1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Examples
1093 Fade in first 15 seconds of audio:
1095 afade=t=in:ss=0:d=15
1099 Fade out last 25 seconds of a 900 seconds audio:
1101 afade=t=out:st=875:d=25
1106 Denoise audio samples with FFT.
1108 A description of the accepted parameters follows.
1112 Set the noise reduction in dB, allowed range is 0.01 to 97.
1113 Default value is 12 dB.
1116 Set the noise floor in dB, allowed range is -80 to -20.
1117 Default value is -50 dB.
1122 It accepts the following values:
1131 Select shellac noise.
1134 Select custom noise, defined in @code{bn} option.
1136 Default value is white noise.
1140 Set custom band noise for every one of 15 bands.
1141 Bands are separated by ' ' or '|'.
1144 Set the residual floor in dB, allowed range is -80 to -20.
1145 Default value is -38 dB.
1148 Enable noise tracking. By default is disabled.
1149 With this enabled, noise floor is automatically adjusted.
1152 Enable residual tracking. By default is disabled.
1155 Set the output mode.
1157 It accepts the following values:
1160 Pass input unchanged.
1163 Pass noise filtered out.
1168 Default value is @var{o}.
1172 @subsection Commands
1174 This filter supports the following commands:
1176 @item sample_noise, sn
1177 Start or stop measuring noise profile.
1178 Syntax for the command is : "start" or "stop" string.
1179 After measuring noise profile is stopped it will be
1180 automatically applied in filtering.
1182 @item noise_reduction, nr
1183 Change noise reduction. Argument is single float number.
1184 Syntax for the command is : "@var{noise_reduction}"
1186 @item noise_floor, nf
1187 Change noise floor. Argument is single float number.
1188 Syntax for the command is : "@var{noise_floor}"
1190 @item output_mode, om
1191 Change output mode operation.
1192 Syntax for the command is : "i", "o" or "n" string.
1196 Apply arbitrary expressions to samples in frequency domain.
1200 Set frequency domain real expression for each separate channel separated
1201 by '|'. Default is "re".
1202 If the number of input channels is greater than the number of
1203 expressions, the last specified expression is used for the remaining
1207 Set frequency domain imaginary expression for each separate channel
1208 separated by '|'. Default is "im".
1210 Each expression in @var{real} and @var{imag} can contain the following
1211 constants and functions:
1218 current frequency bin number
1221 number of available bins
1224 channel number of the current expression
1233 current real part of frequency bin of current channel
1236 current imaginary part of frequency bin of current channel
1239 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1242 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1246 Set window size. Allowed range is from 16 to 131072.
1247 Default is @code{4096}
1250 Set window function. Default is @code{hann}.
1253 Set window overlap. If set to 1, the recommended overlap for selected
1254 window function will be picked. Default is @code{0.75}.
1257 @subsection Examples
1261 Leave almost only low frequencies in audio:
1263 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1267 Apply robotize effect:
1269 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1273 Apply whisper effect:
1275 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1282 Apply an arbitrary Finite Impulse Response filter.
1284 This filter is designed for applying long FIR filters,
1285 up to 60 seconds long.
1287 It can be used as component for digital crossover filters,
1288 room equalization, cross talk cancellation, wavefield synthesis,
1289 auralization, ambiophonics, ambisonics and spatialization.
1291 This filter uses the streams higher than first one as FIR coefficients.
1292 If the non-first stream holds a single channel, it will be used
1293 for all input channels in the first stream, otherwise
1294 the number of channels in the non-first stream must be same as
1295 the number of channels in the first stream.
1297 It accepts the following parameters:
1301 Set dry gain. This sets input gain.
1304 Set wet gain. This sets final output gain.
1307 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1310 Enable applying gain measured from power of IR.
1312 Set which approach to use for auto gain measurement.
1316 Do not apply any gain.
1319 select peak gain, very conservative approach. This is default value.
1322 select DC gain, limited application.
1325 select gain to noise approach, this is most popular one.
1329 Set gain to be applied to IR coefficients before filtering.
1330 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1333 Set format of IR stream. Can be @code{mono} or @code{input}.
1334 Default is @code{input}.
1337 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1338 Allowed range is 0.1 to 60 seconds.
1341 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1342 By default it is disabled.
1345 Set for which IR channel to display frequency response. By default is first channel
1346 displayed. This option is used only when @var{response} is enabled.
1349 Set video stream size. This option is used only when @var{response} is enabled.
1352 Set video stream frame rate. This option is used only when @var{response} is enabled.
1355 Set minimal partition size used for convolution. Default is @var{8192}.
1356 Allowed range is from @var{1} to @var{32768}.
1357 Lower values decreases latency at cost of higher CPU usage.
1360 Set maximal partition size used for convolution. Default is @var{8192}.
1361 Allowed range is from @var{8} to @var{32768}.
1362 Lower values may increase CPU usage.
1365 Set number of input impulse responses streams which will be switchable at runtime.
1366 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1369 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1370 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1371 This option can be changed at runtime via @ref{commands}.
1374 @subsection Examples
1378 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1380 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1387 Set output format constraints for the input audio. The framework will
1388 negotiate the most appropriate format to minimize conversions.
1390 It accepts the following parameters:
1393 @item sample_fmts, f
1394 A '|'-separated list of requested sample formats.
1396 @item sample_rates, r
1397 A '|'-separated list of requested sample rates.
1399 @item channel_layouts, cl
1400 A '|'-separated list of requested channel layouts.
1402 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1403 for the required syntax.
1406 If a parameter is omitted, all values are allowed.
1408 Force the output to either unsigned 8-bit or signed 16-bit stereo
1410 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1414 Apply frequency shift to input audio samples.
1416 The filter accepts the following options:
1420 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1421 Default value is 0.0.
1424 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1425 Default value is 1.0.
1428 @subsection Commands
1430 This filter supports the all above options as @ref{commands}.
1434 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1435 processing reduces disturbing noise between useful signals.
1437 Gating is done by detecting the volume below a chosen level @var{threshold}
1438 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1439 floor is set via @var{range}. Because an exact manipulation of the signal
1440 would cause distortion of the waveform the reduction can be levelled over
1441 time. This is done by setting @var{attack} and @var{release}.
1443 @var{attack} determines how long the signal has to fall below the threshold
1444 before any reduction will occur and @var{release} sets the time the signal
1445 has to rise above the threshold to reduce the reduction again.
1446 Shorter signals than the chosen attack time will be left untouched.
1450 Set input level before filtering.
1451 Default is 1. Allowed range is from 0.015625 to 64.
1454 Set the mode of operation. Can be @code{upward} or @code{downward}.
1455 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1456 will be amplified, expanding dynamic range in upward direction.
1457 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1460 Set the level of gain reduction when the signal is below the threshold.
1461 Default is 0.06125. Allowed range is from 0 to 1.
1462 Setting this to 0 disables reduction and then filter behaves like expander.
1465 If a signal rises above this level the gain reduction is released.
1466 Default is 0.125. Allowed range is from 0 to 1.
1469 Set a ratio by which the signal is reduced.
1470 Default is 2. Allowed range is from 1 to 9000.
1473 Amount of milliseconds the signal has to rise above the threshold before gain
1475 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1478 Amount of milliseconds the signal has to fall below the threshold before the
1479 reduction is increased again. Default is 250 milliseconds.
1480 Allowed range is from 0.01 to 9000.
1483 Set amount of amplification of signal after processing.
1484 Default is 1. Allowed range is from 1 to 64.
1487 Curve the sharp knee around the threshold to enter gain reduction more softly.
1488 Default is 2.828427125. Allowed range is from 1 to 8.
1491 Choose if exact signal should be taken for detection or an RMS like one.
1492 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1495 Choose if the average level between all channels or the louder channel affects
1497 Default is @code{average}. Can be @code{average} or @code{maximum}.
1500 @subsection Commands
1502 This filter supports the all above options as @ref{commands}.
1506 Apply an arbitrary Infinite Impulse Response filter.
1508 It accepts the following parameters:
1512 Set B/numerator/zeros/reflection coefficients.
1515 Set A/denominator/poles/ladder coefficients.
1527 Set coefficients format.
1531 lattice-ladder function
1533 analog transfer function
1535 digital transfer function
1537 Z-plane zeros/poles, cartesian (default)
1539 Z-plane zeros/poles, polar radians
1541 Z-plane zeros/poles, polar degrees
1547 Set type of processing.
1559 Set filtering precision.
1563 double-precision floating-point (default)
1565 single-precision floating-point
1573 Normalize filter coefficients, by default is enabled.
1574 Enabling it will normalize magnitude response at DC to 0dB.
1577 How much to use filtered signal in output. Default is 1.
1578 Range is between 0 and 1.
1581 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1582 By default it is disabled.
1585 Set for which IR channel to display frequency response. By default is first channel
1586 displayed. This option is used only when @var{response} is enabled.
1589 Set video stream size. This option is used only when @var{response} is enabled.
1592 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1595 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1596 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1599 Different coefficients and gains can be provided for every channel, in such case
1600 use '|' to separate coefficients or gains. Last provided coefficients will be
1601 used for all remaining channels.
1603 @subsection Examples
1607 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1609 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1613 Same as above but in @code{zp} format:
1615 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1619 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1621 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1627 The limiter prevents an input signal from rising over a desired threshold.
1628 This limiter uses lookahead technology to prevent your signal from distorting.
1629 It means that there is a small delay after the signal is processed. Keep in mind
1630 that the delay it produces is the attack time you set.
1632 The filter accepts the following options:
1636 Set input gain. Default is 1.
1639 Set output gain. Default is 1.
1642 Don't let signals above this level pass the limiter. Default is 1.
1645 The limiter will reach its attenuation level in this amount of time in
1646 milliseconds. Default is 5 milliseconds.
1649 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1650 Default is 50 milliseconds.
1653 When gain reduction is always needed ASC takes care of releasing to an
1654 average reduction level rather than reaching a reduction of 0 in the release
1658 Select how much the release time is affected by ASC, 0 means nearly no changes
1659 in release time while 1 produces higher release times.
1662 Auto level output signal. Default is enabled.
1663 This normalizes audio back to 0dB if enabled.
1666 Depending on picked setting it is recommended to upsample input 2x or 4x times
1667 with @ref{aresample} before applying this filter.
1671 Apply a two-pole all-pass filter with central frequency (in Hz)
1672 @var{frequency}, and filter-width @var{width}.
1673 An all-pass filter changes the audio's frequency to phase relationship
1674 without changing its frequency to amplitude relationship.
1676 The filter accepts the following options:
1680 Set frequency in Hz.
1683 Set method to specify band-width of filter.
1698 Specify the band-width of a filter in width_type units.
1701 How much to use filtered signal in output. Default is 1.
1702 Range is between 0 and 1.
1705 Specify which channels to filter, by default all available are filtered.
1708 Normalize biquad coefficients, by default is disabled.
1709 Enabling it will normalize magnitude response at DC to 0dB.
1712 Set the filter order, can be 1 or 2. Default is 2.
1715 Set transform type of IIR filter.
1724 Set precison of filtering.
1727 Pick automatic sample format depending on surround filters.
1729 Always use signed 16-bit.
1731 Always use signed 32-bit.
1733 Always use float 32-bit.
1735 Always use float 64-bit.
1739 @subsection Commands
1741 This filter supports the following commands:
1744 Change allpass frequency.
1745 Syntax for the command is : "@var{frequency}"
1748 Change allpass width_type.
1749 Syntax for the command is : "@var{width_type}"
1752 Change allpass width.
1753 Syntax for the command is : "@var{width}"
1757 Syntax for the command is : "@var{mix}"
1764 The filter accepts the following options:
1768 Set the number of loops. Setting this value to -1 will result in infinite loops.
1772 Set maximal number of samples. Default is 0.
1775 Set first sample of loop. Default is 0.
1781 Merge two or more audio streams into a single multi-channel stream.
1783 The filter accepts the following options:
1788 Set the number of inputs. Default is 2.
1792 If the channel layouts of the inputs are disjoint, and therefore compatible,
1793 the channel layout of the output will be set accordingly and the channels
1794 will be reordered as necessary. If the channel layouts of the inputs are not
1795 disjoint, the output will have all the channels of the first input then all
1796 the channels of the second input, in that order, and the channel layout of
1797 the output will be the default value corresponding to the total number of
1800 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1801 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1802 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1803 first input, b1 is the first channel of the second input).
1805 On the other hand, if both input are in stereo, the output channels will be
1806 in the default order: a1, a2, b1, b2, and the channel layout will be
1807 arbitrarily set to 4.0, which may or may not be the expected value.
1809 All inputs must have the same sample rate, and format.
1811 If inputs do not have the same duration, the output will stop with the
1814 @subsection Examples
1818 Merge two mono files into a stereo stream:
1820 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1824 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1826 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1832 Mixes multiple audio inputs into a single output.
1834 Note that this filter only supports float samples (the @var{amerge}
1835 and @var{pan} audio filters support many formats). If the @var{amix}
1836 input has integer samples then @ref{aresample} will be automatically
1837 inserted to perform the conversion to float samples.
1841 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1843 will mix 3 input audio streams to a single output with the same duration as the
1844 first input and a dropout transition time of 3 seconds.
1846 It accepts the following parameters:
1850 The number of inputs. If unspecified, it defaults to 2.
1853 How to determine the end-of-stream.
1857 The duration of the longest input. (default)
1860 The duration of the shortest input.
1863 The duration of the first input.
1867 @item dropout_transition
1868 The transition time, in seconds, for volume renormalization when an input
1869 stream ends. The default value is 2 seconds.
1872 Specify weight of each input audio stream as sequence.
1873 Each weight is separated by space. By default all inputs have same weight.
1876 @subsection Commands
1878 This filter supports the following commands:
1881 Syntax is same as option with same name.
1886 Multiply first audio stream with second audio stream and store result
1887 in output audio stream. Multiplication is done by multiplying each
1888 sample from first stream with sample at same position from second stream.
