions or commercials. Output lines consist of
7261 the frame number of the detected frame, the percentage of blackness,
7262 the position in the file if known or -1 and the timestamp in seconds.
7264 In order to display the output lines, you need to set the loglevel at
7265 least to the AV_LOG_INFO value.
7267 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7268 The value represents the percentage of pixels in the picture that
7269 are below the threshold value.
7271 It accepts the following parameters:
7276 The percentage of the pixels that have to be below the threshold; it defaults to
7279 @item threshold, thresh
7280 The threshold below which a pixel value is considered black; it defaults to
7288 Blend two video frames into each other.
7290 The @code{blend} filter takes two input streams and outputs one
7291 stream, the first input is the "top" layer and second input is
7292 "bottom" layer. By default, the output terminates when the longest input terminates.
7294 The @code{tblend} (time blend) filter takes two consecutive frames
7295 from one single stream, and outputs the result obtained by blending
7296 the new frame on top of the old frame.
7298 A description of the accepted options follows.
7306 Set blend mode for specific pixel component or all pixel components in case
7307 of @var{all_mode}. Default value is @code{normal}.
7309 Available values for component modes are:
7351 Set blend opacity for specific pixel component or all pixel components in case
7352 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7359 Set blend expression for specific pixel component or all pixel components in case
7360 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7362 The expressions can use the following variables:
7366 The sequential number of the filtered frame, starting from @code{0}.
7370 the coordinates of the current sample
7374 the width and height of currently filtered plane
7378 Width and height scale for the plane being filtered. It is the
7379 ratio between the dimensions of the current plane to the luma plane,
7380 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7381 the luma plane and @code{0.5,0.5} for the chroma planes.
7384 Time of the current frame, expressed in seconds.
7387 Value of pixel component at current location for first video frame (top layer).
7390 Value of pixel component at current location for second video frame (bottom layer).
7394 The @code{blend} filter also supports the @ref{framesync} options.
7396 @subsection Examples
7400 Apply transition from bottom layer to top layer in first 10 seconds:
7402 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7406 Apply linear horizontal transition from top layer to bottom layer:
7408 blend=all_expr='A*(X/W)+B*(1-X/W)'
7412 Apply 1x1 checkerboard effect:
7414 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7418 Apply uncover left effect:
7420 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7424 Apply uncover down effect:
7426 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7430 Apply uncover up-left effect:
7432 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7436 Split diagonally video and shows top and bottom layer on each side:
7438 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7442 Display differences between the current and the previous frame:
7444 tblend=all_mode=grainextract
7450 Denoise frames using Block-Matching 3D algorithm.
7452 The filter accepts the following options.
7456 Set denoising strength. Default value is 1.
7457 Allowed range is from 0 to 999.9.
7458 The denoising algorithm is very sensitive to sigma, so adjust it
7459 according to the source.
7462 Set local patch size. This sets dimensions in 2D.
7465 Set sliding step for processing blocks. Default value is 4.
7466 Allowed range is from 1 to 64.
7467 Smaller values allows processing more reference blocks and is slower.
7470 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7471 When set to 1, no block matching is done. Larger values allows more blocks
7473 Allowed range is from 1 to 256.
7476 Set radius for search block matching. Default is 9.
7477 Allowed range is from 1 to INT32_MAX.
7480 Set step between two search locations for block matching. Default is 1.
7481 Allowed range is from 1 to 64. Smaller is slower.
7484 Set threshold of mean square error for block matching. Valid range is 0 to
7488 Set thresholding parameter for hard thresholding in 3D transformed domain.
7489 Larger values results in stronger hard-thresholding filtering in frequency
7493 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7494 Default is @code{basic}.
7497 If enabled, filter will use 2nd stream for block matching.
7498 Default is disabled for @code{basic} value of @var{estim} option,
7499 and always enabled if value of @var{estim} is @code{final}.
7502 Set planes to filter. Default is all available except alpha.
7505 @subsection Examples
7509 Basic filtering with bm3d:
7511 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7515 Same as above, but filtering only luma:
7517 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7521 Same as above, but with both estimation modes:
7523 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
7527 Same as above, but prefilter with @ref{nlmeans} filter instead:
7529 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
7535 Apply a boxblur algorithm to the input video.
7537 It accepts the following parameters:
7541 @item luma_radius, lr
7542 @item luma_power, lp
7543 @item chroma_radius, cr
7544 @item chroma_power, cp
7545 @item alpha_radius, ar
7546 @item alpha_power, ap
7550 A description of the accepted options follows.
7553 @item luma_radius, lr
7554 @item chroma_radius, cr
7555 @item alpha_radius, ar
7556 Set an expression for the box radius in pixels used for blurring the
7557 corresponding input plane.
7559 The radius value must be a non-negative number, and must not be
7560 greater than the value of the expression @code{min(w,h)/2} for the
7561 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7564 Default value for @option{luma_radius} is "2". If not specified,
7565 @option{chroma_radius} and @option{alpha_radius} default to the
7566 corresponding value set for @option{luma_radius}.
7568 The expressions can contain the following constants:
7572 The input width and height in pixels.
7576 The input chroma image width and height in pixels.
7580 The horizontal and vertical chroma subsample values. For example, for the
7581 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7584 @item luma_power, lp
7585 @item chroma_power, cp
7586 @item alpha_power, ap
7587 Specify how many times the boxblur filter is applied to the
7588 corresponding plane.
7590 Default value for @option{luma_power} is 2. If not specified,
7591 @option{chroma_power} and @option{alpha_power} default to the
7592 corresponding value set for @option{luma_power}.
7594 A value of 0 will disable the effect.
7597 @subsection Examples
7601 Apply a boxblur filter with the luma, chroma, and alpha radii
7604 boxblur=luma_radius=2:luma_power=1
7609 Set the luma radius to 2, and alpha and chroma radius to 0:
7611 boxblur=2:1:cr=0:ar=0
7615 Set the luma and chroma radii to a fraction of the video dimension:
7617 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7623 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7624 Deinterlacing Filter").
7626 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7627 interpolation algorithms.
7628 It accepts the following parameters:
7632 The interlacing mode to adopt. It accepts one of the following values:
7636 Output one frame for each frame.
7638 Output one frame for each field.
7641 The default value is @code{send_field}.
7644 The picture field parity assumed for the input interlaced video. It accepts one
7645 of the following values:
7649 Assume the top field is first.
7651 Assume the bottom field is first.
7653 Enable automatic detection of field parity.
7656 The default value is @code{auto}.
7657 If the interlacing is unknown or the decoder does not export this information,
7658 top field first will be assumed.
7661 Specify which frames to deinterlace. Accepts one of the following
7666 Deinterlace all frames.
7668 Only deinterlace frames marked as interlaced.
7671 The default value is @code{all}.
7676 Apply Contrast Adaptive Sharpen filter to video stream.
7678 The filter accepts the following options:
7682 Set the sharpening strength. Default value is 0.
7685 Set planes to filter. Default value is to filter all
7686 planes except alpha plane.
7689 @subsection Commands
7690 This filter supports same @ref{commands} as options.
7693 Remove all color information for all colors except for certain one.
7695 The filter accepts the following options:
7699 The color which will not be replaced with neutral chroma.
7702 Similarity percentage with the above color.
7703 0.01 matches only the exact key color, while 1.0 matches everything.
7707 0.0 makes pixels either fully gray, or not gray at all.
7708 Higher values result in more preserved color.
7711 Signals that the color passed is already in YUV instead of RGB.
7713 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7714 This can be used to pass exact YUV values as hexadecimal numbers.
7717 @subsection Commands
7718 This filter supports same @ref{commands} as options.
7719 The command accepts the same syntax of the corresponding option.
7721 If the specified expression is not valid, it is kept at its current
7725 YUV colorspace color/chroma keying.
7727 The filter accepts the following options:
7731 The color which will be replaced with transparency.
7734 Similarity percentage with the key color.
7736 0.01 matches only the exact key color, while 1.0 matches everything.
7741 0.0 makes pixels either fully transparent, or not transparent at all.
7743 Higher values result in semi-transparent pixels, with a higher transparency
7744 the more similar the pixels color is to the key color.
7747 Signals that the color passed is already in YUV instead of RGB.
7749 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7750 This can be used to pass exact YUV values as hexadecimal numbers.
7753 @subsection Commands
7754 This filter supports same @ref{commands} as options.
7755 The command accepts the same syntax of the corresponding option.
7757 If the specified expression is not valid, it is kept at its current
7760 @subsection Examples
7764 Make every green pixel in the input image transparent:
7766 ffmpeg -i input.png -vf chromakey=green out.png
7770 Overlay a greenscreen-video on top of a static black background.
7772 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
7777 Reduce chrominance noise.
7779 The filter accepts the following options:
7783 Set threshold for averaging chrominance values.
7784 Sum of absolute difference of Y, U and V pixel components of current
7785 pixel and neighbour pixels lower than this threshold will be used in
7786 averaging. Luma component is left unchanged and is copied to output.
7787 Default value is 30. Allowed range is from 1 to 200.
7790 Set horizontal radius of rectangle used for averaging.
7791 Allowed range is from 1 to 100. Default value is 5.
7794 Set vertical radius of rectangle used for averaging.
7795 Allowed range is from 1 to 100. Default value is 5.
7798 Set horizontal step when averaging. Default value is 1.
7799 Allowed range is from 1 to 50.
7800 Mostly useful to speed-up filtering.
7803 Set vertical step when averaging. Default value is 1.
7804 Allowed range is from 1 to 50.
7805 Mostly useful to speed-up filtering.
7808 Set Y threshold for averaging chrominance values.
7809 Set finer control for max allowed difference between Y components
7810 of current pixel and neigbour pixels.
7811 Default value is 200. Allowed range is from 1 to 200.
7814 Set U threshold for averaging chrominance values.
7815 Set finer control for max allowed difference between U components
7816 of current pixel and neigbour pixels.
7817 Default value is 200. Allowed range is from 1 to 200.
7820 Set V threshold for averaging chrominance values.
7821 Set finer control for max allowed difference between V components
7822 of current pixel and neigbour pixels.
7823 Default value is 200. Allowed range is from 1 to 200.
7826 @subsection Commands
7827 This filter supports same @ref{commands} as options.
7828 The command accepts the same syntax of the corresponding option.
7830 @section chromashift
7831 Shift chroma pixels horizontally and/or vertically.
7833 The filter accepts the following options:
7836 Set amount to shift chroma-blue horizontally.
7838 Set amount to shift chroma-blue vertically.
7840 Set amount to shift chroma-red horizontally.
7842 Set amount to shift chroma-red vertically.
7844 Set edge mode, can be @var{smear}, default, or @var{warp}.
7847 @subsection Commands
7849 This filter supports the all above options as @ref{commands}.
7853 Display CIE color diagram with pixels overlaid onto it.
7855 The filter accepts the following options:
7870 @item uhdtv, rec2020
7884 Set what gamuts to draw.
7886 See @code{system} option for available values.
7889 Set ciescope size, by default set to 512.
7892 Set intensity used to map input pixel values to CIE diagram.
7895 Set contrast used to draw tongue colors that are out of active color system gamut.
7898 Correct gamma displayed on scope, by default enabled.
7901 Show white point on CIE diagram, by default disabled.
7904 Set input gamma. Used only with XYZ input color space.
7909 Visualize information exported by some codecs.
7911 Some codecs can export information through frames using side-data or other
7912 means. For example, some MPEG based codecs export motion vectors through the
7913 @var{export_mvs} flag in the codec @option{flags2} option.
7915 The filter accepts the following option:
7919 Set motion vectors to visualize.
7921 Available flags for @var{mv} are:
7925 forward predicted MVs of P-frames
7927 forward predicted MVs of B-frames
7929 backward predicted MVs of B-frames
7933 Display quantization parameters using the chroma planes.
7936 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7938 Available flags for @var{mv_type} are:
7942 forward predicted MVs
7944 backward predicted MVs
7947 @item frame_type, ft
7948 Set frame type to visualize motion vectors of.
7950 Available flags for @var{frame_type} are:
7954 intra-coded frames (I-frames)
7956 predicted frames (P-frames)
7958 bi-directionally predicted frames (B-frames)
7962 @subsection Examples
7966 Visualize forward predicted MVs of all frames using @command{ffplay}:
7968 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7972 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7974 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7978 @section colorbalance
7979 Modify intensity of primary colors (red, green and blue) of input frames.
7981 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7982 regions for the red-cyan, green-magenta or blue-yellow balance.
7984 A positive adjustment value shifts the balance towards the primary color, a negative
7985 value towards the complementary color.
7987 The filter accepts the following options:
7993 Adjust red, green and blue shadows (darkest pixels).
7998 Adjust red, green and blue midtones (medium pixels).
8003 Adjust red, green and blue highlights (brightest pixels).
8005 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8008 Preserve lightness when changing color balance. Default is disabled.
8011 @subsection Examples
8015 Add red color cast to shadows:
8021 @subsection Commands
8023 This filter supports the all above options as @ref{commands}.
8025 @section colorchannelmixer
8027 Adjust video input frames by re-mixing color channels.
8029 This filter modifies a color channel by adding the values associated to
8030 the other channels of the same pixels. For example if the value to
8031 modify is red, the output value will be:
8033 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8036 The filter accepts the following options:
8043 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8044 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8050 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8051 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8057 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8058 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8064 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8065 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8067 Allowed ranges for options are @code{[-2.0, 2.0]}.
8070 Preserve lightness when changing colors. Default is disabled.
8073 @subsection Examples
8077 Convert source to grayscale:
8079 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8082 Simulate sepia tones:
8084 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8088 @subsection Commands
8090 This filter supports the all above options as @ref{commands}.
8093 RGB colorspace color keying.
8095 The filter accepts the following options:
8099 The color which will be replaced with transparency.
8102 Similarity percentage with the key color.
8104 0.01 matches only the exact key color, while 1.0 matches everything.
8109 0.0 makes pixels either fully transparent, or not transparent at all.
8111 Higher values result in semi-transparent pixels, with a higher transparency
8112 the more similar the pixels color is to the key color.
8115 @subsection Examples
8119 Make every green pixel in the input image transparent:
8121 ffmpeg -i input.png -vf colorkey=green out.png
8125 Overlay a greenscreen-video on top of a static background image.
8127 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
8131 @subsection Commands
8132 This filter supports same @ref{commands} as options.
8133 The command accepts the same syntax of the corresponding option.
8135 If the specified expression is not valid, it is kept at its current
8139 Remove all color information for all RGB colors except for certain one.
8141 The filter accepts the following options:
8145 The color which will not be replaced with neutral gray.
8148 Similarity percentage with the above color.
8149 0.01 matches only the exact key color, while 1.0 matches everything.
8152 Blend percentage. 0.0 makes pixels fully gray.
8153 Higher values result in more preserved color.
8156 @subsection Commands
8157 This filter supports same @ref{commands} as options.
8158 The command accepts the same syntax of the corresponding option.
8160 If the specified expression is not valid, it is kept at its current
8163 @section colorlevels
8165 Adjust video input frames using levels.
8167 The filter accepts the following options:
8174 Adjust red, green, blue and alpha input black point.
8175 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8181 Adjust red, green, blue and alpha input white point.
8182 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8184 Input levels are used to lighten highlights (bright tones), darken shadows
8185 (dark tones), change the balance of bright and dark tones.
8191 Adjust red, green, blue and alpha output black point.
8192 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8198 Adjust red, green, blue and alpha output white point.
8199 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8201 Output levels allows manual selection of a constrained output level range.
8204 @subsection Examples
8208 Make video output darker:
8210 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8216 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8220 Make video output lighter:
8222 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8226 Increase brightness:
8228 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8232 @subsection Commands
8234 This filter supports the all above options as @ref{commands}.
8236 @section colormatrix
8238 Convert color matrix.
8240 The filter accepts the following options:
8245 Specify the source and destination color matrix. Both values must be
8248 The accepted values are:
8276 For example to convert from BT.601 to SMPTE-240M, use the command:
8278 colormatrix=bt601:smpte240m
8283 Convert colorspace, transfer characteristics or color primaries.
8284 Input video needs to have an even size.
8286 The filter accepts the following options:
8291 Specify all color properties at once.
8293 The accepted values are:
8323 Specify output colorspace.
8325 The accepted values are:
8334 BT.470BG or BT.601-6 625
8337 SMPTE-170M or BT.601-6 525
8346 BT.2020 with non-constant luminance
8352 Specify output transfer characteristics.
8354 The accepted values are:
8366 Constant gamma of 2.2
8369 Constant gamma of 2.8
8372 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8390 BT.2020 for 10-bits content
8393 BT.2020 for 12-bits content
8399 Specify output color primaries.
8401 The accepted values are:
8410 BT.470BG or BT.601-6 625
8413 SMPTE-170M or BT.601-6 525
8437 Specify output color range.
8439 The accepted values are:
8442 TV (restricted) range
8445 MPEG (restricted) range
8456 Specify output color format.
8458 The accepted values are:
8461 YUV 4:2:0 planar 8-bits
8464 YUV 4:2:0 planar 10-bits
8467 YUV 4:2:0 planar 12-bits
8470 YUV 4:2:2 planar 8-bits
8473 YUV 4:2:2 planar 10-bits
8476 YUV 4:2:2 planar 12-bits
8479 YUV 4:4:4 planar 8-bits
8482 YUV 4:4:4 planar 10-bits
8485 YUV 4:4:4 planar 12-bits
8490 Do a fast conversion, which skips gamma/primary correction. This will take
8491 significantly less CPU, but will be mathematically incorrect. To get output
8492 compatible with that produced by the colormatrix filter, use fast=1.
8495 Specify dithering mode.
8497 The accepted values are:
8503 Floyd-Steinberg dithering
8507 Whitepoint adaptation mode.
8509 The accepted values are:
8512 Bradford whitepoint adaptation
8515 von Kries whitepoint adaptation
8518 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8522 Override all input properties at once. Same accepted values as @ref{all}.
8525 Override input colorspace. Same accepted values as @ref{space}.
8528 Override input color primaries. Same accepted values as @ref{primaries}.
8531 Override input transfer characteristics. Same accepted values as @ref{trc}.
8534 Override input color range. Same accepted values as @ref{range}.
8538 The filter converts the transfer characteristics, color space and color
8539 primaries to the specified user values. The output value, if not specified,
8540 is set to a default value based on the "all" property. If that property is
8541 also not specified, the filter will log an error. The output color range and
8542 format default to the same value as the input color range and format. The
8543 input transfer characteristics, color space, color primaries and color range
8544 should be set on the input data. If any of these are missing, the filter will
8545 log an error and no conversion will take place.
8547 For example to convert the input to SMPTE-240M, use the command:
8549 colorspace=smpte240m
8552 @section convolution
8554 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8556 The filter accepts the following options:
8563 Set matrix for each plane.
8564 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8565 and from 1 to 49 odd number of signed integers in @var{row} mode.
8571 Set multiplier for calculated value for each plane.
8572 If unset or 0, it will be sum of all matrix elements.
8578 Set bias for each plane. This value is added to the result of the multiplication.
8579 Useful for making the overall image brighter or darker. Default is 0.0.
8585 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8586 Default is @var{square}.
8589 @subsection Commands
8591 This filter supports the all above options as @ref{commands}.
8593 @subsection Examples
8599 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"
8605 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"
8611 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"
8617 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"
8621 Apply laplacian edge detector which includes diagonals:
8623 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"
8629 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"
8635 Apply 2D convolution of video stream in frequency domain using second stream
8638 The filter accepts the following options:
8642 Set which planes to process.
8645 Set which impulse video frames will be processed, can be @var{first}
8646 or @var{all}. Default is @var{all}.
8649 The @code{convolve} filter also supports the @ref{framesync} options.
8653 Copy the input video source unchanged to the output. This is mainly useful for
8658 Video filtering on GPU using Apple's CoreImage API on OSX.
8660 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8661 processed by video hardware. However, software-based OpenGL implementations
8662 exist which means there is no guarantee for hardware processing. It depends on
8665 There are many filters and image generators provided by Apple that come with a
8666 large variety of options. The filter has to be referenced by its name along
8669 The coreimage filter accepts the following options:
8672 List all available filters and generators along with all their respective
8673 options as well as possible minimum and maximum values along with the default
8680 Specify all filters by their respective name and options.
8681 Use @var{list_filters} to determine all valid filter names and options.
8682 Numerical options are specified by a float value and are automatically clamped
8683 to their respective value range. Vector and color options have to be specified
8684 by a list of space separated float values. Character escaping has to be done.
8685 A special option name @code{default} is available to use default options for a
8688 It is required to specify either @code{default} or at least one of the filter options.
8689 All omitted options are used with their default values.
8690 The syntax of the filter string is as follows:
8692 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8696 Specify a rectangle where the output of the filter chain is copied into the
8697 input image. It is given by a list of space separated float values:
8699 output_rect=x\ y\ width\ height
8701 If not given, the output rectangle equals the dimensions of the input image.
8702 The output rectangle is automatically cropped at the borders of the input
8703 image. Negative values are valid for each component.
8705 output_rect=25\ 25\ 100\ 100
8709 Several filters can be chained for successive processing without GPU-HOST
8710 transfers allowing for fast processing of complex filter chains.
8711 Currently, only filters with zero (generators) or exactly one (filters) input
8712 image and one output image are supported. Also, transition filters are not yet
8715 Some filters generate output images with additional padding depending on the
8716 respective filter kernel. The padding is automatically removed to ensure the
8717 filter output has the same size as the input image.
8719 For image generators, the size of the output image is determined by the
8720 previous output image of the filter chain or the input image of the whole
8721 filterchain, respectively. The generators do not use the pixel information of
8722 this image to generate their output. However, the generated output is
8723 blended onto this image, resulting in partial or complete coverage of the
8726 The @ref{coreimagesrc} video source can be used for generating input images
8727 which are directly fed into the filter chain. By using it, providing input
8728 images by another video source or an input video is not required.
8730 @subsection Examples
8735 List all filters available:
8737 coreimage=list_filters=true
8741 Use the CIBoxBlur filter with default options to blur an image:
8743 coreimage=filter=CIBoxBlur@@default
8747 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8748 its center at 100x100 and a radius of 50 pixels:
8750 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8754 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8755 given as complete and escaped command-line for Apple's standard bash shell:
8757 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8763 Cover a rectangular object
8765 It accepts the following options:
8769 Filepath of the optional cover image, needs to be in yuv420.
8774 It accepts the following values:
8777 cover it by the supplied image
8779 cover it by interpolating the surrounding pixels
8782 Default value is @var{blur}.
8785 @subsection Examples
8789 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8791 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8797 Crop the input video to given dimensions.
8799 It accepts the following parameters:
8803 The width of the output video. It defaults to @code{iw}.
8804 This expression is evaluated only once during the filter
8805 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8808 The height of the output video. It defaults to @code{ih}.
8809 This expression is evaluated only once during the filter
8810 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8813 The horizontal position, in the input video, of the left edge of the output
8814 video. It defaults to @code{(in_w-out_w)/2}.
8815 This expression is evaluated per-frame.
8818 The vertical position, in the input video, of the top edge of the output video.
8819 It defaults to @code{(in_h-out_h)/2}.
8820 This expression is evaluated per-frame.
8823 If set to 1 will force the output display aspect ratio
8824 to be the same of the input, by changing the output sample aspect
8825 ratio. It defaults to 0.
8828 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8829 width/height/x/y as specified and will not be rounded to nearest smaller value.
8833 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8834 expressions containing the following constants:
8839 The computed values for @var{x} and @var{y}. They are evaluated for
8844 The input width and height.
8848 These are the same as @var{in_w} and @var{in_h}.
8852 The output (cropped) width and height.
8856 These are the same as @var{out_w} and @var{out_h}.
8859 same as @var{iw} / @var{ih}
8862 input sample aspect ratio
8865 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8869 horizontal and vertical chroma subsample values. For example for the
8870 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8873 The number of the input frame, starting from 0.
8876 the position in the file of the input frame, NAN if unknown
8879 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8883 The expression for @var{out_w} may depend on the value of @var{out_h},
8884 and the expression for @var{out_h} may depend on @var{out_w}, but they
8885 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8886 evaluated after @var{out_w} and @var{out_h}.
8888 The @var{x} and @var{y} parameters specify the expressions for the
8889 position of the top-left corner of the output (non-cropped) area. They
8890 are evaluated for each frame. If the evaluated value is not valid, it
8891 is approximated to the nearest valid value.
8893 The expression for @var{x} may depend on @var{y}, and the expression
8894 for @var{y} may depend on @var{x}.
8896 @subsection Examples
8900 Crop area with size 100x100 at position (12,34).
8905 Using named options, the example above becomes:
8907 crop=w=100:h=100:x=12:y=34
8911 Crop the central input area with size 100x100:
8917 Crop the central input area with size 2/3 of the input video:
8919 crop=2/3*in_w:2/3*in_h
8923 Crop the input video central square:
8930 Delimit the rectangle with the top-left corner placed at position
8931 100:100 and the right-bottom corner corresponding to the right-bottom
8932 corner of the input image.
8934 crop=in_w-100:in_h-100:100:100
8938 Crop 10 pixels from the left and right borders, and 20 pixels from
8939 the top and bottom borders
8941 crop=in_w-2*10:in_h-2*20
8945 Keep only the bottom right quarter of the input image:
8947 crop=in_w/2:in_h/2:in_w/2:in_h/2
8951 Crop height for getting Greek harmony:
8953 crop=in_w:1/PHI*in_w
8957 Apply trembling effect:
8959 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)
8963 Apply erratic camera effect depending on timestamp:
8965 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)"
8969 Set x depending on the value of y:
8971 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8975 @subsection Commands
8977 This filter supports the following commands:
8983 Set width/height of the output video and the horizontal/vertical position
8985 The command accepts the same syntax of the corresponding option.
8987 If the specified expression is not valid, it is kept at its current
8993 Auto-detect the crop size.
8995 It calculates the necessary cropping parameters and prints the
8996 recommended parameters via the logging system. The detected dimensions
8997 correspond to the non-black area of the input video.
8999 It accepts the following parameters:
9004 Set higher black value threshold, which can be optionally specified
9005 from nothing (0) to everything (255 for 8-bit based formats). An intensity
9006 value greater to the set value is considered non-black. It defaults to 24.
9007 You can also specify a value between 0.0 and 1.0 which will be scaled depending
9008 on the bitdepth of the pixel format.
9011 The value which the width/height should be divisible by. It defaults to
9012 16. The offset is automatically adjusted to center the video. Use 2 to
9013 get only even dimensions (needed for 4:2:2 video). 16 is best when
9014 encoding to most video codecs.
9017 Set the number of initial frames for which evaluation is skipped.
9018 Default is 2. Range is 0 to INT_MAX.
9020 @item reset_count, reset
9021 Set the counter that determines after how many frames cropdetect will
9022 reset the previously detected largest video area and start over to
9023 detect the current optimal crop area. Default value is 0.
9025 This can be useful when channel logos distort the video area. 0
9026 indicates 'never reset', and returns the largest area encountered during
9033 Delay video filtering until a given wallclock timestamp. The filter first
9034 passes on @option{preroll} amount of frames, then it buffers at most
9035 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9036 it forwards the buffered frames and also any subsequent frames coming in its
9039 The filter can be used synchronize the output of multiple ffmpeg processes for
9040 realtime output devices like decklink. By putting the delay in the filtering
9041 chain and pre-buffering frames the process can pass on data to output almost
9042 immediately after the target wallclock timestamp is reached.
9044 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9050 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9053 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9056 The maximum duration of content to buffer before waiting for the cue expressed
9057 in seconds. Default is 0.
9064 Apply color adjustments using curves.
9066 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9067 component (red, green and blue) has its values defined by @var{N} key points
9068 tied from each other using a smooth curve. The x-axis represents the pixel
9069 values from the input frame, and the y-axis the new pixel values to be set for
9072 By default, a component curve is defined by the two points @var{(0;0)} and
9073 @var{(1;1)}. This creates a straight line where each original pixel value is
9074 "adjusted" to its own value, which means no change to the image.
9076 The filter allows you to redefine these two points and add some more. A new
9077 curve (using a natural cubic spline interpolation) will be define to pass
9078 smoothly through all these new coordinates. The new defined points needs to be
9079 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9080 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9081 the vector spaces, the values will be clipped accordingly.
9083 The filter accepts the following options:
9087 Select one of the available color presets. This option can be used in addition
9088 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9089 options takes priority on the preset values.
9090 Available presets are:
9093 @item color_negative
9096 @item increase_contrast
9098 @item linear_contrast
9099 @item medium_contrast
9101 @item strong_contrast
9104 Default is @code{none}.
9106 Set the master key points. These points will define a second pass mapping. It
9107 is sometimes called a "luminance" or "value" mapping. It can be used with
9108 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9109 post-processing LUT.
9111 Set the key points for the red component.
9113 Set the key points for the green component.
9115 Set the key points for the blue component.
9117 Set the key points for all components (not including master).
9118 Can be used in addition to the other key points component
9119 options. In this case, the unset component(s) will fallback on this
9120 @option{all} setting.
9122 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9124 Save Gnuplot script of the curves in specified file.
9127 To avoid some filtergraph syntax conflicts, each key points list need to be
9128 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9130 @subsection Examples
9134 Increase slightly the middle level of blue:
9136 curves=blue='0/0 0.5/0.58 1/1'
9142 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'
9144 Here we obtain the following coordinates for each components:
9147 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9149 @code{(0;0) (0.50;0.48) (1;1)}
9151 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9155 The previous example can also be achieved with the associated built-in preset:
9157 curves=preset=vintage
9167 Use a Photoshop preset and redefine the points of the green component:
9169 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9173 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9174 and @command{gnuplot}:
9176 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9177 gnuplot -p /tmp/curves.plt
9183 Video data analysis filter.
9185 This filter shows hexadecimal pixel values of part of video.
9187 The filter accepts the following options:
9191 Set output video size.
9194 Set x offset from where to pick pixels.
9197 Set y offset from where to pick pixels.
9200 Set scope mode, can be one of the following:
9203 Draw hexadecimal pixel values with white color on black background.
9206 Draw hexadecimal pixel values with input video pixel color on black
9210 Draw hexadecimal pixel values on color background picked from input video,
9211 the text color is picked in such way so its always visible.
9215 Draw rows and columns numbers on left and top of video.
9218 Set background opacity.
9221 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9224 Set pixel components to display. By default all pixel components are displayed.
9228 Apply Directional blur filter.
9230 The filter accepts the following options:
9234 Set angle of directional blur. Default is @code{45}.
9237 Set radius of directional blur. Default is @code{5}.
9240 Set which planes to filter. By default all planes are filtered.
9243 @subsection Commands
9244 This filter supports same @ref{commands} as options.
9245 The command accepts the same syntax of the corresponding option.
9247 If the specified expression is not valid, it is kept at its current
9252 Denoise frames using 2D DCT (frequency domain filtering).
9254 This filter is not designed for real time.
9256 The filter accepts the following options:
9260 Set the noise sigma constant.
9262 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9263 coefficient (absolute value) below this threshold with be dropped.
9265 If you need a more advanced filtering, see @option{expr}.
9267 Default is @code{0}.
9270 Set number overlapping pixels for each block. Since the filter can be slow, you
9271 may want to reduce this value, at the cost of a less effective filter and the
9272 risk of various artefacts.
9274 If the overlapping value doesn't permit processing the whole input width or
9275 height, a warning will be displayed and according borders won't be denoised.
9277 Default value is @var{blocksize}-1, which is the best possible setting.
9280 Set the coefficient factor expression.
9282 For each coefficient of a DCT block, this expression will be evaluated as a
9283 multiplier value for the coefficient.
9285 If this is option is set, the @option{sigma} option will be ignored.
9287 The absolute value of the coefficient can be accessed through the @var{c}
9291 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9292 @var{blocksize}, which is the width and height of the processed blocks.
9294 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9295 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9296 on the speed processing. Also, a larger block size does not necessarily means a
9300 @subsection Examples
9302 Apply a denoise with a @option{sigma} of @code{4.5}:
9307 The same operation can be achieved using the expression system:
9309 dctdnoiz=e='gte(c, 4.5*3)'
9312 Violent denoise using a block size of @code{16x16}:
9319 Remove banding artifacts from input video.
9320 It works by replacing banded pixels with average value of referenced pixels.
9322 The filter accepts the following options:
9329 Set banding detection threshold for each plane. Default is 0.02.
9330 Valid range is 0.00003 to 0.5.
9331 If difference between current pixel and reference pixel is less than threshold,
9332 it will be considered as banded.
9335 Banding detection range in pixels. Default is 16. If positive, random number
9336 in range 0 to set value will be used. If negative, exact absolute value
9338 The range defines square of four pixels around current pixel.
9341 Set direction in radians from which four pixel will be compared. If positive,
9342 random direction from 0 to set direction will be picked. If negative, exact of
9343 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9344 will pick only pixels on same row and -PI/2 will pick only pixels on same
9348 If enabled, current pixel is compared with average value of all four
9349 surrounding pixels. The default is enabled. If disabled current pixel is
9350 compared with all four surrounding pixels. The pixel is considered banded
9351 if only all four differences with surrounding pixels are less than threshold.
9354 If enabled, current pixel is changed if and only if all pixel components are banded,
9355 e.g. banding detection threshold is triggered for all color components.
9356 The default is disabled.
9361 Remove blocking artifacts from input video.
9363 The filter accepts the following options:
9367 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9368 This controls what kind of deblocking is applied.
9371 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9377 Set blocking detection thresholds. Allowed range is 0 to 1.
9378 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9379 Using higher threshold gives more deblocking strength.
9380 Setting @var{alpha} controls threshold detection at exact edge of block.
9381 Remaining options controls threshold detection near the edge. Each one for
9382 below/above or left/right. Setting any of those to @var{0} disables
9386 Set planes to filter. Default is to filter all available planes.
9389 @subsection Examples
9393 Deblock using weak filter and block size of 4 pixels.
9395 deblock=filter=weak:block=4
9399 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9400 deblocking more edges.
9402 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9406 Similar as above, but filter only first plane.
9408 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9412 Similar as above, but filter only second and third plane.
9414 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9421 Drop duplicated frames at regular intervals.
9423 The filter accepts the following options:
9427 Set the number of frames from which one will be dropped. Setting this to
9428 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9429 Default is @code{5}.
9432 Set the threshold for duplicate detection. If the difference metric for a frame
9433 is less than or equal to this value, then it is declared as duplicate. Default
9437 Set scene change threshold. Default is @code{15}.
9441 Set the size of the x and y-axis blocks used during metric calculations.
9442 Larger blocks give better noise suppression, but also give worse detection of
9443 small movements. Must be a power of two. Default is @code{32}.
9446 Mark main input as a pre-processed input and activate clean source input
9447 stream. This allows the input to be pre-processed with various filters to help
9448 the metrics calculation while keeping the frame selection lossless. When set to
9449 @code{1}, the first stream is for the pre-processed input, and the second
9450 stream is the clean source from where the kept frames are chosen. Default is
9454 Set whether or not chroma is considered in the metric calculations. Default is
9460 Apply 2D deconvolution of video stream in frequency domain using second stream
9463 The filter accepts the following options:
9467 Set which planes to process.
9470 Set which impulse video frames will be processed, can be @var{first}
9471 or @var{all}. Default is @var{all}.
9474 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9475 and height are not same and not power of 2 or if stream prior to convolving
9479 The @code{deconvolve} filter also supports the @ref{framesync} options.
9483 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9485 It accepts the following options:
9489 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9490 @var{rainbows} for cross-color reduction.
9493 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9496 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9499 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9502 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9507 Apply deflate effect to the video.
9509 This filter replaces the pixel by the local(3x3) average by taking into account
9510 only values lower than the pixel.
9512 It accepts the following options:
9519 Limit the maximum change for each plane, default is 65535.
9520 If 0, plane will remain unchanged.
9523 @subsection Commands
9525 This filter supports the all above options as @ref{commands}.
9529 Remove temporal frame luminance variations.
9531 It accepts the following options:
9535 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9538 Set averaging mode to smooth temporal luminance variations.
9540 Available values are:
9565 Do not actually modify frame. Useful when one only wants metadata.
9570 Remove judder produced by partially interlaced telecined content.
9572 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9573 source was partially telecined content then the output of @code{pullup,dejudder}
9574 will have a variable frame rate. May change the recorded frame rate of the
9575 container. Aside from that change, this filter will not affect constant frame
9578 The option available in this filter is:
9582 Specify the length of the window over which the judder repeats.
9584 Accepts any integer greater than 1. Useful values are:
9588 If the original was telecined from 24 to 30 fps (Film to NTSC).
9591 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9594 If a mixture of the two.
9597 The default is @samp{4}.
9602 Suppress a TV station logo by a simple interpolation of the surrounding
9603 pixels. Just set a rectangle covering the logo and watch it disappear
9604 (and sometimes something even uglier appear - your mileage may vary).
9606 It accepts the following parameters:
9611 Specify the top left corner coordinates of the logo. They must be
9616 Specify the width and height of the logo to clear. They must be
9620 Specify the thickness of the fuzzy edge of the rectangle (added to
9621 @var{w} and @var{h}). The default value is 1. This option is
9622 deprecated, setting higher values should no longer be necessary and
9626 When set to 1, a green rectangle is drawn on the screen to simplify
9627 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9628 The default value is 0.
9630 The rectangle is drawn on the outermost pixels which will be (partly)
9631 replaced with interpolated values. The values of the next pixels
9632 immediately outside this rectangle in each direction will be used to
9633 compute the interpolated pixel values inside the rectangle.
9637 @subsection Examples
9641 Set a rectangle covering the area with top left corner coordinates 0,0
9642 and size 100x77, and a band of size 10:
9644 delogo=x=0:y=0:w=100:h=77:band=10
9652 Remove the rain in the input image/video by applying the derain methods based on
9653 convolutional neural networks. Supported models:
9657 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9658 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9661 Training as well as model generation scripts are provided in
9662 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9664 Native model files (.model) can be generated from TensorFlow model
9665 files (.pb) by using tools/python/convert.py
9667 The filter accepts the following options:
9671 Specify which filter to use. This option accepts the following values:
9675 Derain filter. To conduct derain filter, you need to use a derain model.
9678 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9680 Default value is @samp{derain}.
9683 Specify which DNN backend to use for model loading and execution. This option accepts
9684 the following values:
9688 Native implementation of DNN loading and execution.
9691 TensorFlow backend. To enable this backend you
9692 need to install the TensorFlow for C library (see
9693 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9694 @code{--enable-libtensorflow}
9696 Default value is @samp{native}.
9699 Set path to model file specifying network architecture and its parameters.
9700 Note that different backends use different file formats. TensorFlow and native
9701 backend can load files for only its format.
9704 It can also be finished with @ref{dnn_processing} filter.
9708 Attempt to fix small changes in horizontal and/or vertical shift. This
9709 filter helps remove camera shake from hand-holding a camera, bumping a
9710 tripod, moving on a vehicle, etc.
9712 The filter accepts the following options:
9720 Specify a rectangular area where to limit the search for motion
9722 If desired the search for motion vectors can be limited to a
9723 rectangular area of the frame defined by its top left corner, width
9724 and height. These parameters have the same meaning as the drawbox
9725 filter which can be used to visualise the position of the bounding
9728 This is useful when simultaneous movement of subjects within the frame
9729 might be confused for camera motion by the motion vector search.
9731 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9732 then the full frame is used. This allows later options to be set
9733 without specifying the bounding box for the motion vector search.
9735 Default - search the whole frame.
9739 Specify the maximum extent of movement in x and y directions in the
9740 range 0-64 pixels. Default 16.
9743 Specify how to generate pixels to fill blanks at the edge of the
9744 frame. Available values are:
9747 Fill zeroes at blank locations
9749 Original image at blank locations
9751 Extruded edge value at blank locations
9753 Mirrored edge at blank locations
9755 Default value is @samp{mirror}.
9758 Specify the blocksize to use for motion search. Range 4-128 pixels,
9762 Specify the contrast threshold for blocks. Only blocks with more than
9763 the specified contrast (difference between darkest and lightest
9764 pixels) will be considered. Range 1-255, default 125.
9767 Specify the search strategy. Available values are:
9770 Set exhaustive search
9772 Set less exhaustive search.
9774 Default value is @samp{exhaustive}.
9777 If set then a detailed log of the motion search is written to the
9784 Remove unwanted contamination of foreground colors, caused by reflected color of
9785 greenscreen or bluescreen.
9787 This filter accepts the following options:
9791 Set what type of despill to use.
9794 Set how spillmap will be generated.
9797 Set how much to get rid of still remaining spill.
9800 Controls amount of red in spill area.
9803 Controls amount of green in spill area.
9804 Should be -1 for greenscreen.
9807 Controls amount of blue in spill area.
9808 Should be -1 for bluescreen.
9811 Controls brightness of spill area, preserving colors.
9814 Modify alpha from generated spillmap.
9817 @subsection Commands
9819 This filter supports the all above options as @ref{commands}.
9823 Apply an exact inverse of the telecine operation. It requires a predefined
9824 pattern specified using the pattern option which must be the same as that passed
9825 to the telecine filter.
9827 This filter accepts the following options:
9836 The default value is @code{top}.
9840 A string of numbers representing the pulldown pattern you wish to apply.
9841 The default value is @code{23}.
9844 A number representing position of the first frame with respect to the telecine
9845 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9850 Apply dilation effect to the video.
9852 This filter replaces the pixel by the local(3x3) maximum.
9854 It accepts the following options:
9861 Limit the maximum change for each plane, default is 65535.
9862 If 0, plane will remain unchanged.
9865 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9868 Flags to local 3x3 coordinates maps like this:
9875 @subsection Commands
9877 This filter supports the all above options as @ref{commands}.
9881 Displace pixels as indicated by second and third input stream.
9883 It takes three input streams and outputs one stream, the first input is the
9884 source, and second and third input are displacement maps.
9886 The second input specifies how much to displace pixels along the
9887 x-axis, while the third input specifies how much to displace pixels
9889 If one of displacement map streams terminates, last frame from that
9890 displacement map will be used.
9892 Note that once generated, displacements maps can be reused over and over again.
9894 A description of the accepted options follows.
9898 Set displace behavior for pixels that are out of range.
9900 Available values are:
9903 Missing pixels are replaced by black pixels.
9906 Adjacent pixels will spread out to replace missing pixels.
9909 Out of range pixels are wrapped so they point to pixels of other side.
9912 Out of range pixels will be replaced with mirrored pixels.
9914 Default is @samp{smear}.
9918 @subsection Examples
9922 Add ripple effect to rgb input of video size hd720:
9924 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
9928 Add wave effect to rgb input of video size hd720:
9930 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
9934 @anchor{dnn_processing}
9935 @section dnn_processing
9937 Do image processing with deep neural networks. It works together with another filter
9938 which converts the pixel format of the Frame to what the dnn network requires.