1890 With this element-wise multiplication one can create amplitude fades and
1891 amplitude modulations.
1893 @section anequalizer
1895 High-order parametric multiband equalizer for each channel.
1897 It accepts the following parameters:
1901 This option string is in format:
1902 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1903 Each equalizer band is separated by '|'.
1907 Set channel number to which equalization will be applied.
1908 If input doesn't have that channel the entry is ignored.
1911 Set central frequency for band.
1912 If input doesn't have that frequency the entry is ignored.
1915 Set band width in Hertz.
1918 Set band gain in dB.
1921 Set filter type for band, optional, can be:
1925 Butterworth, this is default.
1936 With this option activated frequency response of anequalizer is displayed
1940 Set video stream size. Only useful if curves option is activated.
1943 Set max gain that will be displayed. Only useful if curves option is activated.
1944 Setting this to a reasonable value makes it possible to display gain which is derived from
1945 neighbour bands which are too close to each other and thus produce higher gain
1946 when both are activated.
1949 Set frequency scale used to draw frequency response in video output.
1950 Can be linear or logarithmic. Default is logarithmic.
1953 Set color for each channel curve which is going to be displayed in video stream.
1954 This is list of color names separated by space or by '|'.
1955 Unrecognised or missing colors will be replaced by white color.
1958 @subsection Examples
1962 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1963 for first 2 channels using Chebyshev type 1 filter:
1965 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1969 @subsection Commands
1971 This filter supports the following commands:
1974 Alter existing filter parameters.
1975 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1977 @var{fN} is existing filter number, starting from 0, if no such filter is available
1979 @var{freq} set new frequency parameter.
1980 @var{width} set new width parameter in Hertz.
1981 @var{gain} set new gain parameter in dB.
1983 Full filter invocation with asendcmd may look like this:
1984 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1989 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1991 Each sample is adjusted by looking for other samples with similar contexts. This
1992 context similarity is defined by comparing their surrounding patches of size
1993 @option{p}. Patches are searched in an area of @option{r} around the sample.
1995 The filter accepts the following options:
1999 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2002 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2003 Default value is 2 milliseconds.
2006 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2007 Default value is 6 milliseconds.
2010 Set the output mode.
2012 It accepts the following values:
2015 Pass input unchanged.
2018 Pass noise filtered out.
2023 Default value is @var{o}.
2027 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2030 @subsection Commands
2032 This filter supports the all above options as @ref{commands}.
2035 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2037 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2038 relate to producing the least mean square of the error signal (difference between the desired,
2039 2nd input audio stream and the actual signal, the 1st input audio stream).
2041 A description of the accepted options follows.
2054 Set the filter leakage.
2057 It accepts the following values:
2066 Pass filtered samples.
2069 Pass difference between desired and filtered samples.
2071 Default value is @var{o}.
2075 @subsection Examples
2079 One of many usages of this filter is noise reduction, input audio is filtered
2080 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2082 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2086 @subsection Commands
2088 This filter supports the same commands as options, excluding option @code{order}.
2092 Pass the audio source unchanged to the output.
2096 Pad the end of an audio stream with silence.
2098 This can be used together with @command{ffmpeg} @option{-shortest} to
2099 extend audio streams to the same length as the video stream.
2101 A description of the accepted options follows.
2105 Set silence packet size. Default value is 4096.
2108 Set the number of samples of silence to add to the end. After the
2109 value is reached, the stream is terminated. This option is mutually
2110 exclusive with @option{whole_len}.
2113 Set the minimum total number of samples in the output audio stream. If
2114 the value is longer than the input audio length, silence is added to
2115 the end, until the value is reached. This option is mutually exclusive
2116 with @option{pad_len}.
2119 Specify the duration of samples of silence to add. See
2120 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2121 for the accepted syntax. Used only if set to non-zero value.
2124 Specify the minimum total duration in the output audio stream. See
2125 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2126 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2127 the input audio length, silence is added to the end, until the value is reached.
2128 This option is mutually exclusive with @option{pad_dur}
2131 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2132 nor @option{whole_dur} option is set, the filter will add silence to the end of
2133 the input stream indefinitely.
2135 @subsection Examples
2139 Add 1024 samples of silence to the end of the input:
2145 Make sure the audio output will contain at least 10000 samples, pad
2146 the input with silence if required:
2148 apad=whole_len=10000
2152 Use @command{ffmpeg} to pad the audio input with silence, so that the
2153 video stream will always result the shortest and will be converted
2154 until the end in the output file when using the @option{shortest}
2157 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2162 Add a phasing effect to the input audio.
2164 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2165 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2167 A description of the accepted parameters follows.
2171 Set input gain. Default is 0.4.
2174 Set output gain. Default is 0.74
2177 Set delay in milliseconds. Default is 3.0.
2180 Set decay. Default is 0.4.
2183 Set modulation speed in Hz. Default is 0.5.
2186 Set modulation type. Default is triangular.
2188 It accepts the following values:
2195 @section aphaseshift
2196 Apply phase shift to input audio samples.
2198 The filter accepts the following options:
2202 Specify phase shift. Allowed range is from -1.0 to 1.0.
2203 Default value is 0.0.
2206 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2207 Default value is 1.0.
2210 @subsection Commands
2212 This filter supports the all above options as @ref{commands}.
2216 Audio pulsator is something between an autopanner and a tremolo.
2217 But it can produce funny stereo effects as well. Pulsator changes the volume
2218 of the left and right channel based on a LFO (low frequency oscillator) with
2219 different waveforms and shifted phases.
2220 This filter have the ability to define an offset between left and right
2221 channel. An offset of 0 means that both LFO shapes match each other.
2222 The left and right channel are altered equally - a conventional tremolo.
2223 An offset of 50% means that the shape of the right channel is exactly shifted
2224 in phase (or moved backwards about half of the frequency) - pulsator acts as
2225 an autopanner. At 1 both curves match again. Every setting in between moves the
2226 phase shift gapless between all stages and produces some "bypassing" sounds with
2227 sine and triangle waveforms. The more you set the offset near 1 (starting from
2228 the 0.5) the faster the signal passes from the left to the right speaker.
2230 The filter accepts the following options:
2234 Set input gain. By default it is 1. Range is [0.015625 - 64].
2237 Set output gain. By default it is 1. Range is [0.015625 - 64].
2240 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2241 sawup or sawdown. Default is sine.
2244 Set modulation. Define how much of original signal is affected by the LFO.
2247 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2250 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2253 Set pulse width. Default is 1. Allowed range is [0 - 2].
2256 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2259 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2263 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2267 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2268 if timing is set to hz.
2274 Resample the input audio to the specified parameters, using the
2275 libswresample library. If none are specified then the filter will
2276 automatically convert between its input and output.
2278 This filter is also able to stretch/squeeze the audio data to make it match
2279 the timestamps or to inject silence / cut out audio to make it match the
2280 timestamps, do a combination of both or do neither.
2282 The filter accepts the syntax
2283 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2284 expresses a sample rate and @var{resampler_options} is a list of
2285 @var{key}=@var{value} pairs, separated by ":". See the
2286 @ref{Resampler Options,,"Resampler Options" section in the
2287 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2288 for the complete list of supported options.
2290 @subsection Examples
2294 Resample the input audio to 44100Hz:
2300 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2301 samples per second compensation:
2303 aresample=async=1000
2309 Reverse an audio clip.
2311 Warning: This filter requires memory to buffer the entire clip, so trimming
2314 @subsection Examples
2318 Take the first 5 seconds of a clip, and reverse it.
2320 atrim=end=5,areverse
2326 Reduce noise from speech using Recurrent Neural Networks.
2328 This filter accepts the following options:
2332 Set train model file to load. This option is always required.
2335 Set how much to mix filtered samples into final output.
2336 Allowed range is from -1 to 1. Default value is 1.
2337 Negative values are special, they set how much to keep filtered noise
2338 in the final filter output. Set this option to -1 to hear actual
2339 noise removed from input signal.
2342 @section asetnsamples
2344 Set the number of samples per each output audio frame.
2346 The last output packet may contain a different number of samples, as
2347 the filter will flush all the remaining samples when the input audio
2350 The filter accepts the following options:
2354 @item nb_out_samples, n
2355 Set the number of frames per each output audio frame. The number is
2356 intended as the number of samples @emph{per each channel}.
2357 Default value is 1024.
2360 If set to 1, the filter will pad the last audio frame with zeroes, so
2361 that the last frame will contain the same number of samples as the
2362 previous ones. Default value is 1.
2365 For example, to set the number of per-frame samples to 1234 and
2366 disable padding for the last frame, use:
2368 asetnsamples=n=1234:p=0
2373 Set the sample rate without altering the PCM data.
2374 This will result in a change of speed and pitch.
2376 The filter accepts the following options:
2379 @item sample_rate, r
2380 Set the output sample rate. Default is 44100 Hz.
2385 Show a line containing various information for each input audio frame.
2386 The input audio is not modified.
2388 The shown line contains a sequence of key/value pairs of the form
2389 @var{key}:@var{value}.
2391 The following values are shown in the output:
2395 The (sequential) number of the input frame, starting from 0.
2398 The presentation timestamp of the input frame, in time base units; the time base
2399 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2402 The presentation timestamp of the input frame in seconds.
2405 position of the frame in the input stream, -1 if this information in
2406 unavailable and/or meaningless (for example in case of synthetic audio)
2415 The sample rate for the audio frame.
2418 The number of samples (per channel) in the frame.
2421 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2422 audio, the data is treated as if all the planes were concatenated.
2424 @item plane_checksums
2425 A list of Adler-32 checksums for each data plane.
2429 Apply audio soft clipping.
2431 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2432 along a smooth curve, rather than the abrupt shape of hard-clipping.
2434 This filter accepts the following options:
2438 Set type of soft-clipping.
2440 It accepts the following values:
2454 Set additional parameter which controls sigmoid function.
2457 Set oversampling factor.
2460 @subsection Commands
2462 This filter supports the all above options as @ref{commands}.
2465 Automatic Speech Recognition
2467 This filter uses PocketSphinx for speech recognition. To enable
2468 compilation of this filter, you need to configure FFmpeg with
2469 @code{--enable-pocketsphinx}.
2471 It accepts the following options:
2475 Set sampling rate of input audio. Defaults is @code{16000}.
2476 This need to match speech models, otherwise one will get poor results.
2479 Set dictionary containing acoustic model files.
2482 Set pronunciation dictionary.
2485 Set language model file.
2488 Set language model set.
2491 Set which language model to use.
2494 Set output for log messages.
2497 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2502 Display time domain statistical information about the audio channels.
2503 Statistics are calculated and displayed for each audio channel and,
2504 where applicable, an overall figure is also given.
2506 It accepts the following option:
2509 Short window length in seconds, used for peak and trough RMS measurement.
2510 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2514 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2515 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2518 Available keys for each channel are:
2564 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2565 this @code{lavfi.astats.Overall.Peak_count}.
2567 For description what each key means read below.
2570 Set number of frame after which stats are going to be recalculated.
2571 Default is disabled.
2573 @item measure_perchannel
2574 Select the entries which need to be measured per channel. The metadata keys can
2575 be used as flags, default is @option{all} which measures everything.
2576 @option{none} disables all per channel measurement.
2578 @item measure_overall
2579 Select the entries which need to be measured overall. The metadata keys can
2580 be used as flags, default is @option{all} which measures everything.
2581 @option{none} disables all overall measurement.
2585 A description of each shown parameter follows:
2589 Mean amplitude displacement from zero.
2592 Minimal sample level.
2595 Maximal sample level.
2597 @item Min difference
2598 Minimal difference between two consecutive samples.
2600 @item Max difference
2601 Maximal difference between two consecutive samples.
2603 @item Mean difference
2604 Mean difference between two consecutive samples.
2605 The average of each difference between two consecutive samples.
2607 @item RMS difference
2608 Root Mean Square difference between two consecutive samples.
2612 Standard peak and RMS level measured in dBFS.
2616 Peak and trough values for RMS level measured over a short window.
2619 Standard ratio of peak to RMS level (note: not in dB).
2622 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2623 (i.e. either @var{Min level} or @var{Max level}).
2626 Number of occasions (not the number of samples) that the signal attained either
2627 @var{Min level} or @var{Max level}.
2629 @item Noise floor dB
2630 Minimum local peak measured in dBFS over a short window.
2632 @item Noise floor count
2633 Number of occasions (not the number of samples) that the signal attained
2637 Overall bit depth of audio. Number of bits used for each sample.
2640 Measured dynamic range of audio in dB.
2642 @item Zero crossings
2643 Number of points where the waveform crosses the zero level axis.
2645 @item Zero crossings rate
2646 Rate of Zero crossings and number of audio samples.
2650 Boost subwoofer frequencies.
2652 The filter accepts the following options:
2656 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2657 Default value is 0.7.
2660 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2661 Default value is 0.7.
2664 Set delay line decay gain value. Allowed range is from 0 to 1.
2665 Default value is 0.7.
2668 Set delay line feedback gain value. Allowed range is from 0 to 1.
2669 Default value is 0.9.
2672 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2673 Default value is 100.
2676 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2677 Default value is 0.5.
2680 Set delay. Allowed range is from 1 to 100.
2681 Default value is 20.
2684 @subsection Commands
2686 This filter supports the all above options as @ref{commands}.
2689 Cut subwoofer frequencies.
2691 This filter allows to set custom, steeper
2692 roll off than highpass filter, and thus is able to more attenuate
2693 frequency content in stop-band.
2695 The filter accepts the following options:
2699 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2700 Default value is 20.
2703 Set filter order. Available values are from 3 to 20.
2704 Default value is 10.
2707 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2710 @subsection Commands
2712 This filter supports the all above options as @ref{commands}.
2715 Cut super frequencies.
2717 The filter accepts the following options:
2721 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2722 Default value is 20000.