9940 The filter accepts the following options:
9944 Specify which DNN backend to use for model loading and execution. This option accepts
9945 the following values:
9949 Native implementation of DNN loading and execution.
9952 TensorFlow backend. To enable this backend you
9953 need to install the TensorFlow for C library (see
9954 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9955 @code{--enable-libtensorflow}
9958 OpenVINO backend. To enable this backend you
9959 need to build and install the OpenVINO for C library (see
9960 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9961 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9962 be needed if the header files and libraries are not installed into system path)
9966 Default value is @samp{native}.
9969 Set path to model file specifying network architecture and its parameters.
9970 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9971 backend can load files for only its format.
9973 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9976 Set the input name of the dnn network.
9979 Set the output name of the dnn network.
9982 use DNN async execution if set (default: set),
9983 roll back to sync execution if the backend does not support async.
9987 @subsection Examples
9991 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9993 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9997 Halve the pixel value of the frame with format gray32f:
9999 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
10003 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
10005 ./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
10009 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
10011 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
10018 Draw a colored box on the input image.
10020 It accepts the following parameters:
10025 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10029 The expressions which specify the width and height of the box; if 0 they are interpreted as
10030 the input width and height. It defaults to 0.
10033 Specify the color of the box to write. For the general syntax of this option,
10034 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10035 value @code{invert} is used, the box edge color is the same as the
10036 video with inverted luma.
10039 The expression which sets the thickness of the box edge.
10040 A value of @code{fill} will create a filled box. Default value is @code{3}.
10042 See below for the list of accepted constants.
10045 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10046 will overwrite the video's color and alpha pixels.
10047 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10050 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10051 following constants:
10055 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10059 horizontal and vertical chroma subsample values. For example for the
10060 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10064 The input width and height.
10067 The input sample aspect ratio.
10071 The x and y offset coordinates where the box is drawn.
10075 The width and height of the drawn box.
10078 The thickness of the drawn box.
10080 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10081 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10085 @subsection Examples
10089 Draw a black box around the edge of the input image:
10095 Draw a box with color red and an opacity of 50%:
10097 drawbox=10:20:200:60:red@@0.5
10100 The previous example can be specified as:
10102 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10106 Fill the box with pink color:
10108 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10112 Draw a 2-pixel red 2.40:1 mask:
10114 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
10118 @subsection Commands
10119 This filter supports same commands as options.
10120 The command accepts the same syntax of the corresponding option.
10122 If the specified expression is not valid, it is kept at its current
10127 Draw a graph using input video metadata.
10129 It accepts the following parameters:
10133 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10136 Set 1st foreground color expression.
10139 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10142 Set 2nd foreground color expression.
10145 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10148 Set 3rd foreground color expression.
10151 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10154 Set 4th foreground color expression.
10157 Set minimal value of metadata value.
10160 Set maximal value of metadata value.
10163 Set graph background color. Default is white.
10168 Available values for mode is:
10175 Default is @code{line}.
10180 Available values for slide is:
10183 Draw new frame when right border is reached.
10186 Replace old columns with new ones.
10189 Scroll from right to left.
10192 Scroll from left to right.
10195 Draw single picture.
10198 Default is @code{frame}.
10201 Set size of graph video. For the syntax of this option, check the
10202 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10203 The default value is @code{900x256}.
10206 Set the output frame rate. Default value is @code{25}.
10208 The foreground color expressions can use the following variables:
10211 Minimal value of metadata value.
10214 Maximal value of metadata value.
10217 Current metadata key value.
10220 The color is defined as 0xAABBGGRR.
10223 Example using metadata from @ref{signalstats} filter:
10225 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10228 Example using metadata from @ref{ebur128} filter:
10230 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10235 Draw a grid on the input image.
10237 It accepts the following parameters:
10242 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10246 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10247 input width and height, respectively, minus @code{thickness}, so image gets
10248 framed. Default to 0.
10251 Specify the color of the grid. For the general syntax of this option,
10252 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10253 value @code{invert} is used, the grid color is the same as the
10254 video with inverted luma.
10257 The expression which sets the thickness of the grid line. Default value is @code{1}.
10259 See below for the list of accepted constants.
10262 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10263 will overwrite the video's color and alpha pixels.
10264 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10267 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10268 following constants:
10272 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10276 horizontal and vertical chroma subsample values. For example for the
10277 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10281 The input grid cell width and height.
10284 The input sample aspect ratio.
10288 The x and y coordinates of some point of grid intersection (meant to configure offset).
10292 The width and height of the drawn cell.
10295 The thickness of the drawn cell.
10297 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10298 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10302 @subsection Examples
10306 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10308 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10312 Draw a white 3x3 grid with an opacity of 50%:
10314 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10318 @subsection Commands
10319 This filter supports same commands as options.
10320 The command accepts the same syntax of the corresponding option.
10322 If the specified expression is not valid, it is kept at its current
10328 Draw a text string or text from a specified file on top of a video, using the
10329 libfreetype library.
10331 To enable compilation of this filter, you need to configure FFmpeg with
10332 @code{--enable-libfreetype}.
10333 To enable default font fallback and the @var{font} option you need to
10334 configure FFmpeg with @code{--enable-libfontconfig}.
10335 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10336 @code{--enable-libfribidi}.
10340 It accepts the following parameters:
10345 Used to draw a box around text using the background color.
10346 The value must be either 1 (enable) or 0 (disable).
10347 The default value of @var{box} is 0.
10350 Set the width of the border to be drawn around the box using @var{boxcolor}.
10351 The default value of @var{boxborderw} is 0.
10354 The color to be used for drawing box around text. For the syntax of this
10355 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10357 The default value of @var{boxcolor} is "white".
10360 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10361 The default value of @var{line_spacing} is 0.
10364 Set the width of the border to be drawn around the text using @var{bordercolor}.
10365 The default value of @var{borderw} is 0.
10368 Set the color to be used for drawing border around text. For the syntax of this
10369 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10371 The default value of @var{bordercolor} is "black".
10374 Select how the @var{text} is expanded. Can be either @code{none},
10375 @code{strftime} (deprecated) or
10376 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10380 Set a start time for the count. Value is in microseconds. Only applied
10381 in the deprecated strftime expansion mode. To emulate in normal expansion
10382 mode use the @code{pts} function, supplying the start time (in seconds)
10383 as the second argument.
10386 If true, check and fix text coords to avoid clipping.
10389 The color to be used for drawing fonts. For the syntax of this option, check
10390 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10392 The default value of @var{fontcolor} is "black".
10394 @item fontcolor_expr
10395 String which is expanded the same way as @var{text} to obtain dynamic
10396 @var{fontcolor} value. By default this option has empty value and is not
10397 processed. When this option is set, it overrides @var{fontcolor} option.
10400 The font family to be used for drawing text. By default Sans.
10403 The font file to be used for drawing text. The path must be included.
10404 This parameter is mandatory if the fontconfig support is disabled.
10407 Draw the text applying alpha blending. The value can
10408 be a number between 0.0 and 1.0.
10409 The expression accepts the same variables @var{x, y} as well.
10410 The default value is 1.
10411 Please see @var{fontcolor_expr}.
10414 The font size to be used for drawing text.
10415 The default value of @var{fontsize} is 16.
10418 If set to 1, attempt to shape the text (for example, reverse the order of
10419 right-to-left text and join Arabic characters) before drawing it.
10420 Otherwise, just draw the text exactly as given.
10421 By default 1 (if supported).
10423 @item ft_load_flags
10424 The flags to be used for loading the fonts.
10426 The flags map the corresponding flags supported by libfreetype, and are
10427 a combination of the following values:
10434 @item vertical_layout
10435 @item force_autohint
10438 @item ignore_global_advance_width
10440 @item ignore_transform
10442 @item linear_design
10446 Default value is "default".
10448 For more information consult the documentation for the FT_LOAD_*
10452 The color to be used for drawing a shadow behind the drawn text. For the
10453 syntax of this option, check the @ref{color syntax,,"Color" section in the
10454 ffmpeg-utils manual,ffmpeg-utils}.
10456 The default value of @var{shadowcolor} is "black".
10460 The x and y offsets for the text shadow position with respect to the
10461 position of the text. They can be either positive or negative
10462 values. The default value for both is "0".
10465 The starting frame number for the n/frame_num variable. The default value
10469 The size in number of spaces to use for rendering the tab.
10470 Default value is 4.
10473 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10474 format. It can be used with or without text parameter. @var{timecode_rate}
10475 option must be specified.
10477 @item timecode_rate, rate, r
10478 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10479 integer. Minimum value is "1".
10480 Drop-frame timecode is supported for frame rates 30 & 60.
10483 If set to 1, the output of the timecode option will wrap around at 24 hours.
10484 Default is 0 (disabled).
10487 The text string to be drawn. The text must be a sequence of UTF-8
10488 encoded characters.
10489 This parameter is mandatory if no file is specified with the parameter
10493 A text file containing text to be drawn. The text must be a sequence
10494 of UTF-8 encoded characters.
10496 This parameter is mandatory if no text string is specified with the
10497 parameter @var{text}.
10499 If both @var{text} and @var{textfile} are specified, an error is thrown.
10502 If set to 1, the @var{textfile} will be reloaded before each frame.
10503 Be sure to update it atomically, or it may be read partially, or even fail.
10507 The expressions which specify the offsets where text will be drawn
10508 within the video frame. They are relative to the top/left border of the
10511 The default value of @var{x} and @var{y} is "0".
10513 See below for the list of accepted constants and functions.
10516 The parameters for @var{x} and @var{y} are expressions containing the
10517 following constants and functions:
10521 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10525 horizontal and vertical chroma subsample values. For example for the
10526 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10529 the height of each text line
10537 @item max_glyph_a, ascent
10538 the maximum distance from the baseline to the highest/upper grid
10539 coordinate used to place a glyph outline point, for all the rendered
10541 It is a positive value, due to the grid's orientation with the Y axis
10544 @item max_glyph_d, descent
10545 the maximum distance from the baseline to the lowest grid coordinate
10546 used to place a glyph outline point, for all the rendered glyphs.
10547 This is a negative value, due to the grid's orientation, with the Y axis
10551 maximum glyph height, that is the maximum height for all the glyphs
10552 contained in the rendered text, it is equivalent to @var{ascent} -
10556 maximum glyph width, that is the maximum width for all the glyphs
10557 contained in the rendered text
10560 the number of input frame, starting from 0
10562 @item rand(min, max)
10563 return a random number included between @var{min} and @var{max}
10566 The input sample aspect ratio.
10569 timestamp expressed in seconds, NAN if the input timestamp is unknown
10572 the height of the rendered text
10575 the width of the rendered text
10579 the x and y offset coordinates where the text is drawn.
10581 These parameters allow the @var{x} and @var{y} expressions to refer
10582 to each other, so you can for example specify @code{y=x/dar}.
10585 A one character description of the current frame's picture type.
10588 The current packet's position in the input file or stream
10589 (in bytes, from the start of the input). A value of -1 indicates
10590 this info is not available.
10593 The current packet's duration, in seconds.
10596 The current packet's size (in bytes).
10599 @anchor{drawtext_expansion}
10600 @subsection Text expansion
10602 If @option{expansion} is set to @code{strftime},
10603 the filter recognizes strftime() sequences in the provided text and
10604 expands them accordingly. Check the documentation of strftime(). This
10605 feature is deprecated.
10607 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10609 If @option{expansion} is set to @code{normal} (which is the default),
10610 the following expansion mechanism is used.
10612 The backslash character @samp{\}, followed by any character, always expands to
10613 the second character.
10615 Sequences of the form @code{%@{...@}} are expanded. The text between the
10616 braces is a function name, possibly followed by arguments separated by ':'.
10617 If the arguments contain special characters or delimiters (':' or '@}'),
10618 they should be escaped.
10620 Note that they probably must also be escaped as the value for the
10621 @option{text} option in the filter argument string and as the filter
10622 argument in the filtergraph description, and possibly also for the shell,
10623 that makes up to four levels of escaping; using a text file avoids these
10626 The following functions are available:
10631 The expression evaluation result.
10633 It must take one argument specifying the expression to be evaluated,
10634 which accepts the same constants and functions as the @var{x} and
10635 @var{y} values. Note that not all constants should be used, for
10636 example the text size is not known when evaluating the expression, so
10637 the constants @var{text_w} and @var{text_h} will have an undefined
10640 @item expr_int_format, eif
10641 Evaluate the expression's value and output as formatted integer.
10643 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10644 The second argument specifies the output format. Allowed values are @samp{x},
10645 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10646 @code{printf} function.
10647 The third parameter is optional and sets the number of positions taken by the output.
10648 It can be used to add padding with zeros from the left.
10651 The time at which the filter is running, expressed in UTC.
10652 It can accept an argument: a strftime() format string.
10655 The time at which the filter is running, expressed in the local time zone.
10656 It can accept an argument: a strftime() format string.
10659 Frame metadata. Takes one or two arguments.
10661 The first argument is mandatory and specifies the metadata key.
10663 The second argument is optional and specifies a default value, used when the
10664 metadata key is not found or empty.
10666 Available metadata can be identified by inspecting entries
10667 starting with TAG included within each frame section
10668 printed by running @code{ffprobe -show_frames}.
10670 String metadata generated in filters leading to
10671 the drawtext filter are also available.
10674 The frame number, starting from 0.
10677 A one character description of the current picture type.
10680 The timestamp of the current frame.
10681 It can take up to three arguments.
10683 The first argument is the format of the timestamp; it defaults to @code{flt}
10684 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10685 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10686 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10687 @code{localtime} stands for the timestamp of the frame formatted as
10688 local time zone time.
10690 The second argument is an offset added to the timestamp.
10692 If the format is set to @code{hms}, a third argument @code{24HH} may be
10693 supplied to present the hour part of the formatted timestamp in 24h format
10696 If the format is set to @code{localtime} or @code{gmtime},
10697 a third argument may be supplied: a strftime() format string.
10698 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10701 @subsection Commands
10703 This filter supports altering parameters via commands:
10706 Alter existing filter parameters.
10708 Syntax for the argument is the same as for filter invocation, e.g.
10711 fontsize=56:fontcolor=green:text='Hello World'
10714 Full filter invocation with sendcmd would look like this:
10717 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10721 If the entire argument can't be parsed or applied as valid values then the filter will
10722 continue with its existing parameters.
10724 @subsection Examples
10728 Draw "Test Text" with font FreeSerif, using the default values for the
10729 optional parameters.
10732 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10736 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10737 and y=50 (counting from the top-left corner of the screen), text is
10738 yellow with a red box around it. Both the text and the box have an
10742 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10743 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10746 Note that the double quotes are not necessary if spaces are not used
10747 within the parameter list.
10750 Show the text at the center of the video frame:
10752 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10756 Show the text at a random position, switching to a new position every 30 seconds:
10758 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)"
10762 Show a text line sliding from right to left in the last row of the video
10763 frame. The file @file{LONG_LINE} is assumed to contain a single line
10766 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10770 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10772 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10776 Draw a single green letter "g", at the center of the input video.
10777 The glyph baseline is placed at half screen height.
10779 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10783 Show text for 1 second every 3 seconds:
10785 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10789 Use fontconfig to set the font. Note that the colons need to be escaped.
10791 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10795 Draw "Test Text" with font size dependent on height of the video.
10797 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10801 Print the date of a real-time encoding (see strftime(3)):
10803 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10807 Show text fading in and out (appearing/disappearing):
10810 DS=1.0 # display start
10811 DE=10.0 # display end
10812 FID=1.5 # fade in duration
10813 FOD=5 # fade out duration
10814 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 @}"
10818 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10819 and the @option{fontsize} value are included in the @option{y} offset.
10821 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10822 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10826 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10827 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10828 must have option @option{-export_path_metadata 1} for the special metadata fields
10829 to be available for filters.
10831 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10836 For more information about libfreetype, check:
10837 @url{http://www.freetype.org/}.
10839 For more information about fontconfig, check:
10840 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10842 For more information about libfribidi, check:
10843 @url{http://fribidi.org/}.
10845 @section edgedetect
10847 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10849 The filter accepts the following options:
10854 Set low and high threshold values used by the Canny thresholding
10857 The high threshold selects the "strong" edge pixels, which are then
10858 connected through 8-connectivity with the "weak" edge pixels selected
10859 by the low threshold.
10861 @var{low} and @var{high} threshold values must be chosen in the range
10862 [0,1], and @var{low} should be lesser or equal to @var{high}.
10864 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10868 Define the drawing mode.
10872 Draw white/gray wires on black background.
10875 Mix the colors to create a paint/cartoon effect.
10878 Apply Canny edge detector on all selected planes.
10880 Default value is @var{wires}.
10883 Select planes for filtering. By default all available planes are filtered.
10886 @subsection Examples
10890 Standard edge detection with custom values for the hysteresis thresholding:
10892 edgedetect=low=0.1:high=0.4
10896 Painting effect without thresholding:
10898 edgedetect=mode=colormix:high=0
10904 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10906 For each input image, the filter will compute the optimal mapping from
10907 the input to the output given the codebook length, that is the number
10908 of distinct output colors.
10910 This filter accepts the following options.
10913 @item codebook_length, l
10914 Set codebook length. The value must be a positive integer, and
10915 represents the number of distinct output colors. Default value is 256.
10918 Set the maximum number of iterations to apply for computing the optimal
10919 mapping. The higher the value the better the result and the higher the
10920 computation time. Default value is 1.
10923 Set a random seed, must be an integer included between 0 and
10924 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10925 will try to use a good random seed on a best effort basis.
10928 Set pal8 output pixel format. This option does not work with codebook
10929 length greater than 256.
10934 Measure graylevel entropy in histogram of color channels of video frames.
10936 It accepts the following parameters:
10940 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10942 @var{diff} mode measures entropy of histogram delta values, absolute differences
10943 between neighbour histogram values.
10947 Set brightness, contrast, saturation and approximate gamma adjustment.
10949 The filter accepts the following options:
10953 Set the contrast expression. The value must be a float value in range
10954 @code{-1000.0} to @code{1000.0}. The default value is "1".
10957 Set the brightness expression. The value must be a float value in
10958 range @code{-1.0} to @code{1.0}. The default value is "0".
10961 Set the saturation expression. The value must be a float in
10962 range @code{0.0} to @code{3.0}. The default value is "1".
10965 Set the gamma expression. The value must be a float in range
10966 @code{0.1} to @code{10.0}. The default value is "1".
10969 Set the gamma expression for red. The value must be a float in
10970 range @code{0.1} to @code{10.0}. The default value is "1".
10973 Set the gamma expression for green. The value must be a float in range
10974 @code{0.1} to @code{10.0}. The default value is "1".
10977 Set the gamma expression for blue. The value must be a float in range
10978 @code{0.1} to @code{10.0}. The default value is "1".
10981 Set the gamma weight expression. It can be used to reduce the effect
10982 of a high gamma value on bright image areas, e.g. keep them from
10983 getting overamplified and just plain white. The value must be a float
10984 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10985 gamma correction all the way down while @code{1.0} leaves it at its
10986 full strength. Default is "1".
10989 Set when the expressions for brightness, contrast, saturation and
10990 gamma expressions are evaluated.
10992 It accepts the following values:
10995 only evaluate expressions once during the filter initialization or
10996 when a command is processed
10999 evaluate expressions for each incoming frame
11002 Default value is @samp{init}.
11005 The expressions accept the following parameters:
11008 frame count of the input frame starting from 0
11011 byte position of the corresponding packet in the input file, NAN if
11015 frame rate of the input video, NAN if the input frame rate is unknown
11018 timestamp expressed in seconds, NAN if the input timestamp is unknown
11021 @subsection Commands
11022 The filter supports the following commands:
11026 Set the contrast expression.
11029 Set the brightness expression.
11032 Set the saturation expression.
11035 Set the gamma expression.
11038 Set the gamma_r expression.
11041 Set gamma_g expression.
11044 Set gamma_b expression.
11047 Set gamma_weight expression.
11049 The command accepts the same syntax of the corresponding option.
11051 If the specified expression is not valid, it is kept at its current
11058 Apply erosion effect to the video.
11060 This filter replaces the pixel by the local(3x3) minimum.
11062 It accepts the following options:
11069 Limit the maximum change for each plane, default is 65535.
11070 If 0, plane will remain unchanged.
11073 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11076 Flags to local 3x3 coordinates maps like this:
11083 @subsection Commands
11085 This filter supports the all above options as @ref{commands}.
11089 Deinterlace the input video ("estdif" stands for "Edge Slope
11090 Tracing Deinterlacing Filter").
11092 Spatial only filter that uses edge slope tracing algorithm
11093 to interpolate missing lines.
11094 It accepts the following parameters:
11098 The interlacing mode to adopt. It accepts one of the following values:
11102 Output one frame for each frame.
11104 Output one frame for each field.
11107 The default value is @code{field}.
11110 The picture field parity assumed for the input interlaced video. It accepts one
11111 of the following values:
11115 Assume the top field is first.
11117 Assume the bottom field is first.
11119 Enable automatic detection of field parity.
11122 The default value is @code{auto}.
11123 If the interlacing is unknown or the decoder does not export this information,
11124 top field first will be assumed.
11127 Specify which frames to deinterlace. Accepts one of the following
11132 Deinterlace all frames.
11134 Only deinterlace frames marked as interlaced.
11137 The default value is @code{all}.
11140 Specify the search radius for edge slope tracing. Default value is 1.
11141 Allowed range is from 1 to 15.
11144 Specify the search radius for best edge matching. Default value is 2.
11145 Allowed range is from 0 to 15.
11148 Specify the interpolation used. Default is 4-point interpolation. It accepts one
11149 of the following values:
11153 Two-point interpolation.
11155 Four-point interpolation.
11157 Six-point interpolation.
11161 @subsection Commands
11162 This filter supports same @ref{commands} as options.
11164 @section extractplanes
11166 Extract color channel components from input video stream into
11167 separate grayscale video streams.
11169 The filter accepts the following option:
11173 Set plane(s) to extract.
11175 Available values for planes are:
11186 Choosing planes not available in the input will result in an error.
11187 That means you cannot select @code{r}, @code{g}, @code{b} planes
11188 with @code{y}, @code{u}, @code{v} planes at same time.
11191 @subsection Examples
11195 Extract luma, u and v color channel component from input video frame
11196 into 3 grayscale outputs:
11198 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
11204 Apply a fade-in/out effect to the input video.
11206 It accepts the following parameters:
11210 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11212 Default is @code{in}.
11214 @item start_frame, s
11215 Specify the number of the frame to start applying the fade
11216 effect at. Default is 0.
11219 The number of frames that the fade effect lasts. At the end of the
11220 fade-in effect, the output video will have the same intensity as the input video.
11221 At the end of the fade-out transition, the output video will be filled with the
11222 selected @option{color}.
11226 If set to 1, fade only alpha channel, if one exists on the input.
11227 Default value is 0.
11229 @item start_time, st
11230 Specify the timestamp (in seconds) of the frame to start to apply the fade
11231 effect. If both start_frame and start_time are specified, the fade will start at
11232 whichever comes last. Default is 0.
11235 The number of seconds for which the fade effect has to last. At the end of the
11236 fade-in effect the output video will have the same intensity as the input video,
11237 at the end of the fade-out transition the output video will be filled with the
11238 selected @option{color}.
11239 If both duration and nb_frames are specified, duration is used. Default is 0
11240 (nb_frames is used by default).
11243 Specify the color of the fade. Default is "black".
11246 @subsection Examples
11250 Fade in the first 30 frames of video:
11255 The command above is equivalent to:
11261 Fade out the last 45 frames of a 200-frame video:
11264 fade=type=out:start_frame=155:nb_frames=45
11268 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11270 fade=in:0:25, fade=out:975:25
11274 Make the first 5 frames yellow, then fade in from frame 5-24:
11276 fade=in:5:20:color=yellow
11280 Fade in alpha over first 25 frames of video:
11282 fade=in:0:25:alpha=1
11286 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11288 fade=t=in:st=5.5:d=0.5
11294 Denoise frames using 3D FFT (frequency domain filtering).
11296 The filter accepts the following options:
11300 Set the noise sigma constant. This sets denoising strength.
11301 Default value is 1. Allowed range is from 0 to 30.
11302 Using very high sigma with low overlap may give blocking artifacts.
11305 Set amount of denoising. By default all detected noise is reduced.
11306 Default value is 1. Allowed range is from 0 to 1.
11309 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11310 Actual size of block in pixels is 2 to power of @var{block}, so by default
11311 block size in pixels is 2^4 which is 16.
11314 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11317 Set number of previous frames to use for denoising. By default is set to 0.
11320 Set number of next frames to to use for denoising. By default is set to 0.
11323 Set planes which will be filtered, by default are all available filtered
11328 Apply arbitrary expressions to samples in frequency domain
11332 Adjust the dc value (gain) of the luma plane of the image. The filter
11333 accepts an integer value in range @code{0} to @code{1000}. The default
11334 value is set to @code{0}.
11337 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11338 filter accepts an integer value in range @code{0} to @code{1000}. The
11339 default value is set to @code{0}.
11342 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11343 filter accepts an integer value in range @code{0} to @code{1000}. The
11344 default value is set to @code{0}.
11347 Set the frequency domain weight expression for the luma plane.
11350 Set the frequency domain weight expression for the 1st chroma plane.
11353 Set the frequency domain weight expression for the 2nd chroma plane.
11356 Set when the expressions are evaluated.
11358 It accepts the following values:
11361 Only evaluate expressions once during the filter initialization.
11364 Evaluate expressions for each incoming frame.
11367 Default value is @samp{init}.
11369 The filter accepts the following variables:
11372 The coordinates of the current sample.
11376 The width and height of the image.
11379 The number of input frame, starting from 0.
11382 @subsection Examples
11388 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11394 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11400 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11406 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11413 Extract a single field from an interlaced image using stride
11414 arithmetic to avoid wasting CPU time. The output frames are marked as
11417 The filter accepts the following options:
11421 Specify whether to extract the top (if the value is @code{0} or
11422 @code{top}) or the bottom field (if the value is @code{1} or
11428 Create new frames by copying the top and bottom fields from surrounding frames
11429 supplied as numbers by the hint file.
11433 Set file containing hints: absolute/relative frame numbers.
11435 There must be one line for each frame in a clip. Each line must contain two
11436 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11437 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11438 is current frame number for @code{absolute} mode or out of [-1, 1] range
11439 for @code{relative} mode. First number tells from which frame to pick up top
11440 field and second number tells from which frame to pick up bottom field.
11442 If optionally followed by @code{+} output frame will be marked as interlaced,
11443 else if followed by @code{-} output frame will be marked as progressive, else
11444 it will be marked same as input frame.
11445 If optionally followed by @code{t} output frame will use only top field, or in
11446 case of @code{b} it will use only bottom field.
11447 If line starts with @code{#} or @code{;} that line is skipped.
11450 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11453 Example of first several lines of @code{hint} file for @code{relative} mode:
11455 0,0 - # first frame
11456 1,0 - # second frame, use third's frame top field and second's frame bottom field
11457 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11472 @section fieldmatch
11474 Field matching filter for inverse telecine. It is meant to reconstruct the
11475 progressive frames from a telecined stream. The filter does not drop duplicated
11476 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11477 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11479 The separation of the field matching and the decimation is notably motivated by
11480 the possibility of inserting a de-interlacing filter fallback between the two.
11481 If the source has mixed telecined and real interlaced content,
11482 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11483 But these remaining combed frames will be marked as interlaced, and thus can be
11484 de-interlaced by a later filter such as @ref{yadif} before decimation.
11486 In addition to the various configuration options, @code{fieldmatch} can take an
11487 optional second stream, activated through the @option{ppsrc} option. If
11488 enabled, the frames reconstruction will be based on the fields and frames from
11489 this second stream. This allows the first input to be pre-processed in order to
11490 help the various algorithms of the filter, while keeping the output lossless
11491 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11492 or brightness/contrast adjustments can help.
11494 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11495 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11496 which @code{fieldmatch} is based on. While the semantic and usage are very
11497 close, some behaviour and options names can differ.
11499 The @ref{decimate} filter currently only works for constant frame rate input.
11500 If your input has mixed telecined (30fps) and progressive content with a lower
11501 framerate like 24fps use the following filterchain to produce the necessary cfr
11502 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11504 The filter accepts the following options:
11508 Specify the assumed field order of the input stream. Available values are:
11512 Auto detect parity (use FFmpeg's internal parity value).
11514 Assume bottom field first.
11516 Assume top field first.
11519 Note that it is sometimes recommended not to trust the parity announced by the
11522 Default value is @var{auto}.
11525 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11526 sense that it won't risk creating jerkiness due to duplicate frames when
11527 possible, but if there are bad edits or blended fields it will end up
11528 outputting combed frames when a good match might actually exist. On the other
11529 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11530 but will almost always find a good frame if there is one. The other values are
11531 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11532 jerkiness and creating duplicate frames versus finding good matches in sections
11533 with bad edits, orphaned fields, blended fields, etc.
11535 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11537 Available values are:
11541 2-way matching (p/c)
11543 2-way matching, and trying 3rd match if still combed (p/c + n)
11545 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11547 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11548 still combed (p/c + n + u/b)
11550 3-way matching (p/c/n)
11552 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11553 detected as combed (p/c/n + u/b)
11556 The parenthesis at the end indicate the matches that would be used for that
11557 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11560 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11563 Default value is @var{pc_n}.
11566 Mark the main input stream as a pre-processed input, and enable the secondary
11567 input stream as the clean source to pick the fields from. See the filter
11568 introduction for more details. It is similar to the @option{clip2} feature from
11571 Default value is @code{0} (disabled).
11574 Set the field to match from. It is recommended to set this to the same value as
11575 @option{order} unless you experience matching failures with that setting. In
11576 certain circumstances changing the field that is used to match from can have a
11577 large impact on matching performance. Available values are:
11581 Automatic (same value as @option{order}).
11583 Match from the bottom field.
11585 Match from the top field.
11588 Default value is @var{auto}.
11591 Set whether or not chroma is included during the match comparisons. In most
11592 cases it is recommended to leave this enabled. You should set this to @code{0}
11593 only if your clip has bad chroma problems such as heavy rainbowing or other
11594 artifacts. Setting this to @code{0} could also be used to speed things up at
11595 the cost of some accuracy.
11597 Default value is @code{1}.
11601 These define an exclusion band which excludes the lines between @option{y0} and
11602 @option{y1} from being included in the field matching decision. An exclusion
11603 band can be used to ignore subtitles, a logo, or other things that may
11604 interfere with the matching. @option{y0} sets the starting scan line and
11605 @option{y1} sets the ending line; all lines in between @option{y0} and
11606 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11607 @option{y0} and @option{y1} to the same value will disable the feature.
11608 @option{y0} and @option{y1} defaults to @code{0}.
11611 Set the scene change detection threshold as a percentage of maximum change on
11612 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11613 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11614 @option{scthresh} is @code{[0.0, 100.0]}.
11616 Default value is @code{12.0}.
11619 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11620 account the combed scores of matches when deciding what match to use as the
11621 final match. Available values are:
11625 No final matching based on combed scores.
11627 Combed scores are only used when a scene change is detected.
11629 Use combed scores all the time.
11632 Default is @var{sc}.
11635 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11636 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11637 Available values are:
11641 No forced calculation.
11643 Force p/c/n calculations.
11645 Force p/c/n/u/b calculations.
11648 Default value is @var{none}.
11651 This is the area combing threshold used for combed frame detection. This
11652 essentially controls how "strong" or "visible" combing must be to be detected.
11653 Larger values mean combing must be more visible and smaller values mean combing
11654 can be less visible or strong and still be detected. Valid settings are from
11655 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11656 be detected as combed). This is basically a pixel difference value. A good
11657 range is @code{[8, 12]}.
11659 Default value is @code{9}.
11662 Sets whether or not chroma is considered in the combed frame decision. Only
11663 disable this if your source has chroma problems (rainbowing, etc.) that are
11664 causing problems for the combed frame detection with chroma enabled. Actually,
11665 using @option{chroma}=@var{0} is usually more reliable, except for the case
11666 where there is chroma only combing in the source.
11668 Default value is @code{0}.
11672 Respectively set the x-axis and y-axis size of the window used during combed
11673 frame detection. This has to do with the size of the area in which
11674 @option{combpel} pixels are required to be detected as combed for a frame to be
11675 declared combed. See the @option{combpel} parameter description for more info.
11676 Possible values are any number that is a power of 2 starting at 4 and going up
11679 Default value is @code{16}.
11682 The number of combed pixels inside any of the @option{blocky} by
11683 @option{blockx} size blocks on the frame for the frame to be detected as
11684 combed. While @option{cthresh} controls how "visible" the combing must be, this
11685 setting controls "how much" combing there must be in any localized area (a
11686 window defined by the @option{blockx} and @option{blocky} settings) on the
11687 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11688 which point no frames will ever be detected as combed). This setting is known
11689 as @option{MI} in TFM/VFM vocabulary.
11691 Default value is @code{80}.
11694 @anchor{p/c/n/u/b meaning}
11695 @subsection p/c/n/u/b meaning
11697 @subsubsection p/c/n
11699 We assume the following telecined stream:
11702 Top fields: 1 2 2 3 4
11703 Bottom fields: 1 2 3 4 4
11706 The numbers correspond to the progressive frame the fields relate to. Here, the
11707 first two frames are progressive, the 3rd and 4th are combed, and so on.
11709 When @code{fieldmatch} is configured to run a matching from bottom
11710 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11715 B 1 2 3 4 4 <-- matching reference
11724 As a result of the field matching, we can see that some frames get duplicated.
11725 To perform a complete inverse telecine, you need to rely on a decimation filter
11726 after this operation. See for instance the @ref{decimate} filter.
11728 The same operation now matching from top fields (@option{field}=@var{top})
11733 T 1 2 2 3 4 <-- matching reference
11743 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11744 basically, they refer to the frame and field of the opposite parity:
11747 @item @var{p} matches the field of the opposite parity in the previous frame
11748 @item @var{c} matches the field of the opposite parity in the current frame
11749 @item @var{n} matches the field of the opposite parity in the next frame
11754 The @var{u} and @var{b} matching are a bit special in the sense that they match
11755 from the opposite parity flag. In the following examples, we assume that we are
11756 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11757 'x' is placed above and below each matched fields.
11759 With bottom matching (@option{field}=@var{bottom}):
11764 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11765 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11773 With top matching (@option{field}=@var{top}):
11778 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11779 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11787 @subsection Examples
11789 Simple IVTC of a top field first telecined stream:
11791 fieldmatch=order=tff:combmatch=none, decimate
11794 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11796 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11799 @section fieldorder
11801 Transform the field order of the input video.
11803 It accepts the following parameters:
11808 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11809 for bottom field first.
11812 The default value is @samp{tff}.
11814 The transformation is done by shifting the picture content up or down
11815 by one line, and filling the remaining line with appropriate picture content.
11816 This method is consistent with most broadcast field order converters.
11818 If the input video is not flagged as being interlaced, or it is already
11819 flagged as being of the required output field order, then this filter does
11820 not alter the incoming video.
11822 It is very useful when converting to or from PAL DV material,
11823 which is bottom field first.
11827 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11830 @section fifo, afifo
11832 Buffer input images and send them when they are requested.
11834 It is mainly useful when auto-inserted by the libavfilter
11837 It does not take parameters.
11839 @section fillborders
11841 Fill borders of the input video, without changing video stream dimensions.
11842 Sometimes video can have garbage at the four edges and you may not want to
11843 crop video input to keep size multiple of some number.
11845 This filter accepts the following options:
11849 Number of pixels to fill from left border.
11852 Number of pixels to fill from right border.
11855 Number of pixels to fill from top border.
11858 Number of pixels to fill from bottom border.
11863 It accepts the following values:
11866 fill pixels using outermost pixels
11869 fill pixels using mirroring (half sample symmetric)
11872 fill pixels with constant value
11875 fill pixels using reflecting (whole sample symmetric)
11878 fill pixels using wrapping
11881 fade pixels to constant value
11884 Default is @var{smear}.
11887 Set color for pixels in fixed or fade mode. Default is @var{black}.
11890 @subsection Commands
11891 This filter supports same @ref{commands} as options.
11892 The command accepts the same syntax of the corresponding option.
11894 If the specified expression is not valid, it is kept at its current
11899 Find a rectangular object
11901 It accepts the following options:
11905 Filepath of the object image, needs to be in gray8.
11908 Detection threshold, default is 0.5.
11911 Number of mipmaps, default is 3.
11913 @item xmin, ymin, xmax, ymax
11914 Specifies the rectangle in which to search.
11917 @subsection Examples
11921 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11923 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11929 Flood area with values of same pixel components with another values.
11931 It accepts the following options:
11934 Set pixel x coordinate.
11937 Set pixel y coordinate.
11940 Set source #0 component value.
11943 Set source #1 component value.
11946 Set source #2 component value.
11949 Set source #3 component value.
11952 Set destination #0 component value.
11955 Set destination #1 component value.
11958 Set destination #2 component value.
11961 Set destination #3 component value.
11967 Convert the input video to one of the specified pixel formats.
11968 Libavfilter will try to pick one that is suitable as input to
11971 It accepts the following parameters:
11975 A '|'-separated list of pixel format names, such as
11976 "pix_fmts=yuv420p|monow|rgb24".
11980 @subsection Examples
11984 Convert the input video to the @var{yuv420p} format
11986 format=pix_fmts=yuv420p
11989 Convert the input video to any of the formats in the list
11991 format=pix_fmts=yuv420p|yuv444p|yuv410p
11998 Convert the video to specified constant frame rate by duplicating or dropping
11999 frames as necessary.
12001 It accepts the following parameters:
12005 The desired output frame rate. The default is @code{25}.
12008 Assume the first PTS should be the given value, in seconds. This allows for
12009 padding/trimming at the start of stream. By default, no assumption is made
12010 about the first frame's expected PTS, so no padding or trimming is done.
12011 For example, this could be set to 0 to pad the beginning with duplicates of
12012 the first frame if a video stream starts after the audio stream or to trim any
12013 frames with a negative PTS.
12016 Timestamp (PTS) rounding method.
12018 Possible values are:
12025 round towards -infinity
12027 round towards +infinity
12031 The default is @code{near}.
12034 Action performed when reading the last frame.
12036 Possible values are:
12039 Use same timestamp rounding method as used for other frames.
12041 Pass through last frame if input duration has not been reached yet.
12043 The default is @code{round}.
12047 Alternatively, the options can be specified as a flat string:
12048 @var{fps}[:@var{start_time}[:@var{round}]].
12050 See also the @ref{setpts} filter.
12052 @subsection Examples
12056 A typical usage in order to set the fps to 25:
12062 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12064 fps=fps=film:round=near
12070 Pack two different video streams into a stereoscopic video, setting proper
12071 metadata on supported codecs. The two views should have the same size and
12072 framerate and processing will stop when the shorter video ends. Please note
12073 that you may conveniently adjust view properties with the @ref{scale} and
12076 It accepts the following parameters:
12080 The desired packing format. Supported values are:
12085 The views are next to each other (default).
12088 The views are on top of each other.
12091 The views are packed by line.
12094 The views are packed by column.
12097 The views are temporally interleaved.
12106 # Convert left and right views into a frame-sequential video
12107 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12109 # Convert views into a side-by-side video with the same output resolution as the input
12110 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
12115 Change the frame rate by interpolating new video output frames from the source
12118 This filter is not designed to function correctly with interlaced media. If
12119 you wish to change the frame rate of interlaced media then you are required
12120 to deinterlace before this filter and re-interlace after this filter.
12122 A description of the accepted options follows.
12126 Specify the output frames per second. This option can also be specified
12127 as a value alone. The default is @code{50}.
12130 Specify the start of a range where the output frame will be created as a
12131 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12132 the default is @code{15}.
12135 Specify the end of a range where the output frame will be created as a
12136 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12137 the default is @code{240}.
12140 Specify the level at which a scene change is detected as a value between
12141 0 and 100 to indicate a new scene; a low value reflects a low
12142 probability for the current frame to introduce a new scene, while a higher
12143 value means the current frame is more likely to be one.
12144 The default is @code{8.2}.
12147 Specify flags influencing the filter process.
12149 Available value for @var{flags} is:
12152 @item scene_change_detect, scd
12153 Enable scene change detection using the value of the option @var{scene}.
12154 This flag is enabled by default.
12160 Select one frame every N-th frame.
12162 This filter accepts the following option:
12165 Select frame after every @code{step} frames.
12166 Allowed values are positive integers higher than 0. Default value is @code{1}.
12169 @section freezedetect
12171 Detect frozen video.
12173 This filter logs a message and sets frame metadata when it detects that the
12174 input video has no significant change in content during a specified duration.
12175 Video freeze detection calculates the mean average absolute difference of all
12176 the components of video frames and compares it to a noise floor.
12178 The printed times and duration are expressed in seconds. The
12179 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12180 whose timestamp equals or exceeds the detection duration and it contains the
12181 timestamp of the first frame of the freeze. The
12182 @code{lavfi.freezedetect.freeze_duration} and
12183 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12186 The filter accepts the following options:
12190 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12191 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12195 Set freeze duration until notification (default is 2 seconds).
12198 @section freezeframes
12200 Freeze video frames.
12202 This filter freezes video frames using frame from 2nd input.
12204 The filter accepts the following options:
12208 Set number of first frame from which to start freeze.
12211 Set number of last frame from which to end freeze.
12214 Set number of frame from 2nd input which will be used instead of replaced frames.
12220 Apply a frei0r effect to the input video.
12222 To enable the compilation of this filter, you need to install the frei0r
12223 header and configure FFmpeg with @code{--enable-frei0r}.
12225 It accepts the following parameters:
12230 The name of the frei0r effect to load. If the environment variable
12231 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12232 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12233 Otherwise, the standard frei0r paths are searched, in this order:
12234 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12235 @file{/usr/lib/frei0r-1/}.
12237 @item filter_params
12238 A '|'-separated list of parameters to pass to the frei0r effect.
12242 A frei0r effect parameter can be a boolean (its value is either
12243 "y" or "n"), a double, a color (specified as
12244 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12245 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12246 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12247 a position (specified as @var{X}/@var{Y}, where
12248 @var{X} and @var{Y} are floating point numbers) and/or a string.
12250 The number and types of parameters depend on the loaded effect. If an
12251 effect parameter is not specified, the default value is set.
12253 @subsection Examples
12257 Apply the distort0r effect, setting the first two double parameters:
12259 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12263 Apply the colordistance effect, taking a color as the first parameter:
12265 frei0r=colordistance:0.2/0.3/0.4
12266 frei0r=colordistance:violet
12267 frei0r=colordistance:0x112233
12271 Apply the perspective effect, specifying the top left and top right image
12274 frei0r=perspective:0.2/0.2|0.8/0.2
12278 For more information, see
12279 @url{http://frei0r.dyne.org}
12281 @subsection Commands
12283 This filter supports the @option{filter_params} option as @ref{commands}.
12287 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12289 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12290 processing filter, one of them is performed once per block, not per pixel.