2725 Set filter order. Available values are from 3 to 20.
2726 Default value is 10.
2729 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2732 @subsection Commands
2734 This filter supports the all above options as @ref{commands}.
2740 The filter accepts exactly one parameter, the audio tempo. If not
2741 specified then the filter will assume nominal 1.0 tempo. Tempo must
2742 be in the [0.5, 100.0] range.
2744 Note that tempo greater than 2 will skip some samples rather than
2745 blend them in. If for any reason this is a concern it is always
2746 possible to daisy-chain several instances of atempo to achieve the
2747 desired product tempo.
2749 @subsection Examples
2753 Slow down audio to 80% tempo:
2759 To speed up audio to 300% tempo:
2765 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2767 atempo=sqrt(3),atempo=sqrt(3)
2771 @subsection Commands
2773 This filter supports the following commands:
2776 Change filter tempo scale factor.
2777 Syntax for the command is : "@var{tempo}"
2782 Trim the input so that the output contains one continuous subpart of the input.
2784 It accepts the following parameters:
2787 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2788 sample with the timestamp @var{start} will be the first sample in the output.
2791 Specify time of the first audio sample that will be dropped, i.e. the
2792 audio sample immediately preceding the one with the timestamp @var{end} will be
2793 the last sample in the output.
2796 Same as @var{start}, except this option sets the start timestamp in samples
2800 Same as @var{end}, except this option sets the end timestamp in samples instead
2804 The maximum duration of the output in seconds.
2807 The number of the first sample that should be output.
2810 The number of the first sample that should be dropped.
2813 @option{start}, @option{end}, and @option{duration} are expressed as time
2814 duration specifications; see
2815 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2817 Note that the first two sets of the start/end options and the @option{duration}
2818 option look at the frame timestamp, while the _sample options simply count the
2819 samples that pass through the filter. So start/end_pts and start/end_sample will
2820 give different results when the timestamps are wrong, inexact or do not start at
2821 zero. Also note that this filter does not modify the timestamps. If you wish
2822 to have the output timestamps start at zero, insert the asetpts filter after the
2825 If multiple start or end options are set, this filter tries to be greedy and
2826 keep all samples that match at least one of the specified constraints. To keep
2827 only the part that matches all the constraints at once, chain multiple atrim
2830 The defaults are such that all the input is kept. So it is possible to set e.g.
2831 just the end values to keep everything before the specified time.
2836 Drop everything except the second minute of input:
2838 ffmpeg -i INPUT -af atrim=60:120
2842 Keep only the first 1000 samples:
2844 ffmpeg -i INPUT -af atrim=end_sample=1000
2849 @section axcorrelate
2850 Calculate normalized cross-correlation between two input audio streams.
2852 Resulted samples are always between -1 and 1 inclusive.
2853 If result is 1 it means two input samples are highly correlated in that selected segment.
2854 Result 0 means they are not correlated at all.
2855 If result is -1 it means two input samples are out of phase, which means they cancel each
2858 The filter accepts the following options:
2862 Set size of segment over which cross-correlation is calculated.
2863 Default is 256. Allowed range is from 2 to 131072.
2866 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2867 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2868 are always zero and thus need much less calculations to make.
2869 This is generally not true, but is valid for typical audio streams.
2872 @subsection Examples
2876 Calculate correlation between channels in stereo audio stream:
2878 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2884 Apply a two-pole Butterworth band-pass filter with central
2885 frequency @var{frequency}, and (3dB-point) band-width width.
2886 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2887 instead of the default: constant 0dB peak gain.
2888 The filter roll off at 6dB per octave (20dB per decade).
2890 The filter accepts the following options:
2894 Set the filter's central frequency. Default is @code{3000}.
2897 Constant skirt gain if set to 1. Defaults to 0.
2900 Set method to specify band-width of filter.
2915 Specify the band-width of a filter in width_type units.
2918 How much to use filtered signal in output. Default is 1.
2919 Range is between 0 and 1.
2922 Specify which channels to filter, by default all available are filtered.
2925 Normalize biquad coefficients, by default is disabled.
2926 Enabling it will normalize magnitude response at DC to 0dB.
2929 Set transform type of IIR filter.
2938 Set precison of filtering.
2941 Pick automatic sample format depending on surround filters.
2943 Always use signed 16-bit.
2945 Always use signed 32-bit.
2947 Always use float 32-bit.
2949 Always use float 64-bit.
2953 @subsection Commands
2955 This filter supports the following commands:
2958 Change bandpass frequency.
2959 Syntax for the command is : "@var{frequency}"
2962 Change bandpass width_type.
2963 Syntax for the command is : "@var{width_type}"
2966 Change bandpass width.
2967 Syntax for the command is : "@var{width}"
2970 Change bandpass mix.
2971 Syntax for the command is : "@var{mix}"
2976 Apply a two-pole Butterworth band-reject filter with central
2977 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2978 The filter roll off at 6dB per octave (20dB per decade).
2980 The filter accepts the following options:
2984 Set the filter's central frequency. Default is @code{3000}.
2987 Set method to specify band-width of filter.
3002 Specify the band-width of a filter in width_type units.
3005 How much to use filtered signal in output. Default is 1.
3006 Range is between 0 and 1.
3009 Specify which channels to filter, by default all available are filtered.
3012 Normalize biquad coefficients, by default is disabled.
3013 Enabling it will normalize magnitude response at DC to 0dB.
3016 Set transform type of IIR filter.
3025 Set precison of filtering.
3028 Pick automatic sample format depending on surround filters.
3030 Always use signed 16-bit.
3032 Always use signed 32-bit.
3034 Always use float 32-bit.
3036 Always use float 64-bit.
3040 @subsection Commands
3042 This filter supports the following commands:
3045 Change bandreject frequency.
3046 Syntax for the command is : "@var{frequency}"
3049 Change bandreject width_type.
3050 Syntax for the command is : "@var{width_type}"
3053 Change bandreject width.
3054 Syntax for the command is : "@var{width}"
3057 Change bandreject mix.
3058 Syntax for the command is : "@var{mix}"
3061 @section bass, lowshelf
3063 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3064 shelving filter with a response similar to that of a standard
3065 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3067 The filter accepts the following options:
3071 Give the gain at 0 Hz. Its useful range is about -20
3072 (for a large cut) to +20 (for a large boost).
3073 Beware of clipping when using a positive gain.
3076 Set the filter's central frequency and so can be used
3077 to extend or reduce the frequency range to be boosted or cut.
3078 The default value is @code{100} Hz.
3081 Set method to specify band-width of filter.
3096 Determine how steep is the filter's shelf transition.
3099 How much to use filtered signal in output. Default is 1.
3100 Range is between 0 and 1.
3103 Specify which channels to filter, by default all available are filtered.
3106 Normalize biquad coefficients, by default is disabled.
3107 Enabling it will normalize magnitude response at DC to 0dB.
3110 Set transform type of IIR filter.
3119 Set precison of filtering.
3122 Pick automatic sample format depending on surround filters.
3124 Always use signed 16-bit.
3126 Always use signed 32-bit.
3128 Always use float 32-bit.
3130 Always use float 64-bit.
3134 @subsection Commands
3136 This filter supports the following commands:
3139 Change bass frequency.
3140 Syntax for the command is : "@var{frequency}"
3143 Change bass width_type.
3144 Syntax for the command is : "@var{width_type}"
3148 Syntax for the command is : "@var{width}"
3152 Syntax for the command is : "@var{gain}"
3156 Syntax for the command is : "@var{mix}"
3161 Apply a biquad IIR filter with the given coefficients.
3162 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3163 are the numerator and denominator coefficients respectively.
3164 and @var{channels}, @var{c} specify which channels to filter, by default all
3165 available are filtered.
3167 @subsection Commands
3169 This filter supports the following commands:
3177 Change biquad parameter.
3178 Syntax for the command is : "@var{value}"
3181 How much to use filtered signal in output. Default is 1.
3182 Range is between 0 and 1.
3185 Specify which channels to filter, by default all available are filtered.
3188 Normalize biquad coefficients, by default is disabled.
3189 Enabling it will normalize magnitude response at DC to 0dB.
3192 Set transform type of IIR filter.
3201 Set precison of filtering.
3204 Pick automatic sample format depending on surround filters.
3206 Always use signed 16-bit.
3208 Always use signed 32-bit.
3210 Always use float 32-bit.
3212 Always use float 64-bit.
3217 Bauer stereo to binaural transformation, which improves headphone listening of
3218 stereo audio records.
3220 To enable compilation of this filter you need to configure FFmpeg with
3221 @code{--enable-libbs2b}.
3223 It accepts the following parameters:
3227 Pre-defined crossfeed level.
3231 Default level (fcut=700, feed=50).
3234 Chu Moy circuit (fcut=700, feed=60).
3237 Jan Meier circuit (fcut=650, feed=95).
3242 Cut frequency (in Hz).
3251 Remap input channels to new locations.
3253 It accepts the following parameters:
3256 Map channels from input to output. The argument is a '|'-separated list of
3257 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3258 @var{in_channel} form. @var{in_channel} can be either the name of the input
3259 channel (e.g. FL for front left) or its index in the input channel layout.
3260 @var{out_channel} is the name of the output channel or its index in the output
3261 channel layout. If @var{out_channel} is not given then it is implicitly an
3262 index, starting with zero and increasing by one for each mapping.
3264 @item channel_layout
3265 The channel layout of the output stream.
3268 If no mapping is present, the filter will implicitly map input channels to
3269 output channels, preserving indices.
3271 @subsection Examples
3275 For example, assuming a 5.1+downmix input MOV file,
3277 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3279 will create an output WAV file tagged as stereo from the downmix channels of
3283 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3285 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3289 @section channelsplit
3291 Split each channel from an input audio stream into a separate output stream.
3293 It accepts the following parameters:
3295 @item channel_layout
3296 The channel layout of the input stream. The default is "stereo".
3298 A channel layout describing the channels to be extracted as separate output streams
3299 or "all" to extract each input channel as a separate stream. The default is "all".
3301 Choosing channels not present in channel layout in the input will result in an error.
3304 @subsection Examples
3308 For example, assuming a stereo input MP3 file,
3310 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3312 will create an output Matroska file with two audio streams, one containing only
3313 the left channel and the other the right channel.
3316 Split a 5.1 WAV file into per-channel files:
3318 ffmpeg -i in.wav -filter_complex
3319 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3320 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3321 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3326 Extract only LFE from a 5.1 WAV file:
3328 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3329 -map '[LFE]' lfe.wav
3334 Add a chorus effect to the audio.
3336 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3338 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3339 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3340 The modulation depth defines the range the modulated delay is played before or after
3341 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3342 sound tuned around the original one, like in a chorus where some vocals are slightly
3345 It accepts the following parameters:
3348 Set input gain. Default is 0.4.
3351 Set output gain. Default is 0.4.
3354 Set delays. A typical delay is around 40ms to 60ms.
3366 @subsection Examples
3372 chorus=0.7:0.9:55:0.4:0.25:2
3378 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3382 Fuller sounding chorus with three delays:
3384 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
3389 Compress or expand the audio's dynamic range.
3391 It accepts the following parameters:
3397 A list of times in seconds for each channel over which the instantaneous level
3398 of the input signal is averaged to determine its volume. @var{attacks} refers to
3399 increase of volume and @var{decays} refers to decrease of volume. For most
3400 situations, the attack time (response to the audio getting louder) should be
3401 shorter than the decay time, because the human ear is more sensitive to sudden
3402 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3403 a typical value for decay is 0.8 seconds.
3404 If specified number of attacks & decays is lower than number of channels, the last
3405 set attack/decay will be used for all remaining channels.
3408 A list of points for the transfer function, specified in dB relative to the
3409 maximum possible signal amplitude. Each key points list must be defined using
3410 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3411 @code{x0/y0 x1/y1 x2/y2 ....}
3413 The input values must be in strictly increasing order but the transfer function
3414 does not have to be monotonically rising. The point @code{0/0} is assumed but
3415 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3416 function are @code{-70/-70|-60/-20|1/0}.
3419 Set the curve radius in dB for all joints. It defaults to 0.01.
3422 Set the additional gain in dB to be applied at all points on the transfer
3423 function. This allows for easy adjustment of the overall gain.
3427 Set an initial volume, in dB, to be assumed for each channel when filtering
3428 starts. This permits the user to supply a nominal level initially, so that, for
3429 example, a very large gain is not applied to initial signal levels before the
3430 companding has begun to operate. A typical value for audio which is initially
3431 quiet is -90 dB. It defaults to 0.
3434 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3435 delayed before being fed to the volume adjuster. Specifying a delay
3436 approximately equal to the attack/decay times allows the filter to effectively
3437 operate in predictive rather than reactive mode. It defaults to 0.
3441 @subsection Examples
3445 Make music with both quiet and loud passages suitable for listening to in a
3448 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3451 Another example for audio with whisper and explosion parts:
3453 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3457 A noise gate for when the noise is at a lower level than the signal:
3459 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3463 Here is another noise gate, this time for when the noise is at a higher level
3464 than the signal (making it, in some ways, similar to squelch):
3466 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3470 2:1 compression starting at -6dB:
3472 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3476 2:1 compression starting at -9dB:
3478 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3482 2:1 compression starting at -12dB:
3484 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3488 2:1 compression starting at -18dB:
3490 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3494 3:1 compression starting at -15dB:
3496 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3502 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3508 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
3512 Hard limiter at -6dB:
3514 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3518 Hard limiter at -12dB:
3520 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3524 Hard noise gate at -35 dB:
3526 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3532 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3536 @section compensationdelay
3538 Compensation Delay Line is a metric based delay to compensate differing
3539 positions of microphones or speakers.