12291 This allows for much higher speed.
12293 The filter accepts the following options:
12297 Set quality. This option defines the number of levels for averaging. It accepts
12298 an integer in the range 4-5. Default value is @code{4}.
12301 Force a constant quantization parameter. It accepts an integer in range 0-63.
12302 If not set, the filter will use the QP from the video stream (if available).
12305 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12306 more details but also more artifacts, while higher values make the image smoother
12307 but also blurrier. Default value is @code{0} − PSNR optimal.
12309 @item use_bframe_qp
12310 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12311 option may cause flicker since the B-Frames have often larger QP. Default is
12312 @code{0} (not enabled).
12318 Apply Gaussian blur filter.
12320 The filter accepts the following options:
12324 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12327 Set number of steps for Gaussian approximation. Default is @code{1}.
12330 Set which planes to filter. By default all planes are filtered.
12333 Set vertical sigma, if negative it will be same as @code{sigma}.
12334 Default is @code{-1}.
12337 @subsection Commands
12338 This filter supports same commands as options.
12339 The command accepts the same syntax of the corresponding option.
12341 If the specified expression is not valid, it is kept at its current
12346 Apply generic equation to each pixel.
12348 The filter accepts the following options:
12351 @item lum_expr, lum
12352 Set the luminance expression.
12354 Set the chrominance blue expression.
12356 Set the chrominance red expression.
12357 @item alpha_expr, a
12358 Set the alpha expression.
12360 Set the red expression.
12361 @item green_expr, g
12362 Set the green expression.
12364 Set the blue expression.
12367 The colorspace is selected according to the specified options. If one
12368 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12369 options is specified, the filter will automatically select a YCbCr
12370 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12371 @option{blue_expr} options is specified, it will select an RGB
12374 If one of the chrominance expression is not defined, it falls back on the other
12375 one. If no alpha expression is specified it will evaluate to opaque value.
12376 If none of chrominance expressions are specified, they will evaluate
12377 to the luminance expression.
12379 The expressions can use the following variables and functions:
12383 The sequential number of the filtered frame, starting from @code{0}.
12387 The coordinates of the current sample.
12391 The width and height of the image.
12395 Width and height scale depending on the currently filtered plane. It is the
12396 ratio between the corresponding luma plane number of pixels and the current
12397 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12398 @code{0.5,0.5} for chroma planes.
12401 Time of the current frame, expressed in seconds.
12404 Return the value of the pixel at location (@var{x},@var{y}) of the current
12408 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12412 Return the value of the pixel at location (@var{x},@var{y}) of the
12413 blue-difference chroma plane. Return 0 if there is no such plane.
12416 Return the value of the pixel at location (@var{x},@var{y}) of the
12417 red-difference chroma plane. Return 0 if there is no such plane.
12422 Return the value of the pixel at location (@var{x},@var{y}) of the
12423 red/green/blue component. Return 0 if there is no such component.
12426 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12427 plane. Return 0 if there is no such plane.
12429 @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)
12430 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12431 sums of samples within a rectangle. See the functions without the sum postfix.
12433 @item interpolation
12434 Set one of interpolation methods:
12439 Default is bilinear.
12442 For functions, if @var{x} and @var{y} are outside the area, the value will be
12443 automatically clipped to the closer edge.
12445 Please note that this filter can use multiple threads in which case each slice
12446 will have its own expression state. If you want to use only a single expression
12447 state because your expressions depend on previous state then you should limit
12448 the number of filter threads to 1.
12450 @subsection Examples
12454 Flip the image horizontally:
12460 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12461 wavelength of 100 pixels:
12463 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12467 Generate a fancy enigmatic moving light:
12469 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
12473 Generate a quick emboss effect:
12475 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12479 Modify RGB components depending on pixel position:
12481 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12485 Create a radial gradient that is the same size as the input (also see
12486 the @ref{vignette} filter):
12488 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12494 Fix the banding artifacts that are sometimes introduced into nearly flat
12495 regions by truncation to 8-bit color depth.
12496 Interpolate the gradients that should go where the bands are, and
12499 It is designed for playback only. Do not use it prior to
12500 lossy compression, because compression tends to lose the dither and
12501 bring back the bands.
12503 It accepts the following parameters:
12508 The maximum amount by which the filter will change any one pixel. This is also
12509 the threshold for detecting nearly flat regions. Acceptable values range from
12510 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12514 The neighborhood to fit the gradient to. A larger radius makes for smoother
12515 gradients, but also prevents the filter from modifying the pixels near detailed
12516 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12517 values will be clipped to the valid range.
12521 Alternatively, the options can be specified as a flat string:
12522 @var{strength}[:@var{radius}]
12524 @subsection Examples
12528 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12534 Specify radius, omitting the strength (which will fall-back to the default
12542 @anchor{graphmonitor}
12543 @section graphmonitor
12544 Show various filtergraph stats.
12546 With this filter one can debug complete filtergraph.
12547 Especially issues with links filling with queued frames.
12549 The filter accepts the following options:
12553 Set video output size. Default is @var{hd720}.
12556 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12559 Set output mode, can be @var{fulll} or @var{compact}.
12560 In @var{compact} mode only filters with some queued frames have displayed stats.
12563 Set flags which enable which stats are shown in video.
12565 Available values for flags are:
12568 Display number of queued frames in each link.
12570 @item frame_count_in
12571 Display number of frames taken from filter.
12573 @item frame_count_out
12574 Display number of frames given out from filter.
12577 Display current filtered frame pts.
12580 Display current filtered frame time.
12583 Display time base for filter link.
12586 Display used format for filter link.
12589 Display video size or number of audio channels in case of audio used by filter link.
12592 Display video frame rate or sample rate in case of audio used by filter link.
12595 Display link output status.
12599 Set upper limit for video rate of output stream, Default value is @var{25}.
12600 This guarantee that output video frame rate will not be higher than this value.
12604 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12605 and corrects the scene colors accordingly.
12607 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12609 The filter accepts the following options:
12613 The order of differentiation to be applied on the scene. Must be chosen in the range
12614 [0,2] and default value is 1.
12617 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12618 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12619 max value instead of calculating Minkowski distance.
12622 The standard deviation of Gaussian blur to be applied on the scene. Must be
12623 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12624 can't be equal to 0 if @var{difford} is greater than 0.
12627 @subsection Examples
12633 greyedge=difford=1:minknorm=5:sigma=2
12639 greyedge=difford=1:minknorm=0:sigma=2
12647 Apply a Hald CLUT to a video stream.
12649 First input is the video stream to process, and second one is the Hald CLUT.
12650 The Hald CLUT input can be a simple picture or a complete video stream.
12652 The filter accepts the following options:
12656 Force termination when the shortest input terminates. Default is @code{0}.
12658 Continue applying the last CLUT after the end of the stream. A value of
12659 @code{0} disable the filter after the last frame of the CLUT is reached.
12660 Default is @code{1}.
12663 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12664 filters share the same internals).
12666 This filter also supports the @ref{framesync} options.
12668 More information about the Hald CLUT can be found on Eskil Steenberg's website
12669 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12671 @subsection Workflow examples
12673 @subsubsection Hald CLUT video stream
12675 Generate an identity Hald CLUT stream altered with various effects:
12677 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
12680 Note: make sure you use a lossless codec.
12682 Then use it with @code{haldclut} to apply it on some random stream:
12684 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12687 The Hald CLUT will be applied to the 10 first seconds (duration of
12688 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12689 to the remaining frames of the @code{mandelbrot} stream.
12691 @subsubsection Hald CLUT with preview
12693 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12694 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12695 biggest possible square starting at the top left of the picture. The remaining
12696 padding pixels (bottom or right) will be ignored. This area can be used to add
12697 a preview of the Hald CLUT.
12699 Typically, the following generated Hald CLUT will be supported by the
12700 @code{haldclut} filter:
12703 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12704 pad=iw+320 [padded_clut];
12705 smptebars=s=320x256, split [a][b];
12706 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12707 [main][b] overlay=W-320" -frames:v 1 clut.png
12710 It contains the original and a preview of the effect of the CLUT: SMPTE color
12711 bars are displayed on the right-top, and below the same color bars processed by
12714 Then, the effect of this Hald CLUT can be visualized with:
12716 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12721 Flip the input video horizontally.
12723 For example, to horizontally flip the input video with @command{ffmpeg}:
12725 ffmpeg -i in.avi -vf "hflip" out.avi
12729 This filter applies a global color histogram equalization on a
12732 It can be used to correct video that has a compressed range of pixel
12733 intensities. The filter redistributes the pixel intensities to
12734 equalize their distribution across the intensity range. It may be
12735 viewed as an "automatically adjusting contrast filter". This filter is
12736 useful only for correcting degraded or poorly captured source
12739 The filter accepts the following options:
12743 Determine the amount of equalization to be applied. As the strength
12744 is reduced, the distribution of pixel intensities more-and-more
12745 approaches that of the input frame. The value must be a float number
12746 in the range [0,1] and defaults to 0.200.
12749 Set the maximum intensity that can generated and scale the output
12750 values appropriately. The strength should be set as desired and then
12751 the intensity can be limited if needed to avoid washing-out. The value
12752 must be a float number in the range [0,1] and defaults to 0.210.
12755 Set the antibanding level. If enabled the filter will randomly vary
12756 the luminance of output pixels by a small amount to avoid banding of
12757 the histogram. Possible values are @code{none}, @code{weak} or
12758 @code{strong}. It defaults to @code{none}.
12764 Compute and draw a color distribution histogram for the input video.
12766 The computed histogram is a representation of the color component
12767 distribution in an image.
12769 Standard histogram displays the color components distribution in an image.
12770 Displays color graph for each color component. Shows distribution of
12771 the Y, U, V, A or R, G, B components, depending on input format, in the
12772 current frame. Below each graph a color component scale meter is shown.
12774 The filter accepts the following options:
12778 Set height of level. Default value is @code{200}.
12779 Allowed range is [50, 2048].
12782 Set height of color scale. Default value is @code{12}.
12783 Allowed range is [0, 40].
12787 It accepts the following values:
12790 Per color component graphs are placed below each other.
12793 Per color component graphs are placed side by side.
12796 Presents information identical to that in the @code{parade}, except
12797 that the graphs representing color components are superimposed directly
12800 Default is @code{stack}.
12803 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12804 Default is @code{linear}.
12807 Set what color components to display.
12808 Default is @code{7}.
12811 Set foreground opacity. Default is @code{0.7}.
12814 Set background opacity. Default is @code{0.5}.
12817 @subsection Examples
12822 Calculate and draw histogram:
12824 ffplay -i input -vf histogram
12832 This is a high precision/quality 3d denoise filter. It aims to reduce
12833 image noise, producing smooth images and making still images really
12834 still. It should enhance compressibility.
12836 It accepts the following optional parameters:
12840 A non-negative floating point number which specifies spatial luma strength.
12841 It defaults to 4.0.
12843 @item chroma_spatial
12844 A non-negative floating point number which specifies spatial chroma strength.
12845 It defaults to 3.0*@var{luma_spatial}/4.0.
12848 A floating point number which specifies luma temporal strength. It defaults to
12849 6.0*@var{luma_spatial}/4.0.
12852 A floating point number which specifies chroma temporal strength. It defaults to
12853 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12856 @subsection Commands
12857 This filter supports same @ref{commands} as options.
12858 The command accepts the same syntax of the corresponding option.
12860 If the specified expression is not valid, it is kept at its current
12863 @anchor{hwdownload}
12864 @section hwdownload
12866 Download hardware frames to system memory.
12868 The input must be in hardware frames, and the output a non-hardware format.
12869 Not all formats will be supported on the output - it may be necessary to insert
12870 an additional @option{format} filter immediately following in the graph to get
12871 the output in a supported format.
12875 Map hardware frames to system memory or to another device.
12877 This filter has several different modes of operation; which one is used depends
12878 on the input and output formats:
12881 Hardware frame input, normal frame output
12883 Map the input frames to system memory and pass them to the output. If the
12884 original hardware frame is later required (for example, after overlaying
12885 something else on part of it), the @option{hwmap} filter can be used again
12886 in the next mode to retrieve it.
12888 Normal frame input, hardware frame output
12890 If the input is actually a software-mapped hardware frame, then unmap it -
12891 that is, return the original hardware frame.
12893 Otherwise, a device must be provided. Create new hardware surfaces on that
12894 device for the output, then map them back to the software format at the input
12895 and give those frames to the preceding filter. This will then act like the
12896 @option{hwupload} filter, but may be able to avoid an additional copy when
12897 the input is already in a compatible format.
12899 Hardware frame input and output
12901 A device must be supplied for the output, either directly or with the
12902 @option{derive_device} option. The input and output devices must be of
12903 different types and compatible - the exact meaning of this is
12904 system-dependent, but typically it means that they must refer to the same
12905 underlying hardware context (for example, refer to the same graphics card).
12907 If the input frames were originally created on the output device, then unmap
12908 to retrieve the original frames.
12910 Otherwise, map the frames to the output device - create new hardware frames
12911 on the output corresponding to the frames on the input.
12914 The following additional parameters are accepted:
12918 Set the frame mapping mode. Some combination of:
12921 The mapped frame should be readable.
12923 The mapped frame should be writeable.
12925 The mapping will always overwrite the entire frame.
12927 This may improve performance in some cases, as the original contents of the
12928 frame need not be loaded.
12930 The mapping must not involve any copying.
12932 Indirect mappings to copies of frames are created in some cases where either
12933 direct mapping is not possible or it would have unexpected properties.
12934 Setting this flag ensures that the mapping is direct and will fail if that is
12937 Defaults to @var{read+write} if not specified.
12939 @item derive_device @var{type}
12940 Rather than using the device supplied at initialisation, instead derive a new
12941 device of type @var{type} from the device the input frames exist on.
12944 In a hardware to hardware mapping, map in reverse - create frames in the sink
12945 and map them back to the source. This may be necessary in some cases where
12946 a mapping in one direction is required but only the opposite direction is
12947 supported by the devices being used.
12949 This option is dangerous - it may break the preceding filter in undefined
12950 ways if there are any additional constraints on that filter's output.
12951 Do not use it without fully understanding the implications of its use.
12957 Upload system memory frames to hardware surfaces.
12959 The device to upload to must be supplied when the filter is initialised. If
12960 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12961 option or with the @option{derive_device} option. The input and output devices
12962 must be of different types and compatible - the exact meaning of this is
12963 system-dependent, but typically it means that they must refer to the same
12964 underlying hardware context (for example, refer to the same graphics card).
12966 The following additional parameters are accepted:
12969 @item derive_device @var{type}
12970 Rather than using the device supplied at initialisation, instead derive a new
12971 device of type @var{type} from the device the input frames exist on.
12974 @anchor{hwupload_cuda}
12975 @section hwupload_cuda
12977 Upload system memory frames to a CUDA device.
12979 It accepts the following optional parameters:
12983 The number of the CUDA device to use
12988 Apply a high-quality magnification filter designed for pixel art. This filter
12989 was originally created by Maxim Stepin.
12991 It accepts the following option:
12995 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12996 @code{hq3x} and @code{4} for @code{hq4x}.
12997 Default is @code{3}.
13001 Stack input videos horizontally.
13003 All streams must be of same pixel format and of same height.
13005 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
13006 to create same output.
13008 The filter accepts the following option:
13012 Set number of input streams. Default is 2.
13015 If set to 1, force the output to terminate when the shortest input
13016 terminates. Default value is 0.
13021 Modify the hue and/or the saturation of the input.
13023 It accepts the following parameters:
13027 Specify the hue angle as a number of degrees. It accepts an expression,
13028 and defaults to "0".
13031 Specify the saturation in the [-10,10] range. It accepts an expression and
13035 Specify the hue angle as a number of radians. It accepts an
13036 expression, and defaults to "0".
13039 Specify the brightness in the [-10,10] range. It accepts an expression and
13043 @option{h} and @option{H} are mutually exclusive, and can't be
13044 specified at the same time.
13046 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13047 expressions containing the following constants:
13051 frame count of the input frame starting from 0
13054 presentation timestamp of the input frame expressed in time base units
13057 frame rate of the input video, NAN if the input frame rate is unknown
13060 timestamp expressed in seconds, NAN if the input timestamp is unknown
13063 time base of the input video
13066 @subsection Examples
13070 Set the hue to 90 degrees and the saturation to 1.0:
13076 Same command but expressing the hue in radians:
13082 Rotate hue and make the saturation swing between 0
13083 and 2 over a period of 1 second:
13085 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13089 Apply a 3 seconds saturation fade-in effect starting at 0:
13091 hue="s=min(t/3\,1)"
13094 The general fade-in expression can be written as:
13096 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13100 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13102 hue="s=max(0\, min(1\, (8-t)/3))"
13105 The general fade-out expression can be written as:
13107 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13112 @subsection Commands
13114 This filter supports the following commands:
13120 Modify the hue and/or the saturation and/or brightness of the input video.
13121 The command accepts the same syntax of the corresponding option.
13123 If the specified expression is not valid, it is kept at its current
13127 @section hysteresis
13129 Grow first stream into second stream by connecting components.
13130 This makes it possible to build more robust edge masks.
13132 This filter accepts the following options:
13136 Set which planes will be processed as bitmap, unprocessed planes will be
13137 copied from first stream.
13138 By default value 0xf, all planes will be processed.
13141 Set threshold which is used in filtering. If pixel component value is higher than
13142 this value filter algorithm for connecting components is activated.
13143 By default value is 0.
13146 The @code{hysteresis} filter also supports the @ref{framesync} options.
13150 Detect video interlacing type.
13152 This filter tries to detect if the input frames are interlaced, progressive,
13153 top or bottom field first. It will also try to detect fields that are
13154 repeated between adjacent frames (a sign of telecine).
13156 Single frame detection considers only immediately adjacent frames when classifying each frame.
13157 Multiple frame detection incorporates the classification history of previous frames.
13159 The filter will log these metadata values:
13162 @item single.current_frame
13163 Detected type of current frame using single-frame detection. One of:
13164 ``tff'' (top field first), ``bff'' (bottom field first),
13165 ``progressive'', or ``undetermined''
13168 Cumulative number of frames detected as top field first using single-frame detection.
13171 Cumulative number of frames detected as top field first using multiple-frame detection.
13174 Cumulative number of frames detected as bottom field first using single-frame detection.
13176 @item multiple.current_frame
13177 Detected type of current frame using multiple-frame detection. One of:
13178 ``tff'' (top field first), ``bff'' (bottom field first),
13179 ``progressive'', or ``undetermined''
13182 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13184 @item single.progressive
13185 Cumulative number of frames detected as progressive using single-frame detection.
13187 @item multiple.progressive
13188 Cumulative number of frames detected as progressive using multiple-frame detection.
13190 @item single.undetermined
13191 Cumulative number of frames that could not be classified using single-frame detection.
13193 @item multiple.undetermined
13194 Cumulative number of frames that could not be classified using multiple-frame detection.
13196 @item repeated.current_frame
13197 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13199 @item repeated.neither
13200 Cumulative number of frames with no repeated field.
13203 Cumulative number of frames with the top field repeated from the previous frame's top field.
13205 @item repeated.bottom
13206 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13209 The filter accepts the following options:
13213 Set interlacing threshold.
13215 Set progressive threshold.
13217 Threshold for repeated field detection.
13219 Number of frames after which a given frame's contribution to the
13220 statistics is halved (i.e., it contributes only 0.5 to its
13221 classification). The default of 0 means that all frames seen are given
13222 full weight of 1.0 forever.
13223 @item analyze_interlaced_flag
13224 When this is not 0 then idet will use the specified number of frames to determine
13225 if the interlaced flag is accurate, it will not count undetermined frames.
13226 If the flag is found to be accurate it will be used without any further
13227 computations, if it is found to be inaccurate it will be cleared without any
13228 further computations. This allows inserting the idet filter as a low computational
13229 method to clean up the interlaced flag
13234 Deinterleave or interleave fields.
13236 This filter allows one to process interlaced images fields without
13237 deinterlacing them. Deinterleaving splits the input frame into 2
13238 fields (so called half pictures). Odd lines are moved to the top
13239 half of the output image, even lines to the bottom half.
13240 You can process (filter) them independently and then re-interleave them.
13242 The filter accepts the following options:
13246 @item chroma_mode, c
13247 @item alpha_mode, a
13248 Available values for @var{luma_mode}, @var{chroma_mode} and
13249 @var{alpha_mode} are:
13255 @item deinterleave, d
13256 Deinterleave fields, placing one above the other.
13258 @item interleave, i
13259 Interleave fields. Reverse the effect of deinterleaving.
13261 Default value is @code{none}.
13263 @item luma_swap, ls
13264 @item chroma_swap, cs
13265 @item alpha_swap, as
13266 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13269 @subsection Commands
13271 This filter supports the all above options as @ref{commands}.
13275 Apply inflate effect to the video.
13277 This filter replaces the pixel by the local(3x3) average by taking into account
13278 only values higher than the pixel.
13280 It accepts the following options:
13287 Limit the maximum change for each plane, default is 65535.
13288 If 0, plane will remain unchanged.
13291 @subsection Commands
13293 This filter supports the all above options as @ref{commands}.
13297 Simple interlacing filter from progressive contents. This interleaves upper (or
13298 lower) lines from odd frames with lower (or upper) lines from even frames,
13299 halving the frame rate and preserving image height.
13302 Original Original New Frame
13303 Frame 'j' Frame 'j+1' (tff)
13304 ========== =========== ==================
13305 Line 0 --------------------> Frame 'j' Line 0
13306 Line 1 Line 1 ----> Frame 'j+1' Line 1
13307 Line 2 ---------------------> Frame 'j' Line 2
13308 Line 3 Line 3 ----> Frame 'j+1' Line 3
13310 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13313 It accepts the following optional parameters:
13317 This determines whether the interlaced frame is taken from the even
13318 (tff - default) or odd (bff) lines of the progressive frame.
13321 Vertical lowpass filter to avoid twitter interlacing and
13322 reduce moire patterns.
13326 Disable vertical lowpass filter
13329 Enable linear filter (default)
13332 Enable complex filter. This will slightly less reduce twitter and moire
13333 but better retain detail and subjective sharpness impression.
13340 Deinterlace input video by applying Donald Graft's adaptive kernel
13341 deinterling. Work on interlaced parts of a video to produce
13342 progressive frames.
13344 The description of the accepted parameters follows.
13348 Set the threshold which affects the filter's tolerance when
13349 determining if a pixel line must be processed. It must be an integer
13350 in the range [0,255] and defaults to 10. A value of 0 will result in
13351 applying the process on every pixels.
13354 Paint pixels exceeding the threshold value to white if set to 1.
13358 Set the fields order. Swap fields if set to 1, leave fields alone if
13362 Enable additional sharpening if set to 1. Default is 0.
13365 Enable twoway sharpening if set to 1. Default is 0.
13368 @subsection Examples
13372 Apply default values:
13374 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13378 Enable additional sharpening:
13384 Paint processed pixels in white:
13392 Slowly update darker pixels.
13394 This filter makes short flashes of light appear longer.
13395 This filter accepts the following options:
13399 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13402 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13405 @section lenscorrection
13407 Correct radial lens distortion
13409 This filter can be used to correct for radial distortion as can result from the use
13410 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13411 one can use tools available for example as part of opencv or simply trial-and-error.
13412 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13413 and extract the k1 and k2 coefficients from the resulting matrix.
13415 Note that effectively the same filter is available in the open-source tools Krita and
13416 Digikam from the KDE project.
13418 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13419 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13420 brightness distribution, so you may want to use both filters together in certain
13421 cases, though you will have to take care of ordering, i.e. whether vignetting should
13422 be applied before or after lens correction.
13424 @subsection Options
13426 The filter accepts the following options:
13430 Relative x-coordinate of the focal point of the image, and thereby the center of the
13431 distortion. This value has a range [0,1] and is expressed as fractions of the image
13432 width. Default is 0.5.
13434 Relative y-coordinate of the focal point of the image, and thereby the center of the
13435 distortion. This value has a range [0,1] and is expressed as fractions of the image
13436 height. Default is 0.5.
13438 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13439 no correction. Default is 0.
13441 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13442 0 means no correction. Default is 0.
13444 Set interpolation type. Can be @code{nearest} or @code{bilinear}.
13445 Default is @code{nearest}.
13448 The formula that generates the correction is:
13450 @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)
13452 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13453 distances from the focal point in the source and target images, respectively.
13457 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13459 The @code{lensfun} filter requires the camera make, camera model, and lens model
13460 to apply the lens correction. The filter will load the lensfun database and
13461 query it to find the corresponding camera and lens entries in the database. As
13462 long as these entries can be found with the given options, the filter can
13463 perform corrections on frames. Note that incomplete strings will result in the
13464 filter choosing the best match with the given options, and the filter will
13465 output the chosen camera and lens models (logged with level "info"). You must
13466 provide the make, camera model, and lens model as they are required.
13468 The filter accepts the following options:
13472 The make of the camera (for example, "Canon"). This option is required.
13475 The model of the camera (for example, "Canon EOS 100D"). This option is
13479 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13480 option is required.
13483 The type of correction to apply. The following values are valid options:
13487 Enables fixing lens vignetting.
13490 Enables fixing lens geometry. This is the default.
13493 Enables fixing chromatic aberrations.
13496 Enables fixing lens vignetting and lens geometry.
13499 Enables fixing lens vignetting and chromatic aberrations.
13502 Enables fixing both lens geometry and chromatic aberrations.
13505 Enables all possible corrections.
13509 The focal length of the image/video (zoom; expected constant for video). For
13510 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13511 range should be chosen when using that lens. Default 18.
13514 The aperture of the image/video (expected constant for video). Note that
13515 aperture is only used for vignetting correction. Default 3.5.
13517 @item focus_distance
13518 The focus distance of the image/video (expected constant for video). Note that
13519 focus distance is only used for vignetting and only slightly affects the
13520 vignetting correction process. If unknown, leave it at the default value (which
13524 The scale factor which is applied after transformation. After correction the
13525 video is no longer necessarily rectangular. This parameter controls how much of
13526 the resulting image is visible. The value 0 means that a value will be chosen
13527 automatically such that there is little or no unmapped area in the output
13528 image. 1.0 means that no additional scaling is done. Lower values may result
13529 in more of the corrected image being visible, while higher values may avoid
13530 unmapped areas in the output.
13532 @item target_geometry
13533 The target geometry of the output image/video. The following values are valid
13537 @item rectilinear (default)
13540 @item equirectangular
13541 @item fisheye_orthographic
13542 @item fisheye_stereographic
13543 @item fisheye_equisolid
13544 @item fisheye_thoby
13547 Apply the reverse of image correction (instead of correcting distortion, apply
13550 @item interpolation
13551 The type of interpolation used when correcting distortion. The following values
13556 @item linear (default)
13561 @subsection Examples
13565 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13566 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13570 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
13574 Apply the same as before, but only for the first 5 seconds of video.
13577 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
13584 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13585 score between two input videos.
13587 The obtained VMAF score is printed through the logging system.
13589 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13590 After installing the library it can be enabled using:
13591 @code{./configure --enable-libvmaf}.
13592 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13594 The filter has following options:
13598 Set the model path which is to be used for SVM.
13599 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13602 Set the file path to be used to store logs.
13605 Set the format of the log file (csv, json or xml).
13607 @item enable_transform
13608 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13609 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13610 Default value: @code{false}
13613 Invokes the phone model which will generate VMAF scores higher than in the
13614 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13615 Default value: @code{false}
13618 Enables computing psnr along with vmaf.
13619 Default value: @code{false}
13622 Enables computing ssim along with vmaf.
13623 Default value: @code{false}
13626 Enables computing ms_ssim along with vmaf.
13627 Default value: @code{false}
13630 Set the pool method to be used for computing vmaf.
13631 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13634 Set number of threads to be used when computing vmaf.
13635 Default value: @code{0}, which makes use of all available logical processors.
13638 Set interval for frame subsampling used when computing vmaf.
13639 Default value: @code{1}
13641 @item enable_conf_interval
13642 Enables confidence interval.
13643 Default value: @code{false}
13646 This filter also supports the @ref{framesync} options.
13648 @subsection Examples
13651 On the below examples the input file @file{main.mpg} being processed is
13652 compared with the reference file @file{ref.mpg}.
13655 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13659 Example with options:
13661 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13665 Example with options and different containers:
13667 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 -
13673 Limits the pixel components values to the specified range [min, max].
13675 The filter accepts the following options:
13679 Lower bound. Defaults to the lowest allowed value for the input.
13682 Upper bound. Defaults to the highest allowed value for the input.
13685 Specify which planes will be processed. Defaults to all available.
13688 @subsection Commands
13690 This filter supports the all above options as @ref{commands}.
13696 The filter accepts the following options:
13700 Set the number of loops. Setting this value to -1 will result in infinite loops.
13704 Set maximal size in number of frames. Default is 0.
13707 Set first frame of loop. Default is 0.
13710 @subsection Examples
13714 Loop single first frame infinitely:
13716 loop=loop=-1:size=1:start=0
13720 Loop single first frame 10 times:
13722 loop=loop=10:size=1:start=0
13726 Loop 10 first frames 5 times:
13728 loop=loop=5:size=10:start=0
13734 Apply a 1D LUT to an input video.
13736 The filter accepts the following options:
13740 Set the 1D LUT file name.
13742 Currently supported formats:
13751 Select interpolation mode.
13753 Available values are:
13757 Use values from the nearest defined point.
13759 Interpolate values using the linear interpolation.
13761 Interpolate values using the cosine interpolation.
13763 Interpolate values using the cubic interpolation.
13765 Interpolate values using the spline interpolation.
13772 Apply a 3D LUT to an input video.
13774 The filter accepts the following options:
13778 Set the 3D LUT file name.
13780 Currently supported formats:
13794 Select interpolation mode.
13796 Available values are:
13800 Use values from the nearest defined point.
13802 Interpolate values using the 8 points defining a cube.
13804 Interpolate values using a tetrahedron.
13810 Turn certain luma values into transparency.
13812 The filter accepts the following options:
13816 Set the luma which will be used as base for transparency.
13817 Default value is @code{0}.
13820 Set the range of luma values to be keyed out.
13821 Default value is @code{0.01}.
13824 Set the range of softness. Default value is @code{0}.
13825 Use this to control gradual transition from zero to full transparency.
13828 @subsection Commands
13829 This filter supports same @ref{commands} as options.
13830 The command accepts the same syntax of the corresponding option.
13832 If the specified expression is not valid, it is kept at its current
13835 @section lut, lutrgb, lutyuv
13837 Compute a look-up table for binding each pixel component input value
13838 to an output value, and apply it to the input video.
13840 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13841 to an RGB input video.
13843 These filters accept the following parameters:
13846 set first pixel component expression
13848 set second pixel component expression
13850 set third pixel component expression
13852 set fourth pixel component expression, corresponds to the alpha component
13855 set red component expression
13857 set green component expression
13859 set blue component expression
13861 alpha component expression
13864 set Y/luminance component expression
13866 set U/Cb component expression
13868 set V/Cr component expression
13871 Each of them specifies the expression to use for computing the lookup table for
13872 the corresponding pixel component values.
13874 The exact component associated to each of the @var{c*} options depends on the
13877 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13878 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13880 The expressions can contain the following constants and functions:
13885 The input width and height.
13888 The input value for the pixel component.
13891 The input value, clipped to the @var{minval}-@var{maxval} range.
13894 The maximum value for the pixel component.
13897 The minimum value for the pixel component.
13900 The negated value for the pixel component value, clipped to the
13901 @var{minval}-@var{maxval} range; it corresponds to the expression
13902 "maxval-clipval+minval".
13905 The computed value in @var{val}, clipped to the
13906 @var{minval}-@var{maxval} range.
13908 @item gammaval(gamma)
13909 The computed gamma correction value of the pixel component value,
13910 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13912 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13916 All expressions default to "val".
13918 @subsection Examples
13922 Negate input video:
13924 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13925 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13928 The above is the same as:
13930 lutrgb="r=negval:g=negval:b=negval"
13931 lutyuv="y=negval:u=negval:v=negval"
13941 Remove chroma components, turning the video into a graytone image:
13943 lutyuv="u=128:v=128"
13947 Apply a luma burning effect:
13953 Remove green and blue components:
13959 Set a constant alpha channel value on input:
13961 format=rgba,lutrgb=a="maxval-minval/2"
13965 Correct luminance gamma by a factor of 0.5:
13967 lutyuv=y=gammaval(0.5)
13971 Discard least significant bits of luma:
13973 lutyuv=y='bitand(val, 128+64+32)'
13977 Technicolor like effect:
13979 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13983 @section lut2, tlut2
13985 The @code{lut2} filter takes two input streams and outputs one
13988 The @code{tlut2} (time lut2) filter takes two consecutive frames
13989 from one single stream.
13991 This filter accepts the following parameters:
13994 set first pixel component expression
13996 set second pixel component expression
13998 set third pixel component expression
14000 set fourth pixel component expression, corresponds to the alpha component
14003 set output bit depth, only available for @code{lut2} filter. By default is 0,
14004 which means bit depth is automatically picked from first input format.
14007 The @code{lut2} filter also supports the @ref{framesync} options.
14009 Each of them specifies the expression to use for computing the lookup table for
14010 the corresponding pixel component values.
14012 The exact component associated to each of the @var{c*} options depends on the
14015 The expressions can contain the following constants:
14020 The input width and height.
14023 The first input value for the pixel component.
14026 The second input value for the pixel component.
14029 The first input video bit depth.
14032 The second input video bit depth.
14035 All expressions default to "x".
14037 @subsection Examples
14041 Highlight differences between two RGB video streams:
14043 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)'
14047 Highlight differences between two YUV video streams:
14049 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)'
14053 Show max difference between two video streams:
14055 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)))'
14059 @section maskedclamp
14061 Clamp the first input stream with the second input and third input stream.
14063 Returns the value of first stream to be between second input
14064 stream - @code{undershoot} and third input stream + @code{overshoot}.
14066 This filter accepts the following options:
14069 Default value is @code{0}.
14072 Default value is @code{0}.
14075 Set which planes will be processed as bitmap, unprocessed planes will be
14076 copied from first stream.
14077 By default value 0xf, all planes will be processed.
14080 @subsection Commands
14082 This filter supports the all above options as @ref{commands}.
14086 Merge the second and third input stream into output stream using absolute differences
14087 between second input stream and first input stream and absolute difference between
14088 third input stream and first input stream. The picked value will be from second input
14089 stream if second absolute difference is greater than first one or from third input stream
14092 This filter accepts the following options:
14095 Set which planes will be processed as bitmap, unprocessed planes will be
14096 copied from first stream.
14097 By default value 0xf, all planes will be processed.
14100 @subsection Commands
14102 This filter supports the all above options as @ref{commands}.
14104 @section maskedmerge
14106 Merge the first input stream with the second input stream using per pixel
14107 weights in the third input stream.
14109 A value of 0 in the third stream pixel component means that pixel component
14110 from first stream is returned unchanged, while maximum value (eg. 255 for
14111 8-bit videos) means that pixel component from second stream is returned
14112 unchanged. Intermediate values define the amount of merging between both
14113 input stream's pixel components.
14115 This filter accepts the following options:
14118 Set which planes will be processed as bitmap, unprocessed planes will be
14119 copied from first stream.
14120 By default value 0xf, all planes will be processed.
14123 @subsection Commands
14125 This filter supports the all above options as @ref{commands}.
14129 Merge the second and third input stream into output stream using absolute differences
14130 between second input stream and first input stream and absolute difference between
14131 third input stream and first input stream. The picked value will be from second input
14132 stream if second absolute difference is less than first one or from third input stream
14135 This filter accepts the following options:
14138 Set which planes will be processed as bitmap, unprocessed planes will be
14139 copied from first stream.
14140 By default value 0xf, all planes will be processed.
14143 @subsection Commands
14145 This filter supports the all above options as @ref{commands}.
14147 @section maskedthreshold
14148 Pick pixels comparing absolute difference of two video streams with fixed
14151 If absolute difference between pixel component of first and second video
14152 stream is equal or lower than user supplied threshold than pixel component
14153 from first video stream is picked, otherwise pixel component from second
14154 video stream is picked.
14156 This filter accepts the following options:
14159 Set threshold used when picking pixels from absolute difference from two input
14163 Set which planes will be processed as bitmap, unprocessed planes will be
14164 copied from second stream.
14165 By default value 0xf, all planes will be processed.
14168 @subsection Commands
14170 This filter supports the all above options as @ref{commands}.
14173 Create mask from input video.
14175 For example it is useful to create motion masks after @code{tblend} filter.
14177 This filter accepts the following options:
14181 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14184 Set high threshold. Any pixel component higher than this value will be set to max value
14185 allowed for current pixel format.
14188 Set planes to filter, by default all available planes are filtered.
14191 Fill all frame pixels with this value.
14194 Set max average pixel value for frame. If sum of all pixel components is higher that this
14195 average, output frame will be completely filled with value set by @var{fill} option.
14196 Typically useful for scene changes when used in combination with @code{tblend} filter.
14201 Apply motion-compensation deinterlacing.
14203 It needs one field per frame as input and must thus be used together
14204 with yadif=1/3 or equivalent.
14206 This filter accepts the following options:
14209 Set the deinterlacing mode.
14211 It accepts one of the following values:
14216 use iterative motion estimation
14218 like @samp{slow}, but use multiple reference frames.
14220 Default value is @samp{fast}.
14223 Set the picture field parity assumed for the input video. It must be
14224 one of the following values:
14228 assume top field first
14230 assume bottom field first
14233 Default value is @samp{bff}.
14236 Set per-block quantization parameter (QP) used by the internal
14239 Higher values should result in a smoother motion vector field but less
14240 optimal individual vectors. Default value is 1.
14245 Pick median pixel from certain rectangle defined by radius.
14247 This filter accepts the following options:
14251 Set horizontal radius size. Default value is @code{1}.
14252 Allowed range is integer from 1 to 127.
14255 Set which planes to process. Default is @code{15}, which is all available planes.
14258 Set vertical radius size. Default value is @code{0}.
14259 Allowed range is integer from 0 to 127.
14260 If it is 0, value will be picked from horizontal @code{radius} option.
14263 Set median percentile. Default value is @code{0.5}.
14264 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14265 minimum values, and @code{1} maximum values.
14268 @subsection Commands
14269 This filter supports same @ref{commands} as options.
14270 The command accepts the same syntax of the corresponding option.
14272 If the specified expression is not valid, it is kept at its current
14275 @section mergeplanes
14277 Merge color channel components from several video streams.
14279 The filter accepts up to 4 input streams, and merge selected input
14280 planes to the output video.
14282 This filter accepts the following options:
14285 Set input to output plane mapping. Default is @code{0}.
14287 The mappings is specified as a bitmap. It should be specified as a
14288 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14289 mapping for the first plane of the output stream. 'A' sets the number of
14290 the input stream to use (from 0 to 3), and 'a' the plane number of the
14291 corresponding input to use (from 0 to 3). The rest of the mappings is
14292 similar, 'Bb' describes the mapping for the output stream second
14293 plane, 'Cc' describes the mapping for the output stream third plane and
14294 'Dd' describes the mapping for the output stream fourth plane.
14297 Set output pixel format. Default is @code{yuva444p}.
14300 @subsection Examples
14304 Merge three gray video streams of same width and height into single video stream:
14306 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14310 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14312 [a0][a1]mergeplanes=0x00010210:yuva444p
14316 Swap Y and A plane in yuva444p stream:
14318 format=yuva444p,mergeplanes=0x03010200:yuva444p
14322 Swap U and V plane in yuv420p stream:
14324 format=yuv420p,mergeplanes=0x000201:yuv420p
14328 Cast a rgb24 clip to yuv444p:
14330 format=rgb24,mergeplanes=0x000102:yuv444p
14336 Estimate and export motion vectors using block matching algorithms.
14337 Motion vectors are stored in frame side data to be used by other filters.
14339 This filter accepts the following options:
14342 Specify the motion estimation method. Accepts one of the following values:
14346 Exhaustive search algorithm.
14348 Three step search algorithm.
14350 Two dimensional logarithmic search algorithm.
14352 New three step search algorithm.
14354 Four step search algorithm.
14356 Diamond search algorithm.
14358 Hexagon-based search algorithm.
14360 Enhanced predictive zonal search algorithm.
14362 Uneven multi-hexagon search algorithm.
14364 Default value is @samp{esa}.
14367 Macroblock size. Default @code{16}.
14370 Search parameter. Default @code{7}.
14373 @section midequalizer
14375 Apply Midway Image Equalization effect using two video streams.
14377 Midway Image Equalization adjusts a pair of images to have the same
14378 histogram, while maintaining their dynamics as much as possible. It's
14379 useful for e.g. matching exposures from a pair of stereo cameras.
14381 This filter has two inputs and one output, which must be of same pixel format, but
14382 may be of different sizes. The output of filter is first input adjusted with
14383 midway histogram of both inputs.
14385 This filter accepts the following option:
14389 Set which planes to process. Default is @code{15}, which is all available planes.
14392 @section minterpolate
14394 Convert the video to specified frame rate using motion interpolation.
14396 This filter accepts the following options:
14399 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}.
14402 Motion interpolation mode. Following values are accepted:
14405 Duplicate previous or next frame for interpolating new ones.
14407 Blend source frames. Interpolated frame is mean of previous and next frames.
14409 Motion compensated interpolation. Following options are effective when this mode is selected:
14413 Motion compensation mode. Following values are accepted:
14416 Overlapped block motion compensation.
14418 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14420 Default mode is @samp{obmc}.
14423 Motion estimation mode. Following values are accepted:
14426 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14428 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14430 Default mode is @samp{bilat}.
14433 The algorithm to be used for motion estimation. Following values are accepted:
14436 Exhaustive search algorithm.
14438 Three step search algorithm.
14440 Two dimensional logarithmic search algorithm.
14442 New three step search algorithm.
14444 Four step search algorithm.
14446 Diamond search algorithm.
14448 Hexagon-based search algorithm.
14450 Enhanced predictive zonal search algorithm.
14452 Uneven multi-hexagon search algorithm.
14454 Default algorithm is @samp{epzs}.
14457 Macroblock size. Default @code{16}.
14460 Motion estimation search parameter. Default @code{32}.
14463 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).
14468 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:
14471 Disable scene change detection.
14473 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14475 Default method is @samp{fdiff}.
14477 @item scd_threshold
14478 Scene change detection threshold. Default is @code{10.}.
14483 Mix several video input streams into one video stream.
14485 A description of the accepted options follows.
14489 The number of inputs. If unspecified, it defaults to 2.
14492 Specify weight of each input video stream as sequence.
14493 Each weight is separated by space. If number of weights
14494 is smaller than number of @var{frames} last specified
14495 weight will be used for all remaining unset weights.
14498 Specify scale, if it is set it will be multiplied with sum
14499 of each weight multiplied with pixel values to give final destination
14500 pixel value. By default @var{scale} is auto scaled to sum of weights.
14503 Specify how end of stream is determined.
14506 The duration of the longest input. (default)
14509 The duration of the shortest input.
14512 The duration of the first input.
14516 @section mpdecimate
14518 Drop frames that do not differ greatly from the previous frame in
14519 order to reduce frame rate.
14521 The main use of this filter is for very-low-bitrate encoding
14522 (e.g. streaming over dialup modem), but it could in theory be used for
14523 fixing movies that were inverse-telecined incorrectly.
14525 A description of the accepted options follows.
14529 Set the maximum number of consecutive frames which can be dropped (if
14530 positive), or the minimum interval between dropped frames (if
14531 negative). If the value is 0, the frame is dropped disregarding the
14532 number of previous sequentially dropped frames.
14534 Default value is 0.
14539 Set the dropping threshold values.
14541 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14542 represent actual pixel value differences, so a threshold of 64
14543 corresponds to 1 unit of difference for each pixel, or the same spread
14544 out differently over the block.
14546 A frame is a candidate for dropping if no 8x8 blocks differ by more
14547 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14548 meaning the whole image) differ by more than a threshold of @option{lo}.
14550 Default value for @option{hi} is 64*12, default value for @option{lo} is
14551 64*5, and default value for @option{frac} is 0.33.
14557 Negate (invert) the input video.
14559 It accepts the following option:
14564 With value 1, it negates the alpha component, if present. Default value is 0.
14570 Denoise frames using Non-Local Means algorithm.
14572 Each pixel is adjusted by looking for other pixels with similar contexts. This
14573 context similarity is defined by comparing their surrounding patches of size
14574 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14577 Note that the research area defines centers for patches, which means some
14578 patches will be made of pixels outside that research area.
14580 The filter accepts the following options.
14584 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14587 Set patch size. Default is 7. Must be odd number in range [0, 99].
14590 Same as @option{p} but for chroma planes.
14592 The default value is @var{0} and means automatic.
14595 Set research size. Default is 15. Must be odd number in range [0, 99].
14598 Same as @option{r} but for chroma planes.
14600 The default value is @var{0} and means automatic.
14605 Deinterlace video using neural network edge directed interpolation.
14607 This filter accepts the following options:
14611 Mandatory option, without binary file filter can not work.
14612 Currently file can be found here:
14613 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14616 Set which frames to deinterlace, by default it is @code{all}.
14617 Can be @code{all} or @code{interlaced}.
14620 Set mode of operation.
14622 Can be one of the following:
14626 Use frame flags, both fields.
14628 Use frame flags, single field.
14630 Use top field only.
14632 Use bottom field only.
14634 Use both fields, top first.
14636 Use both fields, bottom first.
14640 Set which planes to process, by default filter process all frames.
14643 Set size of local neighborhood around each pixel, used by the predictor neural
14646 Can be one of the following:
14659 Set the number of neurons in predictor neural network.
14660 Can be one of the following:
14671 Controls the number of different neural network predictions that are blended
14672 together to compute the final output value. Can be @code{fast}, default or
14676 Set which set of weights to use in the predictor.
14677 Can be one of the following:
14681 weights trained to minimize absolute error
14683 weights trained to minimize squared error
14687 Controls whether or not the prescreener neural network is used to decide
14688 which pixels should be processed by the predictor neural network and which
14689 can be handled by simple cubic interpolation.
14690 The prescreener is trained to know whether cubic interpolation will be
14691 sufficient for a pixel or whether it should be predicted by the predictor nn.
14692 The computational complexity of the prescreener nn is much less than that of
14693 the predictor nn. Since most pixels can be handled by cubic interpolation,
14694 using the prescreener generally results in much faster processing.
14695 The prescreener is pretty accurate, so the difference between using it and not
14696 using it is almost always unnoticeable.
14698 Can be one of the following:
14708 Default is @code{new}.
14711 @subsection Commands
14712 This filter supports same @ref{commands} as options, excluding @var{weights} option.
14716 Force libavfilter not to use any of the specified pixel formats for the
14717 input to the next filter.
14719 It accepts the following parameters:
14723 A '|'-separated list of pixel format names, such as
14724 pix_fmts=yuv420p|monow|rgb24".
14728 @subsection Examples
14732 Force libavfilter to use a format different from @var{yuv420p} for the
14733 input to the vflip filter:
14735 noformat=pix_fmts=yuv420p,vflip
14739 Convert the input video to any of the formats not contained in the list:
14741 noformat=yuv420p|yuv444p|yuv410p
14747 Add noise on video input frame.
14749 The filter accepts the following options:
14757 Set noise seed for specific pixel component or all pixel components in case
14758 of @var{all_seed}. Default value is @code{123457}.
14760 @item all_strength, alls
14761 @item c0_strength, c0s
14762 @item c1_strength, c1s
14763 @item c2_strength, c2s
14764 @item c3_strength, c3s
14765 Set noise strength for specific pixel component or all pixel components in case
14766 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14768 @item all_flags, allf
14769 @item c0_flags, c0f
14770 @item c1_flags, c1f
14771 @item c2_flags, c2f
14772 @item c3_flags, c3f
14773 Set pixel component flags or set flags for all components if @var{all_flags}.
14774 Available values for component flags are:
14777 averaged temporal noise (smoother)
14779 mix random noise with a (semi)regular pattern
14781 temporal noise (noise pattern changes between frames)
14783 uniform noise (gaussian otherwise)
14787 @subsection Examples
14789 Add temporal and uniform noise to input video:
14791 noise=alls=20:allf=t+u
14796 Normalize RGB video (aka histogram stretching, contrast stretching).
14797 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14799 For each channel of each frame, the filter computes the input range and maps
14800 it linearly to the user-specified output range. The output range defaults
14801 to the full dynamic range from pure black to pure white.
14803 Temporal smoothing can be used on the input range to reduce flickering (rapid
14804 changes in brightness) caused when small dark or bright objects enter or leave
14805 the scene. This is similar to the auto-exposure (automatic gain control) on a
14806 video camera, and, like a video camera, it may cause a period of over- or
14807 under-exposure of the video.
14809 The R,G,B channels can be normalized independently, which may cause some
14810 color shifting, or linked together as a single channel, which prevents
14811 color shifting. Linked normalization preserves hue. Independent normalization
14812 does not, so it can be used to remove some color casts. Independent and linked
14813 normalization can be combined in any ratio.
14815 The normalize filter accepts the following options:
14820 Colors which define the output range. The minimum input value is mapped to
14821 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14822 The defaults are black and white respectively. Specifying white for
14823 @var{blackpt} and black for @var{whitept} will give color-inverted,
14824 normalized video. Shades of grey can be used to reduce the dynamic range
14825 (contrast). Specifying saturated colors here can create some interesting
14829 The number of previous frames to use for temporal smoothing. The input range
14830 of each channel is smoothed using a rolling average over the current frame
14831 and the @var{smoothing} previous frames. The default is 0 (no temporal
14835 Controls the ratio of independent (color shifting) channel normalization to
14836 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14837 independent. Defaults to 1.0 (fully independent).
14840 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14841 expensive no-op. Defaults to 1.0 (full strength).
14845 @subsection Commands
14846 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14847 The command accepts the same syntax of the corresponding option.
14849 If the specified expression is not valid, it is kept at its current
14852 @subsection Examples
14854 Stretch video contrast to use the full dynamic range, with no temporal
14855 smoothing; may flicker depending on the source content:
14857 normalize=blackpt=black:whitept=white:smoothing=0
14860 As above, but with 50 frames of temporal smoothing; flicker should be
14861 reduced, depending on the source content:
14863 normalize=blackpt=black:whitept=white:smoothing=50
14866 As above, but with hue-preserving linked channel normalization:
14868 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14871 As above, but with half strength:
14873 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14876 Map the darkest input color to red, the brightest input color to cyan:
14878 normalize=blackpt=red:whitept=cyan
14883 Pass the video source unchanged to the output.
14886 Optical Character Recognition
14888 This filter uses Tesseract for optical character recognition. To enable
14889 compilation of this filter, you need to configure FFmpeg with
14890 @code{--enable-libtesseract}.
14892 It accepts the following options:
14896 Set datapath to tesseract data. Default is to use whatever was
14897 set at installation.
14900 Set language, default is "eng".
14903 Set character whitelist.
14906 Set character blacklist.
14909 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14910 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14914 Apply a video transform using libopencv.
14916 To enable this filter, install the libopencv library and headers and
14917 configure FFmpeg with @code{--enable-libopencv}.
14919 It accepts the following parameters:
14924 The name of the libopencv filter to apply.
14926 @item filter_params
14927 The parameters to pass to the libopencv filter. If not specified, the default
14928 values are assumed.
14932 Refer to the official libopencv documentation for more precise
14934 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14936 Several libopencv filters are supported; see the following subsections.
14941 Dilate an image by using a specific structuring element.
14942 It corresponds to the libopencv function @code{cvDilate}.
14944 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14946 @var{struct_el} represents a structuring element, and has the syntax:
14947 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14949 @var{cols} and @var{rows} represent the number of columns and rows of
14950 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14951 point, and @var{shape} the shape for the structuring element. @var{shape}
14952 must be "rect", "cross", "ellipse", or "custom".
14954 If the value for @var{shape} is "custom", it must be followed by a
14955 string of the form "=@var{filename}". The file with name
14956 @var{filename} is assumed to represent a binary image, with each
14957 printable character corresponding to a bright pixel. When a custom
14958 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14959 or columns and rows of the read file are assumed instead.
14961 The default value for @var{struct_el} is "3x3+0x0/rect".
14963 @var{nb_iterations} specifies the number of times the transform is
14964 applied to the image, and defaults to 1.
14968 # Use the default values
14971 # Dilate using a structuring element with a 5x5 cross, iterating two times
14972 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14974 # Read the shape from the file diamond.shape, iterating two times.
14975 # The file diamond.shape may contain a pattern of characters like this
14981 # The specified columns and rows are ignored
14982 # but the anchor point coordinates are not
14983 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14988 Erode an image by using a specific structuring element.
14989 It corresponds to the libopencv function @code{cvErode}.
14991 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14992 with the same syntax and semantics as the @ref{dilate} filter.
14996 Smooth the input video.
14998 The filter takes the following parameters:
14999 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
15001 @var{type} is the type of smooth filter to apply, and must be one of
15002 the following values: "blur", "blur_no_scale", "median", "gaussian",
15003 or "bilateral". The default value is "gaussian".
15005 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
15006 depends on the smooth type. @var{param1} and
15007 @var{param2} accept integer positive values or 0. @var{param3} and
15008 @var{param4} accept floating point values.
15010 The default value for @var{param1} is 3. The default value for the
15011 other parameters is 0.
15013 These parameters correspond to the parameters assigned to the
15014 libopencv function @code{cvSmooth}.
15016 @section oscilloscope
15018 2D Video Oscilloscope.
15020 Useful to measure spatial impulse, step responses, chroma delays, etc.
15022 It accepts the following parameters:
15026 Set scope center x position.
15029 Set scope center y position.
15032 Set scope size, relative to frame diagonal.
15035 Set scope tilt/rotation.
15041 Set trace center x position.
15044 Set trace center y position.
15047 Set trace width, relative to width of frame.
15050 Set trace height, relative to height of frame.
15053 Set which components to trace. By default it traces first three components.
15056 Draw trace grid. By default is enabled.
15059 Draw some statistics. By default is enabled.
15062 Draw scope. By default is enabled.
15065 @subsection Commands
15066 This filter supports same @ref{commands} as options.
15067 The command accepts the same syntax of the corresponding option.
15069 If the specified expression is not valid, it is kept at its current
15072 @subsection Examples
15076 Inspect full first row of video frame.
15078 oscilloscope=x=0.5:y=0:s=1
15082 Inspect full last row of video frame.
15084 oscilloscope=x=0.5:y=1:s=1
15088 Inspect full 5th line of video frame of height 1080.
15090 oscilloscope=x=0.5:y=5/1080:s=1
15094 Inspect full last column of video frame.
15096 oscilloscope=x=1:y=0.5:s=1:t=1
15104 Overlay one video on top of another.
15106 It takes two inputs and has one output. The first input is the "main"
15107 video on which the second input is overlaid.
15109 It accepts the following parameters:
15111 A description of the accepted options follows.
15116 Set the expression for the x and y coordinates of the overlaid video
15117 on the main video. Default value is "0" for both expressions. In case
15118 the expression is invalid, it is set to a huge value (meaning that the
15119 overlay will not be displayed within the output visible area).
15122 See @ref{framesync}.
15125 Set when the expressions for @option{x}, and @option{y} are evaluated.
15127 It accepts the following values:
15130 only evaluate expressions once during the filter initialization or
15131 when a command is processed
15134 evaluate expressions for each incoming frame
15137 Default value is @samp{frame}.
15140 See @ref{framesync}.
15143 Set the format for the output video.
15145 It accepts the following values:
15148 force YUV420 output
15151 force YUV420p10 output
15154 force YUV422 output
15157 force YUV422p10 output
15160 force YUV444 output
15163 force packed RGB output
15166 force planar RGB output
15169 automatically pick format
15172 Default value is @samp{yuv420}.
15175 See @ref{framesync}.
15178 Set format of alpha of the overlaid video, it can be @var{straight} or
15179 @var{premultiplied}. Default is @var{straight}.
15182 The @option{x}, and @option{y} expressions can contain the following
15188 The main input width and height.
15192 The overlay input width and height.
15196 The computed values for @var{x} and @var{y}. They are evaluated for
15201 horizontal and vertical chroma subsample values of the output
15202 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15206 the number of input frame, starting from 0
15209 the position in the file of the input frame, NAN if unknown
15212 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15216 This filter also supports the @ref{framesync} options.
15218 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15219 when evaluation is done @emph{per frame}, and will evaluate to NAN
15220 when @option{eval} is set to @samp{init}.
15222 Be aware that frames are taken from each input video in timestamp
15223 order, hence, if their initial timestamps differ, it is a good idea
15224 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15225 have them begin in the same zero timestamp, as the example for
15226 the @var{movie} filter does.
15228 You can chain together more overlays but you should test the
15229 efficiency of such approach.
15231 @subsection Commands
15233 This filter supports the following commands:
15237 Modify the x and y of the overlay input.
15238 The command accepts the same syntax of the corresponding option.
15240 If the specified expression is not valid, it is kept at its current
15244 @subsection Examples
15248 Draw the overlay at 10 pixels from the bottom right corner of the main
15251 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15254 Using named options the example above becomes:
15256 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15260 Insert a transparent PNG logo in the bottom left corner of the input,
15261 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15263 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15267 Insert 2 different transparent PNG logos (second logo on bottom
15268 right corner) using the @command{ffmpeg} tool:
15270 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
15274 Add a transparent color layer on top of the main video; @code{WxH}
15275 must specify the size of the main input to the overlay filter:
15277 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15281 Play an original video and a filtered version (here with the deshake
15282 filter) side by side using the @command{ffplay} tool:
15284 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15287 The above command is the same as:
15289 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15293 Make a sliding overlay appearing from the left to the right top part of the
15294 screen starting since time 2:
15296 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15300 Compose output by putting two input videos side to side:
15302 ffmpeg -i left.avi -i right.avi -filter_complex "
15303 nullsrc=size=200x100 [background];
15304 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15305 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15306 [background][left] overlay=shortest=1 [background+left];
15307 [background+left][right] overlay=shortest=1:x=100 [left+right]
15312 Mask 10-20 seconds of a video by applying the delogo filter to a section
15314 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15315 -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]'
15320 Chain several overlays in cascade:
15322 nullsrc=s=200x200 [bg];
15323 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15324 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15325 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15326 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15327 [in3] null, [mid2] overlay=100:100 [out0]
15332 @anchor{overlay_cuda}
15333 @section overlay_cuda
15335 Overlay one video on top of another.
15337 This is the CUDA variant of the @ref{overlay} filter.
15338 It only accepts CUDA frames. The underlying input pixel formats have to match.
15340 It takes two inputs and has one output. The first input is the "main"
15341 video on which the second input is overlaid.
15343 It accepts the following parameters:
15348 Set the x and y coordinates of the overlaid video on the main video.
15349 Default value is "0" for both expressions.
15352 See @ref{framesync}.
15355 See @ref{framesync}.
15358 See @ref{framesync}.
15362 This filter also supports the @ref{framesync} options.
15366 Apply Overcomplete Wavelet denoiser.
15368 The filter accepts the following options:
15374 Larger depth values will denoise lower frequency components more, but
15375 slow down filtering.
15377 Must be an int in the range 8-16, default is @code{8}.
15379 @item luma_strength, ls
15382 Must be a double value in the range 0-1000, default is @code{1.0}.
15384 @item chroma_strength, cs
15385 Set chroma strength.
15387 Must be a double value in the range 0-1000, default is @code{1.0}.
15393 Add paddings to the input image, and place the original input at the
15394 provided @var{x}, @var{y} coordinates.
15396 It accepts the following parameters:
15401 Specify an expression for the size of the output image with the
15402 paddings added. If the value for @var{width} or @var{height} is 0, the
15403 corresponding input size is used for the output.
15405 The @var{width} expression can reference the value set by the
15406 @var{height} expression, and vice versa.
15408 The default value of @var{width} and @var{height} is 0.
15412 Specify the offsets to place the input image at within the padded area,
15413 with respect to the top/left border of the output image.
15415 The @var{x} expression can reference the value set by the @var{y}
15416 expression, and vice versa.
15418 The default value of @var{x} and @var{y} is 0.
15420 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15421 so the input image is centered on the padded area.
15424 Specify the color of the padded area. For the syntax of this option,
15425 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15426 manual,ffmpeg-utils}.
15428 The default value of @var{color} is "black".
15431 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15433 It accepts the following values:
15437 Only evaluate expressions once during the filter initialization or when
15438 a command is processed.
15441 Evaluate expressions for each incoming frame.
15445 Default value is @samp{init}.
15448 Pad to aspect instead to a resolution.
15452 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15453 options are expressions containing the following constants:
15458 The input video width and height.
15462 These are the same as @var{in_w} and @var{in_h}.
15466 The output width and height (the size of the padded area), as
15467 specified by the @var{width} and @var{height} expressions.
15471 These are the same as @var{out_w} and @var{out_h}.
15475 The x and y offsets as specified by the @var{x} and @var{y}
15476 expressions, or NAN if not yet specified.
15479 same as @var{iw} / @var{ih}
15482 input sample aspect ratio
15485 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15489 The horizontal and vertical chroma subsample values. For example for the
15490 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15493 @subsection Examples
15497 Add paddings with the color "violet" to the input video. The output video
15498 size is 640x480, and the top-left corner of the input video is placed at
15501 pad=640:480:0:40:violet
15504 The example above is equivalent to the following command:
15506 pad=width=640:height=480:x=0:y=40:color=violet
15510 Pad the input to get an output with dimensions increased by 3/2,
15511 and put the input video at the center of the padded area:
15513 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15517 Pad the input to get a squared output with size equal to the maximum
15518 value between the input width and height, and put the input video at
15519 the center of the padded area:
15521 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15525 Pad the input to get a final w/h ratio of 16:9:
15527 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15531 In case of anamorphic video, in order to set the output display aspect
15532 correctly, it is necessary to use @var{sar} in the expression,
15533 according to the relation:
15535 (ih * X / ih) * sar = output_dar
15536 X = output_dar / sar
15539 Thus the previous example needs to be modified to:
15541 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15545 Double the output size and put the input video in the bottom-right
15546 corner of the output padded area:
15548 pad="2*iw:2*ih:ow-iw:oh-ih"
15552 @anchor{palettegen}
15553 @section palettegen
15555 Generate one palette for a whole video stream.
15557 It accepts the following options:
15561 Set the maximum number of colors to quantize in the palette.
15562 Note: the palette will still contain 256 colors; the unused palette entries
15565 @item reserve_transparent
15566 Create a palette of 255 colors maximum and reserve the last one for
15567 transparency. Reserving the transparency color is useful for GIF optimization.
15568 If not set, the maximum of colors in the palette will be 256. You probably want
15569 to disable this option for a standalone image.
15572 @item transparency_color
15573 Set the color that will be used as background for transparency.
15576 Set statistics mode.
15578 It accepts the following values:
15581 Compute full frame histograms.
15583 Compute histograms only for the part that differs from previous frame. This
15584 might be relevant to give more importance to the moving part of your input if
15585 the background is static.
15587 Compute new histogram for each frame.
15590 Default value is @var{full}.
15593 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15594 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15595 color quantization of the palette. This information is also visible at
15596 @var{info} logging level.
15598 @subsection Examples
15602 Generate a representative palette of a given video using @command{ffmpeg}:
15604 ffmpeg -i input.mkv -vf palettegen palette.png
15608 @section paletteuse
15610 Use a palette to downsample an input video stream.
15612 The filter takes two inputs: one video stream and a palette. The palette must
15613 be a 256 pixels image.
15615 It accepts the following options:
15619 Select dithering mode. Available algorithms are:
15622 Ordered 8x8 bayer dithering (deterministic)
15624 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15625 Note: this dithering is sometimes considered "wrong" and is included as a
15627 @item floyd_steinberg
15628 Floyd and Steingberg dithering (error diffusion)
15630 Frankie Sierra dithering v2 (error diffusion)
15632 Frankie Sierra dithering v2 "Lite" (error diffusion)
15635 Default is @var{sierra2_4a}.
15638 When @var{bayer} dithering is selected, this option defines the scale of the
15639 pattern (how much the crosshatch pattern is visible). A low value means more
15640 visible pattern for less banding, and higher value means less visible pattern
15641 at the cost of more banding.
15643 The option must be an integer value in the range [0,5]. Default is @var{2}.
15646 If set, define the zone to process
15650 Only the changing rectangle will be reprocessed. This is similar to GIF
15651 cropping/offsetting compression mechanism. This option can be useful for speed
15652 if only a part of the image is changing, and has use cases such as limiting the
15653 scope of the error diffusal @option{dither} to the rectangle that bounds the
15654 moving scene (it leads to more deterministic output if the scene doesn't change
15655 much, and as a result less moving noise and better GIF compression).
15658 Default is @var{none}.
15661 Take new palette for each output frame.
15663 @item alpha_threshold
15664 Sets the alpha threshold for transparency. Alpha values above this threshold
15665 will be treated as completely opaque, and values below this threshold will be
15666 treated as completely transparent.
15668 The option must be an integer value in the range [0,255]. Default is @var{128}.
15671 @subsection Examples
15675 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15676 using @command{ffmpeg}:
15678 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15682 @section perspective
15684 Correct perspective of video not recorded perpendicular to the screen.
15686 A description of the accepted parameters follows.
15697 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15698 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15699 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15700 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15701 then the corners of the source will be sent to the specified coordinates.
15703 The expressions can use the following variables:
15708 the width and height of video frame.
15712 Output frame count.
15715 @item interpolation
15716 Set interpolation for perspective correction.
15718 It accepts the following values:
15724 Default value is @samp{linear}.
15727 Set interpretation of coordinate options.
15729 It accepts the following values:
15733 Send point in the source specified by the given coordinates to
15734 the corners of the destination.
15736 @item 1, destination
15738 Send the corners of the source to the point in the destination specified
15739 by the given coordinates.
15741 Default value is @samp{source}.
15745 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15747 It accepts the following values:
15750 only evaluate expressions once during the filter initialization or
15751 when a command is processed
15754 evaluate expressions for each incoming frame
15757 Default value is @samp{init}.
15762 Delay interlaced video by one field time so that the field order changes.
15764 The intended use is to fix PAL movies that have been captured with the
15765 opposite field order to the film-to-video transfer.
15767 A description of the accepted parameters follows.
15773 It accepts the following values:
15776 Capture field order top-first, transfer bottom-first.
15777 Filter will delay the bottom field.
15780 Capture field order bottom-first, transfer top-first.
15781 Filter will delay the top field.
15784 Capture and transfer with the same field order. This mode only exists
15785 for the documentation of the other options to refer to, but if you
15786 actually select it, the filter will faithfully do nothing.
15789 Capture field order determined automatically by field flags, transfer
15791 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15792 basis using field flags. If no field information is available,
15793 then this works just like @samp{u}.
15796 Capture unknown or varying, transfer opposite.
15797 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15798 analyzing the images and selecting the alternative that produces best
15799 match between the fields.
15802 Capture top-first, transfer unknown or varying.
15803 Filter selects among @samp{t} and @samp{p} using image analysis.
15806 Capture bottom-first, transfer unknown or varying.
15807 Filter selects among @samp{b} and @samp{p} using image analysis.
15810 Capture determined by field flags, transfer unknown or varying.
15811 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15812 image analysis. If no field information is available, then this works just
15813 like @samp{U}. This is the default mode.
15816 Both capture and transfer unknown or varying.
15817 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15821 @subsection Commands
15823 This filter supports the all above options as @ref{commands}.
15825 @section photosensitivity
15826 Reduce various flashes in video, so to help users with epilepsy.
15828 It accepts the following options:
15831 Set how many frames to use when filtering. Default is 30.
15834 Set detection threshold factor. Default is 1.
15838 Set how many pixels to skip when sampling frames. Default is 1.
15839 Allowed range is from 1 to 1024.
15842 Leave frames unchanged. Default is disabled.
15845 @section pixdesctest
15847 Pixel format descriptor test filter, mainly useful for internal
15848 testing. The output video should be equal to the input video.
15852 format=monow, pixdesctest
15855 can be used to test the monowhite pixel format descriptor definition.
15859 Display sample values of color channels. Mainly useful for checking color
15860 and levels. Minimum supported resolution is 640x480.
15862 The filters accept the following options:
15866 Set scope X position, relative offset on X axis.
15869 Set scope Y position, relative offset on Y axis.
15878 Set window opacity. This window also holds statistics about pixel area.
15881 Set window X position, relative offset on X axis.
15884 Set window Y position, relative offset on Y axis.
15889 Enable the specified chain of postprocessing subfilters using libpostproc. This
15890 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15891 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15892 Each subfilter and some options have a short and a long name that can be used
15893 interchangeably, i.e. dr/dering are the same.
15895 The filters accept the following options:
15899 Set postprocessing subfilters string.
15902 All subfilters share common options to determine their scope:
15906 Honor the quality commands for this subfilter.
15909 Do chrominance filtering, too (default).
15912 Do luminance filtering only (no chrominance).
15915 Do chrominance filtering only (no luminance).
15918 These options can be appended after the subfilter name, separated by a '|'.
15920 Available subfilters are:
15923 @item hb/hdeblock[|difference[|flatness]]
15924 Horizontal deblocking filter
15927 Difference factor where higher values mean more deblocking (default: @code{32}).
15929 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15932 @item vb/vdeblock[|difference[|flatness]]
15933 Vertical deblocking filter
15936 Difference factor where higher values mean more deblocking (default: @code{32}).
15938 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15941 @item ha/hadeblock[|difference[|flatness]]
15942 Accurate horizontal deblocking filter
15945 Difference factor where higher values mean more deblocking (default: @code{32}).
15947 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15950 @item va/vadeblock[|difference[|flatness]]
15951 Accurate vertical deblocking filter
15954 Difference factor where higher values mean more deblocking (default: @code{32}).
15956 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15960 The horizontal and vertical deblocking filters share the difference and
15961 flatness values so you cannot set different horizontal and vertical
15965 @item h1/x1hdeblock
15966 Experimental horizontal deblocking filter
15968 @item v1/x1vdeblock
15969 Experimental vertical deblocking filter
15974 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15977 larger -> stronger filtering
15979 larger -> stronger filtering
15981 larger -> stronger filtering
15984 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15987 Stretch luminance to @code{0-255}.
15990 @item lb/linblenddeint
15991 Linear blend deinterlacing filter that deinterlaces the given block by
15992 filtering all lines with a @code{(1 2 1)} filter.
15994 @item li/linipoldeint
15995 Linear interpolating deinterlacing filter that deinterlaces the given block by
15996 linearly interpolating every second line.
15998 @item ci/cubicipoldeint
15999 Cubic interpolating deinterlacing filter deinterlaces the given block by
16000 cubically interpolating every second line.
16002 @item md/mediandeint
16003 Median deinterlacing filter that deinterlaces the given block by applying a
16004 median filter to every second line.
16006 @item fd/ffmpegdeint
16007 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
16008 second line with a @code{(-1 4 2 4 -1)} filter.
16011 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
16012 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
16014 @item fq/forceQuant[|quantizer]
16015 Overrides the quantizer table from the input with the constant quantizer you
16023 Default pp filter combination (@code{hb|a,vb|a,dr|a})
16026 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
16029 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
16032 @subsection Examples
16036 Apply horizontal and vertical deblocking, deringing and automatic
16037 brightness/contrast:
16043 Apply default filters without brightness/contrast correction:
16049 Apply default filters and temporal denoiser:
16051 pp=default/tmpnoise|1|2|3
16055 Apply deblocking on luminance only, and switch vertical deblocking on or off
16056 automatically depending on available CPU time:
16063 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16064 similar to spp = 6 with 7 point DCT, where only the center sample is
16067 The filter accepts the following options:
16071 Force a constant quantization parameter. It accepts an integer in range
16072 0 to 63. If not set, the filter will use the QP from the video stream
16076 Set thresholding mode. Available modes are:
16080 Set hard thresholding.
16082 Set soft thresholding (better de-ringing effect, but likely blurrier).
16084 Set medium thresholding (good results, default).
16088 @section premultiply
16089 Apply alpha premultiply effect to input video stream using first plane
16090 of second stream as alpha.
16092 Both streams must have same dimensions and same pixel format.
16094 The filter accepts the following option:
16098 Set which planes will be processed, unprocessed planes will be copied.
16099 By default value 0xf, all planes will be processed.
16102 Do not require 2nd input for processing, instead use alpha plane from input stream.
16106 Apply prewitt operator to input video stream.
16108 The filter accepts the following option:
16112 Set which planes will be processed, unprocessed planes will be copied.
16113 By default value 0xf, all planes will be processed.
16116 Set value which will be multiplied with filtered result.
16119 Set value which will be added to filtered result.
16122 @subsection Commands
16124 This filter supports the all above options as @ref{commands}.
16126 @section pseudocolor
16128 Alter frame colors in video with pseudocolors.
16130 This filter accepts the following options:
16134 set pixel first component expression
16137 set pixel second component expression
16140 set pixel third component expression
16143 set pixel fourth component expression, corresponds to the alpha component
16146 set component to use as base for altering colors
16149 Each of them specifies the expression to use for computing the lookup table for
16150 the corresponding pixel component values.
16152 The expressions can contain the following constants and functions:
16157 The input width and height.
16160 The input value for the pixel component.
16162 @item ymin, umin, vmin, amin
16163 The minimum allowed component value.
16165 @item ymax, umax, vmax, amax
16166 The maximum allowed component value.
16169 All expressions default to "val".
16171 @subsection Examples
16175 Change too high luma values to gradient:
16177 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'"
16183 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16184 Ratio) between two input videos.
16186 This filter takes in input two input videos, the first input is
16187 considered the "main" source and is passed unchanged to the
16188 output. The second input is used as a "reference" video for computing
16191 Both video inputs must have the same resolution and pixel format for
16192 this filter to work correctly. Also it assumes that both inputs
16193 have the same number of frames, which are compared one by one.
16195 The obtained average PSNR is printed through the logging system.
16197 The filter stores the accumulated MSE (mean squared error) of each
16198 frame, and at the end of the processing it is averaged across all frames
16199 equally, and the following formula is applied to obtain the PSNR:
16202 PSNR = 10*log10(MAX^2/MSE)
16205 Where MAX is the average of the maximum values of each component of the
16208 The description of the accepted parameters follows.
16211 @item stats_file, f
16212 If specified the filter will use the named file to save the PSNR of
16213 each individual frame. When filename equals "-" the data is sent to
16216 @item stats_version
16217 Specifies which version of the stats file format to use. Details of
16218 each format are written below.
16219 Default value is 1.
16221 @item stats_add_max
16222 Determines whether the max value is output to the stats log.
16223 Default value is 0.
16224 Requires stats_version >= 2. If this is set and stats_version < 2,
16225 the filter will return an error.
16228 This filter also supports the @ref{framesync} options.
16230 The file printed if @var{stats_file} is selected, contains a sequence of
16231 key/value pairs of the form @var{key}:@var{value} for each compared
16234 If a @var{stats_version} greater than 1 is specified, a header line precedes
16235 the list of per-frame-pair stats, with key value pairs following the frame
16236 format with the following parameters:
16239 @item psnr_log_version
16240 The version of the log file format. Will match @var{stats_version}.
16243 A comma separated list of the per-frame-pair parameters included in
16247 A description of each shown per-frame-pair parameter follows:
16251 sequential number of the input frame, starting from 1
16254 Mean Square Error pixel-by-pixel average difference of the compared
16255 frames, averaged over all the image components.
16257 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16258 Mean Square Error pixel-by-pixel average difference of the compared
16259 frames for the component specified by the suffix.
16261 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16262 Peak Signal to Noise ratio of the compared frames for the component
16263 specified by the suffix.
16265 @item max_avg, max_y, max_u, max_v
16266 Maximum allowed value for each channel, and average over all
16270 @subsection Examples
16275 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16276 [main][ref] psnr="stats_file=stats.log" [out]
16279 On this example the input file being processed is compared with the
16280 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16281 is stored in @file{stats.log}.
16284 Another example with different containers:
16286 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 -
16293 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16294 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16297 The pullup filter is designed to take advantage of future context in making
16298 its decisions. This filter is stateless in the sense that it does not lock
16299 onto a pattern to follow, but it instead looks forward to the following
16300 fields in order to identify matches and rebuild progressive frames.
16302 To produce content with an even framerate, insert the fps filter after
16303 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16304 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16306 The filter accepts the following options:
16313 These options set the amount of "junk" to ignore at the left, right, top, and
16314 bottom of the image, respectively. Left and right are in units of 8 pixels,
16315 while top and bottom are in units of 2 lines.
16316 The default is 8 pixels on each side.
16319 Set the strict breaks. Setting this option to 1 will reduce the chances of
16320 filter generating an occasional mismatched frame, but it may also cause an
16321 excessive number of frames to be dropped during high motion sequences.
16322 Conversely, setting it to -1 will make filter match fields more easily.
16323 This may help processing of video where there is slight blurring between
16324 the fields, but may also cause there to be interlaced frames in the output.
16325 Default value is @code{0}.
16328 Set the metric plane to use. It accepts the following values:
16334 Use chroma blue plane.
16337 Use chroma red plane.
16340 This option may be set to use chroma plane instead of the default luma plane
16341 for doing filter's computations. This may improve accuracy on very clean
16342 source material, but more likely will decrease accuracy, especially if there
16343 is chroma noise (rainbow effect) or any grayscale video.
16344 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16345 load and make pullup usable in realtime on slow machines.
16348 For best results (without duplicated frames in the output file) it is
16349 necessary to change the output frame rate. For example, to inverse
16350 telecine NTSC input:
16352 ffmpeg -i input -vf pullup -r 24000/1001 ...
16357 Change video quantization parameters (QP).
16359 The filter accepts the following option:
16363 Set expression for quantization parameter.
16366 The expression is evaluated through the eval API and can contain, among others,
16367 the following constants:
16371 1 if index is not 129, 0 otherwise.
16374 Sequential index starting from -129 to 128.
16377 @subsection Examples
16381 Some equation like:
16389 Flush video frames from internal cache of frames into a random order.
16390 No frame is discarded.
16391 Inspired by @ref{frei0r} nervous filter.
16395 Set size in number of frames of internal cache, in range from @code{2} to
16396 @code{512}. Default is @code{30}.
16399 Set seed for random number generator, must be an integer included between
16400 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16401 less than @code{0}, the filter will try to use a good random seed on a
16405 @section readeia608
16407 Read closed captioning (EIA-608) information from the top lines of a video frame.
16409 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16410 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16411 with EIA-608 data (starting from 0). A description of each metadata value follows:
16414 @item lavfi.readeia608.X.cc
16415 The two bytes stored as EIA-608 data (printed in hexadecimal).
16417 @item lavfi.readeia608.X.line
16418 The number of the line on which the EIA-608 data was identified and read.
16421 This filter accepts the following options:
16425 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16428 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16431 Set the ratio of width reserved for sync code detection.
16432 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16435 Enable checking the parity bit. In the event of a parity error, the filter will output
16436 @code{0x00} for that character. Default is false.
16439 Lowpass lines prior to further processing. Default is enabled.
16442 @subsection Commands
16444 This filter supports the all above options as @ref{commands}.
16446 @subsection Examples
16450 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16452 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
16458 Read vertical interval timecode (VITC) information from the top lines of a
16461 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16462 timecode value, if a valid timecode has been detected. Further metadata key
16463 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16464 timecode data has been found or not.
16466 This filter accepts the following options:
16470 Set the maximum number of lines to scan for VITC data. If the value is set to
16471 @code{-1} the full video frame is scanned. Default is @code{45}.
16474 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16475 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16478 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16479 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16482 @subsection Examples
16486 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16487 draw @code{--:--:--:--} as a placeholder:
16489 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16495 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16497 Destination pixel at position (X, Y) will be picked from source (x, y) position
16498 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16499 value for pixel will be used for destination pixel.
16501 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16502 will have Xmap/Ymap video stream dimensions.
16503 Xmap and Ymap input video streams are 16bit depth, single channel.
16507 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16508 Default is @code{color}.
16511 Specify the color of the unmapped pixels. For the syntax of this option,
16512 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16513 manual,ffmpeg-utils}. Default color is @code{black}.
16516 @section removegrain
16518 The removegrain filter is a spatial denoiser for progressive video.
16522 Set mode for the first plane.
16525 Set mode for the second plane.
16528 Set mode for the third plane.
16531 Set mode for the fourth plane.
16534 Range of mode is from 0 to 24. Description of each mode follows:
16538 Leave input plane unchanged. Default.
16541 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16544 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16547 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16550 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16551 This is equivalent to a median filter.
16554 Line-sensitive clipping giving the minimal change.
16557 Line-sensitive clipping, intermediate.
16560 Line-sensitive clipping, intermediate.
16563 Line-sensitive clipping, intermediate.
16566 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16569 Replaces the target pixel with the closest neighbour.
16572 [1 2 1] horizontal and vertical kernel blur.
16578 Bob mode, interpolates top field from the line where the neighbours
16579 pixels are the closest.
16582 Bob mode, interpolates bottom field from the line where the neighbours
16583 pixels are the closest.
16586 Bob mode, interpolates top field. Same as 13 but with a more complicated
16587 interpolation formula.
16590 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16591 interpolation formula.
16594 Clips the pixel with the minimum and maximum of respectively the maximum and
16595 minimum of each pair of opposite neighbour pixels.
16598 Line-sensitive clipping using opposite neighbours whose greatest distance from
16599 the current pixel is minimal.
16602 Replaces the pixel with the average of its 8 neighbours.
16605 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16608 Clips pixels using the averages of opposite neighbour.
16611 Same as mode 21 but simpler and faster.
16614 Small edge and halo removal, but reputed useless.
16620 @section removelogo
16622 Suppress a TV station logo, using an image file to determine which
16623 pixels comprise the logo. It works by filling in the pixels that
16624 comprise the logo with neighboring pixels.
16626 The filter accepts the following options:
16630 Set the filter bitmap file, which can be any image format supported by
16631 libavformat. The width and height of the image file must match those of the
16632 video stream being processed.
16635 Pixels in the provided bitmap image with a value of zero are not
16636 considered part of the logo, non-zero pixels are considered part of
16637 the logo. If you use white (255) for the logo and black (0) for the
16638 rest, you will be safe. For making the filter bitmap, it is
16639 recommended to take a screen capture of a black frame with the logo
16640 visible, and then using a threshold filter followed by the erode
16641 filter once or twice.
16643 If needed, little splotches can be fixed manually. Remember that if
16644 logo pixels are not covered, the filter quality will be much
16645 reduced. Marking too many pixels as part of the logo does not hurt as
16646 much, but it will increase the amount of blurring needed to cover over
16647 the image and will destroy more information than necessary, and extra
16648 pixels will slow things down on a large logo.
16650 @section repeatfields
16652 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16653 fields based on its value.
16657 Reverse a video clip.
16659 Warning: This filter requires memory to buffer the entire clip, so trimming
16662 @subsection Examples
16666 Take the first 5 seconds of a clip, and reverse it.
16673 Shift R/G/B/A pixels horizontally and/or vertically.
16675 The filter accepts the following options:
16678 Set amount to shift red horizontally.
16680 Set amount to shift red vertically.
16682 Set amount to shift green horizontally.
16684 Set amount to shift green vertically.
16686 Set amount to shift blue horizontally.
16688 Set amount to shift blue vertically.
16690 Set amount to shift alpha horizontally.
16692 Set amount to shift alpha vertically.
16694 Set edge mode, can be @var{smear}, default, or @var{warp}.
16697 @subsection Commands
16699 This filter supports the all above options as @ref{commands}.
16702 Apply roberts cross operator to input video stream.
16704 The filter accepts the following option:
16708 Set which planes will be processed, unprocessed planes will be copied.
16709 By default value 0xf, all planes will be processed.
16712 Set value which will be multiplied with filtered result.
16715 Set value which will be added to filtered result.
16718 @subsection Commands
16720 This filter supports the all above options as @ref{commands}.
16724 Rotate video by an arbitrary angle expressed in radians.
16726 The filter accepts the following options:
16728 A description of the optional parameters follows.
16731 Set an expression for the angle by which to rotate the input video
16732 clockwise, expressed as a number of radians. A negative value will
16733 result in a counter-clockwise rotation. By default it is set to "0".
16735 This expression is evaluated for each frame.
16738 Set the output width expression, default value is "iw".
16739 This expression is evaluated just once during configuration.
16742 Set the output height expression, default value is "ih".
16743 This expression is evaluated just once during configuration.
16746 Enable bilinear interpolation if set to 1, a value of 0 disables
16747 it. Default value is 1.
16750 Set the color used to fill the output area not covered by the rotated
16751 image. For the general syntax of this option, check the
16752 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16753 If the special value "none" is selected then no
16754 background is printed (useful for example if the background is never shown).
16756 Default value is "black".
16759 The expressions for the angle and the output size can contain the
16760 following constants and functions:
16764 sequential number of the input frame, starting from 0. It is always NAN
16765 before the first frame is filtered.
16768 time in seconds of the input frame, it is set to 0 when the filter is
16769 configured. It is always NAN before the first frame is filtered.
16773 horizontal and vertical chroma subsample values. For example for the
16774 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16778 the input video width and height
16782 the output width and height, that is the size of the padded area as
16783 specified by the @var{width} and @var{height} expressions
16787 the minimal width/height required for completely containing the input
16788 video rotated by @var{a} radians.
16790 These are only available when computing the @option{out_w} and
16791 @option{out_h} expressions.