3541 For example, you have recorded guitar with two microphones placed in
3542 different locations. Because the front of sound wave has fixed speed in
3543 normal conditions, the phasing of microphones can vary and depends on
3544 their location and interposition. The best sound mix can be achieved when
3545 these microphones are in phase (synchronized). Note that a distance of
3546 ~30 cm between microphones makes one microphone capture the signal in
3547 antiphase to the other microphone. That makes the final mix sound moody.
3548 This filter helps to solve phasing problems by adding different delays
3549 to each microphone track and make them synchronized.
3551 The best result can be reached when you take one track as base and
3552 synchronize other tracks one by one with it.
3553 Remember that synchronization/delay tolerance depends on sample rate, too.
3554 Higher sample rates will give more tolerance.
3556 The filter accepts the following parameters:
3560 Set millimeters distance. This is compensation distance for fine tuning.
3564 Set cm distance. This is compensation distance for tightening distance setup.
3568 Set meters distance. This is compensation distance for hard distance setup.
3572 Set dry amount. Amount of unprocessed (dry) signal.
3576 Set wet amount. Amount of processed (wet) signal.
3580 Set temperature in degrees Celsius. This is the temperature of the environment.
3585 Apply headphone crossfeed filter.
3587 Crossfeed is the process of blending the left and right channels of stereo
3589 It is mainly used to reduce extreme stereo separation of low frequencies.
3591 The intent is to produce more speaker like sound to the listener.
3593 The filter accepts the following options:
3597 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3598 This sets gain of low shelf filter for side part of stereo image.
3599 Default is -6dB. Max allowed is -30db when strength is set to 1.
3602 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3603 This sets cut off frequency of low shelf filter. Default is cut off near
3604 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3607 Set curve slope of low shelf filter. Default is 0.5.
3608 Allowed range is from 0.01 to 1.
3611 Set input gain. Default is 0.9.
3614 Set output gain. Default is 1.
3617 @subsection Commands
3619 This filter supports the all above options as @ref{commands}.
3621 @section crystalizer
3622 Simple algorithm to expand audio dynamic range.
3624 The filter accepts the following options:
3628 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3629 (unchanged sound) to 10.0 (maximum effect).
3632 Enable clipping. By default is enabled.
3635 @subsection Commands
3637 This filter supports the all above options as @ref{commands}.
3640 Apply a DC shift to the audio.
3642 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3643 in the recording chain) from the audio. The effect of a DC offset is reduced
3644 headroom and hence volume. The @ref{astats} filter can be used to determine if
3645 a signal has a DC offset.
3649 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3653 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3654 used to prevent clipping.
3659 Apply de-essing to the audio samples.
3663 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3667 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3671 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3675 Set the output mode.
3677 It accepts the following values:
3680 Pass input unchanged.
3683 Pass ess filtered out.
3688 Default value is @var{o}.
3694 Measure audio dynamic range.
3696 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3697 is found in transition material. And anything less that 8 have very poor dynamics
3698 and is very compressed.
3700 The filter accepts the following options:
3704 Set window length in seconds used to split audio into segments of equal length.
3705 Default is 3 seconds.
3709 Dynamic Audio Normalizer.
3711 This filter applies a certain amount of gain to the input audio in order
3712 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3713 contrast to more "simple" normalization algorithms, the Dynamic Audio
3714 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3715 This allows for applying extra gain to the "quiet" sections of the audio
3716 while avoiding distortions or clipping the "loud" sections. In other words:
3717 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3718 sections, in the sense that the volume of each section is brought to the
3719 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3720 this goal *without* applying "dynamic range compressing". It will retain 100%
3721 of the dynamic range *within* each section of the audio file.
3725 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3726 Default is 500 milliseconds.
3727 The Dynamic Audio Normalizer processes the input audio in small chunks,
3728 referred to as frames. This is required, because a peak magnitude has no
3729 meaning for just a single sample value. Instead, we need to determine the
3730 peak magnitude for a contiguous sequence of sample values. While a "standard"
3731 normalizer would simply use the peak magnitude of the complete file, the
3732 Dynamic Audio Normalizer determines the peak magnitude individually for each
3733 frame. The length of a frame is specified in milliseconds. By default, the
3734 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3735 been found to give good results with most files.
3736 Note that the exact frame length, in number of samples, will be determined
3737 automatically, based on the sampling rate of the individual input audio file.
3740 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3741 number. Default is 31.
3742 Probably the most important parameter of the Dynamic Audio Normalizer is the
3743 @code{window size} of the Gaussian smoothing filter. The filter's window size
3744 is specified in frames, centered around the current frame. For the sake of
3745 simplicity, this must be an odd number. Consequently, the default value of 31
3746 takes into account the current frame, as well as the 15 preceding frames and
3747 the 15 subsequent frames. Using a larger window results in a stronger
3748 smoothing effect and thus in less gain variation, i.e. slower gain
3749 adaptation. Conversely, using a smaller window results in a weaker smoothing
3750 effect and thus in more gain variation, i.e. faster gain adaptation.
3751 In other words, the more you increase this value, the more the Dynamic Audio
3752 Normalizer will behave like a "traditional" normalization filter. On the
3753 contrary, the more you decrease this value, the more the Dynamic Audio
3754 Normalizer will behave like a dynamic range compressor.
3757 Set the target peak value. This specifies the highest permissible magnitude
3758 level for the normalized audio input. This filter will try to approach the
3759 target peak magnitude as closely as possible, but at the same time it also
3760 makes sure that the normalized signal will never exceed the peak magnitude.
3761 A frame's maximum local gain factor is imposed directly by the target peak
3762 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3763 It is not recommended to go above this value.
3766 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3767 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3768 factor for each input frame, i.e. the maximum gain factor that does not
3769 result in clipping or distortion. The maximum gain factor is determined by
3770 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3771 additionally bounds the frame's maximum gain factor by a predetermined
3772 (global) maximum gain factor. This is done in order to avoid excessive gain
3773 factors in "silent" or almost silent frames. By default, the maximum gain
3774 factor is 10.0, For most inputs the default value should be sufficient and
3775 it usually is not recommended to increase this value. Though, for input
3776 with an extremely low overall volume level, it may be necessary to allow even
3777 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3778 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3779 Instead, a "sigmoid" threshold function will be applied. This way, the
3780 gain factors will smoothly approach the threshold value, but never exceed that
3784 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3785 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3786 This means that the maximum local gain factor for each frame is defined
3787 (only) by the frame's highest magnitude sample. This way, the samples can
3788 be amplified as much as possible without exceeding the maximum signal
3789 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3790 Normalizer can also take into account the frame's root mean square,
3791 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3792 determine the power of a time-varying signal. It is therefore considered
3793 that the RMS is a better approximation of the "perceived loudness" than
3794 just looking at the signal's peak magnitude. Consequently, by adjusting all
3795 frames to a constant RMS value, a uniform "perceived loudness" can be
3796 established. If a target RMS value has been specified, a frame's local gain
3797 factor is defined as the factor that would result in exactly that RMS value.
3798 Note, however, that the maximum local gain factor is still restricted by the
3799 frame's highest magnitude sample, in order to prevent clipping.
3802 Enable channels coupling. By default is enabled.
3803 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3804 amount. This means the same gain factor will be applied to all channels, i.e.
3805 the maximum possible gain factor is determined by the "loudest" channel.
3806 However, in some recordings, it may happen that the volume of the different
3807 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3808 In this case, this option can be used to disable the channel coupling. This way,
3809 the gain factor will be determined independently for each channel, depending
3810 only on the individual channel's highest magnitude sample. This allows for
3811 harmonizing the volume of the different channels.
3814 Enable DC bias correction. By default is disabled.
3815 An audio signal (in the time domain) is a sequence of sample values.
3816 In the Dynamic Audio Normalizer these sample values are represented in the
3817 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3818 audio signal, or "waveform", should be centered around the zero point.
3819 That means if we calculate the mean value of all samples in a file, or in a
3820 single frame, then the result should be 0.0 or at least very close to that
3821 value. If, however, there is a significant deviation of the mean value from
3822 0.0, in either positive or negative direction, this is referred to as a
3823 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3824 Audio Normalizer provides optional DC bias correction.
3825 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3826 the mean value, or "DC correction" offset, of each input frame and subtract
3827 that value from all of the frame's sample values which ensures those samples
3828 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3829 boundaries, the DC correction offset values will be interpolated smoothly
3830 between neighbouring frames.
3832 @item altboundary, b
3833 Enable alternative boundary mode. By default is disabled.
3834 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3835 around each frame. This includes the preceding frames as well as the
3836 subsequent frames. However, for the "boundary" frames, located at the very
3837 beginning and at the very end of the audio file, not all neighbouring
3838 frames are available. In particular, for the first few frames in the audio
3839 file, the preceding frames are not known. And, similarly, for the last few
3840 frames in the audio file, the subsequent frames are not known. Thus, the
3841 question arises which gain factors should be assumed for the missing frames
3842 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3843 to deal with this situation. The default boundary mode assumes a gain factor
3844 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3845 "fade out" at the beginning and at the end of the input, respectively.
3848 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3849 By default, the Dynamic Audio Normalizer does not apply "traditional"
3850 compression. This means that signal peaks will not be pruned and thus the
3851 full dynamic range will be retained within each local neighbourhood. However,
3852 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3853 normalization algorithm with a more "traditional" compression.
3854 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3855 (thresholding) function. If (and only if) the compression feature is enabled,
3856 all input frames will be processed by a soft knee thresholding function prior
3857 to the actual normalization process. Put simply, the thresholding function is
3858 going to prune all samples whose magnitude exceeds a certain threshold value.
3859 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3860 value. Instead, the threshold value will be adjusted for each individual
3862 In general, smaller parameters result in stronger compression, and vice versa.
3863 Values below 3.0 are not recommended, because audible distortion may appear.
3866 Set the target threshold value. This specifies the lowest permissible
3867 magnitude level for the audio input which will be normalized.
3868 If input frame volume is above this value frame will be normalized.
3869 Otherwise frame may not be normalized at all. The default value is set
3870 to 0, which means all input frames will be normalized.
3871 This option is mostly useful if digital noise is not wanted to be amplified.
3874 @subsection Commands
3876 This filter supports the all above options as @ref{commands}.
3880 Make audio easier to listen to on headphones.
3882 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3883 so that when listened to on headphones the stereo image is moved from
3884 inside your head (standard for headphones) to outside and in front of
3885 the listener (standard for speakers).
3891 Apply a two-pole peaking equalisation (EQ) filter. With this
3892 filter, the signal-level at and around a selected frequency can
3893 be increased or decreased, whilst (unlike bandpass and bandreject
3894 filters) that at all other frequencies is unchanged.
3896 In order to produce complex equalisation curves, this filter can
3897 be given several times, each with a different central frequency.
3899 The filter accepts the following options:
3903 Set the filter's central frequency in Hz.
3906 Set method to specify band-width of filter.
3921 Specify the band-width of a filter in width_type units.
3924 Set the required gain or attenuation in dB.
3925 Beware of clipping when using a positive gain.
3928 How much to use filtered signal in output. Default is 1.
3929 Range is between 0 and 1.
3932 Specify which channels to filter, by default all available are filtered.
3935 Normalize biquad coefficients, by default is disabled.
3936 Enabling it will normalize magnitude response at DC to 0dB.
3939 Set transform type of IIR filter.
3948 Set precison of filtering.
3951 Pick automatic sample format depending on surround filters.
3953 Always use signed 16-bit.
3955 Always use signed 32-bit.
3957 Always use float 32-bit.
3959 Always use float 64-bit.
3963 @subsection Examples
3966 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3968 equalizer=f=1000:t=h:width=200:g=-10
3972 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3974 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3978 @subsection Commands
3980 This filter supports the following commands:
3983 Change equalizer frequency.
3984 Syntax for the command is : "@var{frequency}"
3987 Change equalizer width_type.
3988 Syntax for the command is : "@var{width_type}"
3991 Change equalizer width.
3992 Syntax for the command is : "@var{width}"
3995 Change equalizer gain.
3996 Syntax for the command is : "@var{gain}"
3999 Change equalizer mix.
4000 Syntax for the command is : "@var{mix}"
4003 @section extrastereo
4005 Linearly increases the difference between left and right channels which
4006 adds some sort of "live" effect to playback.
4008 The filter accepts the following options:
4012 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4013 (average of both channels), with 1.0 sound will be unchanged, with
4014 -1.0 left and right channels will be swapped.
4017 Enable clipping. By default is enabled.
4020 @subsection Commands
4022 This filter supports the all above options as @ref{commands}.
4024 @section firequalizer
4025 Apply FIR Equalization using arbitrary frequency response.
4027 The filter accepts the following option:
4031 Set gain curve equation (in dB). The expression can contain variables:
4034 the evaluated frequency
4038 channel number, set to 0 when multichannels evaluation is disabled
4040 channel id, see libavutil/channel_layout.h, set to the first channel id when
4041 multichannels evaluation is disabled
4045 channel_layout, see libavutil/channel_layout.h
4050 @item gain_interpolate(f)
4051 interpolate gain on frequency f based on gain_entry
4052 @item cubic_interpolate(f)
4053 same as gain_interpolate, but smoother
4055 This option is also available as command. Default is @code{gain_interpolate(f)}.
4058 Set gain entry for gain_interpolate function. The expression can
4062 store gain entry at frequency f with value g
4064 This option is also available as command.
4067 Set filter delay in seconds. Higher value means more accurate.
4068 Default is @code{0.01}.
4071 Set filter accuracy in Hz. Lower value means more accurate.
4072 Default is @code{5}.
4075 Set window function. Acceptable values are:
4078 rectangular window, useful when gain curve is already smooth
4080 hann window (default)
4086 3-terms continuous 1st derivative nuttall window
4088 minimum 3-terms discontinuous nuttall window
4090 4-terms continuous 1st derivative nuttall window
4092 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4094 blackman-harris window
4100 If enabled, use fixed number of audio samples. This improves speed when
4101 filtering with large delay. Default is disabled.