16794 @subsection Examples
16798 Rotate the input by PI/6 radians clockwise:
16804 Rotate the input by PI/6 radians counter-clockwise:
16810 Rotate the input by 45 degrees clockwise:
16816 Apply a constant rotation with period T, starting from an angle of PI/3:
16818 rotate=PI/3+2*PI*t/T
16822 Make the input video rotation oscillating with a period of T
16823 seconds and an amplitude of A radians:
16825 rotate=A*sin(2*PI/T*t)
16829 Rotate the video, output size is chosen so that the whole rotating
16830 input video is always completely contained in the output:
16832 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16836 Rotate the video, reduce the output size so that no background is ever
16839 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16843 @subsection Commands
16845 The filter supports the following commands:
16849 Set the angle expression.
16850 The command accepts the same syntax of the corresponding option.
16852 If the specified expression is not valid, it is kept at its current
16858 Apply Shape Adaptive Blur.
16860 The filter accepts the following options:
16863 @item luma_radius, lr
16864 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16865 value is 1.0. A greater value will result in a more blurred image, and
16866 in slower processing.
16868 @item luma_pre_filter_radius, lpfr
16869 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16872 @item luma_strength, ls
16873 Set luma maximum difference between pixels to still be considered, must
16874 be a value in the 0.1-100.0 range, default value is 1.0.
16876 @item chroma_radius, cr
16877 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16878 greater value will result in a more blurred image, and in slower
16881 @item chroma_pre_filter_radius, cpfr
16882 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16884 @item chroma_strength, cs
16885 Set chroma maximum difference between pixels to still be considered,
16886 must be a value in the -0.9-100.0 range.
16889 Each chroma option value, if not explicitly specified, is set to the
16890 corresponding luma option value.
16895 Scale (resize) the input video, using the libswscale library.
16897 The scale filter forces the output display aspect ratio to be the same
16898 of the input, by changing the output sample aspect ratio.
16900 If the input image format is different from the format requested by
16901 the next filter, the scale filter will convert the input to the
16904 @subsection Options
16905 The filter accepts the following options, or any of the options
16906 supported by the libswscale scaler.
16908 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16909 the complete list of scaler options.
16914 Set the output video dimension expression. Default value is the input
16917 If the @var{width} or @var{w} value is 0, the input width is used for
16918 the output. If the @var{height} or @var{h} value is 0, the input height
16919 is used for the output.
16921 If one and only one of the values is -n with n >= 1, the scale filter
16922 will use a value that maintains the aspect ratio of the input image,
16923 calculated from the other specified dimension. After that it will,
16924 however, make sure that the calculated dimension is divisible by n and
16925 adjust the value if necessary.
16927 If both values are -n with n >= 1, the behavior will be identical to
16928 both values being set to 0 as previously detailed.
16930 See below for the list of accepted constants for use in the dimension
16934 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16938 Only evaluate expressions once during the filter initialization or when a command is processed.
16941 Evaluate expressions for each incoming frame.
16945 Default value is @samp{init}.
16949 Set the interlacing mode. It accepts the following values:
16953 Force interlaced aware scaling.
16956 Do not apply interlaced scaling.
16959 Select interlaced aware scaling depending on whether the source frames
16960 are flagged as interlaced or not.
16963 Default value is @samp{0}.
16966 Set libswscale scaling flags. See
16967 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16968 complete list of values. If not explicitly specified the filter applies
16972 @item param0, param1
16973 Set libswscale input parameters for scaling algorithms that need them. See
16974 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16975 complete documentation. If not explicitly specified the filter applies
16981 Set the video size. For the syntax of this option, check the
16982 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16984 @item in_color_matrix
16985 @item out_color_matrix
16986 Set in/output YCbCr color space type.
16988 This allows the autodetected value to be overridden as well as allows forcing
16989 a specific value used for the output and encoder.
16991 If not specified, the color space type depends on the pixel format.
16997 Choose automatically.
17000 Format conforming to International Telecommunication Union (ITU)
17001 Recommendation BT.709.
17004 Set color space conforming to the United States Federal Communications
17005 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
17010 Set color space conforming to:
17014 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
17017 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
17020 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
17025 Set color space conforming to SMPTE ST 240:1999.
17028 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
17033 Set in/output YCbCr sample range.
17035 This allows the autodetected value to be overridden as well as allows forcing
17036 a specific value used for the output and encoder. If not specified, the
17037 range depends on the pixel format. Possible values:
17041 Choose automatically.
17044 Set full range (0-255 in case of 8-bit luma).
17046 @item mpeg/limited/tv
17047 Set "MPEG" range (16-235 in case of 8-bit luma).
17050 @item force_original_aspect_ratio
17051 Enable decreasing or increasing output video width or height if necessary to
17052 keep the original aspect ratio. Possible values:
17056 Scale the video as specified and disable this feature.
17059 The output video dimensions will automatically be decreased if needed.
17062 The output video dimensions will automatically be increased if needed.
17066 One useful instance of this option is that when you know a specific device's
17067 maximum allowed resolution, you can use this to limit the output video to
17068 that, while retaining the aspect ratio. For example, device A allows
17069 1280x720 playback, and your video is 1920x800. Using this option (set it to
17070 decrease) and specifying 1280x720 to the command line makes the output
17073 Please note that this is a different thing than specifying -1 for @option{w}
17074 or @option{h}, you still need to specify the output resolution for this option
17077 @item force_divisible_by
17078 Ensures that both the output dimensions, width and height, are divisible by the
17079 given integer when used together with @option{force_original_aspect_ratio}. This
17080 works similar to using @code{-n} in the @option{w} and @option{h} options.
17082 This option respects the value set for @option{force_original_aspect_ratio},
17083 increasing or decreasing the resolution accordingly. The video's aspect ratio
17084 may be slightly modified.
17086 This option can be handy if you need to have a video fit within or exceed
17087 a defined resolution using @option{force_original_aspect_ratio} but also have
17088 encoder restrictions on width or height divisibility.
17092 The values of the @option{w} and @option{h} options are expressions
17093 containing the following constants:
17098 The input width and height
17102 These are the same as @var{in_w} and @var{in_h}.
17106 The output (scaled) width and height
17110 These are the same as @var{out_w} and @var{out_h}
17113 The same as @var{iw} / @var{ih}
17116 input sample aspect ratio
17119 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17123 horizontal and vertical input chroma subsample values. For example for the
17124 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17128 horizontal and vertical output chroma subsample values. For example for the
17129 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17132 The (sequential) number of the input frame, starting from 0.
17133 Only available with @code{eval=frame}.
17136 The presentation timestamp of the input frame, expressed as a number of
17137 seconds. Only available with @code{eval=frame}.
17140 The position (byte offset) of the frame in the input stream, or NaN if
17141 this information is unavailable and/or meaningless (for example in case of synthetic video).
17142 Only available with @code{eval=frame}.
17145 @subsection Examples
17149 Scale the input video to a size of 200x100
17154 This is equivalent to:
17165 Specify a size abbreviation for the output size:
17170 which can also be written as:
17176 Scale the input to 2x:
17178 scale=w=2*iw:h=2*ih
17182 The above is the same as:
17184 scale=2*in_w:2*in_h
17188 Scale the input to 2x with forced interlaced scaling:
17190 scale=2*iw:2*ih:interl=1
17194 Scale the input to half size:
17196 scale=w=iw/2:h=ih/2
17200 Increase the width, and set the height to the same size:
17206 Seek Greek harmony:
17213 Increase the height, and set the width to 3/2 of the height:
17215 scale=w=3/2*oh:h=3/5*ih
17219 Increase the size, making the size a multiple of the chroma
17222 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17226 Increase the width to a maximum of 500 pixels,
17227 keeping the same aspect ratio as the input:
17229 scale=w='min(500\, iw*3/2):h=-1'
17233 Make pixels square by combining scale and setsar:
17235 scale='trunc(ih*dar):ih',setsar=1/1
17239 Make pixels square by combining scale and setsar,
17240 making sure the resulting resolution is even (required by some codecs):
17242 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17246 @subsection Commands
17248 This filter supports the following commands:
17252 Set the output video dimension expression.
17253 The command accepts the same syntax of the corresponding option.
17255 If the specified expression is not valid, it is kept at its current
17261 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17262 format conversion on CUDA video frames. Setting the output width and height
17263 works in the same way as for the @var{scale} filter.
17265 The following additional options are accepted:
17268 The pixel format of the output CUDA frames. If set to the string "same" (the
17269 default), the input format will be kept. Note that automatic format negotiation
17270 and conversion is not yet supported for hardware frames
17273 The interpolation algorithm used for resizing. One of the following:
17280 @item cubic2p_bspline
17281 2-parameter cubic (B=1, C=0)
17283 @item cubic2p_catmullrom
17284 2-parameter cubic (B=0, C=1/2)
17286 @item cubic2p_b05c03
17287 2-parameter cubic (B=1/2, C=3/10)
17295 @item force_original_aspect_ratio
17296 Enable decreasing or increasing output video width or height if necessary to
17297 keep the original aspect ratio. Possible values:
17301 Scale the video as specified and disable this feature.
17304 The output video dimensions will automatically be decreased if needed.
17307 The output video dimensions will automatically be increased if needed.
17311 One useful instance of this option is that when you know a specific device's
17312 maximum allowed resolution, you can use this to limit the output video to
17313 that, while retaining the aspect ratio. For example, device A allows
17314 1280x720 playback, and your video is 1920x800. Using this option (set it to
17315 decrease) and specifying 1280x720 to the command line makes the output
17318 Please note that this is a different thing than specifying -1 for @option{w}
17319 or @option{h}, you still need to specify the output resolution for this option
17322 @item force_divisible_by
17323 Ensures that both the output dimensions, width and height, are divisible by the
17324 given integer when used together with @option{force_original_aspect_ratio}. This
17325 works similar to using @code{-n} in the @option{w} and @option{h} options.
17327 This option respects the value set for @option{force_original_aspect_ratio},
17328 increasing or decreasing the resolution accordingly. The video's aspect ratio
17329 may be slightly modified.
17331 This option can be handy if you need to have a video fit within or exceed
17332 a defined resolution using @option{force_original_aspect_ratio} but also have
17333 encoder restrictions on width or height divisibility.
17339 Scale (resize) the input video, based on a reference video.
17341 See the scale filter for available options, scale2ref supports the same but
17342 uses the reference video instead of the main input as basis. scale2ref also
17343 supports the following additional constants for the @option{w} and
17344 @option{h} options:
17349 The main input video's width and height
17352 The same as @var{main_w} / @var{main_h}
17355 The main input video's sample aspect ratio
17357 @item main_dar, mdar
17358 The main input video's display aspect ratio. Calculated from
17359 @code{(main_w / main_h) * main_sar}.
17363 The main input video's horizontal and vertical chroma subsample values.
17364 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17368 The (sequential) number of the main input frame, starting from 0.
17369 Only available with @code{eval=frame}.
17372 The presentation timestamp of the main input frame, expressed as a number of
17373 seconds. Only available with @code{eval=frame}.
17376 The position (byte offset) of the frame in the main input stream, or NaN if
17377 this information is unavailable and/or meaningless (for example in case of synthetic video).
17378 Only available with @code{eval=frame}.
17381 @subsection Examples
17385 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17387 'scale2ref[b][a];[a][b]overlay'
17391 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17393 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17397 @subsection Commands
17399 This filter supports the following commands:
17403 Set the output video dimension expression.
17404 The command accepts the same syntax of the corresponding option.
17406 If the specified expression is not valid, it is kept at its current
17411 Scroll input video horizontally and/or vertically by constant speed.
17413 The filter accepts the following options:
17415 @item horizontal, h
17416 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17417 Negative values changes scrolling direction.
17420 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17421 Negative values changes scrolling direction.
17424 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17427 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17430 @subsection Commands
17432 This filter supports the following @ref{commands}:
17434 @item horizontal, h
17435 Set the horizontal scrolling speed.
17437 Set the vertical scrolling speed.
17443 Detect video scene change.
17445 This filter sets frame metadata with mafd between frame, the scene score, and
17446 forward the frame to the next filter, so they can use these metadata to detect
17447 scene change or others.
17449 In addition, this filter logs a message and sets frame metadata when it detects
17450 a scene change by @option{threshold}.
17452 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17454 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17455 to detect scene change.
17457 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17458 detect scene change with @option{threshold}.
17460 The filter accepts the following options:
17464 Set the scene change detection threshold as a percentage of maximum change. Good
17465 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17468 Default value is @code{10.}.
17471 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17472 You can enable it if you want to get snapshot of scene change frames only.
17475 @anchor{selectivecolor}
17476 @section selectivecolor
17478 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17479 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17480 by the "purity" of the color (that is, how saturated it already is).
17482 This filter is similar to the Adobe Photoshop Selective Color tool.
17484 The filter accepts the following options:
17487 @item correction_method
17488 Select color correction method.
17490 Available values are:
17493 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17496 Specified adjustments are relative to the original component value.
17498 Default is @code{absolute}.
17500 Adjustments for red pixels (pixels where the red component is the maximum)
17502 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17504 Adjustments for green pixels (pixels where the green component is the maximum)
17506 Adjustments for cyan pixels (pixels where the red component is the minimum)
17508 Adjustments for blue pixels (pixels where the blue component is the maximum)
17510 Adjustments for magenta pixels (pixels where the green component is the minimum)
17512 Adjustments for white pixels (pixels where all components are greater than 128)
17514 Adjustments for all pixels except pure black and pure white
17516 Adjustments for black pixels (pixels where all components are lesser than 128)
17518 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17521 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17522 4 space separated floating point adjustment values in the [-1,1] range,
17523 respectively to adjust the amount of cyan, magenta, yellow and black for the
17524 pixels of its range.
17526 @subsection Examples
17530 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17531 increase magenta by 27% in blue areas:
17533 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17537 Use a Photoshop selective color preset:
17539 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17543 @anchor{separatefields}
17544 @section separatefields
17546 The @code{separatefields} takes a frame-based video input and splits
17547 each frame into its components fields, producing a new half height clip
17548 with twice the frame rate and twice the frame count.
17550 This filter use field-dominance information in frame to decide which
17551 of each pair of fields to place first in the output.
17552 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17554 @section setdar, setsar
17556 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17559 This is done by changing the specified Sample (aka Pixel) Aspect
17560 Ratio, according to the following equation:
17562 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17565 Keep in mind that the @code{setdar} filter does not modify the pixel
17566 dimensions of the video frame. Also, the display aspect ratio set by
17567 this filter may be changed by later filters in the filterchain,
17568 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17571 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17572 the filter output video.
17574 Note that as a consequence of the application of this filter, the
17575 output display aspect ratio will change according to the equation
17578 Keep in mind that the sample aspect ratio set by the @code{setsar}
17579 filter may be changed by later filters in the filterchain, e.g. if
17580 another "setsar" or a "setdar" filter is applied.
17582 It accepts the following parameters:
17585 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17586 Set the aspect ratio used by the filter.
17588 The parameter can be a floating point number string, an expression, or
17589 a string of the form @var{num}:@var{den}, where @var{num} and
17590 @var{den} are the numerator and denominator of the aspect ratio. If
17591 the parameter is not specified, it is assumed the value "0".
17592 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17596 Set the maximum integer value to use for expressing numerator and
17597 denominator when reducing the expressed aspect ratio to a rational.
17598 Default value is @code{100}.
17602 The parameter @var{sar} is an expression containing
17603 the following constants:
17607 These are approximated values for the mathematical constants e
17608 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17611 The input width and height.
17614 These are the same as @var{w} / @var{h}.
17617 The input sample aspect ratio.
17620 The input display aspect ratio. It is the same as
17621 (@var{w} / @var{h}) * @var{sar}.
17624 Horizontal and vertical chroma subsample values. For example, for the
17625 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17628 @subsection Examples
17633 To change the display aspect ratio to 16:9, specify one of the following:
17640 To change the sample aspect ratio to 10:11, specify:
17646 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17647 1000 in the aspect ratio reduction, use the command:
17649 setdar=ratio=16/9:max=1000
17657 Force field for the output video frame.
17659 The @code{setfield} filter marks the interlace type field for the
17660 output frames. It does not change the input frame, but only sets the
17661 corresponding property, which affects how the frame is treated by
17662 following filters (e.g. @code{fieldorder} or @code{yadif}).
17664 The filter accepts the following options:
17669 Available values are:
17673 Keep the same field property.
17676 Mark the frame as bottom-field-first.
17679 Mark the frame as top-field-first.
17682 Mark the frame as progressive.
17689 Force frame parameter for the output video frame.
17691 The @code{setparams} filter marks interlace and color range for the
17692 output frames. It does not change the input frame, but only sets the
17693 corresponding property, which affects how the frame is treated by
17698 Available values are:
17702 Keep the same field property (default).
17705 Mark the frame as bottom-field-first.
17708 Mark the frame as top-field-first.
17711 Mark the frame as progressive.
17715 Available values are:
17719 Keep the same color range property (default).
17721 @item unspecified, unknown
17722 Mark the frame as unspecified color range.
17724 @item limited, tv, mpeg
17725 Mark the frame as limited range.
17727 @item full, pc, jpeg
17728 Mark the frame as full range.
17731 @item color_primaries
17732 Set the color primaries.
17733 Available values are:
17737 Keep the same color primaries property (default).
17754 Set the color transfer.
17755 Available values are:
17759 Keep the same color trc property (default).
17781 Set the colorspace.
17782 Available values are:
17786 Keep the same colorspace property (default).
17799 @item chroma-derived-nc
17800 @item chroma-derived-c
17807 Show a line containing various information for each input video frame.
17808 The input video is not modified.
17810 This filter supports the following options:
17814 Calculate checksums of each plane. By default enabled.
17817 The shown line contains a sequence of key/value pairs of the form
17818 @var{key}:@var{value}.
17820 The following values are shown in the output:
17824 The (sequential) number of the input frame, starting from 0.
17827 The Presentation TimeStamp of the input frame, expressed as a number of
17828 time base units. The time base unit depends on the filter input pad.
17831 The Presentation TimeStamp of the input frame, expressed as a number of
17835 The position of the frame in the input stream, or -1 if this information is
17836 unavailable and/or meaningless (for example in case of synthetic video).
17839 The pixel format name.
17842 The sample aspect ratio of the input frame, expressed in the form
17843 @var{num}/@var{den}.
17846 The size of the input frame. For the syntax of this option, check the
17847 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17850 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17851 for bottom field first).
17854 This is 1 if the frame is a key frame, 0 otherwise.
17857 The picture type of the input frame ("I" for an I-frame, "P" for a
17858 P-frame, "B" for a B-frame, or "?" for an unknown type).
17859 Also refer to the documentation of the @code{AVPictureType} enum and of
17860 the @code{av_get_picture_type_char} function defined in
17861 @file{libavutil/avutil.h}.
17864 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17866 @item plane_checksum
17867 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17868 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17871 The mean value of pixels in each plane of the input frame, expressed in the form
17872 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17875 The standard deviation of pixel values in each plane of the input frame, expressed
17876 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17880 @section showpalette
17882 Displays the 256 colors palette of each frame. This filter is only relevant for
17883 @var{pal8} pixel format frames.
17885 It accepts the following option:
17889 Set the size of the box used to represent one palette color entry. Default is
17890 @code{30} (for a @code{30x30} pixel box).
17893 @section shuffleframes
17895 Reorder and/or duplicate and/or drop video frames.
17897 It accepts the following parameters:
17901 Set the destination indexes of input frames.
17902 This is space or '|' separated list of indexes that maps input frames to output
17903 frames. Number of indexes also sets maximal value that each index may have.
17904 '-1' index have special meaning and that is to drop frame.
17907 The first frame has the index 0. The default is to keep the input unchanged.
17909 @subsection Examples
17913 Swap second and third frame of every three frames of the input:
17915 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17919 Swap 10th and 1st frame of every ten frames of the input:
17921 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17925 @section shufflepixels
17927 Reorder pixels in video frames.
17929 This filter accepts the following options:
17933 Set shuffle direction. Can be forward or inverse direction.
17934 Default direction is forward.
17937 Set shuffle mode. Can be horizontal, vertical or block mode.
17941 Set shuffle block_size. In case of horizontal shuffle mode only width
17942 part of size is used, and in case of vertical shuffle mode only height
17943 part of size is used.
17946 Set random seed used with shuffling pixels. Mainly useful to set to be able
17947 to reverse filtering process to get original input.
17948 For example, to reverse forward shuffle you need to use same parameters
17949 and exact same seed and to set direction to inverse.
17952 @section shuffleplanes
17954 Reorder and/or duplicate video planes.
17956 It accepts the following parameters:
17961 The index of the input plane to be used as the first output plane.
17964 The index of the input plane to be used as the second output plane.
17967 The index of the input plane to be used as the third output plane.
17970 The index of the input plane to be used as the fourth output plane.
17974 The first plane has the index 0. The default is to keep the input unchanged.
17976 @subsection Examples
17980 Swap the second and third planes of the input:
17982 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17986 @anchor{signalstats}
17987 @section signalstats
17988 Evaluate various visual metrics that assist in determining issues associated
17989 with the digitization of analog video media.
17991 By default the filter will log these metadata values:
17995 Display the minimal Y value contained within the input frame. Expressed in
17999 Display the Y value at the 10% percentile within the input frame. Expressed in
18003 Display the average Y value within the input frame. Expressed in range of
18007 Display the Y value at the 90% percentile within the input frame. Expressed in
18011 Display the maximum Y value contained within the input frame. Expressed in
18015 Display the minimal U value contained within the input frame. Expressed in
18019 Display the U value at the 10% percentile within the input frame. Expressed in
18023 Display the average U value within the input frame. Expressed in range of
18027 Display the U value at the 90% percentile within the input frame. Expressed in
18031 Display the maximum U value contained within the input frame. Expressed in
18035 Display the minimal V value contained within the input frame. Expressed in
18039 Display the V value at the 10% percentile within the input frame. Expressed in
18043 Display the average V value within the input frame. Expressed in range of
18047 Display the V value at the 90% percentile within the input frame. Expressed in
18051 Display the maximum V value contained within the input frame. Expressed in
18055 Display the minimal saturation value contained within the input frame.
18056 Expressed in range of [0-~181.02].
18059 Display the saturation value at the 10% percentile within the input frame.
18060 Expressed in range of [0-~181.02].
18063 Display the average saturation value within the input frame. Expressed in range
18067 Display the saturation value at the 90% percentile within the input frame.
18068 Expressed in range of [0-~181.02].
18071 Display the maximum saturation value contained within the input frame.
18072 Expressed in range of [0-~181.02].
18075 Display the median value for hue within the input frame. Expressed in range of
18079 Display the average value for hue within the input frame. Expressed in range of
18083 Display the average of sample value difference between all values of the Y
18084 plane in the current frame and corresponding values of the previous input frame.
18085 Expressed in range of [0-255].
18088 Display the average of sample value difference between all values of the U
18089 plane in the current frame and corresponding values of the previous input frame.
18090 Expressed in range of [0-255].
18093 Display the average of sample value difference between all values of the V
18094 plane in the current frame and corresponding values of the previous input frame.
18095 Expressed in range of [0-255].
18098 Display bit depth of Y plane in current frame.
18099 Expressed in range of [0-16].
18102 Display bit depth of U plane in current frame.
18103 Expressed in range of [0-16].
18106 Display bit depth of V plane in current frame.
18107 Expressed in range of [0-16].
18110 The filter accepts the following options:
18116 @option{stat} specify an additional form of image analysis.
18117 @option{out} output video with the specified type of pixel highlighted.
18119 Both options accept the following values:
18123 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18124 unlike the neighboring pixels of the same field. Examples of temporal outliers
18125 include the results of video dropouts, head clogs, or tape tracking issues.
18128 Identify @var{vertical line repetition}. Vertical line repetition includes
18129 similar rows of pixels within a frame. In born-digital video vertical line
18130 repetition is common, but this pattern is uncommon in video digitized from an
18131 analog source. When it occurs in video that results from the digitization of an
18132 analog source it can indicate concealment from a dropout compensator.
18135 Identify pixels that fall outside of legal broadcast range.
18139 Set the highlight color for the @option{out} option. The default color is
18143 @subsection Examples
18147 Output data of various video metrics:
18149 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18153 Output specific data about the minimum and maximum values of the Y plane per frame:
18155 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18159 Playback video while highlighting pixels that are outside of broadcast range in red.
18161 ffplay example.mov -vf signalstats="out=brng:color=red"
18165 Playback video with signalstats metadata drawn over the frame.
18167 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18170 The contents of signalstat_drawtext.txt used in the command are:
18173 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18174 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18175 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18176 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18184 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18185 input. In this case the matching between the inputs can be calculated additionally.
18186 The filter always passes through the first input. The signature of each stream can
18187 be written into a file.
18189 It accepts the following options:
18193 Enable or disable the matching process.
18195 Available values are:
18199 Disable the calculation of a matching (default).
18201 Calculate the matching for the whole video and output whether the whole video
18202 matches or only parts.
18204 Calculate only until a matching is found or the video ends. Should be faster in
18209 Set the number of inputs. The option value must be a non negative integer.
18210 Default value is 1.
18213 Set the path to which the output is written. If there is more than one input,
18214 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18215 integer), that will be replaced with the input number. If no filename is
18216 specified, no output will be written. This is the default.
18219 Choose the output format.
18221 Available values are:
18225 Use the specified binary representation (default).
18227 Use the specified xml representation.
18231 Set threshold to detect one word as similar. The option value must be an integer
18232 greater than zero. The default value is 9000.
18235 Set threshold to detect all words as similar. The option value must be an integer
18236 greater than zero. The default value is 60000.
18239 Set threshold to detect frames as similar. The option value must be an integer
18240 greater than zero. The default value is 116.
18243 Set the minimum length of a sequence in frames to recognize it as matching
18244 sequence. The option value must be a non negative integer value.
18245 The default value is 0.
18248 Set the minimum relation, that matching frames to all frames must have.
18249 The option value must be a double value between 0 and 1. The default value is 0.5.
18252 @subsection Examples
18256 To calculate the signature of an input video and store it in signature.bin:
18258 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18262 To detect whether two videos match and store the signatures in XML format in
18263 signature0.xml and signature1.xml:
18265 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 -
18273 Blur the input video without impacting the outlines.
18275 It accepts the following options:
18278 @item luma_radius, lr
18279 Set the luma radius. The option value must be a float number in
18280 the range [0.1,5.0] that specifies the variance of the gaussian filter
18281 used to blur the image (slower if larger). Default value is 1.0.
18283 @item luma_strength, ls
18284 Set the luma strength. The option value must be a float number
18285 in the range [-1.0,1.0] that configures the blurring. A value included
18286 in [0.0,1.0] will blur the image whereas a value included in
18287 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18289 @item luma_threshold, lt
18290 Set the luma threshold used as a coefficient to determine
18291 whether a pixel should be blurred or not. The option value must be an
18292 integer in the range [-30,30]. A value of 0 will filter all the image,
18293 a value included in [0,30] will filter flat areas and a value included
18294 in [-30,0] will filter edges. Default value is 0.
18296 @item chroma_radius, cr
18297 Set the chroma radius. The option value must be a float number in
18298 the range [0.1,5.0] that specifies the variance of the gaussian filter
18299 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18301 @item chroma_strength, cs
18302 Set the chroma strength. The option value must be a float number
18303 in the range [-1.0,1.0] that configures the blurring. A value included
18304 in [0.0,1.0] will blur the image whereas a value included in
18305 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18307 @item chroma_threshold, ct
18308 Set the chroma threshold used as a coefficient to determine
18309 whether a pixel should be blurred or not. The option value must be an
18310 integer in the range [-30,30]. A value of 0 will filter all the image,
18311 a value included in [0,30] will filter flat areas and a value included
18312 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18315 If a chroma option is not explicitly set, the corresponding luma value
18319 Apply sobel operator to input video stream.
18321 The filter accepts the following option:
18325 Set which planes will be processed, unprocessed planes will be copied.
18326 By default value 0xf, all planes will be processed.
18329 Set value which will be multiplied with filtered result.
18332 Set value which will be added to filtered result.
18335 @subsection Commands
18337 This filter supports the all above options as @ref{commands}.
18342 Apply a simple postprocessing filter that compresses and decompresses the image
18343 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18344 and average the results.
18346 The filter accepts the following options:
18350 Set quality. This option defines the number of levels for averaging. It accepts
18351 an integer in the range 0-6. If set to @code{0}, the filter will have no
18352 effect. A value of @code{6} means the higher quality. For each increment of
18353 that value the speed drops by a factor of approximately 2. Default value is
18357 Force a constant quantization parameter. If not set, the filter will use the QP
18358 from the video stream (if available).
18361 Set thresholding mode. Available modes are:
18365 Set hard thresholding (default).
18367 Set soft thresholding (better de-ringing effect, but likely blurrier).
18370 @item use_bframe_qp
18371 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18372 option may cause flicker since the B-Frames have often larger QP. Default is
18373 @code{0} (not enabled).
18376 @subsection Commands
18378 This filter supports the following commands:
18380 @item quality, level
18381 Set quality level. The value @code{max} can be used to set the maximum level,
18382 currently @code{6}.
18388 Scale the input by applying one of the super-resolution methods based on
18389 convolutional neural networks. Supported models:
18393 Super-Resolution Convolutional Neural Network model (SRCNN).
18394 See @url{https://arxiv.org/abs/1501.00092}.
18397 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18398 See @url{https://arxiv.org/abs/1609.05158}.
18401 Training scripts as well as scripts for model file (.pb) saving can be found at
18402 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18403 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18405 Native model files (.model) can be generated from TensorFlow model
18406 files (.pb) by using tools/python/convert.py
18408 The filter accepts the following options:
18412 Specify which DNN backend to use for model loading and execution. This option accepts
18413 the following values:
18417 Native implementation of DNN loading and execution.
18420 TensorFlow backend. To enable this backend you
18421 need to install the TensorFlow for C library (see
18422 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18423 @code{--enable-libtensorflow}
18426 Default value is @samp{native}.
18429 Set path to model file specifying network architecture and its parameters.
18430 Note that different backends use different file formats. TensorFlow backend
18431 can load files for both formats, while native backend can load files for only
18435 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18436 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18437 input upscaled using bicubic upscaling with proper scale factor.
18440 This feature can also be finished with @ref{dnn_processing} filter.
18444 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18446 This filter takes in input two input videos, the first input is
18447 considered the "main" source and is passed unchanged to the
18448 output. The second input is used as a "reference" video for computing
18451 Both video inputs must have the same resolution and pixel format for
18452 this filter to work correctly. Also it assumes that both inputs
18453 have the same number of frames, which are compared one by one.
18455 The filter stores the calculated SSIM of each frame.
18457 The description of the accepted parameters follows.
18460 @item stats_file, f
18461 If specified the filter will use the named file to save the SSIM of
18462 each individual frame. When filename equals "-" the data is sent to
18466 The file printed if @var{stats_file} is selected, contains a sequence of
18467 key/value pairs of the form @var{key}:@var{value} for each compared
18470 A description of each shown parameter follows:
18474 sequential number of the input frame, starting from 1
18476 @item Y, U, V, R, G, B
18477 SSIM of the compared frames for the component specified by the suffix.
18480 SSIM of the compared frames for the whole frame.
18483 Same as above but in dB representation.
18486 This filter also supports the @ref{framesync} options.
18488 @subsection Examples
18493 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18494 [main][ref] ssim="stats_file=stats.log" [out]
18497 On this example the input file being processed is compared with the
18498 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18499 is stored in @file{stats.log}.
18502 Another example with both psnr and ssim at same time:
18504 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18508 Another example with different containers:
18510 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 -
18516 Convert between different stereoscopic image formats.
18518 The filters accept the following options:
18522 Set stereoscopic image format of input.
18524 Available values for input image formats are:
18527 side by side parallel (left eye left, right eye right)
18530 side by side crosseye (right eye left, left eye right)
18533 side by side parallel with half width resolution
18534 (left eye left, right eye right)
18537 side by side crosseye with half width resolution
18538 (right eye left, left eye right)
18542 above-below (left eye above, right eye below)
18546 above-below (right eye above, left eye below)
18550 above-below with half height resolution
18551 (left eye above, right eye below)
18555 above-below with half height resolution
18556 (right eye above, left eye below)
18559 alternating frames (left eye first, right eye second)
18562 alternating frames (right eye first, left eye second)
18565 interleaved rows (left eye has top row, right eye starts on next row)
18568 interleaved rows (right eye has top row, left eye starts on next row)
18571 interleaved columns, left eye first
18574 interleaved columns, right eye first
18576 Default value is @samp{sbsl}.
18580 Set stereoscopic image format of output.
18584 side by side parallel (left eye left, right eye right)
18587 side by side crosseye (right eye left, left eye right)
18590 side by side parallel with half width resolution
18591 (left eye left, right eye right)
18594 side by side crosseye with half width resolution
18595 (right eye left, left eye right)
18599 above-below (left eye above, right eye below)
18603 above-below (right eye above, left eye below)
18607 above-below with half height resolution
18608 (left eye above, right eye below)
18612 above-below with half height resolution
18613 (right eye above, left eye below)
18616 alternating frames (left eye first, right eye second)
18619 alternating frames (right eye first, left eye second)
18622 interleaved rows (left eye has top row, right eye starts on next row)
18625 interleaved rows (right eye has top row, left eye starts on next row)
18628 anaglyph red/blue gray
18629 (red filter on left eye, blue filter on right eye)
18632 anaglyph red/green gray
18633 (red filter on left eye, green filter on right eye)
18636 anaglyph red/cyan gray
18637 (red filter on left eye, cyan filter on right eye)
18640 anaglyph red/cyan half colored
18641 (red filter on left eye, cyan filter on right eye)
18644 anaglyph red/cyan color
18645 (red filter on left eye, cyan filter on right eye)
18648 anaglyph red/cyan color optimized with the least squares projection of dubois
18649 (red filter on left eye, cyan filter on right eye)
18652 anaglyph green/magenta gray
18653 (green filter on left eye, magenta filter on right eye)
18656 anaglyph green/magenta half colored
18657 (green filter on left eye, magenta filter on right eye)
18660 anaglyph green/magenta colored
18661 (green filter on left eye, magenta filter on right eye)
18664 anaglyph green/magenta color optimized with the least squares projection of dubois
18665 (green filter on left eye, magenta filter on right eye)
18668 anaglyph yellow/blue gray
18669 (yellow filter on left eye, blue filter on right eye)
18672 anaglyph yellow/blue half colored
18673 (yellow filter on left eye, blue filter on right eye)
18676 anaglyph yellow/blue colored
18677 (yellow filter on left eye, blue filter on right eye)
18680 anaglyph yellow/blue color optimized with the least squares projection of dubois
18681 (yellow filter on left eye, blue filter on right eye)
18684 mono output (left eye only)
18687 mono output (right eye only)
18690 checkerboard, left eye first
18693 checkerboard, right eye first
18696 interleaved columns, left eye first
18699 interleaved columns, right eye first
18705 Default value is @samp{arcd}.
18708 @subsection Examples
18712 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18718 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18724 @section streamselect, astreamselect
18725 Select video or audio streams.
18727 The filter accepts the following options:
18731 Set number of inputs. Default is 2.
18734 Set input indexes to remap to outputs.
18737 @subsection Commands
18739 The @code{streamselect} and @code{astreamselect} filter supports the following
18744 Set input indexes to remap to outputs.
18747 @subsection Examples
18751 Select first 5 seconds 1st stream and rest of time 2nd stream:
18753 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18757 Same as above, but for audio:
18759 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18766 Draw subtitles on top of input video using the libass library.
18768 To enable compilation of this filter you need to configure FFmpeg with
18769 @code{--enable-libass}. This filter also requires a build with libavcodec and
18770 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18771 Alpha) subtitles format.
18773 The filter accepts the following options:
18777 Set the filename of the subtitle file to read. It must be specified.
18779 @item original_size
18780 Specify the size of the original video, the video for which the ASS file
18781 was composed. For the syntax of this option, check the
18782 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18783 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18784 correctly scale the fonts if the aspect ratio has been changed.
18787 Set a directory path containing fonts that can be used by the filter.
18788 These fonts will be used in addition to whatever the font provider uses.
18791 Process alpha channel, by default alpha channel is untouched.
18794 Set subtitles input character encoding. @code{subtitles} filter only. Only
18795 useful if not UTF-8.
18797 @item stream_index, si
18798 Set subtitles stream index. @code{subtitles} filter only.
18801 Override default style or script info parameters of the subtitles. It accepts a
18802 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18805 If the first key is not specified, it is assumed that the first value
18806 specifies the @option{filename}.
18808 For example, to render the file @file{sub.srt} on top of the input
18809 video, use the command:
18814 which is equivalent to:
18816 subtitles=filename=sub.srt
18819 To render the default subtitles stream from file @file{video.mkv}, use:
18821 subtitles=video.mkv
18824 To render the second subtitles stream from that file, use:
18826 subtitles=video.mkv:si=1
18829 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18830 @code{DejaVu Serif}, use:
18832 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18835 @section super2xsai
18837 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18838 Interpolate) pixel art scaling algorithm.
18840 Useful for enlarging pixel art images without reducing sharpness.
18844 Swap two rectangular objects in video.
18846 This filter accepts the following options:
18856 Set 1st rect x coordinate.
18859 Set 1st rect y coordinate.
18862 Set 2nd rect x coordinate.
18865 Set 2nd rect y coordinate.
18867 All expressions are evaluated once for each frame.
18870 The all options are expressions containing the following constants:
18875 The input width and height.
18878 same as @var{w} / @var{h}
18881 input sample aspect ratio
18884 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18887 The number of the input frame, starting from 0.
18890 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18893 the position in the file of the input frame, NAN if unknown
18900 Blend successive video frames.
18906 Apply telecine process to the video.
18908 This filter accepts the following options:
18917 The default value is @code{top}.
18921 A string of numbers representing the pulldown pattern you wish to apply.
18922 The default value is @code{23}.
18926 Some typical patterns:
18931 24p: 2332 (preferred)
18938 24p: 222222222223 ("Euro pulldown")
18943 @section thistogram
18945 Compute and draw a color distribution histogram for the input video across time.
18947 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18948 at certain time, this filter shows also past histograms of number of frames defined
18949 by @code{width} option.
18951 The computed histogram is a representation of the color component
18952 distribution in an image.
18954 The filter accepts the following options:
18958 Set width of single color component output. Default value is @code{0}.
18959 Value of @code{0} means width will be picked from input video.
18960 This also set number of passed histograms to keep.
18961 Allowed range is [0, 8192].
18963 @item display_mode, d
18965 It accepts the following values:
18968 Per color component graphs are placed below each other.
18971 Per color component graphs are placed side by side.
18974 Presents information identical to that in the @code{parade}, except
18975 that the graphs representing color components are superimposed directly
18978 Default is @code{stack}.
18980 @item levels_mode, m
18981 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18982 Default is @code{linear}.
18984 @item components, c
18985 Set what color components to display.
18986 Default is @code{7}.
18989 Set background opacity. Default is @code{0.9}.
18992 Show envelope. Default is disabled.
18995 Set envelope color. Default is @code{gold}.
19000 Available values for slide is:
19003 Draw new frame when right border is reached.
19006 Replace old columns with new ones.
19009 Scroll from right to left.
19012 Scroll from left to right.
19015 Draw single picture.
19018 Default is @code{replace}.
19023 Apply threshold effect to video stream.
19025 This filter needs four video streams to perform thresholding.
19026 First stream is stream we are filtering.
19027 Second stream is holding threshold values, third stream is holding min values,
19028 and last, fourth stream is holding max values.
19030 The filter accepts the following option:
19034 Set which planes will be processed, unprocessed planes will be copied.
19035 By default value 0xf, all planes will be processed.
19038 For example if first stream pixel's component value is less then threshold value
19039 of pixel component from 2nd threshold stream, third stream value will picked,
19040 otherwise fourth stream pixel component value will be picked.
19042 Using color source filter one can perform various types of thresholding:
19044 @subsection Examples
19048 Binary threshold, using gray color as threshold:
19050 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19054 Inverted binary threshold, using gray color as threshold:
19056 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19060 Truncate binary threshold, using gray color as threshold:
19062 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19066 Threshold to zero, using gray color as threshold:
19068 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19072 Inverted threshold to zero, using gray color as threshold:
19074 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19079 Select the most representative frame in a given sequence of consecutive frames.
19081 The filter accepts the following options:
19085 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19086 will pick one of them, and then handle the next batch of @var{n} frames until
19087 the end. Default is @code{100}.
19090 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19091 value will result in a higher memory usage, so a high value is not recommended.
19093 @subsection Examples
19097 Extract one picture each 50 frames:
19103 Complete example of a thumbnail creation with @command{ffmpeg}:
19105 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19112 Tile several successive frames together.
19114 The @ref{untile} filter can do the reverse.
19116 The filter accepts the following options:
19121 Set the grid size (i.e. the number of lines and columns). For the syntax of
19122 this option, check the
19123 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19126 Set the maximum number of frames to render in the given area. It must be less
19127 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19128 the area will be used.
19131 Set the outer border margin in pixels.
19134 Set the inner border thickness (i.e. the number of pixels between frames). For
19135 more advanced padding options (such as having different values for the edges),
19136 refer to the pad video filter.
19139 Specify the color of the unused area. For the syntax of this option, check the
19140 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19141 The default value of @var{color} is "black".
19144 Set the number of frames to overlap when tiling several successive frames together.
19145 The value must be between @code{0} and @var{nb_frames - 1}.
19148 Set the number of frames to initially be empty before displaying first output frame.
19149 This controls how soon will one get first output frame.
19150 The value must be between @code{0} and @var{nb_frames - 1}.
19153 @subsection Examples
19157 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19159 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19161 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19162 duplicating each output frame to accommodate the originally detected frame
19166 Display @code{5} pictures in an area of @code{3x2} frames,
19167 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19168 mixed flat and named options:
19170 tile=3x2:nb_frames=5:padding=7:margin=2
19174 @section tinterlace
19176 Perform various types of temporal field interlacing.
19178 Frames are counted starting from 1, so the first input frame is
19181 The filter accepts the following options:
19186 Specify the mode of the interlacing. This option can also be specified
19187 as a value alone. See below for a list of values for this option.
19189 Available values are:
19193 Move odd frames into the upper field, even into the lower field,
19194 generating a double height frame at half frame rate.
19198 Frame 1 Frame 2 Frame 3 Frame 4
19200 11111 22222 33333 44444
19201 11111 22222 33333 44444
19202 11111 22222 33333 44444
19203 11111 22222 33333 44444
19217 Only output odd frames, even frames are dropped, generating a frame with
19218 unchanged height at half frame rate.
19223 Frame 1 Frame 2 Frame 3 Frame 4
19225 11111 22222 33333 44444
19226 11111 22222 33333 44444
19227 11111 22222 33333 44444
19228 11111 22222 33333 44444
19238 Only output even frames, odd frames are dropped, generating a frame with
19239 unchanged height at half frame rate.
19244 Frame 1 Frame 2 Frame 3 Frame 4
19246 11111 22222 33333 44444
19247 11111 22222 33333 44444
19248 11111 22222 33333 44444
19249 11111 22222 33333 44444
19259 Expand each frame to full height, but pad alternate lines with black,
19260 generating a frame with double height at the same input frame rate.
19265 Frame 1 Frame 2 Frame 3 Frame 4
19267 11111 22222 33333 44444
19268 11111 22222 33333 44444
19269 11111 22222 33333 44444
19270 11111 22222 33333 44444
19273 11111 ..... 33333 .....
19274 ..... 22222 ..... 44444
19275 11111 ..... 33333 .....
19276 ..... 22222 ..... 44444
19277 11111 ..... 33333 .....
19278 ..... 22222 ..... 44444
19279 11111 ..... 33333 .....
19280 ..... 22222 ..... 44444
19284 @item interleave_top, 4
19285 Interleave the upper field from odd frames with the lower field from
19286 even frames, generating a frame with unchanged height at half frame rate.
19291 Frame 1 Frame 2 Frame 3 Frame 4
19293 11111<- 22222 33333<- 44444
19294 11111 22222<- 33333 44444<-
19295 11111<- 22222 33333<- 44444
19296 11111 22222<- 33333 44444<-
19306 @item interleave_bottom, 5
19307 Interleave the lower field from odd frames with the upper field from
19308 even frames, generating a frame with unchanged height at half frame rate.
19313 Frame 1 Frame 2 Frame 3 Frame 4
19315 11111 22222<- 33333 44444<-
19316 11111<- 22222 33333<- 44444
19317 11111 22222<- 33333 44444<-
19318 11111<- 22222 33333<- 44444
19328 @item interlacex2, 6
19329 Double frame rate with unchanged height. Frames are inserted each
19330 containing the second temporal field from the previous input frame and
19331 the first temporal field from the next input frame. This mode relies on
19332 the top_field_first flag. Useful for interlaced video displays with no
19333 field synchronisation.
19338 Frame 1 Frame 2 Frame 3 Frame 4
19340 11111 22222 33333 44444
19341 11111 22222 33333 44444
19342 11111 22222 33333 44444
19343 11111 22222 33333 44444
19346 11111 22222 22222 33333 33333 44444 44444
19347 11111 11111 22222 22222 33333 33333 44444
19348 11111 22222 22222 33333 33333 44444 44444
19349 11111 11111 22222 22222 33333 33333 44444
19354 Move odd frames into the upper field, even into the lower field,
19355 generating a double height frame at same frame rate.
19360 Frame 1 Frame 2 Frame 3 Frame 4
19362 11111 22222 33333 44444
19363 11111 22222 33333 44444
19364 11111 22222 33333 44444
19365 11111 22222 33333 44444
19368 11111 33333 33333 55555
19369 22222 22222 44444 44444
19370 11111 33333 33333 55555
19371 22222 22222 44444 44444
19372 11111 33333 33333 55555
19373 22222 22222 44444 44444
19374 11111 33333 33333 55555
19375 22222 22222 44444 44444
19380 Numeric values are deprecated but are accepted for backward
19381 compatibility reasons.
19383 Default mode is @code{merge}.
19386 Specify flags influencing the filter process.
19388 Available value for @var{flags} is:
19391 @item low_pass_filter, vlpf
19392 Enable linear vertical low-pass filtering in the filter.
19393 Vertical low-pass filtering is required when creating an interlaced
19394 destination from a progressive source which contains high-frequency
19395 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19398 @item complex_filter, cvlpf
19399 Enable complex vertical low-pass filtering.
19400 This will slightly less reduce interlace 'twitter' and Moire
19401 patterning but better retain detail and subjective sharpness impression.
19404 Bypass already interlaced frames, only adjust the frame rate.
19407 Vertical low-pass filtering and bypassing already interlaced frames can only be
19408 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19413 Pick median pixels from several successive input video frames.
19415 The filter accepts the following options:
19419 Set radius of median filter.
19420 Default is 1. Allowed range is from 1 to 127.
19423 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19426 Set median percentile. Default value is @code{0.5}.
19427 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19428 minimum values, and @code{1} maximum values.
19431 @subsection Commands
19433 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19435 @section tmidequalizer
19437 Apply Temporal Midway Video Equalization effect.
19439 Midway Video Equalization adjusts a sequence of video frames to have the same
19440 histograms, while maintaining their dynamics as much as possible. It's
19441 useful for e.g. matching exposures from a video frames sequence.
19443 This filter accepts the following option:
19447 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19450 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19451 Setting this option to 0 effectively does nothing.
19454 Set which planes to process. Default is @code{15}, which is all available planes.
19459 Mix successive video frames.
19461 A description of the accepted options follows.
19465 The number of successive frames to mix. If unspecified, it defaults to 3.
19468 Specify weight of each input video frame.
19469 Each weight is separated by space. If number of weights is smaller than
19470 number of @var{frames} last specified weight will be used for all remaining
19474 Specify scale, if it is set it will be multiplied with sum
19475 of each weight multiplied with pixel values to give final destination
19476 pixel value. By default @var{scale} is auto scaled to sum of weights.
19479 @subsection Examples
19483 Average 7 successive frames:
19485 tmix=frames=7:weights="1 1 1 1 1 1 1"
19489 Apply simple temporal convolution:
19491 tmix=frames=3:weights="-1 3 -1"
19495 Similar as above but only showing temporal differences:
19497 tmix=frames=3:weights="-1 2 -1":scale=1
19503 Tone map colors from different dynamic ranges.
19505 This filter expects data in single precision floating point, as it needs to
19506 operate on (and can output) out-of-range values. Another filter, such as
19507 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19509 The tonemapping algorithms implemented only work on linear light, so input
19510 data should be linearized beforehand (and possibly correctly tagged).
19513 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19516 @subsection Options
19517 The filter accepts the following options.
19521 Set the tone map algorithm to use.
19523 Possible values are:
19526 Do not apply any tone map, only desaturate overbright pixels.
19529 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19530 in-range values, while distorting out-of-range values.
19533 Stretch the entire reference gamut to a linear multiple of the display.
19536 Fit a logarithmic transfer between the tone curves.
19539 Preserve overall image brightness with a simple curve, using nonlinear
19540 contrast, which results in flattening details and degrading color accuracy.
19543 Preserve both dark and bright details better than @var{reinhard}, at the cost
19544 of slightly darkening everything. Use it when detail preservation is more
19545 important than color and brightness accuracy.
19548 Smoothly map out-of-range values, while retaining contrast and colors for
19549 in-range material as much as possible. Use it when color accuracy is more
19550 important than detail preservation.
19556 Tune the tone mapping algorithm.
19558 This affects the following algorithms:
19564 Specifies the scale factor to use while stretching.
19568 Specifies the exponent of the function.
19572 Specify an extra linear coefficient to multiply into the signal before clipping.
19576 Specify the local contrast coefficient at the display peak.
19577 Default to 0.5, which means that in-gamut values will be about half as bright
19584 Specify the transition point from linear to mobius transform. Every value
19585 below this point is guaranteed to be mapped 1:1. The higher the value, the
19586 more accurate the result will be, at the cost of losing bright details.
19587 Default to 0.3, which due to the steep initial slope still preserves in-range
19588 colors fairly accurately.
19592 Apply desaturation for highlights that exceed this level of brightness. The
19593 higher the parameter, the more color information will be preserved. This
19594 setting helps prevent unnaturally blown-out colors for super-highlights, by
19595 (smoothly) turning into white instead. This makes images feel more natural,
19596 at the cost of reducing information about out-of-range colors.
19598 The default of 2.0 is somewhat conservative and will mostly just apply to
19599 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19601 This option works only if the input frame has a supported color tag.
19604 Override signal/nominal/reference peak with this value. Useful when the
19605 embedded peak information in display metadata is not reliable or when tone
19606 mapping from a lower range to a higher range.
19611 Temporarily pad video frames.
19613 The filter accepts the following options:
19617 Specify number of delay frames before input video stream. Default is 0.
19620 Specify number of padding frames after input video stream.
19621 Set to -1 to pad indefinitely. Default is 0.
19624 Set kind of frames added to beginning of stream.
19625 Can be either @var{add} or @var{clone}.
19626 With @var{add} frames of solid-color are added.
19627 With @var{clone} frames are clones of first frame.
19628 Default is @var{add}.
19631 Set kind of frames added to end of stream.
19632 Can be either @var{add} or @var{clone}.
19633 With @var{add} frames of solid-color are added.
19634 With @var{clone} frames are clones of last frame.
19635 Default is @var{add}.
19637 @item start_duration, stop_duration
19638 Specify the duration of the start/stop delay. See
19639 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19640 for the accepted syntax.
19641 These options override @var{start} and @var{stop}. Default is 0.
19644 Specify the color of the padded area. For the syntax of this option,
19645 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19646 manual,ffmpeg-utils}.
19648 The default value of @var{color} is "black".
19654 Transpose rows with columns in the input video and optionally flip it.
19656 It accepts the following parameters:
19661 Specify the transposition direction.
19663 Can assume the following values:
19665 @item 0, 4, cclock_flip
19666 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19674 Rotate by 90 degrees clockwise, that is:
19682 Rotate by 90 degrees counterclockwise, that is:
19689 @item 3, 7, clock_flip
19690 Rotate by 90 degrees clockwise and vertically flip, that is:
19698 For values between 4-7, the transposition is only done if the input
19699 video geometry is portrait and not landscape. These values are
19700 deprecated, the @code{passthrough} option should be used instead.
19702 Numerical values are deprecated, and should be dropped in favor of
19703 symbolic constants.
19706 Do not apply the transposition if the input geometry matches the one
19707 specified by the specified value. It accepts the following values:
19710 Always apply transposition.
19712 Preserve portrait geometry (when @var{height} >= @var{width}).
19714 Preserve landscape geometry (when @var{width} >= @var{height}).
19717 Default value is @code{none}.
19720 For example to rotate by 90 degrees clockwise and preserve portrait
19723 transpose=dir=1:passthrough=portrait
19726 The command above can also be specified as:
19728 transpose=1:portrait
19731 @section transpose_npp
19733 Transpose rows with columns in the input video and optionally flip it.
19734 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19736 It accepts the following parameters:
19741 Specify the transposition direction.
19743 Can assume the following values:
19746 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19749 Rotate by 90 degrees clockwise.
19752 Rotate by 90 degrees counterclockwise.
19755 Rotate by 90 degrees clockwise and vertically flip.
19759 Do not apply the transposition if the input geometry matches the one
19760 specified by the specified value. It accepts the following values:
19763 Always apply transposition. (default)
19765 Preserve portrait geometry (when @var{height} >= @var{width}).
19767 Preserve landscape geometry (when @var{width} >= @var{height}).
19773 Trim the input so that the output contains one continuous subpart of the input.
19775 It accepts the following parameters:
19778 Specify the time of the start of the kept section, i.e. the frame with the
19779 timestamp @var{start} will be the first frame in the output.
19782 Specify the time of the first frame that will be dropped, i.e. the frame
19783 immediately preceding the one with the timestamp @var{end} will be the last
19784 frame in the output.
19787 This is the same as @var{start}, except this option sets the start timestamp
19788 in timebase units instead of seconds.
19791 This is the same as @var{end}, except this option sets the end timestamp
19792 in timebase units instead of seconds.
19795 The maximum duration of the output in seconds.
19798 The number of the first frame that should be passed to the output.
19801 The number of the first frame that should be dropped.
19804 @option{start}, @option{end}, and @option{duration} are expressed as time
19805 duration specifications; see
19806 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19807 for the accepted syntax.
19809 Note that the first two sets of the start/end options and the @option{duration}
19810 option look at the frame timestamp, while the _frame variants simply count the
19811 frames that pass through the filter. Also note that this filter does not modify
19812 the timestamps. If you wish for the output timestamps to start at zero, insert a
19813 setpts filter after the trim filter.
19815 If multiple start or end options are set, this filter tries to be greedy and
19816 keep all the frames that match at least one of the specified constraints. To keep
19817 only the part that matches all the constraints at once, chain multiple trim
19820 The defaults are such that all the input is kept. So it is possible to set e.g.
19821 just the end values to keep everything before the specified time.
19826 Drop everything except the second minute of input:
19828 ffmpeg -i INPUT -vf trim=60:120
19832 Keep only the first second:
19834 ffmpeg -i INPUT -vf trim=duration=1
19839 @section unpremultiply
19840 Apply alpha unpremultiply effect to input video stream using first plane
19841 of second stream as alpha.
19843 Both streams must have same dimensions and same pixel format.
19845 The filter accepts the following option:
19849 Set which planes will be processed, unprocessed planes will be copied.
19850 By default value 0xf, all planes will be processed.
19852 If the format has 1 or 2 components, then luma is bit 0.
19853 If the format has 3 or 4 components:
19854 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19855 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19856 If present, the alpha channel is always the last bit.
19859 Do not require 2nd input for processing, instead use alpha plane from input stream.
19865 Sharpen or blur the input video.
19867 It accepts the following parameters:
19870 @item luma_msize_x, lx
19871 Set the luma matrix horizontal size. It must be an odd integer between
19872 3 and 23. The default value is 5.
19874 @item luma_msize_y, ly
19875 Set the luma matrix vertical size. It must be an odd integer between 3
19876 and 23. The default value is 5.
19878 @item luma_amount, la
19879 Set the luma effect strength. It must be a floating point number, reasonable
19880 values lay between -1.5 and 1.5.
19882 Negative values will blur the input video, while positive values will
19883 sharpen it, a value of zero will disable the effect.
19885 Default value is 1.0.
19887 @item chroma_msize_x, cx
19888 Set the chroma matrix horizontal size. It must be an odd integer
19889 between 3 and 23. The default value is 5.
19891 @item chroma_msize_y, cy
19892 Set the chroma matrix vertical size. It must be an odd integer
19893 between 3 and 23. The default value is 5.
19895 @item chroma_amount, ca
19896 Set the chroma effect strength. It must be a floating point number, reasonable
19897 values lay between -1.5 and 1.5.
19899 Negative values will blur the input video, while positive values will
19900 sharpen it, a value of zero will disable the effect.
19902 Default value is 0.0.
19906 All parameters are optional and default to the equivalent of the
19907 string '5:5:1.0:5:5:0.0'.
19909 @subsection Examples
19913 Apply strong luma sharpen effect:
19915 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19919 Apply a strong blur of both luma and chroma parameters:
19921 unsharp=7:7:-2:7:7:-2
19928 Decompose a video made of tiled images into the individual images.
19930 The frame rate of the output video is the frame rate of the input video
19931 multiplied by the number of tiles.
19933 This filter does the reverse of @ref{tile}.
19935 The filter accepts the following options:
19940 Set the grid size (i.e. the number of lines and columns). For the syntax of
19941 this option, check the
19942 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19945 @subsection Examples
19949 Produce a 1-second video from a still image file made of 25 frames stacked
19950 vertically, like an analogic film reel:
19952 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19958 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19959 the image at several (or - in the case of @option{quality} level @code{8} - all)
19960 shifts and average the results.
19962 The way this differs from the behavior of spp is that uspp actually encodes &
19963 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19964 DCT similar to MJPEG.
19966 The filter accepts the following options:
19970 Set quality. This option defines the number of levels for averaging. It accepts
19971 an integer in the range 0-8. If set to @code{0}, the filter will have no
19972 effect. A value of @code{8} means the higher quality. For each increment of
19973 that value the speed drops by a factor of approximately 2. Default value is
19977 Force a constant quantization parameter. If not set, the filter will use the QP
19978 from the video stream (if available).
19983 Convert 360 videos between various formats.
19985 The filter accepts the following options:
19991 Set format of the input/output video.
19999 Equirectangular projection.
20004 Cubemap with 3x2/6x1/1x6 layout.
20006 Format specific options:
20011 Set padding proportion for the input/output cubemap. Values in decimals.
20018 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)
20021 Default value is @b{@samp{0}}.
20022 Maximum value is @b{@samp{0.1}}.
20026 Set fixed padding for the input/output cubemap. Values in pixels.
20028 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
20032 Set order of faces for the input/output cubemap. Choose one direction for each position.
20034 Designation of directions:
20050 Default value is @b{@samp{rludfb}}.
20054 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20056 Designation of angles:
20059 0 degrees clockwise
20061 90 degrees clockwise
20063 180 degrees clockwise
20065 270 degrees clockwise
20068 Default value is @b{@samp{000000}}.
20072 Equi-Angular Cubemap.
20079 Format specific options:
20084 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20086 If diagonal field of view is set it overrides horizontal and vertical field of view.
20091 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20093 If diagonal field of view is set it overrides horizontal and vertical field of view.
20099 Format specific options:
20104 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20106 If diagonal field of view is set it overrides horizontal and vertical field of view.
20111 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20113 If diagonal field of view is set it overrides horizontal and vertical field of view.
20119 Facebook's 360 formats.
20122 Stereographic format.
20124 Format specific options:
20129 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20131 If diagonal field of view is set it overrides horizontal and vertical field of view.
20136 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20138 If diagonal field of view is set it overrides horizontal and vertical field of view.
20145 Ball format, gives significant distortion toward the back.
20148 Hammer-Aitoff map projection format.
20151 Sinusoidal map projection format.
20154 Fisheye projection.
20156 Format specific options:
20161 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20163 If diagonal field of view is set it overrides horizontal and vertical field of view.
20168 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20170 If diagonal field of view is set it overrides horizontal and vertical field of view.
20174 Pannini projection.
20176 Format specific options:
20179 Set output pannini parameter.
20182 Set input pannini parameter.
20186 Cylindrical projection.
20188 Format specific options:
20193 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20195 If diagonal field of view is set it overrides horizontal and vertical field of view.
20200 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20202 If diagonal field of view is set it overrides horizontal and vertical field of view.
20206 Perspective projection. @i{(output only)}
20208 Format specific options:
20211 Set perspective parameter.
20215 Tetrahedron projection.
20218 Truncated square pyramid projection.
20222 Half equirectangular projection.
20227 Format specific options:
20232 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20234 If diagonal field of view is set it overrides horizontal and vertical field of view.
20239 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20241 If diagonal field of view is set it overrides horizontal and vertical field of view.
20245 Orthographic format.
20247 Format specific options:
20252 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20254 If diagonal field of view is set it overrides horizontal and vertical field of view.
20259 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20261 If diagonal field of view is set it overrides horizontal and vertical field of view.
20265 Octahedron projection.
20269 Set interpolation method.@*
20270 @i{Note: more complex interpolation methods require much more memory to run.}
20280 Bilinear interpolation.
20282 Lagrange9 interpolation.
20285 Bicubic interpolation.
20288 Lanczos interpolation.
20291 Spline16 interpolation.
20294 Gaussian interpolation.
20296 Mitchell interpolation.
20299 Default value is @b{@samp{line}}.
20303 Set the output video resolution.
20305 Default resolution depends on formats.
20309 Set the input/output stereo format.
20320 Default value is @b{@samp{2d}} for input and output format.
20325 Set rotation for the output video. Values in degrees.
20328 Set rotation order for the output video. Choose one item for each position.
20339 Default value is @b{@samp{ypr}}.
20344 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20348 Set if input video is flipped horizontally/vertically. Boolean values.
20351 Set if input video is transposed. Boolean value, by default disabled.
20354 Set if output video needs to be transposed. Boolean value, by default disabled.
20357 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20360 @subsection Examples
20364 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20366 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20369 Extract back view of Equi-Angular Cubemap:
20371 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20374 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20376 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20380 @subsection Commands
20382 This filter supports subset of above options as @ref{commands}.
20384 @section vaguedenoiser
20386 Apply a wavelet based denoiser.
20388 It transforms each frame from the video input into the wavelet domain,
20389 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20390 the obtained coefficients. It does an inverse wavelet transform after.
20391 Due to wavelet properties, it should give a nice smoothed result, and
20392 reduced noise, without blurring picture features.
20394 This filter accepts the following options:
20398 The filtering strength. The higher, the more filtered the video will be.
20399 Hard thresholding can use a higher threshold than soft thresholding
20400 before the video looks overfiltered. Default value is 2.
20403 The filtering method the filter will use.
20405 It accepts the following values:
20408 All values under the threshold will be zeroed.
20411 All values under the threshold will be zeroed. All values above will be
20412 reduced by the threshold.
20415 Scales or nullifies coefficients - intermediary between (more) soft and
20416 (less) hard thresholding.
20419 Default is garrote.
20422 Number of times, the wavelet will decompose the picture. Picture can't
20423 be decomposed beyond a particular point (typically, 8 for a 640x480
20424 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20427 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20430 A list of the planes to process. By default all planes are processed.
20433 The threshold type the filter will use.
20435 It accepts the following values:
20438 Threshold used is same for all decompositions.
20441 Threshold used depends also on each decomposition coefficients.
20444 Default is universal.
20447 @section vectorscope
20449 Display 2 color component values in the two dimensional graph (which is called
20452 This filter accepts the following options:
20456 Set vectorscope mode.
20458 It accepts the following values:
20462 Gray values are displayed on graph, higher brightness means more pixels have
20463 same component color value on location in graph. This is the default mode.
20466 Gray values are displayed on graph. Surrounding pixels values which are not
20467 present in video frame are drawn in gradient of 2 color components which are
20468 set by option @code{x} and @code{y}. The 3rd color component is static.
20471 Actual color components values present in video frame are displayed on graph.
20474 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20475 on graph increases value of another color component, which is luminance by
20476 default values of @code{x} and @code{y}.
20479 Actual colors present in video frame are displayed on graph. If two different
20480 colors map to same position on graph then color with higher value of component
20481 not present in graph is picked.
20484 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20485 component picked from radial gradient.
20489 Set which color component will be represented on X-axis. Default is @code{1}.
20492 Set which color component will be represented on Y-axis. Default is @code{2}.
20495 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20496 of color component which represents frequency of (X, Y) location in graph.
20501 No envelope, this is default.
20504 Instant envelope, even darkest single pixel will be clearly highlighted.
20507 Hold maximum and minimum values presented in graph over time. This way you
20508 can still spot out of range values without constantly looking at vectorscope.
20511 Peak and instant envelope combined together.
20515 Set what kind of graticule to draw.
20524 Set graticule opacity.
20527 Set graticule flags.
20531 Draw graticule for white point.
20534 Draw graticule for black point.
20537 Draw color points short names.
20541 Set background opacity.
20543 @item lthreshold, l
20544 Set low threshold for color component not represented on X or Y axis.
20545 Values lower than this value will be ignored. Default is 0.
20546 Note this value is multiplied with actual max possible value one pixel component
20547 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20550 @item hthreshold, h
20551 Set high threshold for color component not represented on X or Y axis.
20552 Values higher than this value will be ignored. Default is 1.
20553 Note this value is multiplied with actual max possible value one pixel component
20554 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20555 is 0.9 * 255 = 230.
20557 @item colorspace, c
20558 Set what kind of colorspace to use when drawing graticule.
20568 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20569 This means no tint, and output will remain gray.
20572 @anchor{vidstabdetect}
20573 @section vidstabdetect
20575 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20576 @ref{vidstabtransform} for pass 2.
20578 This filter generates a file with relative translation and rotation
20579 transform information about subsequent frames, which is then used by
20580 the @ref{vidstabtransform} filter.
20582 To enable compilation of this filter you need to configure FFmpeg with
20583 @code{--enable-libvidstab}.
20585 This filter accepts the following options:
20589 Set the path to the file used to write the transforms information.
20590 Default value is @file{transforms.trf}.
20593 Set how shaky the video is and how quick the camera is. It accepts an
20594 integer in the range 1-10, a value of 1 means little shakiness, a
20595 value of 10 means strong shakiness. Default value is 5.
20598 Set the accuracy of the detection process. It must be a value in the
20599 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20600 accuracy. Default value is 15.
20603 Set stepsize of the search process. The region around minimum is
20604 scanned with 1 pixel resolution. Default value is 6.
20607 Set minimum contrast. Below this value a local measurement field is
20608 discarded. Must be a floating point value in the range 0-1. Default
20612 Set reference frame number for tripod mode.
20614 If enabled, the motion of the frames is compared to a reference frame
20615 in the filtered stream, identified by the specified number. The idea
20616 is to compensate all movements in a more-or-less static scene and keep
20617 the camera view absolutely still.
20619 If set to 0, it is disabled. The frames are counted starting from 1.
20622 Show fields and transforms in the resulting frames. It accepts an
20623 integer in the range 0-2. Default value is 0, which disables any
20627 @subsection Examples
20631 Use default values:
20637 Analyze strongly shaky movie and put the results in file
20638 @file{mytransforms.trf}:
20640 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20644 Visualize the result of internal transformations in the resulting
20647 vidstabdetect=show=1
20651 Analyze a video with medium shakiness using @command{ffmpeg}:
20653 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20657 @anchor{vidstabtransform}
20658 @section vidstabtransform
20660 Video stabilization/deshaking: pass 2 of 2,
20661 see @ref{vidstabdetect} for pass 1.
20663 Read a file with transform information for each frame and
20664 apply/compensate them. Together with the @ref{vidstabdetect}
20665 filter this can be used to deshake videos. See also
20666 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20667 the @ref{unsharp} filter, see below.
20669 To enable compilation of this filter you need to configure FFmpeg with
20670 @code{--enable-libvidstab}.
20672 @subsection Options
20676 Set path to the file used to read the transforms. Default value is
20677 @file{transforms.trf}.
20680 Set the number of frames (value*2 + 1) used for lowpass filtering the
20681 camera movements. Default value is 10.
20683 For example a number of 10 means that 21 frames are used (10 in the
20684 past and 10 in the future) to smoothen the motion in the video. A
20685 larger value leads to a smoother video, but limits the acceleration of
20686 the camera (pan/tilt movements). 0 is a special case where a static
20687 camera is simulated.
20690 Set the camera path optimization algorithm.
20692 Accepted values are:
20695 gaussian kernel low-pass filter on camera motion (default)
20697 averaging on transformations
20701 Set maximal number of pixels to translate frames. Default value is -1,
20705 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20706 value is -1, meaning no limit.
20709 Specify how to deal with borders that may be visible due to movement
20712 Available values are:
20715 keep image information from previous frame (default)
20717 fill the border black
20721 Invert transforms if set to 1. Default value is 0.
20724 Consider transforms as relative to previous frame if set to 1,
20725 absolute if set to 0. Default value is 0.
20728 Set percentage to zoom. A positive value will result in a zoom-in
20729 effect, a negative value in a zoom-out effect. Default value is 0 (no
20733 Set optimal zooming to avoid borders.
20735 Accepted values are:
20740 optimal static zoom value is determined (only very strong movements
20741 will lead to visible borders) (default)
20743 optimal adaptive zoom value is determined (no borders will be
20744 visible), see @option{zoomspeed}
20747 Note that the value given at zoom is added to the one calculated here.
20750 Set percent to zoom maximally each frame (enabled when
20751 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20755 Specify type of interpolation.
20757 Available values are:
20762 linear only horizontal
20764 linear in both directions (default)
20766 cubic in both directions (slow)
20770 Enable virtual tripod mode if set to 1, which is equivalent to
20771 @code{relative=0:smoothing=0}. Default value is 0.
20773 Use also @code{tripod} option of @ref{vidstabdetect}.
20776 Increase log verbosity if set to 1. Also the detected global motions
20777 are written to the temporary file @file{global_motions.trf}. Default
20781 @subsection Examples
20785 Use @command{ffmpeg} for a typical stabilization with default values:
20787 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20790 Note the use of the @ref{unsharp} filter which is always recommended.
20793 Zoom in a bit more and load transform data from a given file:
20795 vidstabtransform=zoom=5:input="mytransforms.trf"
20799 Smoothen the video even more:
20801 vidstabtransform=smoothing=30
20807 Flip the input video vertically.
20809 For example, to vertically flip a video with @command{ffmpeg}:
20811 ffmpeg -i in.avi -vf "vflip" out.avi
20816 Detect variable frame rate video.
20818 This filter tries to detect if the input is variable or constant frame rate.
20820 At end it will output number of frames detected as having variable delta pts,
20821 and ones with constant delta pts.
20822 If there was frames with variable delta, than it will also show min, max and
20823 average delta encountered.
20827 Boost or alter saturation.
20829 The filter accepts the following options:
20832 Set strength of boost if positive value or strength of alter if negative value.
20833 Default is 0. Allowed range is from -2 to 2.
20836 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20839 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20842 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20845 Set the red luma coefficient.
20848 Set the green luma coefficient.
20851 Set the blue luma coefficient.
20854 If @code{intensity} is negative and this is set to 1, colors will change,
20855 otherwise colors will be less saturated, more towards gray.
20858 @subsection Commands
20860 This filter supports the all above options as @ref{commands}.
20865 Make or reverse a natural vignetting effect.
20867 The filter accepts the following options:
20871 Set lens angle expression as a number of radians.
20873 The value is clipped in the @code{[0,PI/2]} range.
20875 Default value: @code{"PI/5"}
20879 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20883 Set forward/backward mode.
20885 Available modes are:
20888 The larger the distance from the central point, the darker the image becomes.
20891 The larger the distance from the central point, the brighter the image becomes.
20892 This can be used to reverse a vignette effect, though there is no automatic
20893 detection to extract the lens @option{angle} and other settings (yet). It can
20894 also be used to create a burning effect.
20897 Default value is @samp{forward}.
20900 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20902 It accepts the following values:
20905 Evaluate expressions only once during the filter initialization.
20908 Evaluate expressions for each incoming frame. This is way slower than the
20909 @samp{init} mode since it requires all the scalers to be re-computed, but it
20910 allows advanced dynamic expressions.
20913 Default value is @samp{init}.
20916 Set dithering to reduce the circular banding effects. Default is @code{1}
20920 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20921 Setting this value to the SAR of the input will make a rectangular vignetting
20922 following the dimensions of the video.
20924 Default is @code{1/1}.
20927 @subsection Expressions
20929 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20930 following parameters.
20935 input width and height
20938 the number of input frame, starting from 0
20941 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20942 @var{TB} units, NAN if undefined
20945 frame rate of the input video, NAN if the input frame rate is unknown
20948 the PTS (Presentation TimeStamp) of the filtered video frame,
20949 expressed in seconds, NAN if undefined
20952 time base of the input video
20956 @subsection Examples
20960 Apply simple strong vignetting effect:
20966 Make a flickering vignetting:
20968 vignette='PI/4+random(1)*PI/50':eval=frame
20973 @section vmafmotion
20975 Obtain the average VMAF motion score of a video.
20976 It is one of the component metrics of VMAF.
20978 The obtained average motion score is printed through the logging system.
20980 The filter accepts the following options:
20984 If specified, the filter will use the named file to save the motion score of
20985 each frame with respect to the previous frame.
20986 When filename equals "-" the data is sent to standard output.
20991 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20995 Stack input videos vertically.
20997 All streams must be of same pixel format and of same width.
20999 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
21000 to create same output.
21002 The filter accepts the following options:
21006 Set number of input streams. Default is 2.
21009 If set to 1, force the output to terminate when the shortest input
21010 terminates. Default value is 0.
21015 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
21016 Deinterlacing Filter").
21018 Based on the process described by Martin Weston for BBC R&D, and
21019 implemented based on the de-interlace algorithm written by Jim
21020 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
21021 uses filter coefficients calculated by BBC R&D.
21023 This filter uses field-dominance information in frame to decide which
21024 of each pair of fields to place first in the output.
21025 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
21027 There are two sets of filter coefficients, so called "simple"
21028 and "complex". Which set of filter coefficients is used can
21029 be set by passing an optional parameter:
21033 Set the interlacing filter coefficients. Accepts one of the following values:
21037 Simple filter coefficient set.
21039 More-complex filter coefficient set.
21041 Default value is @samp{complex}.
21044 The interlacing mode to adopt. It accepts one of the following values:
21048 Output one frame for each frame.
21050 Output one frame for each field.
21053 The default value is @code{field}.
21056 The picture field parity assumed for the input interlaced video. It accepts one
21057 of the following values:
21061 Assume the top field is first.
21063 Assume the bottom field is first.
21065 Enable automatic detection of field parity.
21068 The default value is @code{auto}.
21069 If the interlacing is unknown or the decoder does not export this information,
21070 top field first will be assumed.
21073 Specify which frames to deinterlace. Accepts one of the following values:
21077 Deinterlace all frames,
21079 Only deinterlace frames marked as interlaced.
21082 Default value is @samp{all}.
21085 @subsection Commands
21086 This filter supports same @ref{commands} as options.
21089 Video waveform monitor.
21091 The waveform monitor plots color component intensity. By default luminance
21092 only. Each column of the waveform corresponds to a column of pixels in the
21095 It accepts the following options:
21099 Can be either @code{row}, or @code{column}. Default is @code{column}.
21100 In row mode, the graph on the left side represents color component value 0 and
21101 the right side represents value = 255. In column mode, the top side represents
21102 color component value = 0 and bottom side represents value = 255.
21105 Set intensity. Smaller values are useful to find out how many values of the same
21106 luminance are distributed across input rows/columns.
21107 Default value is @code{0.04}. Allowed range is [0, 1].
21110 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21111 In mirrored mode, higher values will be represented on the left
21112 side for @code{row} mode and at the top for @code{column} mode. Default is
21113 @code{1} (mirrored).
21117 It accepts the following values:
21120 Presents information identical to that in the @code{parade}, except
21121 that the graphs representing color components are superimposed directly
21124 This display mode makes it easier to spot relative differences or similarities
21125 in overlapping areas of the color components that are supposed to be identical,
21126 such as neutral whites, grays, or blacks.
21129 Display separate graph for the color components side by side in
21130 @code{row} mode or one below the other in @code{column} mode.
21133 Display separate graph for the color components side by side in
21134 @code{column} mode or one below the other in @code{row} mode.
21136 Using this display mode makes it easy to spot color casts in the highlights
21137 and shadows of an image, by comparing the contours of the top and the bottom
21138 graphs of each waveform. Since whites, grays, and blacks are characterized
21139 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21140 should display three waveforms of roughly equal width/height. If not, the
21141 correction is easy to perform by making level adjustments the three waveforms.
21143 Default is @code{stack}.
21145 @item components, c
21146 Set which color components to display. Default is 1, which means only luminance
21147 or red color component if input is in RGB colorspace. If is set for example to
21148 7 it will display all 3 (if) available color components.
21153 No envelope, this is default.
21156 Instant envelope, minimum and maximum values presented in graph will be easily
21157 visible even with small @code{step} value.
21160 Hold minimum and maximum values presented in graph across time. This way you
21161 can still spot out of range values without constantly looking at waveforms.
21164 Peak and instant envelope combined together.
21170 No filtering, this is default.
21173 Luma and chroma combined together.
21176 Similar as above, but shows difference between blue and red chroma.
21179 Similar as above, but use different colors.
21182 Similar as above, but again with different colors.
21185 Displays only chroma.
21188 Displays actual color value on waveform.
21191 Similar as above, but with luma showing frequency of chroma values.
21195 Set which graticule to display.
21199 Do not display graticule.
21202 Display green graticule showing legal broadcast ranges.
21205 Display orange graticule showing legal broadcast ranges.
21208 Display invert graticule showing legal broadcast ranges.
21212 Set graticule opacity.
21215 Set graticule flags.
21219 Draw numbers above lines. By default enabled.
21222 Draw dots instead of lines.
21226 Set scale used for displaying graticule.
21233 Default is digital.
21236 Set background opacity.
21240 Set tint for output.
21241 Only used with lowpass filter and when display is not overlay and input
21242 pixel formats are not RGB.
21245 @section weave, doubleweave
21247 The @code{weave} takes a field-based video input and join
21248 each two sequential fields into single frame, producing a new double
21249 height clip with half the frame rate and half the frame count.
21251 The @code{doubleweave} works same as @code{weave} but without
21252 halving frame rate and frame count.
21254 It accepts the following option:
21258 Set first field. Available values are:
21262 Set the frame as top-field-first.
21265 Set the frame as bottom-field-first.
21269 @subsection Examples
21273 Interlace video using @ref{select} and @ref{separatefields} filter:
21275 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21280 Apply the xBR high-quality magnification filter which is designed for pixel
21281 art. It follows a set of edge-detection rules, see
21282 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21284 It accepts the following option:
21288 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21289 @code{3xBR} and @code{4} for @code{4xBR}.
21290 Default is @code{3}.
21295 Apply cross fade from one input video stream to another input video stream.
21296 The cross fade is applied for specified duration.
21298 The filter accepts the following options:
21302 Set one of available transition effects:
21350 Default transition effect is fade.
21353 Set cross fade duration in seconds.
21354 Default duration is 1 second.
21357 Set cross fade start relative to first input stream in seconds.
21358 Default offset is 0.
21361 Set expression for custom transition effect.
21363 The expressions can use the following variables and functions:
21368 The coordinates of the current sample.
21372 The width and height of the image.
21375 Progress of transition effect.
21378 Currently processed plane.
21381 Return value of first input at current location and plane.
21384 Return value of second input at current location and plane.
21390 Return the value of the pixel at location (@var{x},@var{y}) of the
21391 first/second/third/fourth component of first input.
21397 Return the value of the pixel at location (@var{x},@var{y}) of the
21398 first/second/third/fourth component of second input.
21402 @subsection Examples
21406 Cross fade from one input video to another input video, with fade transition and duration of transition
21407 of 2 seconds starting at offset of 5 seconds:
21409 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21414 Pick median pixels from several input videos.
21416 The filter accepts the following options:
21420 Set number of inputs.
21421 Default is 3. Allowed range is from 3 to 255.
21422 If number of inputs is even number, than result will be mean value between two median values.
21425 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21428 Set median percentile. Default value is @code{0.5}.
21429 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21430 minimum values, and @code{1} maximum values.
21433 @subsection Commands
21435 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21438 Stack video inputs into custom layout.
21440 All streams must be of same pixel format.
21442 The filter accepts the following options:
21446 Set number of input streams. Default is 2.
21449 Specify layout of inputs.
21450 This option requires the desired layout configuration to be explicitly set by the user.
21451 This sets position of each video input in output. Each input
21452 is separated by '|'.
21453 The first number represents the column, and the second number represents the row.
21454 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21455 where X is video input from which to take width or height.
21456 Multiple values can be used when separated by '+'. In such
21457 case values are summed together.
21459 Note that if inputs are of different sizes gaps may appear, as not all of
21460 the output video frame will be filled. Similarly, videos can overlap each
21461 other if their position doesn't leave enough space for the full frame of
21464 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21465 a layout must be set by the user.
21468 If set to 1, force the output to terminate when the shortest input
21469 terminates. Default value is 0.
21472 If set to valid color, all unused pixels will be filled with that color.
21473 By default fill is set to none, so it is disabled.
21476 @subsection Examples
21480 Display 4 inputs into 2x2 grid.
21484 input1(0, 0) | input3(w0, 0)
21485 input2(0, h0) | input4(w0, h0)
21489 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21492 Note that if inputs are of different sizes, gaps or overlaps may occur.
21495 Display 4 inputs into 1x4 grid.
21502 input4(0, h0+h1+h2)
21506 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21509 Note that if inputs are of different widths, unused space will appear.
21512 Display 9 inputs into 3x3 grid.
21516 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21517 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21518 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21522 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
21525 Note that if inputs are of different sizes, gaps or overlaps may occur.
21528 Display 16 inputs into 4x4 grid.
21532 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21533 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21534 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21535 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21539 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|
21540 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
21543 Note that if inputs are of different sizes, gaps or overlaps may occur.
21550 Deinterlace the input video ("yadif" means "yet another deinterlacing
21553 It accepts the following parameters:
21559 The interlacing mode to adopt. It accepts one of the following values:
21562 @item 0, send_frame
21563 Output one frame for each frame.
21564 @item 1, send_field
21565 Output one frame for each field.
21566 @item 2, send_frame_nospatial
21567 Like @code{send_frame}, but it skips the spatial interlacing check.
21568 @item 3, send_field_nospatial
21569 Like @code{send_field}, but it skips the spatial interlacing check.
21572 The default value is @code{send_frame}.
21575 The picture field parity assumed for the input interlaced video. It accepts one
21576 of the following values:
21580 Assume the top field is first.
21582 Assume the bottom field is first.
21584 Enable automatic detection of field parity.
21587 The default value is @code{auto}.
21588 If the interlacing is unknown or the decoder does not export this information,
21589 top field first will be assumed.
21592 Specify which frames to deinterlace. Accepts one of the following
21597 Deinterlace all frames.
21598 @item 1, interlaced
21599 Only deinterlace frames marked as interlaced.
21602 The default value is @code{all}.
21605 @section yadif_cuda
21607 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21608 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21611 It accepts the following parameters:
21617 The interlacing mode to adopt. It accepts one of the following values:
21620 @item 0, send_frame
21621 Output one frame for each frame.
21622 @item 1, send_field
21623 Output one frame for each field.
21624 @item 2, send_frame_nospatial
21625 Like @code{send_frame}, but it skips the spatial interlacing check.
21626 @item 3, send_field_nospatial
21627 Like @code{send_field}, but it skips the spatial interlacing check.
21630 The default value is @code{send_frame}.
21633 The picture field parity assumed for the input interlaced video. It accepts one
21634 of the following values:
21638 Assume the top field is first.
21640 Assume the bottom field is first.
21642 Enable automatic detection of field parity.
21645 The default value is @code{auto}.
21646 If the interlacing is unknown or the decoder does not export this information,
21647 top field first will be assumed.
21650 Specify which frames to deinterlace. Accepts one of the following
21655 Deinterlace all frames.
21656 @item 1, interlaced
21657 Only deinterlace frames marked as interlaced.
21660 The default value is @code{all}.
21665 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21666 The algorithm is described in
21667 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21669 It accepts the following parameters:
21673 Set the window radius. Default value is 3.
21676 Set which planes to filter. Default is only the first plane.
21679 Set blur strength. Default value is 128.
21682 @subsection Commands
21683 This filter supports same @ref{commands} as options.
21687 Apply Zoom & Pan effect.
21689 This filter accepts the following options:
21693 Set the zoom expression. Range is 1-10. Default is 1.
21697 Set the x and y expression. Default is 0.
21700 Set the duration expression in number of frames.
21701 This sets for how many number of frames effect will last for
21702 single input image.
21705 Set the output image size, default is 'hd720'.
21708 Set the output frame rate, default is '25'.
21711 Each expression can contain the following constants:
21730 Output frame count.
21733 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21735 @item out_time, time, ot
21736 The output timestamp expressed in seconds.
21740 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21741 for current input frame.
21745 'x' and 'y' of last output frame of previous input frame or 0 when there was
21746 not yet such frame (first input frame).
21749 Last calculated zoom from 'z' expression for current input frame.
21752 Last calculated zoom of last output frame of previous input frame.
21755 Number of output frames for current input frame. Calculated from 'd' expression
21756 for each input frame.