4104 Enable multichannels evaluation on gain. Default is disabled.
4107 Enable zero phase mode by subtracting timestamp to compensate delay.
4108 Default is disabled.
4111 Set scale used by gain. Acceptable values are:
4114 linear frequency, linear gain
4116 linear frequency, logarithmic (in dB) gain (default)
4118 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4120 logarithmic frequency, logarithmic gain
4124 Set file for dumping, suitable for gnuplot.
4127 Set scale for dumpfile. Acceptable values are same with scale option.
4131 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4132 Default is disabled.
4135 Enable minimum phase impulse response. Default is disabled.
4138 @subsection Examples
4143 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4146 lowpass at 1000 Hz with gain_entry:
4148 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4151 custom equalization:
4153 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4156 higher delay with zero phase to compensate delay:
4158 firequalizer=delay=0.1:fixed=on:zero_phase=on
4161 lowpass on left channel, highpass on right channel:
4163 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4164 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4169 Apply a flanging effect to the audio.
4171 The filter accepts the following options:
4175 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4178 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4181 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4185 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4186 Default value is 71.
4189 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4192 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4193 Default value is @var{sinusoidal}.
4196 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4197 Default value is 25.
4200 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4201 Default is @var{linear}.
4205 Apply Haas effect to audio.
4207 Note that this makes most sense to apply on mono signals.
4208 With this filter applied to mono signals it give some directionality and
4209 stretches its stereo image.
4211 The filter accepts the following options:
4215 Set input level. By default is @var{1}, or 0dB
4218 Set output level. By default is @var{1}, or 0dB.
4221 Set gain applied to side part of signal. By default is @var{1}.
4224 Set kind of middle source. Can be one of the following:
4234 Pick middle part signal of stereo image.
4237 Pick side part signal of stereo image.
4241 Change middle phase. By default is disabled.
4244 Set left channel delay. By default is @var{2.05} milliseconds.
4247 Set left channel balance. By default is @var{-1}.
4250 Set left channel gain. By default is @var{1}.
4253 Change left phase. By default is disabled.
4256 Set right channel delay. By defaults is @var{2.12} milliseconds.
4259 Set right channel balance. By default is @var{1}.
4262 Set right channel gain. By default is @var{1}.
4265 Change right phase. By default is enabled.
4270 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4271 embedded HDCD codes is expanded into a 20-bit PCM stream.
4273 The filter supports the Peak Extend and Low-level Gain Adjustment features
4274 of HDCD, and detects the Transient Filter flag.
4277 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4280 When using the filter with wav, note the default encoding for wav is 16-bit,
4281 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4282 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4284 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4285 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4288 The filter accepts the following options:
4291 @item disable_autoconvert
4292 Disable any automatic format conversion or resampling in the filter graph.
4294 @item process_stereo
4295 Process the stereo channels together. If target_gain does not match between
4296 channels, consider it invalid and use the last valid target_gain.
4299 Set the code detect timer period in ms.
4302 Always extend peaks above -3dBFS even if PE isn't signaled.
4305 Replace audio with a solid tone and adjust the amplitude to signal some
4306 specific aspect of the decoding process. The output file can be loaded in
4307 an audio editor alongside the original to aid analysis.
4309 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4316 Gain adjustment level at each sample
4318 Samples where peak extend occurs
4320 Samples where the code detect timer is active
4322 Samples where the target gain does not match between channels
4328 Apply head-related transfer functions (HRTFs) to create virtual
4329 loudspeakers around the user for binaural listening via headphones.
4330 The HRIRs are provided via additional streams, for each channel
4331 one stereo input stream is needed.
4333 The filter accepts the following options:
4337 Set mapping of input streams for convolution.
4338 The argument is a '|'-separated list of channel names in order as they
4339 are given as additional stream inputs for filter.
4340 This also specify number of input streams. Number of input streams
4341 must be not less than number of channels in first stream plus one.
4344 Set gain applied to audio. Value is in dB. Default is 0.
4347 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4348 processing audio in time domain which is slow.
4349 @var{freq} is processing audio in frequency domain which is fast.
4350 Default is @var{freq}.
4353 Set custom gain for LFE channels. Value is in dB. Default is 0.
4356 Set size of frame in number of samples which will be processed at once.
4357 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4360 Set format of hrir stream.
4361 Default value is @var{stereo}. Alternative value is @var{multich}.
4362 If value is set to @var{stereo}, number of additional streams should
4363 be greater or equal to number of input channels in first input stream.
4364 Also each additional stream should have stereo number of channels.
4365 If value is set to @var{multich}, number of additional streams should
4366 be exactly one. Also number of input channels of additional stream
4367 should be equal or greater than twice number of channels of first input
4371 @subsection Examples
4375 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4376 each amovie filter use stereo file with IR coefficients as input.
4377 The files give coefficients for each position of virtual loudspeaker:
4380 -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"
4385 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4386 but now in @var{multich} @var{hrir} format.
4388 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"
4395 Apply a high-pass filter with 3dB point frequency.
4396 The filter can be either single-pole, or double-pole (the default).
4397 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4399 The filter accepts the following options:
4403 Set frequency in Hz. Default is 3000.
4406 Set number of poles. Default is 2.
4409 Set method to specify band-width of filter.
4424 Specify the band-width of a filter in width_type units.
4425 Applies only to double-pole filter.
4426 The default is 0.707q and gives a Butterworth response.
4429 How much to use filtered signal in output. Default is 1.
4430 Range is between 0 and 1.
4433 Specify which channels to filter, by default all available are filtered.
4436 Normalize biquad coefficients, by default is disabled.
4437 Enabling it will normalize magnitude response at DC to 0dB.
4440 Set transform type of IIR filter.
4449 Set precison of filtering.
4452 Pick automatic sample format depending on surround filters.
4454 Always use signed 16-bit.
4456 Always use signed 32-bit.
4458 Always use float 32-bit.
4460 Always use float 64-bit.
4464 @subsection Commands
4466 This filter supports the following commands:
4469 Change highpass frequency.
4470 Syntax for the command is : "@var{frequency}"
4473 Change highpass width_type.
4474 Syntax for the command is : "@var{width_type}"
4477 Change highpass width.
4478 Syntax for the command is : "@var{width}"
4481 Change highpass mix.
4482 Syntax for the command is : "@var{mix}"
4487 Join multiple input streams into one multi-channel stream.
4489 It accepts the following parameters:
4493 The number of input streams. It defaults to 2.
4495 @item channel_layout
4496 The desired output channel layout. It defaults to stereo.
4499 Map channels from inputs to output. The argument is a '|'-separated list of
4500 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4501 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4502 can be either the name of the input channel (e.g. FL for front left) or its
4503 index in the specified input stream. @var{out_channel} is the name of the output
4507 The filter will attempt to guess the mappings when they are not specified
4508 explicitly. It does so by first trying to find an unused matching input channel
4509 and if that fails it picks the first unused input channel.
4511 Join 3 inputs (with properly set channel layouts):
4513 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4516 Build a 5.1 output from 6 single-channel streams:
4518 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4519 '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'
4525 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4527 To enable compilation of this filter you need to configure FFmpeg with
4528 @code{--enable-ladspa}.
4532 Specifies the name of LADSPA plugin library to load. If the environment
4533 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4534 each one of the directories specified by the colon separated list in
4535 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4536 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4537 @file{/usr/lib/ladspa/}.
4540 Specifies the plugin within the library. Some libraries contain only
4541 one plugin, but others contain many of them. If this is not set filter
4542 will list all available plugins within the specified library.
4545 Set the '|' separated list of controls which are zero or more floating point
4546 values that determine the behavior of the loaded plugin (for example delay,
4548 Controls need to be defined using the following syntax:
4549 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4550 @var{valuei} is the value set on the @var{i}-th control.
4551 Alternatively they can be also defined using the following syntax:
4552 @var{value0}|@var{value1}|@var{value2}|..., where
4553 @var{valuei} is the value set on the @var{i}-th control.
4554 If @option{controls} is set to @code{help}, all available controls and
4555 their valid ranges are printed.
4557 @item sample_rate, s
4558 Specify the sample rate, default to 44100. Only used if plugin have
4562 Set the number of samples per channel per each output frame, default
4563 is 1024. Only used if plugin have zero inputs.
4566 Set the minimum duration of the sourced audio. See
4567 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4568 for the accepted syntax.
4569 Note that the resulting duration may be greater than the specified duration,
4570 as the generated audio is always cut at the end of a complete frame.
4571 If not specified, or the expressed duration is negative, the audio is
4572 supposed to be generated forever.
4573 Only used if plugin have zero inputs.
4576 Enable latency compensation, by default is disabled.
4577 Only used if plugin have inputs.
4580 @subsection Examples
4584 List all available plugins within amp (LADSPA example plugin) library:
4590 List all available controls and their valid ranges for @code{vcf_notch}
4591 plugin from @code{VCF} library:
4593 ladspa=f=vcf:p=vcf_notch:c=help
4597 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4600 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4604 Add reverberation to the audio using TAP-plugins
4605 (Tom's Audio Processing plugins):
4607 ladspa=file=tap_reverb:tap_reverb
4611 Generate white noise, with 0.2 amplitude:
4613 ladspa=file=cmt:noise_source_white:c=c0=.2
4617 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4618 @code{C* Audio Plugin Suite} (CAPS) library:
4620 ladspa=file=caps:Click:c=c1=20'
4624 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4626 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4630 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4631 @code{SWH Plugins} collection:
4633 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4637 Attenuate low frequencies using Multiband EQ from Steve Harris
4638 @code{SWH Plugins} collection:
4640 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4644 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4647 ladspa=caps:Narrower
4651 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4653 ladspa=caps:White:.2
4657 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4659 ladspa=caps:Fractal:c=c1=1
4663 Dynamic volume normalization using @code{VLevel} plugin:
4665 ladspa=vlevel-ladspa:vlevel_mono
4669 @subsection Commands
4671 This filter supports the following commands:
4674 Modify the @var{N}-th control value.
4676 If the specified value is not valid, it is ignored and prior one is kept.
4681 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4682 Support for both single pass (livestreams, files) and double pass (files) modes.
4683 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4684 detect true peaks, the audio stream will be upsampled to 192 kHz.
4685 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4687 The filter accepts the following options:
4691 Set integrated loudness target.
4692 Range is -70.0 - -5.0. Default value is -24.0.
4695 Set loudness range target.
4696 Range is 1.0 - 20.0. Default value is 7.0.
4699 Set maximum true peak.
4700 Range is -9.0 - +0.0. Default value is -2.0.
4702 @item measured_I, measured_i
4703 Measured IL of input file.
4704 Range is -99.0 - +0.0.
4706 @item measured_LRA, measured_lra
4707 Measured LRA of input file.
4708 Range is 0.0 - 99.0.
4710 @item measured_TP, measured_tp
4711 Measured true peak of input file.
4712 Range is -99.0 - +99.0.
4714 @item measured_thresh
4715 Measured threshold of input file.
4716 Range is -99.0 - +0.0.
4719 Set offset gain. Gain is applied before the true-peak limiter.
4720 Range is -99.0 - +99.0. Default is +0.0.
4723 Normalize by linearly scaling the source audio.
4724 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4725 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4726 be lower than source LRA and the change in integrated loudness shouldn't
4727 result in a true peak which exceeds the target TP. If any of these
4728 conditions aren't met, normalization mode will revert to @var{dynamic}.
4729 Options are @code{true} or @code{false}. Default is @code{true}.
4732 Treat mono input files as "dual-mono". If a mono file is intended for playback
4733 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4734 If set to @code{true}, this option will compensate for this effect.
4735 Multi-channel input files are not affected by this option.
4736 Options are true or false. Default is false.
4739 Set print format for stats. Options are summary, json, or none.
4740 Default value is none.
4745 Apply a low-pass filter with 3dB point frequency.
4746 The filter can be either single-pole or double-pole (the default).
4747 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4749 The filter accepts the following options:
4753 Set frequency in Hz. Default is 500.
4756 Set number of poles. Default is 2.
4759 Set method to specify band-width of filter.
4774 Specify the band-width of a filter in width_type units.
4775 Applies only to double-pole filter.
4776 The default is 0.707q and gives a Butterworth response.
4779 How much to use filtered signal in output. Default is 1.
4780 Range is between 0 and 1.
4783 Specify which channels to filter, by default all available are filtered.
4786 Normalize biquad coefficients, by default is disabled.
4787 Enabling it will normalize magnitude response at DC to 0dB.
4790 Set transform type of IIR filter.
4799 Set precison of filtering.
4802 Pick automatic sample format depending on surround filters.
4804 Always use signed 16-bit.
4806 Always use signed 32-bit.
4808 Always use float 32-bit.
4810 Always use float 64-bit.
4814 @subsection Examples
4817 Lowpass only LFE channel, it LFE is not present it does nothing:
4823 @subsection Commands
4825 This filter supports the following commands:
4828 Change lowpass frequency.
4829 Syntax for the command is : "@var{frequency}"
4832 Change lowpass width_type.
4833 Syntax for the command is : "@var{width_type}"
4836 Change lowpass width.
4837 Syntax for the command is : "@var{width}"
4841 Syntax for the command is : "@var{mix}"
4846 Load a LV2 (LADSPA Version 2) plugin.
4848 To enable compilation of this filter you need to configure FFmpeg with
4849 @code{--enable-lv2}.
4853 Specifies the plugin URI. You may need to escape ':'.
4856 Set the '|' separated list of controls which are zero or more floating point
4857 values that determine the behavior of the loaded plugin (for example delay,
4859 If @option{controls} is set to @code{help}, all available controls and
4860 their valid ranges are printed.
4862 @item sample_rate, s
4863 Specify the sample rate, default to 44100. Only used if plugin have
4867 Set the number of samples per channel per each output frame, default
4868 is 1024. Only used if plugin have zero inputs.