21759 number of output frames created for previous input frame
21762 Rational number: input width / input height
21765 sample aspect ratio
21768 display aspect ratio
21772 @subsection Examples
21776 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21778 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
21782 Zoom in up to 1.5x and pan always at center of picture:
21784 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21788 Same as above but without pausing:
21790 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21794 Zoom in 2x into center of picture only for the first second of the input video:
21796 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21803 Scale (resize) the input video, using the z.lib library:
21804 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21805 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21807 The zscale filter forces the output display aspect ratio to be the same
21808 as the input, by changing the output sample aspect ratio.
21810 If the input image format is different from the format requested by
21811 the next filter, the zscale filter will convert the input to the
21814 @subsection Options
21815 The filter accepts the following options.
21820 Set the output video dimension expression. Default value is the input
21823 If the @var{width} or @var{w} value is 0, the input width is used for
21824 the output. If the @var{height} or @var{h} value is 0, the input height
21825 is used for the output.
21827 If one and only one of the values is -n with n >= 1, the zscale filter
21828 will use a value that maintains the aspect ratio of the input image,
21829 calculated from the other specified dimension. After that it will,
21830 however, make sure that the calculated dimension is divisible by n and
21831 adjust the value if necessary.
21833 If both values are -n with n >= 1, the behavior will be identical to
21834 both values being set to 0 as previously detailed.
21836 See below for the list of accepted constants for use in the dimension
21840 Set the video size. For the syntax of this option, check the
21841 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21844 Set the dither type.
21846 Possible values are:
21851 @item error_diffusion
21857 Set the resize filter type.
21859 Possible values are:
21869 Default is bilinear.
21872 Set the color range.
21874 Possible values are:
21881 Default is same as input.
21884 Set the color primaries.
21886 Possible values are:
21896 Default is same as input.
21899 Set the transfer characteristics.
21901 Possible values are:
21915 Default is same as input.
21918 Set the colorspace matrix.
21920 Possible value are:
21931 Default is same as input.
21934 Set the input color range.
21936 Possible values are:
21943 Default is same as input.
21945 @item primariesin, pin
21946 Set the input color primaries.
21948 Possible values are:
21958 Default is same as input.
21960 @item transferin, tin
21961 Set the input transfer characteristics.
21963 Possible values are:
21974 Default is same as input.
21976 @item matrixin, min
21977 Set the input colorspace matrix.
21979 Possible value are:
21991 Set the output chroma location.
21993 Possible values are:
22004 @item chromalin, cin
22005 Set the input chroma location.
22007 Possible values are:
22019 Set the nominal peak luminance.
22022 The values of the @option{w} and @option{h} options are expressions
22023 containing the following constants:
22028 The input width and height
22032 These are the same as @var{in_w} and @var{in_h}.
22036 The output (scaled) width and height
22040 These are the same as @var{out_w} and @var{out_h}
22043 The same as @var{iw} / @var{ih}
22046 input sample aspect ratio
22049 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22053 horizontal and vertical input chroma subsample values. For example for the
22054 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22058 horizontal and vertical output chroma subsample values. For example for the
22059 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22062 @subsection Commands
22064 This filter supports the following commands:
22068 Set the output video dimension expression.
22069 The command accepts the same syntax of the corresponding option.
22071 If the specified expression is not valid, it is kept at its current
22075 @c man end VIDEO FILTERS
22077 @chapter OpenCL Video Filters
22078 @c man begin OPENCL VIDEO FILTERS
22080 Below is a description of the currently available OpenCL video filters.
22082 To enable compilation of these filters you need to configure FFmpeg with
22083 @code{--enable-opencl}.
22085 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22088 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22089 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22090 given device parameters.
22092 @item -filter_hw_device @var{name}
22093 Pass the hardware device called @var{name} to all filters in any filter graph.
22097 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22101 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22103 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22107 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.
22109 @section avgblur_opencl
22111 Apply average blur filter.
22113 The filter accepts the following options:
22117 Set horizontal radius size.
22118 Range is @code{[1, 1024]} and default value is @code{1}.
22121 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22124 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22127 @subsection Example
22131 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.
22133 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22137 @section boxblur_opencl
22139 Apply a boxblur algorithm to the input video.
22141 It accepts the following parameters:
22145 @item luma_radius, lr
22146 @item luma_power, lp
22147 @item chroma_radius, cr
22148 @item chroma_power, cp
22149 @item alpha_radius, ar
22150 @item alpha_power, ap
22154 A description of the accepted options follows.
22157 @item luma_radius, lr
22158 @item chroma_radius, cr
22159 @item alpha_radius, ar
22160 Set an expression for the box radius in pixels used for blurring the
22161 corresponding input plane.
22163 The radius value must be a non-negative number, and must not be
22164 greater than the value of the expression @code{min(w,h)/2} for the
22165 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22168 Default value for @option{luma_radius} is "2". If not specified,
22169 @option{chroma_radius} and @option{alpha_radius} default to the
22170 corresponding value set for @option{luma_radius}.
22172 The expressions can contain the following constants:
22176 The input width and height in pixels.
22180 The input chroma image width and height in pixels.
22184 The horizontal and vertical chroma subsample values. For example, for the
22185 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22188 @item luma_power, lp
22189 @item chroma_power, cp
22190 @item alpha_power, ap
22191 Specify how many times the boxblur filter is applied to the
22192 corresponding plane.
22194 Default value for @option{luma_power} is 2. If not specified,
22195 @option{chroma_power} and @option{alpha_power} default to the
22196 corresponding value set for @option{luma_power}.
22198 A value of 0 will disable the effect.
22201 @subsection Examples
22203 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.
22207 Apply a boxblur filter with the luma, chroma, and alpha radius
22208 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.
22210 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22211 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22215 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.
22217 For the luma plane, a 2x2 box radius will be run once.
22219 For the chroma plane, a 4x4 box radius will be run 5 times.
22221 For the alpha plane, a 3x3 box radius will be run 7 times.
22223 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22227 @section colorkey_opencl
22228 RGB colorspace color keying.
22230 The filter accepts the following options:
22234 The color which will be replaced with transparency.
22237 Similarity percentage with the key color.
22239 0.01 matches only the exact key color, while 1.0 matches everything.
22244 0.0 makes pixels either fully transparent, or not transparent at all.
22246 Higher values result in semi-transparent pixels, with a higher transparency
22247 the more similar the pixels color is to the key color.
22250 @subsection Examples
22254 Make every semi-green pixel in the input transparent with some slight blending:
22256 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22260 @section convolution_opencl
22262 Apply convolution of 3x3, 5x5, 7x7 matrix.
22264 The filter accepts the following options:
22271 Set matrix for each plane.
22272 Matrix is sequence of 9, 25 or 49 signed numbers.
22273 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22279 Set multiplier for calculated value for each plane.
22280 If unset or 0, it will be sum of all matrix elements.
22281 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22287 Set bias for each plane. This value is added to the result of the multiplication.
22288 Useful for making the overall image brighter or darker.
22289 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22293 @subsection Examples
22299 -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
22305 -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
22309 Apply edge enhance:
22311 -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
22317 -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
22321 Apply laplacian edge detector which includes diagonals:
22323 -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
22329 -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
22333 @section erosion_opencl
22335 Apply erosion effect to the video.
22337 This filter replaces the pixel by the local(3x3) minimum.
22339 It accepts the following options:
22346 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22347 If @code{0}, plane will remain unchanged.
22350 Flag which specifies the pixel to refer to.
22351 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22353 Flags to local 3x3 coordinates region centered on @code{x}:
22362 @subsection Example
22366 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.
22368 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22372 @section deshake_opencl
22373 Feature-point based video stabilization filter.
22375 The filter accepts the following options:
22379 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22382 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22384 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22386 Viewing point matches in the output video is only supported for RGB input.
22388 Defaults to @code{0}.
22390 @item adaptive_crop
22391 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22393 Defaults to @code{1}.
22395 @item refine_features
22396 Whether or not feature points should be refined at a sub-pixel level.
22398 This can be turned off for a slight performance gain at the cost of precision.
22400 Defaults to @code{1}.
22402 @item smooth_strength
22403 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22405 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22407 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22409 Defaults to @code{0.0}.
22411 @item smooth_window_multiplier
22412 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22414 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22416 Acceptable values range from @code{0.1} to @code{10.0}.
22418 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22419 potentially improving smoothness, but also increase latency and memory usage.
22421 Defaults to @code{2.0}.
22425 @subsection Examples
22429 Stabilize a video with a fixed, medium smoothing strength:
22431 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22435 Stabilize a video with debugging (both in console and in rendered video):
22437 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22441 @section dilation_opencl
22443 Apply dilation effect to the video.
22445 This filter replaces the pixel by the local(3x3) maximum.
22447 It accepts the following options:
22454 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22455 If @code{0}, plane will remain unchanged.
22458 Flag which specifies the pixel to refer to.
22459 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22461 Flags to local 3x3 coordinates region centered on @code{x}:
22470 @subsection Example
22474 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.
22476 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22480 @section nlmeans_opencl
22482 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22484 @section overlay_opencl
22486 Overlay one video on top of another.
22488 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22489 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22491 The filter accepts the following options:
22496 Set the x coordinate of the overlaid video on the main video.
22497 Default value is @code{0}.
22500 Set the y coordinate of the overlaid video on the main video.
22501 Default value is @code{0}.
22505 @subsection Examples
22509 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22511 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22514 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22516 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22521 @section pad_opencl
22523 Add paddings to the input image, and place the original input at the
22524 provided @var{x}, @var{y} coordinates.
22526 It accepts the following options:
22531 Specify an expression for the size of the output image with the
22532 paddings added. If the value for @var{width} or @var{height} is 0, the
22533 corresponding input size is used for the output.
22535 The @var{width} expression can reference the value set by the
22536 @var{height} expression, and vice versa.
22538 The default value of @var{width} and @var{height} is 0.
22542 Specify the offsets to place the input image at within the padded area,
22543 with respect to the top/left border of the output image.
22545 The @var{x} expression can reference the value set by the @var{y}
22546 expression, and vice versa.
22548 The default value of @var{x} and @var{y} is 0.
22550 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22551 so the input image is centered on the padded area.
22554 Specify the color of the padded area. For the syntax of this option,
22555 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22556 manual,ffmpeg-utils}.
22559 Pad to an aspect instead to a resolution.
22562 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22563 options are expressions containing the following constants:
22568 The input video width and height.
22572 These are the same as @var{in_w} and @var{in_h}.
22576 The output width and height (the size of the padded area), as
22577 specified by the @var{width} and @var{height} expressions.
22581 These are the same as @var{out_w} and @var{out_h}.
22585 The x and y offsets as specified by the @var{x} and @var{y}
22586 expressions, or NAN if not yet specified.
22589 same as @var{iw} / @var{ih}
22592 input sample aspect ratio
22595 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22598 @section prewitt_opencl
22600 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22602 The filter accepts the following option:
22606 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22609 Set value which will be multiplied with filtered result.
22610 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22613 Set value which will be added to filtered result.
22614 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22617 @subsection Example
22621 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22623 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22627 @anchor{program_opencl}
22628 @section program_opencl
22630 Filter video using an OpenCL program.
22635 OpenCL program source file.
22638 Kernel name in program.
22641 Number of inputs to the filter. Defaults to 1.
22644 Size of output frames. Defaults to the same as the first input.
22648 The @code{program_opencl} filter also supports the @ref{framesync} options.
22650 The program source file must contain a kernel function with the given name,
22651 which will be run once for each plane of the output. Each run on a plane
22652 gets enqueued as a separate 2D global NDRange with one work-item for each
22653 pixel to be generated. The global ID offset for each work-item is therefore
22654 the coordinates of a pixel in the destination image.
22656 The kernel function needs to take the following arguments:
22659 Destination image, @var{__write_only image2d_t}.
22661 This image will become the output; the kernel should write all of it.
22663 Frame index, @var{unsigned int}.
22665 This is a counter starting from zero and increasing by one for each frame.
22667 Source images, @var{__read_only image2d_t}.
22669 These are the most recent images on each input. The kernel may read from
22670 them to generate the output, but they can't be written to.
22677 Copy the input to the output (output must be the same size as the input).
22679 __kernel void copy(__write_only image2d_t destination,
22680 unsigned int index,
22681 __read_only image2d_t source)
22683 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22685 int2 location = (int2)(get_global_id(0), get_global_id(1));
22687 float4 value = read_imagef(source, sampler, location);
22689 write_imagef(destination, location, value);
22694 Apply a simple transformation, rotating the input by an amount increasing
22695 with the index counter. Pixel values are linearly interpolated by the
22696 sampler, and the output need not have the same dimensions as the input.
22698 __kernel void rotate_image(__write_only image2d_t dst,
22699 unsigned int index,
22700 __read_only image2d_t src)
22702 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22703 CLK_FILTER_LINEAR);
22705 float angle = (float)index / 100.0f;
22707 float2 dst_dim = convert_float2(get_image_dim(dst));
22708 float2 src_dim = convert_float2(get_image_dim(src));
22710 float2 dst_cen = dst_dim / 2.0f;
22711 float2 src_cen = src_dim / 2.0f;
22713 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22715 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22717 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22718 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22720 src_pos = src_pos * src_dim / dst_dim;
22722 float2 src_loc = src_pos + src_cen;
22724 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22725 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22726 write_imagef(dst, dst_loc, 0.5f);
22728 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22733 Blend two inputs together, with the amount of each input used varying
22734 with the index counter.
22736 __kernel void blend_images(__write_only image2d_t dst,
22737 unsigned int index,
22738 __read_only image2d_t src1,
22739 __read_only image2d_t src2)
22741 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22742 CLK_FILTER_LINEAR);
22744 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22746 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22747 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22748 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22750 float4 val1 = read_imagef(src1, sampler, src1_loc);
22751 float4 val2 = read_imagef(src2, sampler, src2_loc);
22753 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22759 @section roberts_opencl
22760 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22762 The filter accepts the following option:
22766 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22769 Set value which will be multiplied with filtered result.
22770 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22773 Set value which will be added to filtered result.
22774 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22777 @subsection Example
22781 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22783 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22787 @section sobel_opencl
22789 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22791 The filter accepts the following option:
22795 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22798 Set value which will be multiplied with filtered result.
22799 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22802 Set value which will be added to filtered result.
22803 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22806 @subsection Example
22810 Apply sobel operator with scale set to 2 and delta set to 10
22812 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22816 @section tonemap_opencl
22818 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22820 It accepts the following parameters:
22824 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22827 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22830 Apply desaturation for highlights that exceed this level of brightness. The
22831 higher the parameter, the more color information will be preserved. This
22832 setting helps prevent unnaturally blown-out colors for super-highlights, by
22833 (smoothly) turning into white instead. This makes images feel more natural,
22834 at the cost of reducing information about out-of-range colors.
22836 The default value is 0.5, and the algorithm here is a little different from
22837 the cpu version tonemap currently. A setting of 0.0 disables this option.
22840 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22841 is used to detect whether the scene has changed or not. If the distance between
22842 the current frame average brightness and the current running average exceeds
22843 a threshold value, we would re-calculate scene average and peak brightness.
22844 The default value is 0.2.
22847 Specify the output pixel format.
22849 Currently supported formats are:
22856 Set the output color range.
22858 Possible values are:
22864 Default is same as input.
22867 Set the output color primaries.
22869 Possible values are:
22875 Default is same as input.
22878 Set the output transfer characteristics.
22880 Possible values are:
22889 Set the output colorspace matrix.
22891 Possible value are:
22897 Default is same as input.
22901 @subsection Example
22905 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22907 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22911 @section unsharp_opencl
22913 Sharpen or blur the input video.
22915 It accepts the following parameters:
22918 @item luma_msize_x, lx
22919 Set the luma matrix horizontal size.
22920 Range is @code{[1, 23]} and default value is @code{5}.
22922 @item luma_msize_y, ly
22923 Set the luma matrix vertical size.
22924 Range is @code{[1, 23]} and default value is @code{5}.
22926 @item luma_amount, la
22927 Set the luma effect strength.
22928 Range is @code{[-10, 10]} and default value is @code{1.0}.
22930 Negative values will blur the input video, while positive values will
22931 sharpen it, a value of zero will disable the effect.
22933 @item chroma_msize_x, cx
22934 Set the chroma matrix horizontal size.
22935 Range is @code{[1, 23]} and default value is @code{5}.
22937 @item chroma_msize_y, cy
22938 Set the chroma matrix vertical size.
22939 Range is @code{[1, 23]} and default value is @code{5}.
22941 @item chroma_amount, ca
22942 Set the chroma effect strength.
22943 Range is @code{[-10, 10]} and default value is @code{0.0}.
22945 Negative values will blur the input video, while positive values will
22946 sharpen it, a value of zero will disable the effect.
22950 All parameters are optional and default to the equivalent of the
22951 string '5:5:1.0:5:5:0.0'.
22953 @subsection Examples
22957 Apply strong luma sharpen effect:
22959 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22963 Apply a strong blur of both luma and chroma parameters:
22965 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22969 @section xfade_opencl
22971 Cross fade two videos with custom transition effect by using OpenCL.
22973 It accepts the following options:
22977 Set one of possible transition effects.
22981 Select custom transition effect, the actual transition description
22982 will be picked from source and kernel options.
22994 Default transition is fade.
22998 OpenCL program source file for custom transition.
23001 Set name of kernel to use for custom transition from program source file.
23004 Set duration of video transition.
23007 Set time of start of transition relative to first video.
23010 The program source file must contain a kernel function with the given name,
23011 which will be run once for each plane of the output. Each run on a plane
23012 gets enqueued as a separate 2D global NDRange with one work-item for each
23013 pixel to be generated. The global ID offset for each work-item is therefore
23014 the coordinates of a pixel in the destination image.
23016 The kernel function needs to take the following arguments:
23019 Destination image, @var{__write_only image2d_t}.
23021 This image will become the output; the kernel should write all of it.
23024 First Source image, @var{__read_only image2d_t}.
23025 Second Source image, @var{__read_only image2d_t}.
23027 These are the most recent images on each input. The kernel may read from
23028 them to generate the output, but they can't be written to.
23031 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
23038 Apply dots curtain transition effect:
23040 __kernel void blend_images(__write_only image2d_t dst,
23041 __read_only image2d_t src1,
23042 __read_only image2d_t src2,
23045 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23046 CLK_FILTER_LINEAR);
23047 int2 p = (int2)(get_global_id(0), get_global_id(1));
23048 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23049 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23052 float2 dots = (float2)(20.0, 20.0);
23053 float2 center = (float2)(0,0);
23056 float4 val1 = read_imagef(src1, sampler, p);
23057 float4 val2 = read_imagef(src2, sampler, p);
23058 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23060 write_imagef(dst, p, next ? val1 : val2);
23066 @c man end OPENCL VIDEO FILTERS
23068 @chapter VAAPI Video Filters
23069 @c man begin VAAPI VIDEO FILTERS
23071 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23073 To enable compilation of these filters you need to configure FFmpeg with
23074 @code{--enable-vaapi}.
23076 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}
23078 @section tonemap_vaapi
23080 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23081 It maps the dynamic range of HDR10 content to the SDR content.
23082 It currently only accepts HDR10 as input.
23084 It accepts the following parameters:
23088 Specify the output pixel format.
23090 Currently supported formats are:
23099 Set the output color primaries.
23101 Default is same as input.
23104 Set the output transfer characteristics.
23109 Set the output colorspace matrix.
23111 Default is same as input.
23115 @subsection Example
23119 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23121 tonemap_vaapi=format=p010:t=bt2020-10
23125 @c man end VAAPI VIDEO FILTERS
23127 @chapter Video Sources
23128 @c man begin VIDEO SOURCES
23130 Below is a description of the currently available video sources.
23134 Buffer video frames, and make them available to the filter chain.
23136 This source is mainly intended for a programmatic use, in particular
23137 through the interface defined in @file{libavfilter/buffersrc.h}.
23139 It accepts the following parameters:
23144 Specify the size (width and height) of the buffered video frames. For the
23145 syntax of this option, check the
23146 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23149 The input video width.
23152 The input video height.
23155 A string representing the pixel format of the buffered video frames.
23156 It may be a number corresponding to a pixel format, or a pixel format
23160 Specify the timebase assumed by the timestamps of the buffered frames.
23163 Specify the frame rate expected for the video stream.
23165 @item pixel_aspect, sar
23166 The sample (pixel) aspect ratio of the input video.
23169 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23170 to the filtergraph description to specify swscale flags for automatically
23171 inserted scalers. See @ref{Filtergraph syntax}.
23173 @item hw_frames_ctx
23174 When using a hardware pixel format, this should be a reference to an
23175 AVHWFramesContext describing input frames.
23180 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23183 will instruct the source to accept video frames with size 320x240 and
23184 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23185 square pixels (1:1 sample aspect ratio).
23186 Since the pixel format with name "yuv410p" corresponds to the number 6
23187 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23188 this example corresponds to:
23190 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23193 Alternatively, the options can be specified as a flat string, but this
23194 syntax is deprecated:
23196 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23200 Create a pattern generated by an elementary cellular automaton.
23202 The initial state of the cellular automaton can be defined through the
23203 @option{filename} and @option{pattern} options. If such options are
23204 not specified an initial state is created randomly.
23206 At each new frame a new row in the video is filled with the result of
23207 the cellular automaton next generation. The behavior when the whole
23208 frame is filled is defined by the @option{scroll} option.
23210 This source accepts the following options:
23214 Read the initial cellular automaton state, i.e. the starting row, from
23215 the specified file.
23216 In the file, each non-whitespace character is considered an alive
23217 cell, a newline will terminate the row, and further characters in the
23218 file will be ignored.
23221 Read the initial cellular automaton state, i.e. the starting row, from
23222 the specified string.
23224 Each non-whitespace character in the string is considered an alive
23225 cell, a newline will terminate the row, and further characters in the
23226 string will be ignored.
23229 Set the video rate, that is the number of frames generated per second.
23232 @item random_fill_ratio, ratio
23233 Set the random fill ratio for the initial cellular automaton row. It
23234 is a floating point number value ranging from 0 to 1, defaults to
23237 This option is ignored when a file or a pattern is specified.
23239 @item random_seed, seed
23240 Set the seed for filling randomly the initial row, must be an integer
23241 included between 0 and UINT32_MAX. If not specified, or if explicitly
23242 set to -1, the filter will try to use a good random seed on a best
23246 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23247 Default value is 110.
23250 Set the size of the output video. For the syntax of this option, check the
23251 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23253 If @option{filename} or @option{pattern} is specified, the size is set
23254 by default to the width of the specified initial state row, and the
23255 height is set to @var{width} * PHI.
23257 If @option{size} is set, it must contain the width of the specified
23258 pattern string, and the specified pattern will be centered in the
23261 If a filename or a pattern string is not specified, the size value
23262 defaults to "320x518" (used for a randomly generated initial state).
23265 If set to 1, scroll the output upward when all the rows in the output
23266 have been already filled. If set to 0, the new generated row will be
23267 written over the top row just after the bottom row is filled.
23270 @item start_full, full
23271 If set to 1, completely fill the output with generated rows before
23272 outputting the first frame.
23273 This is the default behavior, for disabling set the value to 0.
23276 If set to 1, stitch the left and right row edges together.
23277 This is the default behavior, for disabling set the value to 0.
23280 @subsection Examples
23284 Read the initial state from @file{pattern}, and specify an output of
23287 cellauto=f=pattern:s=200x400
23291 Generate a random initial row with a width of 200 cells, with a fill
23294 cellauto=ratio=2/3:s=200x200
23298 Create a pattern generated by rule 18 starting by a single alive cell
23299 centered on an initial row with width 100:
23301 cellauto=p=@@:s=100x400:full=0:rule=18
23305 Specify a more elaborated initial pattern:
23307 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23312 @anchor{coreimagesrc}
23313 @section coreimagesrc
23314 Video source generated on GPU using Apple's CoreImage API on OSX.
23316 This video source is a specialized version of the @ref{coreimage} video filter.
23317 Use a core image generator at the beginning of the applied filterchain to
23318 generate the content.
23320 The coreimagesrc video source accepts the following options:
23322 @item list_generators
23323 List all available generators along with all their respective options as well as
23324 possible minimum and maximum values along with the default values.
23326 list_generators=true
23330 Specify the size of the sourced video. For the syntax of this option, check the
23331 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23332 The default value is @code{320x240}.
23335 Specify the frame rate of the sourced video, as the number of frames
23336 generated per second. It has to be a string in the format
23337 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23338 number or a valid video frame rate abbreviation. The default value is
23342 Set the sample aspect ratio of the sourced video.
23345 Set the duration of the sourced video. See
23346 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23347 for the accepted syntax.
23349 If not specified, or the expressed duration is negative, the video is
23350 supposed to be generated forever.
23353 Additionally, all options of the @ref{coreimage} video filter are accepted.
23354 A complete filterchain can be used for further processing of the
23355 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23356 and examples for details.
23358 @subsection Examples
23363 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23364 given as complete and escaped command-line for Apple's standard bash shell:
23366 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23368 This example is equivalent to the QRCode example of @ref{coreimage} without the
23369 need for a nullsrc video source.
23374 Generate several gradients.
23378 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23379 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23382 Set frame rate, expressed as number of frames per second. Default
23385 @item c0, c1, c2, c3, c4, c5, c6, c7
23386 Set 8 colors. Default values for colors is to pick random one.
23388 @item x0, y0, y0, y1
23389 Set gradient line source and destination points. If negative or out of range, random ones
23393 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23396 Set seed for picking gradient line points.
23399 Set the duration of the sourced video. See
23400 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23401 for the accepted syntax.
23403 If not specified, or the expressed duration is negative, the video is
23404 supposed to be generated forever.
23407 Set speed of gradients rotation.
23411 @section mandelbrot
23413 Generate a Mandelbrot set fractal, and progressively zoom towards the
23414 point specified with @var{start_x} and @var{start_y}.
23416 This source accepts the following options:
23421 Set the terminal pts value. Default value is 400.
23424 Set the terminal scale value.
23425 Must be a floating point value. Default value is 0.3.
23428 Set the inner coloring mode, that is the algorithm used to draw the
23429 Mandelbrot fractal internal region.
23431 It shall assume one of the following values:
23436 Show time until convergence.
23438 Set color based on point closest to the origin of the iterations.
23443 Default value is @var{mincol}.
23446 Set the bailout value. Default value is 10.0.
23449 Set the maximum of iterations performed by the rendering
23450 algorithm. Default value is 7189.
23453 Set outer coloring mode.
23454 It shall assume one of following values:
23456 @item iteration_count
23457 Set iteration count mode.
23458 @item normalized_iteration_count
23459 set normalized iteration count mode.
23461 Default value is @var{normalized_iteration_count}.
23464 Set frame rate, expressed as number of frames per second. Default
23468 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23469 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23472 Set the initial scale value. Default value is 3.0.
23475 Set the initial x position. Must be a floating point value between
23476 -100 and 100. Default value is -0.743643887037158704752191506114774.
23479 Set the initial y position. Must be a floating point value between
23480 -100 and 100. Default value is -0.131825904205311970493132056385139.
23485 Generate various test patterns, as generated by the MPlayer test filter.
23487 The size of the generated video is fixed, and is 256x256.
23488 This source is useful in particular for testing encoding features.
23490 This source accepts the following options:
23495 Specify the frame rate of the sourced video, as the number of frames
23496 generated per second. It has to be a string in the format
23497 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23498 number or a valid video frame rate abbreviation. The default value is
23502 Set the duration of the sourced video. See
23503 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23504 for the accepted syntax.
23506 If not specified, or the expressed duration is negative, the video is
23507 supposed to be generated forever.
23511 Set the number or the name of the test to perform. Supported tests are:
23525 @item max_frames, m
23526 Set the maximum number of frames generated for each test, default value is 30.
23530 Default value is "all", which will cycle through the list of all tests.
23535 mptestsrc=t=dc_luma
23538 will generate a "dc_luma" test pattern.
23540 @section frei0r_src
23542 Provide a frei0r source.
23544 To enable compilation of this filter you need to install the frei0r
23545 header and configure FFmpeg with @code{--enable-frei0r}.
23547 This source accepts the following parameters:
23552 The size of the video to generate. For the syntax of this option, check the
23553 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23556 The framerate of the generated video. It may be a string of the form
23557 @var{num}/@var{den} or a frame rate abbreviation.
23560 The name to the frei0r source to load. For more information regarding frei0r and
23561 how to set the parameters, read the @ref{frei0r} section in the video filters
23564 @item filter_params
23565 A '|'-separated list of parameters to pass to the frei0r source.
23569 For example, to generate a frei0r partik0l source with size 200x200
23570 and frame rate 10 which is overlaid on the overlay filter main input:
23572 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23577 Generate a life pattern.
23579 This source is based on a generalization of John Conway's life game.
23581 The sourced input represents a life grid, each pixel represents a cell
23582 which can be in one of two possible states, alive or dead. Every cell
23583 interacts with its eight neighbours, which are the cells that are
23584 horizontally, vertically, or diagonally adjacent.
23586 At each interaction the grid evolves according to the adopted rule,
23587 which specifies the number of neighbor alive cells which will make a
23588 cell stay alive or born. The @option{rule} option allows one to specify
23591 This source accepts the following options:
23595 Set the file from which to read the initial grid state. In the file,
23596 each non-whitespace character is considered an alive cell, and newline
23597 is used to delimit the end of each row.
23599 If this option is not specified, the initial grid is generated
23603 Set the video rate, that is the number of frames generated per second.
23606 @item random_fill_ratio, ratio
23607 Set the random fill ratio for the initial random grid. It is a
23608 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23609 It is ignored when a file is specified.
23611 @item random_seed, seed
23612 Set the seed for filling the initial random grid, must be an integer
23613 included between 0 and UINT32_MAX. If not specified, or if explicitly
23614 set to -1, the filter will try to use a good random seed on a best
23620 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23621 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23622 @var{NS} specifies the number of alive neighbor cells which make a
23623 live cell stay alive, and @var{NB} the number of alive neighbor cells
23624 which make a dead cell to become alive (i.e. to "born").
23625 "s" and "b" can be used in place of "S" and "B", respectively.
23627 Alternatively a rule can be specified by an 18-bits integer. The 9
23628 high order bits are used to encode the next cell state if it is alive
23629 for each number of neighbor alive cells, the low order bits specify
23630 the rule for "borning" new cells. Higher order bits encode for an
23631 higher number of neighbor cells.
23632 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23633 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23635 Default value is "S23/B3", which is the original Conway's game of life
23636 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23637 cells, and will born a new cell if there are three alive cells around
23641 Set the size of the output video. For the syntax of this option, check the
23642 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23644 If @option{filename} is specified, the size is set by default to the
23645 same size of the input file. If @option{size} is set, it must contain
23646 the size specified in the input file, and the initial grid defined in
23647 that file is centered in the larger resulting area.
23649 If a filename is not specified, the size value defaults to "320x240"
23650 (used for a randomly generated initial grid).
23653 If set to 1, stitch the left and right grid edges together, and the
23654 top and bottom edges also. Defaults to 1.
23657 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23658 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23659 value from 0 to 255.
23662 Set the color of living (or new born) cells.
23665 Set the color of dead cells. If @option{mold} is set, this is the first color
23666 used to represent a dead cell.
23669 Set mold color, for definitely dead and moldy cells.
23671 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23672 ffmpeg-utils manual,ffmpeg-utils}.
23675 @subsection Examples
23679 Read a grid from @file{pattern}, and center it on a grid of size
23682 life=f=pattern:s=300x300
23686 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23688 life=ratio=2/3:s=200x200
23692 Specify a custom rule for evolving a randomly generated grid:
23698 Full example with slow death effect (mold) using @command{ffplay}:
23700 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23707 @anchor{haldclutsrc}
23710 @anchor{pal100bars}
23711 @anchor{rgbtestsrc}
23713 @anchor{smptehdbars}
23716 @anchor{yuvtestsrc}
23717 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23719 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23721 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23723 The @code{color} source provides an uniformly colored input.
23725 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23726 @ref{haldclut} filter.
23728 The @code{nullsrc} source returns unprocessed video frames. It is
23729 mainly useful to be employed in analysis / debugging tools, or as the
23730 source for filters which ignore the input data.
23732 The @code{pal75bars} source generates a color bars pattern, based on
23733 EBU PAL recommendations with 75% color levels.
23735 The @code{pal100bars} source generates a color bars pattern, based on
23736 EBU PAL recommendations with 100% color levels.
23738 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23739 detecting RGB vs BGR issues. You should see a red, green and blue
23740 stripe from top to bottom.
23742 The @code{smptebars} source generates a color bars pattern, based on
23743 the SMPTE Engineering Guideline EG 1-1990.
23745 The @code{smptehdbars} source generates a color bars pattern, based on
23746 the SMPTE RP 219-2002.
23748 The @code{testsrc} source generates a test video pattern, showing a
23749 color pattern, a scrolling gradient and a timestamp. This is mainly
23750 intended for testing purposes.
23752 The @code{testsrc2} source is similar to testsrc, but supports more
23753 pixel formats instead of just @code{rgb24}. This allows using it as an
23754 input for other tests without requiring a format conversion.
23756 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23757 see a y, cb and cr stripe from top to bottom.
23759 The sources accept the following parameters:
23764 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23765 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23766 pixels to be used as identity matrix for 3D lookup tables. Each component is
23767 coded on a @code{1/(N*N)} scale.
23770 Specify the color of the source, only available in the @code{color}
23771 source. For the syntax of this option, check the
23772 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23775 Specify the size of the sourced video. For the syntax of this option, check the
23776 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23777 The default value is @code{320x240}.
23779 This option is not available with the @code{allrgb}, @code{allyuv}, and
23780 @code{haldclutsrc} filters.
23783 Specify the frame rate of the sourced video, as the number of frames
23784 generated per second. It has to be a string in the format
23785 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23786 number or a valid video frame rate abbreviation. The default value is
23790 Set the duration of the sourced video. See
23791 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23792 for the accepted syntax.
23794 If not specified, or the expressed duration is negative, the video is
23795 supposed to be generated forever.
23797 Since the frame rate is used as time base, all frames including the last one
23798 will have their full duration. If the specified duration is not a multiple
23799 of the frame duration, it will be rounded up.
23802 Set the sample aspect ratio of the sourced video.
23805 Specify the alpha (opacity) of the background, only available in the
23806 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23807 255 (fully opaque, the default).
23810 Set the number of decimals to show in the timestamp, only available in the
23811 @code{testsrc} source.
23813 The displayed timestamp value will correspond to the original
23814 timestamp value multiplied by the power of 10 of the specified
23815 value. Default value is 0.
23818 @subsection Examples
23822 Generate a video with a duration of 5.3 seconds, with size
23823 176x144 and a frame rate of 10 frames per second:
23825 testsrc=duration=5.3:size=qcif:rate=10
23829 The following graph description will generate a red source
23830 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23833 color=c=red@@0.2:s=qcif:r=10
23837 If the input content is to be ignored, @code{nullsrc} can be used. The
23838 following command generates noise in the luminance plane by employing
23839 the @code{geq} filter:
23841 nullsrc=s=256x256, geq=random(1)*255:128:128
23845 @subsection Commands
23847 The @code{color} source supports the following commands:
23851 Set the color of the created image. Accepts the same syntax of the
23852 corresponding @option{color} option.
23857 Generate video using an OpenCL program.
23862 OpenCL program source file.
23865 Kernel name in program.
23868 Size of frames to generate. This must be set.
23871 Pixel format to use for the generated frames. This must be set.
23874 Number of frames generated every second. Default value is '25'.
23878 For details of how the program loading works, see the @ref{program_opencl}
23885 Generate a colour ramp by setting pixel values from the position of the pixel
23886 in the output image. (Note that this will work with all pixel formats, but
23887 the generated output will not be the same.)
23889 __kernel void ramp(__write_only image2d_t dst,
23890 unsigned int index)
23892 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23895 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23897 write_imagef(dst, loc, val);
23902 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23904 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23905 unsigned int index)
23907 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23909 float4 value = 0.0f;
23910 int x = loc.x + index;
23911 int y = loc.y + index;
23912 while (x > 0 || y > 0) {
23913 if (x % 3 == 1 && y % 3 == 1) {
23921 write_imagef(dst, loc, value);
23927 @section sierpinski
23929 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23931 This source accepts the following options:
23935 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23936 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23939 Set frame rate, expressed as number of frames per second. Default
23943 Set seed which is used for random panning.
23946 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23949 Set fractal type, can be default @code{carpet} or @code{triangle}.
23952 @c man end VIDEO SOURCES
23954 @chapter Video Sinks
23955 @c man begin VIDEO SINKS
23957 Below is a description of the currently available video sinks.
23959 @section buffersink
23961 Buffer video frames, and make them available to the end of the filter
23964 This sink is mainly intended for programmatic use, in particular
23965 through the interface defined in @file{libavfilter/buffersink.h}
23966 or the options system.
23968 It accepts a pointer to an AVBufferSinkContext structure, which
23969 defines the incoming buffers' formats, to be passed as the opaque
23970 parameter to @code{avfilter_init_filter} for initialization.
23974 Null video sink: do absolutely nothing with the input video. It is
23975 mainly useful as a template and for use in analysis / debugging
23978 @c man end VIDEO SINKS
23980 @chapter Multimedia Filters
23981 @c man begin MULTIMEDIA FILTERS
23983 Below is a description of the currently available multimedia filters.
23987 Convert input audio to a video output, displaying the audio bit scope.
23989 The filter accepts the following options:
23993 Set frame rate, expressed as number of frames per second. Default
23997 Specify the video size for the output. For the syntax of this option, check the
23998 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23999 Default value is @code{1024x256}.
24002 Specify list of colors separated by space or by '|' which will be used to
24003 draw channels. Unrecognized or missing colors will be replaced
24007 @section adrawgraph
24008 Draw a graph using input audio metadata.
24010 See @ref{drawgraph}
24012 @section agraphmonitor
24014 See @ref{graphmonitor}.
24016 @section ahistogram
24018 Convert input audio to a video output, displaying the volume histogram.
24020 The filter accepts the following options:
24024 Specify how histogram is calculated.
24026 It accepts the following values:
24029 Use single histogram for all channels.
24031 Use separate histogram for each channel.
24033 Default is @code{single}.
24036 Set frame rate, expressed as number of frames per second. Default
24040 Specify the video size for the output. For the syntax of this option, check the
24041 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24042 Default value is @code{hd720}.
24047 It accepts the following values:
24058 reverse logarithmic
24060 Default is @code{log}.
24063 Set amplitude scale.
24065 It accepts the following values:
24072 Default is @code{log}.
24075 Set how much frames to accumulate in histogram.
24076 Default is 1. Setting this to -1 accumulates all frames.
24079 Set histogram ratio of window height.
24082 Set sonogram sliding.
24084 It accepts the following values:
24087 replace old rows with new ones.
24089 scroll from top to bottom.
24091 Default is @code{replace}.
24094 @section aphasemeter
24096 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24097 representing mean phase of current audio frame. A video output can also be produced and is
24098 enabled by default. The audio is passed through as first output.
24100 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24101 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24102 and @code{1} means channels are in phase.
24104 The filter accepts the following options, all related to its video output:
24108 Set the output frame rate. Default value is @code{25}.
24111 Set the video size for the output. For the syntax of this option, check the
24112 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24113 Default value is @code{800x400}.
24118 Specify the red, green, blue contrast. Default values are @code{2},
24119 @code{7} and @code{1}.
24120 Allowed range is @code{[0, 255]}.
24123 Set color which will be used for drawing median phase. If color is
24124 @code{none} which is default, no median phase value will be drawn.
24127 Enable video output. Default is enabled.
24130 @subsection phasing detection
24132 The filter also detects out of phase and mono sequences in stereo streams.
24133 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24135 The filter accepts the following options for this detection:
24139 Enable mono and out of phase detection. Default is disabled.
24142 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24143 Allowed range is @code{[0, 1]}.
24146 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24147 Allowed range is @code{[90, 180]}.
24150 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24153 @subsection Examples
24157 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24159 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24163 @section avectorscope
24165 Convert input audio to a video output, representing the audio vector
24168 The filter is used to measure the difference between channels of stereo
24169 audio stream. A monaural signal, consisting of identical left and right
24170 signal, results in straight vertical line. Any stereo separation is visible
24171 as a deviation from this line, creating a Lissajous figure.
24172 If the straight (or deviation from it) but horizontal line appears this
24173 indicates that the left and right channels are out of phase.
24175 The filter accepts the following options:
24179 Set the vectorscope mode.
24181 Available values are:
24184 Lissajous rotated by 45 degrees.
24187 Same as above but not rotated.
24190 Shape resembling half of circle.
24193 Default value is @samp{lissajous}.
24196 Set the video size for the output. For the syntax of this option, check the
24197 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24198 Default value is @code{400x400}.
24201 Set the output frame rate. Default value is @code{25}.
24207 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24208 @code{160}, @code{80} and @code{255}.
24209 Allowed range is @code{[0, 255]}.
24215 Specify the red, green, blue and alpha fade. Default values are @code{15},
24216 @code{10}, @code{5} and @code{5}.
24217 Allowed range is @code{[0, 255]}.
24220 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24221 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24224 Set the vectorscope drawing mode.
24226 Available values are:
24229 Draw dot for each sample.
24232 Draw line between previous and current sample.
24235 Default value is @samp{dot}.
24238 Specify amplitude scale of audio samples.
24240 Available values are:
24256 Swap left channel axis with right channel axis.
24266 Mirror only x axis.
24269 Mirror only y axis.
24277 @subsection Examples
24281 Complete example using @command{ffplay}:
24283 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24284 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24288 @section bench, abench
24290 Benchmark part of a filtergraph.
24292 The filter accepts the following options:
24296 Start or stop a timer.
24298 Available values are:
24301 Get the current time, set it as frame metadata (using the key
24302 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24305 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24306 the input frame metadata to get the time difference. Time difference, average,
24307 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24308 @code{min}) are then printed. The timestamps are expressed in seconds.
24312 @subsection Examples
24316 Benchmark @ref{selectivecolor} filter:
24318 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24324 Concatenate audio and video streams, joining them together one after the
24327 The filter works on segments of synchronized video and audio streams. All
24328 segments must have the same number of streams of each type, and that will
24329 also be the number of streams at output.
24331 The filter accepts the following options:
24336 Set the number of segments. Default is 2.
24339 Set the number of output video streams, that is also the number of video
24340 streams in each segment. Default is 1.
24343 Set the number of output audio streams, that is also the number of audio
24344 streams in each segment. Default is 0.
24347 Activate unsafe mode: do not fail if segments have a different format.
24351 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24352 @var{a} audio outputs.
24354 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24355 segment, in the same order as the outputs, then the inputs for the second
24358 Related streams do not always have exactly the same duration, for various
24359 reasons including codec frame size or sloppy authoring. For that reason,
24360 related synchronized streams (e.g. a video and its audio track) should be
24361 concatenated at once. The concat filter will use the duration of the longest
24362 stream in each segment (except the last one), and if necessary pad shorter
24363 audio streams with silence.