4871 Set the minimum duration of the sourced audio. See
4872 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4873 for the accepted syntax.
4874 Note that the resulting duration may be greater than the specified duration,
4875 as the generated audio is always cut at the end of a complete frame.
4876 If not specified, or the expressed duration is negative, the audio is
4877 supposed to be generated forever.
4878 Only used if plugin have zero inputs.
4881 @subsection Examples
4885 Apply bass enhancer plugin from Calf:
4887 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4891 Apply vinyl plugin from Calf:
4893 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4897 Apply bit crusher plugin from ArtyFX:
4899 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4904 Multiband Compress or expand the audio's dynamic range.
4906 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4907 This is akin to the crossover of a loudspeaker, and results in flat frequency
4908 response when absent compander action.
4910 It accepts the following parameters:
4914 This option syntax is:
4915 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4916 For explanation of each item refer to compand filter documentation.
4922 Mix channels with specific gain levels. The filter accepts the output
4923 channel layout followed by a set of channels definitions.
4925 This filter is also designed to efficiently remap the channels of an audio
4928 The filter accepts parameters of the form:
4929 "@var{l}|@var{outdef}|@var{outdef}|..."
4933 output channel layout or number of channels
4936 output channel specification, of the form:
4937 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4940 output channel to define, either a channel name (FL, FR, etc.) or a channel
4941 number (c0, c1, etc.)
4944 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4947 input channel to use, see out_name for details; it is not possible to mix
4948 named and numbered input channels
4951 If the `=' in a channel specification is replaced by `<', then the gains for
4952 that specification will be renormalized so that the total is 1, thus
4953 avoiding clipping noise.
4955 @subsection Mixing examples
4957 For example, if you want to down-mix from stereo to mono, but with a bigger
4958 factor for the left channel:
4960 pan=1c|c0=0.9*c0+0.1*c1
4963 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4964 7-channels surround:
4966 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4969 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4970 that should be preferred (see "-ac" option) unless you have very specific
4973 @subsection Remapping examples
4975 The channel remapping will be effective if, and only if:
4978 @item gain coefficients are zeroes or ones,
4979 @item only one input per channel output,
4982 If all these conditions are satisfied, the filter will notify the user ("Pure
4983 channel mapping detected"), and use an optimized and lossless method to do the
4986 For example, if you have a 5.1 source and want a stereo audio stream by
4987 dropping the extra channels:
4989 pan="stereo| c0=FL | c1=FR"
4992 Given the same source, you can also switch front left and front right channels
4993 and keep the input channel layout:
4995 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4998 If the input is a stereo audio stream, you can mute the front left channel (and
4999 still keep the stereo channel layout) with:
5004 Still with a stereo audio stream input, you can copy the right channel in both
5005 front left and right:
5007 pan="stereo| c0=FR | c1=FR"
5012 ReplayGain scanner filter. This filter takes an audio stream as an input and
5013 outputs it unchanged.
5014 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5018 Convert the audio sample format, sample rate and channel layout. It is
5019 not meant to be used directly.
5022 Apply time-stretching and pitch-shifting with librubberband.
5024 To enable compilation of this filter, you need to configure FFmpeg with
5025 @code{--enable-librubberband}.
5027 The filter accepts the following options:
5031 Set tempo scale factor.
5034 Set pitch scale factor.
5037 Set transients detector.
5038 Possible values are:
5047 Possible values are:
5056 Possible values are:
5063 Set processing window size.
5064 Possible values are:
5073 Possible values are:
5080 Enable formant preservation when shift pitching.
5081 Possible values are:
5089 Possible values are:
5098 Possible values are:
5105 @subsection Commands
5107 This filter supports the following commands:
5110 Change filter tempo scale factor.
5111 Syntax for the command is : "@var{tempo}"
5114 Change filter pitch scale factor.
5115 Syntax for the command is : "@var{pitch}"
5118 @section sidechaincompress
5120 This filter acts like normal compressor but has the ability to compress
5121 detected signal using second input signal.
5122 It needs two input streams and returns one output stream.
5123 First input stream will be processed depending on second stream signal.
5124 The filtered signal then can be filtered with other filters in later stages of
5125 processing. See @ref{pan} and @ref{amerge} filter.
5127 The filter accepts the following options:
5131 Set input gain. Default is 1. Range is between 0.015625 and 64.
5134 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5135 Default is @code{downward}.
5138 If a signal of second stream raises above this level it will affect the gain
5139 reduction of first stream.
5140 By default is 0.125. Range is between 0.00097563 and 1.
5143 Set a ratio about which the signal is reduced. 1:2 means that if the level
5144 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5145 Default is 2. Range is between 1 and 20.
5148 Amount of milliseconds the signal has to rise above the threshold before gain
5149 reduction starts. Default is 20. Range is between 0.01 and 2000.
5152 Amount of milliseconds the signal has to fall below the threshold before
5153 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5156 Set the amount by how much signal will be amplified after processing.
5157 Default is 1. Range is from 1 to 64.
5160 Curve the sharp knee around the threshold to enter gain reduction more softly.
5161 Default is 2.82843. Range is between 1 and 8.
5164 Choose if the @code{average} level between all channels of side-chain stream
5165 or the louder(@code{maximum}) channel of side-chain stream affects the
5166 reduction. Default is @code{average}.
5169 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5170 of @code{rms}. Default is @code{rms} which is mainly smoother.
5173 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5176 How much to use compressed signal in output. Default is 1.
5177 Range is between 0 and 1.
5180 @subsection Commands
5182 This filter supports the all above options as @ref{commands}.
5184 @subsection Examples
5188 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5189 depending on the signal of 2nd input and later compressed signal to be
5190 merged with 2nd input:
5192 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5196 @section sidechaingate
5198 A sidechain gate acts like a normal (wideband) gate but has the ability to
5199 filter the detected signal before sending it to the gain reduction stage.
5200 Normally a gate uses the full range signal to detect a level above the
5202 For example: If you cut all lower frequencies from your sidechain signal
5203 the gate will decrease the volume of your track only if not enough highs
5204 appear. With this technique you are able to reduce the resonation of a
5205 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5207 It needs two input streams and returns one output stream.
5208 First input stream will be processed depending on second stream signal.
5210 The filter accepts the following options:
5214 Set input level before filtering.
5215 Default is 1. Allowed range is from 0.015625 to 64.
5218 Set the mode of operation. Can be @code{upward} or @code{downward}.
5219 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5220 will be amplified, expanding dynamic range in upward direction.
5221 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5224 Set the level of gain reduction when the signal is below the threshold.
5225 Default is 0.06125. Allowed range is from 0 to 1.
5226 Setting this to 0 disables reduction and then filter behaves like expander.
5229 If a signal rises above this level the gain reduction is released.
5230 Default is 0.125. Allowed range is from 0 to 1.
5233 Set a ratio about which the signal is reduced.
5234 Default is 2. Allowed range is from 1 to 9000.
5237 Amount of milliseconds the signal has to rise above the threshold before gain
5239 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5242 Amount of milliseconds the signal has to fall below the threshold before the
5243 reduction is increased again. Default is 250 milliseconds.
5244 Allowed range is from 0.01 to 9000.
5247 Set amount of amplification of signal after processing.
5248 Default is 1. Allowed range is from 1 to 64.
5251 Curve the sharp knee around the threshold to enter gain reduction more softly.
5252 Default is 2.828427125. Allowed range is from 1 to 8.
5255 Choose if exact signal should be taken for detection or an RMS like one.
5256 Default is rms. Can be peak or rms.
5259 Choose if the average level between all channels or the louder channel affects
5261 Default is average. Can be average or maximum.
5264 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5267 @subsection Commands
5269 This filter supports the all above options as @ref{commands}.
5271 @section silencedetect
5273 Detect silence in an audio stream.
5275 This filter logs a message when it detects that the input audio volume is less
5276 or equal to a noise tolerance value for a duration greater or equal to the
5277 minimum detected noise duration.
5279 The printed times and duration are expressed in seconds. The
5280 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5281 is set on the first frame whose timestamp equals or exceeds the detection
5282 duration and it contains the timestamp of the first frame of the silence.
5284 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5285 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5286 keys are set on the first frame after the silence. If @option{mono} is
5287 enabled, and each channel is evaluated separately, the @code{.X}
5288 suffixed keys are used, and @code{X} corresponds to the channel number.
5290 The filter accepts the following options:
5294 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5295 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5298 Set silence duration until notification (default is 2 seconds). See
5299 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5300 for the accepted syntax.
5303 Process each channel separately, instead of combined. By default is disabled.
5306 @subsection Examples
5310 Detect 5 seconds of silence with -50dB noise tolerance:
5312 silencedetect=n=-50dB:d=5
5316 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5317 tolerance in @file{silence.mp3}:
5319 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5323 @section silenceremove
5325 Remove silence from the beginning, middle or end of the audio.
5327 The filter accepts the following options:
5331 This value is used to indicate if audio should be trimmed at beginning of
5332 the audio. A value of zero indicates no silence should be trimmed from the
5333 beginning. When specifying a non-zero value, it trims audio up until it
5334 finds non-silence. Normally, when trimming silence from beginning of audio
5335 the @var{start_periods} will be @code{1} but it can be increased to higher
5336 values to trim all audio up to specific count of non-silence periods.
5337 Default value is @code{0}.
5339 @item start_duration
5340 Specify the amount of time that non-silence must be detected before it stops
5341 trimming audio. By increasing the duration, bursts of noises can be treated
5342 as silence and trimmed off. Default value is @code{0}.
5344 @item start_threshold
5345 This indicates what sample value should be treated as silence. For digital
5346 audio, a value of @code{0} may be fine but for audio recorded from analog,
5347 you may wish to increase the value to account for background noise.
5348 Can be specified in dB (in case "dB" is appended to the specified value)
5349 or amplitude ratio. Default value is @code{0}.
5352 Specify max duration of silence at beginning that will be kept after
5353 trimming. Default is 0, which is equal to trimming all samples detected
5357 Specify mode of detection of silence end in start of multi-channel audio.
5358 Can be @var{any} or @var{all}. Default is @var{any}.
5359 With @var{any}, any sample that is detected as non-silence will cause
5360 stopped trimming of silence.
5361 With @var{all}, only if all channels are detected as non-silence will cause
5362 stopped trimming of silence.
5365 Set the count for trimming silence from the end of audio.
5366 To remove silence from the middle of a file, specify a @var{stop_periods}
5367 that is negative. This value is then treated as a positive value and is
5368 used to indicate the effect should restart processing as specified by
5369 @var{start_periods}, making it suitable for removing periods of silence
5370 in the middle of the audio.
5371 Default value is @code{0}.
5374 Specify a duration of silence that must exist before audio is not copied any
5375 more. By specifying a higher duration, silence that is wanted can be left in
5377 Default value is @code{0}.
5379 @item stop_threshold
5380 This is the same as @option{start_threshold} but for trimming silence from
5382 Can be specified in dB (in case "dB" is appended to the specified value)
5383 or amplitude ratio. Default value is @code{0}.
5386 Specify max duration of silence at end that will be kept after
5387 trimming. Default is 0, which is equal to trimming all samples detected
5391 Specify mode of detection of silence start in end of multi-channel audio.
5392 Can be @var{any} or @var{all}. Default is @var{any}.
5393 With @var{any}, any sample that is detected as non-silence will cause
5394 stopped trimming of silence.
5395 With @var{all}, only if all channels are detected as non-silence will cause
5396 stopped trimming of silence.
5399 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5400 and works better with digital silence which is exactly 0.
5401 Default value is @code{rms}.
5404 Set duration in number of seconds used to calculate size of window in number
5405 of samples for detecting silence.
5406 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5409 @subsection Examples
5413 The following example shows how this filter can be used to start a recording
5414 that does not contain the delay at the start which usually occurs between
5415 pressing the record button and the start of the performance:
5417 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5421 Trim all silence encountered from beginning to end where there is more than 1
5422 second of silence in audio:
5424 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5428 Trim all digital silence samples, using peak detection, from beginning to end
5429 where there is more than 0 samples of digital silence in audio and digital
5430 silence is detected in all channels at same positions in stream:
5432 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5438 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5439 loudspeakers around the user for binaural listening via headphones (audio
5440 formats up to 9 channels supported).
5441 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5442 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5443 Austrian Academy of Sciences.
5445 To enable compilation of this filter you need to configure FFmpeg with
5446 @code{--enable-libmysofa}.
5448 The filter accepts the following options:
5452 Set the SOFA file used for rendering.
5455 Set gain applied to audio. Value is in dB. Default is 0.
5458 Set rotation of virtual loudspeakers in deg. Default is 0.
5461 Set elevation of virtual speakers in deg. Default is 0.
5464 Set distance in meters between loudspeakers and the listener with near-field
5465 HRTFs. Default is 1.
5468 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5469 processing audio in time domain which is slow.
5470 @var{freq} is processing audio in frequency domain which is fast.
5471 Default is @var{freq}.
5474 Set custom positions of virtual loudspeakers. Syntax for this option is:
5475 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5476 Each virtual loudspeaker is described with short channel name following with
5477 azimuth and elevation in degrees.
5478 Each virtual loudspeaker description is separated by '|'.
5479 For example to override front left and front right channel positions use:
5480 'speakers=FL 45 15|FR 345 15'.
5481 Descriptions with unrecognised channel names are ignored.
5484 Set custom gain for LFE channels. Value is in dB. Default is 0.
5487 Set custom frame size in number of samples. Default is 1024.
5488 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5489 is set to @var{freq}.
5492 Should all IRs be normalized upon importing SOFA file.
5493 By default is enabled.
5496 Should nearest IRs be interpolated with neighbor IRs if exact position
5497 does not match. By default is disabled.