24365 For this filter to work correctly, all segments must start at timestamp 0.
24367 All corresponding streams must have the same parameters in all segments; the
24368 filtering system will automatically select a common pixel format for video
24369 streams, and a common sample format, sample rate and channel layout for
24370 audio streams, but other settings, such as resolution, must be converted
24371 explicitly by the user.
24373 Different frame rates are acceptable but will result in variable frame rate
24374 at output; be sure to configure the output file to handle it.
24376 @subsection Examples
24380 Concatenate an opening, an episode and an ending, all in bilingual version
24381 (video in stream 0, audio in streams 1 and 2):
24383 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24384 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24385 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24386 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24390 Concatenate two parts, handling audio and video separately, using the
24391 (a)movie sources, and adjusting the resolution:
24393 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24394 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24395 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24397 Note that a desync will happen at the stitch if the audio and video streams
24398 do not have exactly the same duration in the first file.
24402 @subsection Commands
24404 This filter supports the following commands:
24407 Close the current segment and step to the next one
24413 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24414 level. By default, it logs a message at a frequency of 10Hz with the
24415 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24416 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24418 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24419 sample format is double-precision floating point. The input stream will be converted to
24420 this specification, if needed. Users may need to insert aformat and/or aresample filters
24421 after this filter to obtain the original parameters.
24423 The filter also has a video output (see the @var{video} option) with a real
24424 time graph to observe the loudness evolution. The graphic contains the logged
24425 message mentioned above, so it is not printed anymore when this option is set,
24426 unless the verbose logging is set. The main graphing area contains the
24427 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24428 the momentary loudness (400 milliseconds), but can optionally be configured
24429 to instead display short-term loudness (see @var{gauge}).
24431 The green area marks a +/- 1LU target range around the target loudness
24432 (-23LUFS by default, unless modified through @var{target}).
24434 More information about the Loudness Recommendation EBU R128 on
24435 @url{http://tech.ebu.ch/loudness}.
24437 The filter accepts the following options:
24442 Activate the video output. The audio stream is passed unchanged whether this
24443 option is set or no. The video stream will be the first output stream if
24444 activated. Default is @code{0}.
24447 Set the video size. This option is for video only. For the syntax of this
24449 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24450 Default and minimum resolution is @code{640x480}.
24453 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24454 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24455 other integer value between this range is allowed.
24458 Set metadata injection. If set to @code{1}, the audio input will be segmented
24459 into 100ms output frames, each of them containing various loudness information
24460 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24462 Default is @code{0}.
24465 Force the frame logging level.
24467 Available values are:
24470 information logging level
24472 verbose logging level
24475 By default, the logging level is set to @var{info}. If the @option{video} or
24476 the @option{metadata} options are set, it switches to @var{verbose}.
24481 Available modes can be cumulated (the option is a @code{flag} type). Possible
24485 Disable any peak mode (default).
24487 Enable sample-peak mode.
24489 Simple peak mode looking for the higher sample value. It logs a message
24490 for sample-peak (identified by @code{SPK}).
24492 Enable true-peak mode.
24494 If enabled, the peak lookup is done on an over-sampled version of the input
24495 stream for better peak accuracy. It logs a message for true-peak.
24496 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24497 This mode requires a build with @code{libswresample}.
24501 Treat mono input files as "dual mono". If a mono file is intended for playback
24502 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24503 If set to @code{true}, this option will compensate for this effect.
24504 Multi-channel input files are not affected by this option.
24507 Set a specific pan law to be used for the measurement of dual mono files.
24508 This parameter is optional, and has a default value of -3.01dB.
24511 Set a specific target level (in LUFS) used as relative zero in the visualization.
24512 This parameter is optional and has a default value of -23LUFS as specified
24513 by EBU R128. However, material published online may prefer a level of -16LUFS
24514 (e.g. for use with podcasts or video platforms).
24517 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24518 @code{shortterm}. By default the momentary value will be used, but in certain
24519 scenarios it may be more useful to observe the short term value instead (e.g.
24523 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24524 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24525 video output, not the summary or continuous log output.
24528 @subsection Examples
24532 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24534 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24538 Run an analysis with @command{ffmpeg}:
24540 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24544 @section interleave, ainterleave
24546 Temporally interleave frames from several inputs.
24548 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24550 These filters read frames from several inputs and send the oldest
24551 queued frame to the output.
24553 Input streams must have well defined, monotonically increasing frame
24556 In order to submit one frame to output, these filters need to enqueue
24557 at least one frame for each input, so they cannot work in case one
24558 input is not yet terminated and will not receive incoming frames.
24560 For example consider the case when one input is a @code{select} filter
24561 which always drops input frames. The @code{interleave} filter will keep
24562 reading from that input, but it will never be able to send new frames
24563 to output until the input sends an end-of-stream signal.
24565 Also, depending on inputs synchronization, the filters will drop
24566 frames in case one input receives more frames than the other ones, and
24567 the queue is already filled.
24569 These filters accept the following options:
24573 Set the number of different inputs, it is 2 by default.
24576 How to determine the end-of-stream.
24580 The duration of the longest input. (default)
24583 The duration of the shortest input.
24586 The duration of the first input.
24591 @subsection Examples
24595 Interleave frames belonging to different streams using @command{ffmpeg}:
24597 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24601 Add flickering blur effect:
24603 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24607 @section metadata, ametadata
24609 Manipulate frame metadata.
24611 This filter accepts the following options:
24615 Set mode of operation of the filter.
24617 Can be one of the following:
24621 If both @code{value} and @code{key} is set, select frames
24622 which have such metadata. If only @code{key} is set, select
24623 every frame that has such key in metadata.
24626 Add new metadata @code{key} and @code{value}. If key is already available
24630 Modify value of already present key.
24633 If @code{value} is set, delete only keys that have such value.
24634 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24638 Print key and its value if metadata was found. If @code{key} is not set print all
24639 metadata values available in frame.
24643 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24646 Set metadata value which will be used. This option is mandatory for
24647 @code{modify} and @code{add} mode.
24650 Which function to use when comparing metadata value and @code{value}.
24652 Can be one of following:
24656 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24659 Values are interpreted as strings, returns true if metadata value starts with
24660 the @code{value} option string.
24663 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24666 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24669 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24672 Values are interpreted as floats, returns true if expression from option @code{expr}
24676 Values are interpreted as strings, returns true if metadata value ends with
24677 the @code{value} option string.
24681 Set expression which is used when @code{function} is set to @code{expr}.
24682 The expression is evaluated through the eval API and can contain the following
24687 Float representation of @code{value} from metadata key.
24690 Float representation of @code{value} as supplied by user in @code{value} option.
24694 If specified in @code{print} mode, output is written to the named file. Instead of
24695 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24696 for standard output. If @code{file} option is not set, output is written to the log
24697 with AV_LOG_INFO loglevel.
24700 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24704 @subsection Examples
24708 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24711 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24714 Print silencedetect output to file @file{metadata.txt}.
24716 silencedetect,ametadata=mode=print:file=metadata.txt
24719 Direct all metadata to a pipe with file descriptor 4.
24721 metadata=mode=print:file='pipe\:4'
24725 @section perms, aperms
24727 Set read/write permissions for the output frames.
24729 These filters are mainly aimed at developers to test direct path in the
24730 following filter in the filtergraph.
24732 The filters accept the following options:
24736 Select the permissions mode.
24738 It accepts the following values:
24741 Do nothing. This is the default.
24743 Set all the output frames read-only.
24745 Set all the output frames directly writable.
24747 Make the frame read-only if writable, and writable if read-only.
24749 Set each output frame read-only or writable randomly.
24753 Set the seed for the @var{random} mode, must be an integer included between
24754 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24755 @code{-1}, the filter will try to use a good random seed on a best effort
24759 Note: in case of auto-inserted filter between the permission filter and the
24760 following one, the permission might not be received as expected in that
24761 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24762 perms/aperms filter can avoid this problem.
24764 @section realtime, arealtime
24766 Slow down filtering to match real time approximately.
24768 These filters will pause the filtering for a variable amount of time to
24769 match the output rate with the input timestamps.
24770 They are similar to the @option{re} option to @code{ffmpeg}.
24772 They accept the following options:
24776 Time limit for the pauses. Any pause longer than that will be considered
24777 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24779 Speed factor for processing. The value must be a float larger than zero.
24780 Values larger than 1.0 will result in faster than realtime processing,
24781 smaller will slow processing down. The @var{limit} is automatically adapted
24782 accordingly. Default is 1.0.
24784 A processing speed faster than what is possible without these filters cannot
24789 @section select, aselect
24791 Select frames to pass in output.
24793 This filter accepts the following options:
24798 Set expression, which is evaluated for each input frame.
24800 If the expression is evaluated to zero, the frame is discarded.
24802 If the evaluation result is negative or NaN, the frame is sent to the
24803 first output; otherwise it is sent to the output with index
24804 @code{ceil(val)-1}, assuming that the input index starts from 0.
24806 For example a value of @code{1.2} corresponds to the output with index
24807 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24810 Set the number of outputs. The output to which to send the selected
24811 frame is based on the result of the evaluation. Default value is 1.
24814 The expression can contain the following constants:
24818 The (sequential) number of the filtered frame, starting from 0.
24821 The (sequential) number of the selected frame, starting from 0.
24823 @item prev_selected_n
24824 The sequential number of the last selected frame. It's NAN if undefined.
24827 The timebase of the input timestamps.
24830 The PTS (Presentation TimeStamp) of the filtered video frame,
24831 expressed in @var{TB} units. It's NAN if undefined.
24834 The PTS of the filtered video frame,
24835 expressed in seconds. It's NAN if undefined.
24838 The PTS of the previously filtered video frame. It's NAN if undefined.
24840 @item prev_selected_pts
24841 The PTS of the last previously filtered video frame. It's NAN if undefined.
24843 @item prev_selected_t
24844 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24847 The PTS of the first video frame in the video. It's NAN if undefined.
24850 The time of the first video frame in the video. It's NAN if undefined.
24852 @item pict_type @emph{(video only)}
24853 The type of the filtered frame. It can assume one of the following
24865 @item interlace_type @emph{(video only)}
24866 The frame interlace type. It can assume one of the following values:
24869 The frame is progressive (not interlaced).
24871 The frame is top-field-first.
24873 The frame is bottom-field-first.
24876 @item consumed_sample_n @emph{(audio only)}
24877 the number of selected samples before the current frame
24879 @item samples_n @emph{(audio only)}
24880 the number of samples in the current frame
24882 @item sample_rate @emph{(audio only)}
24883 the input sample rate
24886 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24889 the position in the file of the filtered frame, -1 if the information
24890 is not available (e.g. for synthetic video)
24892 @item scene @emph{(video only)}
24893 value between 0 and 1 to indicate a new scene; a low value reflects a low
24894 probability for the current frame to introduce a new scene, while a higher
24895 value means the current frame is more likely to be one (see the example below)
24897 @item concatdec_select
24898 The concat demuxer can select only part of a concat input file by setting an
24899 inpoint and an outpoint, but the output packets may not be entirely contained
24900 in the selected interval. By using this variable, it is possible to skip frames
24901 generated by the concat demuxer which are not exactly contained in the selected
24904 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24905 and the @var{lavf.concat.duration} packet metadata values which are also
24906 present in the decoded frames.
24908 The @var{concatdec_select} variable is -1 if the frame pts is at least
24909 start_time and either the duration metadata is missing or the frame pts is less
24910 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24913 That basically means that an input frame is selected if its pts is within the
24914 interval set by the concat demuxer.
24918 The default value of the select expression is "1".
24920 @subsection Examples
24924 Select all frames in input:
24929 The example above is the same as:
24941 Select only I-frames:
24943 select='eq(pict_type\,I)'
24947 Select one frame every 100:
24949 select='not(mod(n\,100))'
24953 Select only frames contained in the 10-20 time interval:
24955 select=between(t\,10\,20)
24959 Select only I-frames contained in the 10-20 time interval:
24961 select=between(t\,10\,20)*eq(pict_type\,I)
24965 Select frames with a minimum distance of 10 seconds:
24967 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24971 Use aselect to select only audio frames with samples number > 100:
24973 aselect='gt(samples_n\,100)'
24977 Create a mosaic of the first scenes:
24979 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24982 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24986 Send even and odd frames to separate outputs, and compose them:
24988 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24992 Select useful frames from an ffconcat file which is using inpoints and
24993 outpoints but where the source files are not intra frame only.
24995 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24999 @section sendcmd, asendcmd
25001 Send commands to filters in the filtergraph.
25003 These filters read commands to be sent to other filters in the
25006 @code{sendcmd} must be inserted between two video filters,
25007 @code{asendcmd} must be inserted between two audio filters, but apart
25008 from that they act the same way.
25010 The specification of commands can be provided in the filter arguments
25011 with the @var{commands} option, or in a file specified by the
25012 @var{filename} option.
25014 These filters accept the following options:
25017 Set the commands to be read and sent to the other filters.
25019 Set the filename of the commands to be read and sent to the other
25023 @subsection Commands syntax
25025 A commands description consists of a sequence of interval
25026 specifications, comprising a list of commands to be executed when a
25027 particular event related to that interval occurs. The occurring event
25028 is typically the current frame time entering or leaving a given time
25031 An interval is specified by the following syntax:
25033 @var{START}[-@var{END}] @var{COMMANDS};
25036 The time interval is specified by the @var{START} and @var{END} times.
25037 @var{END} is optional and defaults to the maximum time.
25039 The current frame time is considered within the specified interval if
25040 it is included in the interval [@var{START}, @var{END}), that is when
25041 the time is greater or equal to @var{START} and is lesser than
25044 @var{COMMANDS} consists of a sequence of one or more command
25045 specifications, separated by ",", relating to that interval. The
25046 syntax of a command specification is given by:
25048 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25051 @var{FLAGS} is optional and specifies the type of events relating to
25052 the time interval which enable sending the specified command, and must
25053 be a non-null sequence of identifier flags separated by "+" or "|" and
25054 enclosed between "[" and "]".
25056 The following flags are recognized:
25059 The command is sent when the current frame timestamp enters the
25060 specified interval. In other words, the command is sent when the
25061 previous frame timestamp was not in the given interval, and the
25065 The command is sent when the current frame timestamp leaves the
25066 specified interval. In other words, the command is sent when the
25067 previous frame timestamp was in the given interval, and the
25071 The command @var{ARG} is interpreted as expression and result of
25072 expression is passed as @var{ARG}.
25074 The expression is evaluated through the eval API and can contain the following
25079 Original position in the file of the frame, or undefined if undefined
25080 for the current frame.
25083 The presentation timestamp in input.
25086 The count of the input frame for video or audio, starting from 0.
25089 The time in seconds of the current frame.
25092 The start time in seconds of the current command interval.
25095 The end time in seconds of the current command interval.
25098 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25103 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25106 @var{TARGET} specifies the target of the command, usually the name of
25107 the filter class or a specific filter instance name.
25109 @var{COMMAND} specifies the name of the command for the target filter.
25111 @var{ARG} is optional and specifies the optional list of argument for
25112 the given @var{COMMAND}.
25114 Between one interval specification and another, whitespaces, or
25115 sequences of characters starting with @code{#} until the end of line,
25116 are ignored and can be used to annotate comments.
25118 A simplified BNF description of the commands specification syntax
25121 @var{COMMAND_FLAG} ::= "enter" | "leave"
25122 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25123 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25124 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25125 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25126 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25129 @subsection Examples
25133 Specify audio tempo change at second 4:
25135 asendcmd=c='4.0 atempo tempo 1.5',atempo
25139 Target a specific filter instance:
25141 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25145 Specify a list of drawtext and hue commands in a file.
25147 # show text in the interval 5-10
25148 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25149 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25151 # desaturate the image in the interval 15-20
25152 15.0-20.0 [enter] hue s 0,
25153 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25155 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25157 # apply an exponential saturation fade-out effect, starting from time 25
25158 25 [enter] hue s exp(25-t)
25161 A filtergraph allowing to read and process the above command list
25162 stored in a file @file{test.cmd}, can be specified with:
25164 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25169 @section setpts, asetpts
25171 Change the PTS (presentation timestamp) of the input frames.
25173 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25175 This filter accepts the following options:
25180 The expression which is evaluated for each frame to construct its timestamp.
25184 The expression is evaluated through the eval API and can contain the following
25188 @item FRAME_RATE, FR
25189 frame rate, only defined for constant frame-rate video
25192 The presentation timestamp in input
25195 The count of the input frame for video or the number of consumed samples,
25196 not including the current frame for audio, starting from 0.
25198 @item NB_CONSUMED_SAMPLES
25199 The number of consumed samples, not including the current frame (only
25202 @item NB_SAMPLES, S
25203 The number of samples in the current frame (only audio)
25205 @item SAMPLE_RATE, SR
25206 The audio sample rate.
25209 The PTS of the first frame.
25212 the time in seconds of the first frame
25215 State whether the current frame is interlaced.
25218 the time in seconds of the current frame
25221 original position in the file of the frame, or undefined if undefined
25222 for the current frame
25225 The previous input PTS.
25228 previous input time in seconds
25231 The previous output PTS.
25234 previous output time in seconds
25237 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25241 The wallclock (RTC) time at the start of the movie in microseconds.
25244 The timebase of the input timestamps.
25248 @subsection Examples
25252 Start counting PTS from zero
25254 setpts=PTS-STARTPTS
25258 Apply fast motion effect:
25264 Apply slow motion effect:
25270 Set fixed rate of 25 frames per second:
25276 Set fixed rate 25 fps with some jitter:
25278 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25282 Apply an offset of 10 seconds to the input PTS:
25288 Generate timestamps from a "live source" and rebase onto the current timebase:
25290 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25294 Generate timestamps by counting samples:
25303 Force color range for the output video frame.
25305 The @code{setrange} filter marks the color range property for the
25306 output frames. It does not change the input frame, but only sets the
25307 corresponding property, which affects how the frame is treated by
25310 The filter accepts the following options:
25315 Available values are:
25319 Keep the same color range property.
25321 @item unspecified, unknown
25322 Set the color range as unspecified.
25324 @item limited, tv, mpeg
25325 Set the color range as limited.
25327 @item full, pc, jpeg
25328 Set the color range as full.
25332 @section settb, asettb
25334 Set the timebase to use for the output frames timestamps.
25335 It is mainly useful for testing timebase configuration.
25337 It accepts the following parameters:
25342 The expression which is evaluated into the output timebase.
25346 The value for @option{tb} is an arithmetic expression representing a
25347 rational. The expression can contain the constants "AVTB" (the default
25348 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25349 audio only). Default value is "intb".
25351 @subsection Examples
25355 Set the timebase to 1/25:
25361 Set the timebase to 1/10:
25367 Set the timebase to 1001/1000:
25373 Set the timebase to 2*intb:
25379 Set the default timebase value:
25386 Convert input audio to a video output representing frequency spectrum
25387 logarithmically using Brown-Puckette constant Q transform algorithm with
25388 direct frequency domain coefficient calculation (but the transform itself
25389 is not really constant Q, instead the Q factor is actually variable/clamped),
25390 with musical tone scale, from E0 to D#10.
25392 The filter accepts the following options:
25396 Specify the video size for the output. It must be even. For the syntax of this option,
25397 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25398 Default value is @code{1920x1080}.
25401 Set the output frame rate. Default value is @code{25}.
25404 Set the bargraph height. It must be even. Default value is @code{-1} which
25405 computes the bargraph height automatically.
25408 Set the axis height. It must be even. Default value is @code{-1} which computes
25409 the axis height automatically.
25412 Set the sonogram height. It must be even. Default value is @code{-1} which
25413 computes the sonogram height automatically.
25416 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25417 instead. Default value is @code{1}.
25419 @item sono_v, volume
25420 Specify the sonogram volume expression. It can contain variables:
25423 the @var{bar_v} evaluated expression
25424 @item frequency, freq, f
25425 the frequency where it is evaluated
25426 @item timeclamp, tc
25427 the value of @var{timeclamp} option
25431 @item a_weighting(f)
25432 A-weighting of equal loudness
25433 @item b_weighting(f)
25434 B-weighting of equal loudness
25435 @item c_weighting(f)
25436 C-weighting of equal loudness.
25438 Default value is @code{16}.
25440 @item bar_v, volume2
25441 Specify the bargraph volume expression. It can contain variables:
25444 the @var{sono_v} evaluated expression
25445 @item frequency, freq, f
25446 the frequency where it is evaluated
25447 @item timeclamp, tc
25448 the value of @var{timeclamp} option
25452 @item a_weighting(f)
25453 A-weighting of equal loudness
25454 @item b_weighting(f)
25455 B-weighting of equal loudness
25456 @item c_weighting(f)
25457 C-weighting of equal loudness.
25459 Default value is @code{sono_v}.
25461 @item sono_g, gamma
25462 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25463 higher gamma makes the spectrum having more range. Default value is @code{3}.
25464 Acceptable range is @code{[1, 7]}.
25466 @item bar_g, gamma2
25467 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25471 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25472 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25474 @item timeclamp, tc
25475 Specify the transform timeclamp. At low frequency, there is trade-off between
25476 accuracy in time domain and frequency domain. If timeclamp is lower,
25477 event in time domain is represented more accurately (such as fast bass drum),
25478 otherwise event in frequency domain is represented more accurately
25479 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25482 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25483 limits future samples by applying asymmetric windowing in time domain, useful
25484 when low latency is required. Accepted range is @code{[0, 1]}.
25487 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25488 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25491 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25492 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25495 This option is deprecated and ignored.
25498 Specify the transform length in time domain. Use this option to control accuracy
25499 trade-off between time domain and frequency domain at every frequency sample.
25500 It can contain variables:
25502 @item frequency, freq, f
25503 the frequency where it is evaluated
25504 @item timeclamp, tc
25505 the value of @var{timeclamp} option.
25507 Default value is @code{384*tc/(384+tc*f)}.
25510 Specify the transform count for every video frame. Default value is @code{6}.
25511 Acceptable range is @code{[1, 30]}.
25514 Specify the transform count for every single pixel. Default value is @code{0},
25515 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25518 Specify font file for use with freetype to draw the axis. If not specified,
25519 use embedded font. Note that drawing with font file or embedded font is not
25520 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25524 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25525 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25529 Specify font color expression. This is arithmetic expression that should return
25530 integer value 0xRRGGBB. It can contain variables:
25532 @item frequency, freq, f
25533 the frequency where it is evaluated
25534 @item timeclamp, tc
25535 the value of @var{timeclamp} option
25540 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25541 @item r(x), g(x), b(x)
25542 red, green, and blue value of intensity x.
25544 Default value is @code{st(0, (midi(f)-59.5)/12);
25545 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25546 r(1-ld(1)) + b(ld(1))}.
25549 Specify image file to draw the axis. This option override @var{fontfile} and
25550 @var{fontcolor} option.
25553 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25554 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25555 Default value is @code{1}.
25558 Set colorspace. The accepted values are:
25561 Unspecified (default)
25570 BT.470BG or BT.601-6 625
25573 SMPTE-170M or BT.601-6 525
25579 BT.2020 with non-constant luminance
25584 Set spectrogram color scheme. This is list of floating point values with format
25585 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25586 The default is @code{1|0.5|0|0|0.5|1}.
25590 @subsection Examples
25594 Playing audio while showing the spectrum:
25596 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25600 Same as above, but with frame rate 30 fps:
25602 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25606 Playing at 1280x720:
25608 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25612 Disable sonogram display:
25618 A1 and its harmonics: A1, A2, (near)E3, A3:
25620 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),
25621 asplit[a][out1]; [a] showcqt [out0]'
25625 Same as above, but with more accuracy in frequency domain:
25627 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),
25628 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25634 bar_v=10:sono_v=bar_v*a_weighting(f)
25638 Custom gamma, now spectrum is linear to the amplitude.
25644 Custom tlength equation:
25646 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)))'
25650 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25652 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25656 Custom font using fontconfig:
25658 font='Courier New,Monospace,mono|bold'
25662 Custom frequency range with custom axis using image file:
25664 axisfile=myaxis.png:basefreq=40:endfreq=10000
25670 Convert input audio to video output representing the audio power spectrum.
25671 Audio amplitude is on Y-axis while frequency is on X-axis.
25673 The filter accepts the following options:
25677 Specify size of video. For the syntax of this option, check the
25678 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25679 Default is @code{1024x512}.
25683 This set how each frequency bin will be represented.
25685 It accepts the following values:
25691 Default is @code{bar}.
25694 Set amplitude scale.
25696 It accepts the following values:
25710 Default is @code{log}.
25713 Set frequency scale.
25715 It accepts the following values:
25724 Reverse logarithmic scale.
25726 Default is @code{lin}.
25729 Set window size. Allowed range is from 16 to 65536.
25731 Default is @code{2048}
25734 Set windowing function.
25736 It accepts the following values:
25759 Default is @code{hanning}.
25762 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25763 which means optimal overlap for selected window function will be picked.
25766 Set time averaging. Setting this to 0 will display current maximal peaks.
25767 Default is @code{1}, which means time averaging is disabled.
25770 Specify list of colors separated by space or by '|' which will be used to
25771 draw channel frequencies. Unrecognized or missing colors will be replaced
25775 Set channel display mode.
25777 It accepts the following values:
25782 Default is @code{combined}.
25785 Set minimum amplitude used in @code{log} amplitude scaler.
25788 Set data display mode.
25790 It accepts the following values:
25796 Default is @code{magnitude}.
25799 @section showspatial
25801 Convert stereo input audio to a video output, representing the spatial relationship
25802 between two channels.
25804 The filter accepts the following options:
25808 Specify the video size for the output. For the syntax of this option, check the
25809 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25810 Default value is @code{512x512}.
25813 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25816 Set window function.
25818 It accepts the following values:
25843 Default value is @code{hann}.
25846 Set ratio of overlap window. Default value is @code{0.5}.
25847 When value is @code{1} overlap is set to recommended size for specific
25848 window function currently used.
25851 @anchor{showspectrum}
25852 @section showspectrum
25854 Convert input audio to a video output, representing the audio frequency
25857 The filter accepts the following options:
25861 Specify the video size for the output. For the syntax of this option, check the
25862 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25863 Default value is @code{640x512}.
25866 Specify how the spectrum should slide along the window.
25868 It accepts the following values:
25871 the samples start again on the left when they reach the right
25873 the samples scroll from right to left
25875 frames are only produced when the samples reach the right
25877 the samples scroll from left to right
25880 Default value is @code{replace}.
25883 Specify display mode.
25885 It accepts the following values:
25888 all channels are displayed in the same row
25890 all channels are displayed in separate rows
25893 Default value is @samp{combined}.
25896 Specify display color mode.
25898 It accepts the following values:
25901 each channel is displayed in a separate color
25903 each channel is displayed using the same color scheme
25905 each channel is displayed using the rainbow color scheme
25907 each channel is displayed using the moreland color scheme
25909 each channel is displayed using the nebulae color scheme
25911 each channel is displayed using the fire color scheme
25913 each channel is displayed using the fiery color scheme
25915 each channel is displayed using the fruit color scheme
25917 each channel is displayed using the cool color scheme
25919 each channel is displayed using the magma color scheme
25921 each channel is displayed using the green color scheme
25923 each channel is displayed using the viridis color scheme
25925 each channel is displayed using the plasma color scheme
25927 each channel is displayed using the cividis color scheme
25929 each channel is displayed using the terrain color scheme
25932 Default value is @samp{channel}.
25935 Specify scale used for calculating intensity color values.
25937 It accepts the following values:
25942 square root, default
25953 Default value is @samp{sqrt}.
25956 Specify frequency scale.
25958 It accepts the following values:
25966 Default value is @samp{lin}.
25969 Set saturation modifier for displayed colors. Negative values provide
25970 alternative color scheme. @code{0} is no saturation at all.
25971 Saturation must be in [-10.0, 10.0] range.
25972 Default value is @code{1}.
25975 Set window function.
25977 It accepts the following values:
26002 Default value is @code{hann}.
26005 Set orientation of time vs frequency axis. Can be @code{vertical} or
26006 @code{horizontal}. Default is @code{vertical}.
26009 Set ratio of overlap window. Default value is @code{0}.
26010 When value is @code{1} overlap is set to recommended size for specific
26011 window function currently used.
26014 Set scale gain for calculating intensity color values.
26015 Default value is @code{1}.
26018 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
26021 Set color rotation, must be in [-1.0, 1.0] range.
26022 Default value is @code{0}.
26025 Set start frequency from which to display spectrogram. Default is @code{0}.
26028 Set stop frequency to which to display spectrogram. Default is @code{0}.
26031 Set upper frame rate limit. Default is @code{auto}, unlimited.
26034 Draw time and frequency axes and legends. Default is disabled.
26037 The usage is very similar to the showwaves filter; see the examples in that
26040 @subsection Examples
26044 Large window with logarithmic color scaling:
26046 showspectrum=s=1280x480:scale=log
26050 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26052 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26053 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26057 @section showspectrumpic
26059 Convert input audio to a single video frame, representing the audio frequency
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{4096x2048}.
26071 Specify display mode.
26073 It accepts the following values:
26076 all channels are displayed in the same row
26078 all channels are displayed in separate rows
26080 Default value is @samp{combined}.
26083 Specify display color mode.
26085 It accepts the following values:
26088 each channel is displayed in a separate color
26090 each channel is displayed using the same color scheme
26092 each channel is displayed using the rainbow color scheme
26094 each channel is displayed using the moreland color scheme
26096 each channel is displayed using the nebulae color scheme
26098 each channel is displayed using the fire color scheme
26100 each channel is displayed using the fiery color scheme
26102 each channel is displayed using the fruit color scheme
26104 each channel is displayed using the cool color scheme
26106 each channel is displayed using the magma color scheme
26108 each channel is displayed using the green color scheme
26110 each channel is displayed using the viridis color scheme
26112 each channel is displayed using the plasma color scheme
26114 each channel is displayed using the cividis color scheme
26116 each channel is displayed using the terrain color scheme
26118 Default value is @samp{intensity}.
26121 Specify scale used for calculating intensity color values.
26123 It accepts the following values:
26128 square root, default
26138 Default value is @samp{log}.
26141 Specify frequency scale.
26143 It accepts the following values:
26151 Default value is @samp{lin}.
26154 Set saturation modifier for displayed colors. Negative values provide
26155 alternative color scheme. @code{0} is no saturation at all.
26156 Saturation must be in [-10.0, 10.0] range.
26157 Default value is @code{1}.
26160 Set window function.
26162 It accepts the following values:
26186 Default value is @code{hann}.
26189 Set orientation of time vs frequency axis. Can be @code{vertical} or
26190 @code{horizontal}. Default is @code{vertical}.
26193 Set scale gain for calculating intensity color values.
26194 Default value is @code{1}.
26197 Draw time and frequency axes and legends. Default is enabled.
26200 Set color rotation, must be in [-1.0, 1.0] range.
26201 Default value is @code{0}.
26204 Set start frequency from which to display spectrogram. Default is @code{0}.
26207 Set stop frequency to which to display spectrogram. Default is @code{0}.
26210 @subsection Examples
26214 Extract an audio spectrogram of a whole audio track
26215 in a 1024x1024 picture using @command{ffmpeg}:
26217 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26221 @section showvolume
26223 Convert input audio volume to a video output.
26225 The filter accepts the following options:
26232 Set border width, allowed range is [0, 5]. Default is 1.
26235 Set channel width, allowed range is [80, 8192]. Default is 400.
26238 Set channel height, allowed range is [1, 900]. Default is 20.
26241 Set fade, allowed range is [0, 1]. Default is 0.95.
26244 Set volume color expression.
26246 The expression can use the following variables:
26250 Current max volume of channel in dB.
26256 Current channel number, starting from 0.
26260 If set, displays channel names. Default is enabled.
26263 If set, displays volume values. Default is enabled.
26266 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26267 default is @code{h}.
26270 Set step size, allowed range is [0, 5]. Default is 0, which means
26274 Set background opacity, allowed range is [0, 1]. Default is 0.
26277 Set metering mode, can be peak: @code{p} or rms: @code{r},
26278 default is @code{p}.
26281 Set display scale, can be linear: @code{lin} or log: @code{log},
26282 default is @code{lin}.
26286 If set to > 0., display a line for the max level
26287 in the previous seconds.
26288 default is disabled: @code{0.}
26291 The color of the max line. Use when @code{dm} option is set to > 0.
26292 default is: @code{orange}
26297 Convert input audio to a video output, representing the samples waves.
26299 The filter accepts the following options:
26303 Specify the video size for the output. For the syntax of this option, check the
26304 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26305 Default value is @code{600x240}.
26310 Available values are:
26313 Draw a point for each sample.
26316 Draw a vertical line for each sample.
26319 Draw a point for each sample and a line between them.
26322 Draw a centered vertical line for each sample.
26325 Default value is @code{point}.
26328 Set the number of samples which are printed on the same column. A
26329 larger value will decrease the frame rate. Must be a positive
26330 integer. This option can be set only if the value for @var{rate}
26331 is not explicitly specified.
26334 Set the (approximate) output frame rate. This is done by setting the
26335 option @var{n}. Default value is "25".
26337 @item split_channels
26338 Set if channels should be drawn separately or overlap. Default value is 0.
26341 Set colors separated by '|' which are going to be used for drawing of each channel.
26344 Set amplitude scale.
26346 Available values are:
26364 Set the draw mode. This is mostly useful to set for high @var{n}.
26366 Available values are:
26369 Scale pixel values for each drawn sample.
26372 Draw every sample directly.
26375 Default value is @code{scale}.
26378 @subsection Examples
26382 Output the input file audio and the corresponding video representation
26385 amovie=a.mp3,asplit[out0],showwaves[out1]
26389 Create a synthetic signal and show it with showwaves, forcing a
26390 frame rate of 30 frames per second:
26392 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26396 @section showwavespic
26398 Convert input audio to a single video frame, representing the samples waves.
26400 The filter accepts the following options:
26404 Specify the video size for the output. For the syntax of this option, check the
26405 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26406 Default value is @code{600x240}.
26408 @item split_channels
26409 Set if channels should be drawn separately or overlap. Default value is 0.
26412 Set colors separated by '|' which are going to be used for drawing of each channel.
26415 Set amplitude scale.
26417 Available values are:
26437 Available values are:
26440 Scale pixel values for each drawn sample.
26443 Draw every sample directly.
26446 Default value is @code{scale}.
26449 Set the filter mode.
26451 Available values are:
26454 Use average samples values for each drawn sample.
26457 Use peak samples values for each drawn sample.
26460 Default value is @code{average}.
26463 @subsection Examples
26467 Extract a channel split representation of the wave form of a whole audio track
26468 in a 1024x800 picture using @command{ffmpeg}:
26470 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26474 @section sidedata, asidedata
26476 Delete frame side data, or select frames based on it.
26478 This filter accepts the following options:
26482 Set mode of operation of the filter.
26484 Can be one of the following:
26488 Select every frame with side data of @code{type}.
26491 Delete side data of @code{type}. If @code{type} is not set, delete all side
26497 Set side data type used with all modes. Must be set for @code{select} mode. For
26498 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26499 in @file{libavutil/frame.h}. For example, to choose
26500 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26504 @section spectrumsynth
26506 Synthesize audio from 2 input video spectrums, first input stream represents
26507 magnitude across time and second represents phase across time.
26508 The filter will transform from frequency domain as displayed in videos back
26509 to time domain as presented in audio output.
26511 This filter is primarily created for reversing processed @ref{showspectrum}
26512 filter outputs, but can synthesize sound from other spectrograms too.
26513 But in such case results are going to be poor if the phase data is not
26514 available, because in such cases phase data need to be recreated, usually
26515 it's just recreated from random noise.
26516 For best results use gray only output (@code{channel} color mode in
26517 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26518 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26519 @code{data} option. Inputs videos should generally use @code{fullframe}
26520 slide mode as that saves resources needed for decoding video.
26522 The filter accepts the following options:
26526 Specify sample rate of output audio, the sample rate of audio from which
26527 spectrum was generated may differ.
26530 Set number of channels represented in input video spectrums.
26533 Set scale which was used when generating magnitude input spectrum.
26534 Can be @code{lin} or @code{log}. Default is @code{log}.
26537 Set slide which was used when generating inputs spectrums.
26538 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26539 Default is @code{fullframe}.
26542 Set window function used for resynthesis.
26545 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26546 which means optimal overlap for selected window function will be picked.
26549 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26550 Default is @code{vertical}.
26553 @subsection Examples
26557 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26558 then resynthesize videos back to audio with spectrumsynth:
26560 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
26561 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
26562 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26566 @section split, asplit
26568 Split input into several identical outputs.
26570 @code{asplit} works with audio input, @code{split} with video.
26572 The filter accepts a single parameter which specifies the number of outputs. If
26573 unspecified, it defaults to 2.
26575 @subsection Examples
26579 Create two separate outputs from the same input:
26581 [in] split [out0][out1]
26585 To create 3 or more outputs, you need to specify the number of
26588 [in] asplit=3 [out0][out1][out2]
26592 Create two separate outputs from the same input, one cropped and
26595 [in] split [splitout1][splitout2];
26596 [splitout1] crop=100:100:0:0 [cropout];
26597 [splitout2] pad=200:200:100:100 [padout];
26601 Create 5 copies of the input audio with @command{ffmpeg}:
26603 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26609 Receive commands sent through a libzmq client, and forward them to
26610 filters in the filtergraph.
26612 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26613 must be inserted between two video filters, @code{azmq} between two
26614 audio filters. Both are capable to send messages to any filter type.
26616 To enable these filters you need to install the libzmq library and
26617 headers and configure FFmpeg with @code{--enable-libzmq}.
26619 For more information about libzmq see:
26620 @url{http://www.zeromq.org/}
26622 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26623 receives messages sent through a network interface defined by the
26624 @option{bind_address} (or the abbreviation "@option{b}") option.
26625 Default value of this option is @file{tcp://localhost:5555}. You may
26626 want to alter this value to your needs, but do not forget to escape any
26627 ':' signs (see @ref{filtergraph escaping}).
26629 The received message must be in the form:
26631 @var{TARGET} @var{COMMAND} [@var{ARG}]
26634 @var{TARGET} specifies the target of the command, usually the name of
26635 the filter class or a specific filter instance name. The default
26636 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26637 but you can override this by using the @samp{filter_name@@id} syntax
26638 (see @ref{Filtergraph syntax}).
26640 @var{COMMAND} specifies the name of the command for the target filter.
26642 @var{ARG} is optional and specifies the optional argument list for the
26643 given @var{COMMAND}.
26645 Upon reception, the message is processed and the corresponding command
26646 is injected into the filtergraph. Depending on the result, the filter
26647 will send a reply to the client, adopting the format:
26649 @var{ERROR_CODE} @var{ERROR_REASON}
26653 @var{MESSAGE} is optional.
26655 @subsection Examples
26657 Look at @file{tools/zmqsend} for an example of a zmq client which can
26658 be used to send commands processed by these filters.
26660 Consider the following filtergraph generated by @command{ffplay}.
26661 In this example the last overlay filter has an instance name. All other
26662 filters will have default instance names.
26665 ffplay -dumpgraph 1 -f lavfi "
26666 color=s=100x100:c=red [l];
26667 color=s=100x100:c=blue [r];
26668 nullsrc=s=200x100, zmq [bg];
26669 [bg][l] overlay [bg+l];
26670 [bg+l][r] overlay@@my=x=100 "
26673 To change the color of the left side of the video, the following
26674 command can be used:
26676 echo Parsed_color_0 c yellow | tools/zmqsend
26679 To change the right side:
26681 echo Parsed_color_1 c pink | tools/zmqsend
26684 To change the position of the right side:
26686 echo overlay@@my x 150 | tools/zmqsend
26690 @c man end MULTIMEDIA FILTERS
26692 @chapter Multimedia Sources
26693 @c man begin MULTIMEDIA SOURCES
26695 Below is a description of the currently available multimedia sources.
26699 This is the same as @ref{movie} source, except it selects an audio
26705 Read audio and/or video stream(s) from a movie container.
26707 It accepts the following parameters:
26711 The name of the resource to read (not necessarily a file; it can also be a
26712 device or a stream accessed through some protocol).
26714 @item format_name, f
26715 Specifies the format assumed for the movie to read, and can be either
26716 the name of a container or an input device. If not specified, the
26717 format is guessed from @var{movie_name} or by probing.
26719 @item seek_point, sp
26720 Specifies the seek point in seconds. The frames will be output
26721 starting from this seek point. The parameter is evaluated with
26722 @code{av_strtod}, so the numerical value may be suffixed by an IS
26723 postfix. The default value is "0".
26726 Specifies the streams to read. Several streams can be specified,
26727 separated by "+". The source will then have as many outputs, in the
26728 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26729 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26730 respectively the default (best suited) video and audio stream. Default
26731 is "dv", or "da" if the filter is called as "amovie".
26733 @item stream_index, si
26734 Specifies the index of the video stream to read. If the value is -1,
26735 the most suitable video stream will be automatically selected. The default
26736 value is "-1". Deprecated. If the filter is called "amovie", it will select
26737 audio instead of video.
26740 Specifies how many times to read the stream in sequence.
26741 If the value is 0, the stream will be looped infinitely.
26742 Default value is "1".
26744 Note that when the movie is looped the source timestamps are not
26745 changed, so it will generate non monotonically increasing timestamps.
26747 @item discontinuity
26748 Specifies the time difference between frames above which the point is
26749 considered a timestamp discontinuity which is removed by adjusting the later
26753 It allows overlaying a second video on top of the main input of
26754 a filtergraph, as shown in this graph:
26756 input -----------> deltapts0 --> overlay --> output
26759 movie --> scale--> deltapts1 -------+
26761 @subsection Examples
26765 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26766 on top of the input labelled "in":
26768 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26769 [in] setpts=PTS-STARTPTS [main];
26770 [main][over] overlay=16:16 [out]
26774 Read from a video4linux2 device, and overlay it on top of the input
26777 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26778 [in] setpts=PTS-STARTPTS [main];
26779 [main][over] overlay=16:16 [out]
26783 Read the first video stream and the audio stream with id 0x81 from
26784 dvd.vob; the video is connected to the pad named "video" and the audio is
26785 connected to the pad named "audio":
26787 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26791 @subsection Commands
26793 Both movie and amovie support the following commands:
26796 Perform seek using "av_seek_frame".
26797 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26800 @var{stream_index}: If stream_index is -1, a default
26801 stream is selected, and @var{timestamp} is automatically converted
26802 from AV_TIME_BASE units to the stream specific time_base.
26804 @var{timestamp}: Timestamp in AVStream.time_base units
26805 or, if no stream is specified, in AV_TIME_BASE units.
26807 @var{flags}: Flags which select direction and seeking mode.
26811 Get movie duration in AV_TIME_BASE units.
26815 @c man end MULTIMEDIA SOURCES