5500 Minphase all IRs upon loading of SOFA file. By default is disabled.
5503 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5506 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5509 @subsection Examples
5513 Using ClubFritz6 sofa file:
5515 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5519 Using ClubFritz12 sofa file and bigger radius with small rotation:
5521 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5525 Similar as above but with custom speaker positions for front left, front right, back left and back right
5526 and also with custom gain:
5528 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5535 This filter expands or compresses each half-cycle of audio samples
5536 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5537 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5539 The filter accepts the following options:
5543 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5544 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5547 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5548 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5549 would be such that local peak value reaches target peak value but never to surpass it and that
5550 ratio between new and previous peak value does not surpass this option value.
5552 @item compression, c
5553 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5554 This option controls maximum local half-cycle of samples compression. This option is used
5555 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5556 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5557 that peak's half-cycle will be compressed by current compression factor.
5560 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5561 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5562 Any half-cycle samples with their local peak value below or same as this option value will be
5563 compressed by current compression factor, otherwise, if greater than threshold value they will be
5564 expanded with expansion factor so that it could reach peak target value but never surpass it.
5567 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5568 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5569 each new half-cycle until it reaches @option{expansion} value.
5570 Setting this options too high may lead to distortions.
5573 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5574 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5575 each new half-cycle until it reaches @option{compression} value.
5578 Specify which channels to filter, by default all available channels are filtered.
5581 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5582 option. When enabled any half-cycle of samples with their local peak value below or same as
5583 @option{threshold} option will be expanded otherwise it will be compressed.
5586 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5587 When disabled each filtered channel gain calculation is independent, otherwise when this option
5588 is enabled the minimum of all possible gains for each filtered channel is used.
5591 @subsection Commands
5593 This filter supports the all above options as @ref{commands}.
5595 @section stereotools
5597 This filter has some handy utilities to manage stereo signals, for converting
5598 M/S stereo recordings to L/R signal while having control over the parameters
5599 or spreading the stereo image of master track.
5601 The filter accepts the following options:
5605 Set input level before filtering for both channels. Defaults is 1.
5606 Allowed range is from 0.015625 to 64.
5609 Set output level after filtering for both channels. Defaults is 1.
5610 Allowed range is from 0.015625 to 64.
5613 Set input balance between both channels. Default is 0.
5614 Allowed range is from -1 to 1.
5617 Set output balance between both channels. Default is 0.
5618 Allowed range is from -1 to 1.
5621 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5622 clipping. Disabled by default.
5625 Mute the left channel. Disabled by default.
5628 Mute the right channel. Disabled by default.
5631 Change the phase of the left channel. Disabled by default.
5634 Change the phase of the right channel. Disabled by default.
5637 Set stereo mode. Available values are:
5641 Left/Right to Left/Right, this is default.
5644 Left/Right to Mid/Side.
5647 Mid/Side to Left/Right.
5650 Left/Right to Left/Left.
5653 Left/Right to Right/Right.
5656 Left/Right to Left + Right.
5659 Left/Right to Right/Left.
5662 Mid/Side to Left/Left.
5665 Mid/Side to Right/Right.
5668 Mid/Side to Right/Left.
5671 Left/Right to Left - Right.
5675 Set level of side signal. Default is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set balance of side signal. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set level of the middle signal. Default is 1.
5684 Allowed range is from 0.015625 to 64.
5687 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5690 Set stereo base between mono and inversed channels. Default is 0.
5691 Allowed range is from -1 to 1.
5694 Set delay in milliseconds how much to delay left from right channel and
5695 vice versa. Default is 0. Allowed range is from -20 to 20.
5698 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5701 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5703 @item bmode_in, bmode_out
5704 Set balance mode for balance_in/balance_out option.
5706 Can be one of the following:
5710 Classic balance mode. Attenuate one channel at time.
5711 Gain is raised up to 1.
5714 Similar as classic mode above but gain is raised up to 2.
5717 Equal power distribution, from -6dB to +6dB range.
5721 @subsection Commands
5723 This filter supports the all above options as @ref{commands}.
5725 @subsection Examples
5729 Apply karaoke like effect:
5731 stereotools=mlev=0.015625
5735 Convert M/S signal to L/R:
5737 "stereotools=mode=ms>lr"
5741 @section stereowiden
5743 This filter enhance the stereo effect by suppressing signal common to both
5744 channels and by delaying the signal of left into right and vice versa,
5745 thereby widening the stereo effect.
5747 The filter accepts the following options:
5751 Time in milliseconds of the delay of left signal into right and vice versa.
5752 Default is 20 milliseconds.
5755 Amount of gain in delayed signal into right and vice versa. Gives a delay
5756 effect of left signal in right output and vice versa which gives widening
5757 effect. Default is 0.3.
5760 Cross feed of left into right with inverted phase. This helps in suppressing
5761 the mono. If the value is 1 it will cancel all the signal common to both
5762 channels. Default is 0.3.
5765 Set level of input signal of original channel. Default is 0.8.
5768 @subsection Commands
5770 This filter supports the all above options except @code{delay} as @ref{commands}.
5772 @section superequalizer
5773 Apply 18 band equalizer.
5775 The filter accepts the following options:
5782 Set 131Hz band gain.
5784 Set 185Hz band gain.
5786 Set 262Hz band gain.
5788 Set 370Hz band gain.
5790 Set 523Hz band gain.
5792 Set 740Hz band gain.
5794 Set 1047Hz band gain.
5796 Set 1480Hz band gain.
5798 Set 2093Hz band gain.
5800 Set 2960Hz band gain.
5802 Set 4186Hz band gain.
5804 Set 5920Hz band gain.
5806 Set 8372Hz band gain.
5808 Set 11840Hz band gain.
5810 Set 16744Hz band gain.
5812 Set 20000Hz band gain.
5816 Apply audio surround upmix filter.
5818 This filter allows to produce multichannel output from audio stream.
5820 The filter accepts the following options:
5824 Set output channel layout. By default, this is @var{5.1}.
5826 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5827 for the required syntax.
5830 Set input channel layout. By default, this is @var{stereo}.
5832 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5833 for the required syntax.
5836 Set input volume level. By default, this is @var{1}.
5839 Set output volume level. By default, this is @var{1}.
5842 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5845 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5848 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5851 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5852 In @var{add} mode, LFE channel is created from input audio and added to output.
5853 In @var{sub} mode, LFE channel is created from input audio and added to output but
5854 also all non-LFE output channels are subtracted with output LFE channel.
5857 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5858 Default is @var{90}.
5861 Set front center input volume. By default, this is @var{1}.
5864 Set front center output volume. By default, this is @var{1}.
5867 Set front left input volume. By default, this is @var{1}.
5870 Set front left output volume. By default, this is @var{1}.
5873 Set front right input volume. By default, this is @var{1}.
5876 Set front right output volume. By default, this is @var{1}.
5879 Set side left input volume. By default, this is @var{1}.
5882 Set side left output volume. By default, this is @var{1}.
5885 Set side right input volume. By default, this is @var{1}.
5888 Set side right output volume. By default, this is @var{1}.
5891 Set back left input volume. By default, this is @var{1}.
5894 Set back left output volume. By default, this is @var{1}.
5897 Set back right input volume. By default, this is @var{1}.
5900 Set back right output volume. By default, this is @var{1}.
5903 Set back center input volume. By default, this is @var{1}.
5906 Set back center output volume. By default, this is @var{1}.
5909 Set LFE input volume. By default, this is @var{1}.
5912 Set LFE output volume. By default, this is @var{1}.
5915 Set spread usage of stereo image across X axis for all channels.
5918 Set spread usage of stereo image across Y axis for all channels.
5920 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5921 Set spread usage of stereo image across X axis for each channel.
5923 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5924 Set spread usage of stereo image across Y axis for each channel.
5927 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5930 Set window function.
5932 It accepts the following values:
5955 Default is @code{hann}.
5958 Set window overlap. If set to 1, the recommended overlap for selected
5959 window function will be picked. Default is @code{0.5}.
5962 @section treble, highshelf
5964 Boost or cut treble (upper) frequencies of the audio using a two-pole
5965 shelving filter with a response similar to that of a standard
5966 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5968 The filter accepts the following options:
5972 Give the gain at whichever is the lower of ~22 kHz and the
5973 Nyquist frequency. Its useful range is about -20 (for a large cut)
5974 to +20 (for a large boost). Beware of clipping when using a positive gain.
5977 Set the filter's central frequency and so can be used
5978 to extend or reduce the frequency range to be boosted or cut.
5979 The default value is @code{3000} Hz.
5982 Set method to specify band-width of filter.
5997 Determine how steep is the filter's shelf transition.
6000 How much to use filtered signal in output. Default is 1.
6001 Range is between 0 and 1.
6004 Specify which channels to filter, by default all available are filtered.
6007 Normalize biquad coefficients, by default is disabled.
6008 Enabling it will normalize magnitude response at DC to 0dB.
6011 Set transform type of IIR filter.
6020 Set precison of filtering.
6023 Pick automatic sample format depending on surround filters.
6025 Always use signed 16-bit.
6027 Always use signed 32-bit.
6029 Always use float 32-bit.
6031 Always use float 64-bit.
6035 @subsection Commands
6037 This filter supports the following commands:
6040 Change treble frequency.
6041 Syntax for the command is : "@var{frequency}"
6044 Change treble width_type.
6045 Syntax for the command is : "@var{width_type}"
6048 Change treble width.
6049 Syntax for the command is : "@var{width}"
6053 Syntax for the command is : "@var{gain}"
6057 Syntax for the command is : "@var{mix}"
6062 Sinusoidal amplitude modulation.
6064 The filter accepts the following options:
6068 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6069 (20 Hz or lower) will result in a tremolo effect.
6070 This filter may also be used as a ring modulator by specifying
6071 a modulation frequency higher than 20 Hz.
6072 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6075 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6076 Default value is 0.5.
6081 Sinusoidal phase modulation.
6083 The filter accepts the following options:
6087 Modulation frequency in Hertz.
6088 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6091 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6092 Default value is 0.5.
6097 Adjust the input audio volume.
6099 It accepts the following parameters:
6103 Set audio volume expression.
6105 Output values are clipped to the maximum value.
6107 The output audio volume is given by the relation:
6109 @var{output_volume} = @var{volume} * @var{input_volume}
6112 The default value for @var{volume} is "1.0".
6115 This parameter represents the mathematical precision.
6117 It determines which input sample formats will be allowed, which affects the
6118 precision of the volume scaling.
6122 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6124 32-bit floating-point; this limits input sample format to FLT. (default)
6126 64-bit floating-point; this limits input sample format to DBL.
6130 Choose the behaviour on encountering ReplayGain side data in input frames.
6134 Remove ReplayGain side data, ignoring its contents (the default).
6137 Ignore ReplayGain side data, but leave it in the frame.
6140 Prefer the track gain, if present.
6143 Prefer the album gain, if present.
6146 @item replaygain_preamp
6147 Pre-amplification gain in dB to apply to the selected replaygain gain.
6149 Default value for @var{replaygain_preamp} is 0.0.
6151 @item replaygain_noclip
6152 Prevent clipping by limiting the gain applied.
6154 Default value for @var{replaygain_noclip} is 1.
6157 Set when the volume expression is evaluated.
6159 It accepts the following values:
6162 only evaluate expression once during the filter initialization, or
6163 when the @samp{volume} command is sent
6166 evaluate expression for each incoming frame
6169 Default value is @samp{once}.
6172 The volume expression can contain the following parameters.
6176 frame number (starting at zero)
6179 @item nb_consumed_samples
6180 number of samples consumed by the filter
6182 number of samples in the current frame
6184 original frame position in the file
6190 PTS at start of stream
6192 time at start of stream
6198 last set volume value
6201 Note that when @option{eval} is set to @samp{once} only the
6202 @var{sample_rate} and @var{tb} variables are available, all other
6203 variables will evaluate to NAN.
6205 @subsection Commands
6207 This filter supports the following commands:
6210 Modify the volume expression.
6211 The command accepts the same syntax of the corresponding option.
6213 If the specified expression is not valid, it is kept at its current
6217 @subsection Examples
6221 Halve the input audio volume:
6225 volume=volume=-6.0206dB
6228 In all the above example the named key for @option{volume} can be
6229 omitted, for example like in:
6235 Increase input audio power by 6 decibels using fixed-point precision:
6237 volume=volume=6dB:precision=fixed
6241 Fade volume after time 10 with an annihilation period of 5 seconds:
6243 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6247 @section volumedetect
6249 Detect the volume of the input video.
6251 The filter has no parameters. The input is not modified. Statistics about
6252 the volume will be printed in the log when the input stream end is reached.
6254 In particular it will show the mean volume (root mean square), maximum
6255 volume (on a per-sample basis), and the beginning of a histogram of the
6256 registered volume values (from the maximum value to a cumulated 1/1000 of
6259 All volumes are in decibels relative to the maximum PCM value.
6261 @subsection Examples
6263 Here is an excerpt of the output:
6265 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6266 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6267 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6268 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6269 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6270 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6271 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6272 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6273 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6279 The mean square energy is approximately -27 dB, or 10^-2.7.
6281 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6283 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6286 In other words, raising the volume by +4 dB does not cause any clipping,
6287 raising it by +5 dB causes clipping for 6 samples, etc.
6289 @c man end AUDIO FILTERS
6291 @chapter Audio Sources
6292 @c man begin AUDIO SOURCES
6294 Below is a description of the currently available audio sources.
6298 Buffer audio frames, and make them available to the filter chain.
6300 This source is mainly intended for a programmatic use, in particular
6301 through the interface defined in @file{libavfilter/buffersrc.h}.
6303 It accepts the following parameters:
6307 The timebase which will be used for timestamps of submitted frames. It must be
6308 either a floating-point number or in @var{numerator}/@var{denominator} form.
6311 The sample rate of the incoming audio buffers.
6314 The sample format of the incoming audio buffers.
6315 Either a sample format name or its corresponding integer representation from
6316 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6318 @item channel_layout
6319 The channel layout of the incoming audio buffers.
6320 Either a channel layout name from channel_layout_map in
6321 @file{libavutil/channel_layout.c} or its corresponding integer representation
6322 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6325 The number of channels of the incoming audio buffers.
6326 If both @var{channels} and @var{channel_layout} are specified, then they
6331 @subsection Examples
6334 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6337 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6338 Since the sample format with name "s16p" corresponds to the number
6339 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6342 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6347 Generate an audio signal specified by an expression.
6349 This source accepts in input one or more expressions (one for each
6350 channel), which are evaluated and used to generate a corresponding
6353 This source accepts the following options:
6357 Set the '|'-separated expressions list for each separate channel. In case the
6358 @option{channel_layout} option is not specified, the selected channel layout
6359 depends on the number of provided expressions. Otherwise the last
6360 specified expression is applied to the remaining output channels.
6362 @item channel_layout, c
6363 Set the channel layout. The number of channels in the specified layout
6364 must be equal to the number of specified expressions.
6367 Set the minimum duration of the sourced audio. See
6368 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6369 for the accepted syntax.
6370 Note that the resulting duration may be greater than the specified
6371 duration, as the generated audio is always cut at the end of a
6374 If not specified, or the expressed duration is negative, the audio is
6375 supposed to be generated forever.
6378 Set the number of samples per channel per each output frame,
6381 @item sample_rate, s
6382 Specify the sample rate, default to 44100.
6385 Each expression in @var{exprs} can contain the following constants:
6389 number of the evaluated sample, starting from 0
6392 time of the evaluated sample expressed in seconds, starting from 0
6399 @subsection Examples
6409 Generate a sin signal with frequency of 440 Hz, set sample rate to
6412 aevalsrc="sin(440*2*PI*t):s=8000"
6416 Generate a two channels signal, specify the channel layout (Front
6417 Center + Back Center) explicitly:
6419 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6423 Generate white noise:
6425 aevalsrc="-2+random(0)"
6429 Generate an amplitude modulated signal:
6431 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6435 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6437 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6444 Generate a FIR coefficients using frequency sampling method.
6446 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6448 The filter accepts the following options:
6452 Set number of filter coefficents in output audio stream.
6453 Default value is 1025.
6456 Set frequency points from where magnitude and phase are set.
6457 This must be in non decreasing order, and first element must be 0, while last element
6458 must be 1. Elements are separated by white spaces.
6461 Set magnitude value for every frequency point set by @option{frequency}.
6462 Number of values must be same as number of frequency points.
6463 Values are separated by white spaces.
6466 Set phase value for every frequency point set by @option{frequency}.
6467 Number of values must be same as number of frequency points.
6468 Values are separated by white spaces.
6470 @item sample_rate, r
6471 Set sample rate, default is 44100.
6474 Set number of samples per each frame. Default is 1024.
6477 Set window function. Default is blackman.
6482 The null audio source, return unprocessed audio frames. It is mainly useful
6483 as a template and to be employed in analysis / debugging tools, or as
6484 the source for filters which ignore the input data (for example the sox
6487 This source accepts the following options:
6491 @item channel_layout, cl
6493 Specifies the channel layout, and can be either an integer or a string
6494 representing a channel layout. The default value of @var{channel_layout}
6497 Check the channel_layout_map definition in
6498 @file{libavutil/channel_layout.c} for the mapping between strings and
6499 channel layout values.
6501 @item sample_rate, r
6502 Specifies the sample rate, and defaults to 44100.
6505 Set the number of samples per requested frames.
6508 Set the duration of the sourced audio. See
6509 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6510 for the accepted syntax.
6512 If not specified, or the expressed duration is negative, the audio is
6513 supposed to be generated forever.
6516 @subsection Examples
6520 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6522 anullsrc=r=48000:cl=4
6526 Do the same operation with a more obvious syntax:
6528 anullsrc=r=48000:cl=mono
6532 All the parameters need to be explicitly defined.
6536 Synthesize a voice utterance using the libflite library.
6538 To enable compilation of this filter you need to configure FFmpeg with
6539 @code{--enable-libflite}.
6541 Note that versions of the flite library prior to 2.0 are not thread-safe.
6543 The filter accepts the following options:
6548 If set to 1, list the names of the available voices and exit
6549 immediately. Default value is 0.
6552 Set the maximum number of samples per frame. Default value is 512.
6555 Set the filename containing the text to speak.
6558 Set the text to speak.
6561 Set the voice to use for the speech synthesis. Default value is
6562 @code{kal}. See also the @var{list_voices} option.
6565 @subsection Examples
6569 Read from file @file{speech.txt}, and synthesize the text using the
6570 standard flite voice:
6572 flite=textfile=speech.txt
6576 Read the specified text selecting the @code{slt} voice:
6578 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6582 Input text to ffmpeg:
6584 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6588 Make @file{ffplay} speak the specified text, using @code{flite} and
6589 the @code{lavfi} device:
6591 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6595 For more information about libflite, check:
6596 @url{http://www.festvox.org/flite/}
6600 Generate a noise audio signal.
6602 The filter accepts the following options:
6605 @item sample_rate, r
6606 Specify the sample rate. Default value is 48000 Hz.
6609 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6613 Specify the duration of the generated audio stream. Not specifying this option
6614 results in noise with an infinite length.
6616 @item color, colour, c
6617 Specify the color of noise. Available noise colors are white, pink, brown,
6618 blue, violet and velvet. Default color is white.
6621 Specify a value used to seed the PRNG.
6624 Set the number of samples per each output frame, default is 1024.
6627 @subsection Examples
6632 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6634 anoisesrc=d=60:c=pink:r=44100:a=0.5
6640 Generate odd-tap Hilbert transform FIR coefficients.
6642 The resulting stream can be used with @ref{afir} filter for phase-shifting
6643 the signal by 90 degrees.
6645 This is used in many matrix coding schemes and for analytic signal generation.
6646 The process is often written as a multiplication by i (or j), the imaginary unit.
6648 The filter accepts the following options:
6652 @item sample_rate, s
6653 Set sample rate, default is 44100.
6656 Set length of FIR filter, default is 22051.
6659 Set number of samples per each frame.
6662 Set window function to be used when generating FIR coefficients.
6667 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6669 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Set sample rate, default is 44100.
6678 Set number of samples per each frame. Default is 1024.
6681 Set high-pass frequency. Default is 0.
6684 Set low-pass frequency. Default is 0.
6685 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6686 is higher than 0 then filter will create band-pass filter coefficients,
6687 otherwise band-reject filter coefficients.
6690 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6693 Set Kaiser window beta.
6696 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6699 Enable rounding, by default is disabled.
6702 Set number of taps for high-pass filter.
6705 Set number of taps for low-pass filter.
6710 Generate an audio signal made of a sine wave with amplitude 1/8.
6712 The audio signal is bit-exact.
6714 The filter accepts the following options:
6719 Set the carrier frequency. Default is 440 Hz.
6721 @item beep_factor, b
6722 Enable a periodic beep every second with frequency @var{beep_factor} times
6723 the carrier frequency. Default is 0, meaning the beep is disabled.
6725 @item sample_rate, r
6726 Specify the sample rate, default is 44100.
6729 Specify the duration of the generated audio stream.
6731 @item samples_per_frame
6732 Set the number of samples per output frame.
6734 The expression can contain the following constants:
6738 The (sequential) number of the output audio frame, starting from 0.
6741 The PTS (Presentation TimeStamp) of the output audio frame,
6742 expressed in @var{TB} units.
6745 The PTS of the output audio frame, expressed in seconds.
6748 The timebase of the output audio frames.
6751 Default is @code{1024}.
6754 @subsection Examples
6759 Generate a simple 440 Hz sine wave:
6765 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6769 sine=frequency=220:beep_factor=4:duration=5
6773 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6776 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6780 @c man end AUDIO SOURCES
6782 @chapter Audio Sinks
6783 @c man begin AUDIO SINKS
6785 Below is a description of the currently available audio sinks.
6787 @section abuffersink
6789 Buffer audio frames, and make them available to the end of filter chain.
6791 This sink is mainly intended for programmatic use, in particular
6792 through the interface defined in @file{libavfilter/buffersink.h}
6793 or the options system.
6795 It accepts a pointer to an AVABufferSinkContext structure, which
6796 defines the incoming buffers' formats, to be passed as the opaque
6797 parameter to @code{avfilter_init_filter} for initialization.
6800 Null audio sink; do absolutely nothing with the input audio. It is
6801 mainly useful as a template and for use in analysis / debugging
6804 @c man end AUDIO SINKS
6806 @chapter Video Filters
6807 @c man begin VIDEO FILTERS
6809 When you configure your FFmpeg build, you can disable any of the
6810 existing filters using @code{--disable-filters}.
6811 The configure output will show the video filters included in your
6814 Below is a description of the currently available video filters.
6818 Mark a region of interest in a video frame.
6820 The frame data is passed through unchanged, but metadata is attached
6821 to the frame indicating regions of interest which can affect the
6822 behaviour of later encoding. Multiple regions can be marked by
6823 applying the filter multiple times.
6827 Region distance in pixels from the left edge of the frame.
6829 Region distance in pixels from the top edge of the frame.
6831 Region width in pixels.
6833 Region height in pixels.
6835 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6836 and may contain the following variables:
6839 Width of the input frame.
6841 Height of the input frame.
6845 Quantisation offset to apply within the region.
6847 This must be a real value in the range -1 to +1. A value of zero
6848 indicates no quality change. A negative value asks for better quality
6849 (less quantisation), while a positive value asks for worse quality
6850 (greater quantisation).
6852 The range is calibrated so that the extreme values indicate the
6853 largest possible offset - if the rest of the frame is encoded with the
6854 worst possible quality, an offset of -1 indicates that this region
6855 should be encoded with the best possible quality anyway. Intermediate
6856 values are then interpolated in some codec-dependent way.
6858 For example, in 10-bit H.264 the quantisation parameter varies between
6859 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6860 this region should be encoded with a QP around one-tenth of the full
6861 range better than the rest of the frame. So, if most of the frame
6862 were to be encoded with a QP of around 30, this region would get a QP
6863 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6864 An extreme value of -1 would indicate that this region should be
6865 encoded with the best possible quality regardless of the treatment of
6866 the rest of the frame - that is, should be encoded at a QP of -12.
6868 If set to true, remove any existing regions of interest marked on the
6869 frame before adding the new one.
6872 @subsection Examples
6876 Mark the centre quarter of the frame as interesting.
6878 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6881 Mark the 100-pixel-wide region on the left edge of the frame as very
6882 uninteresting (to be encoded at much lower quality than the rest of
6885 addroi=0:0:100:ih:+1/5
6889 @section alphaextract
6891 Extract the alpha component from the input as a grayscale video. This
6892 is especially useful with the @var{alphamerge} filter.
6896 Add or replace the alpha component of the primary input with the
6897 grayscale value of a second input. This is intended for use with
6898 @var{alphaextract} to allow the transmission or storage of frame
6899 sequences that have alpha in a format that doesn't support an alpha
6902 For example, to reconstruct full frames from a normal YUV-encoded video
6903 and a separate video created with @var{alphaextract}, you might use:
6905 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6910 Amplify differences between current pixel and pixels of adjacent frames in
6911 same pixel location.
6913 This filter accepts the following options:
6917 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6918 For example radius of 3 will instruct filter to calculate average of 7 frames.
6921 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6924 Set threshold for difference amplification. Any difference greater or equal to
6925 this value will not alter source pixel. Default is 10.
6926 Allowed range is from 0 to 65535.
6929 Set tolerance for difference amplification. Any difference lower to
6930 this value will not alter source pixel. Default is 0.
6931 Allowed range is from 0 to 65535.
6934 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6935 This option controls maximum possible value that will decrease source pixel value.
6938 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6939 This option controls maximum possible value that will increase source pixel value.
6942 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6945 @subsection Commands
6947 This filter supports the following @ref{commands} that corresponds to option of same name:
6959 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6960 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6961 Substation Alpha) subtitles files.
6963 This filter accepts the following option in addition to the common options from
6964 the @ref{subtitles} filter:
6968 Set the shaping engine
6970 Available values are:
6973 The default libass shaping engine, which is the best available.
6975 Fast, font-agnostic shaper that can do only substitutions
6977 Slower shaper using OpenType for substitutions and positioning
6980 The default is @code{auto}.
6984 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6986 The filter accepts the following options:
6990 Set threshold A for 1st plane. Default is 0.02.
6991 Valid range is 0 to 0.3.
6994 Set threshold B for 1st plane. Default is 0.04.
6995 Valid range is 0 to 5.
6998 Set threshold A for 2nd plane. Default is 0.02.
6999 Valid range is 0 to 0.3.
7002 Set threshold B for 2nd plane. Default is 0.04.
7003 Valid range is 0 to 5.
7006 Set threshold A for 3rd plane. Default is 0.02.
7007 Valid range is 0 to 0.3.
7010 Set threshold B for 3rd plane. Default is 0.04.
7011 Valid range is 0 to 5.
7013 Threshold A is designed to react on abrupt changes in the input signal and
7014 threshold B is designed to react on continuous changes in the input signal.
7017 Set number of frames filter will use for averaging. Default is 9. Must be odd
7018 number in range [5, 129].
7021 Set what planes of frame filter will use for averaging. Default is all.
7024 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7025 Alternatively can be set to @code{s} serial.
7027 Parallel can be faster then serial, while other way around is never true.
7028 Parallel will abort early on first change being greater then thresholds, while serial
7029 will continue processing other side of frames if they are equal or below thresholds.
7032 @subsection Commands
7033 This filter supports same @ref{commands} as options except option @code{s}.
7034 The command accepts the same syntax of the corresponding option.
7038 Apply average blur filter.
7040 The filter accepts the following options:
7044 Set horizontal radius size.
7047 Set which planes to filter. By default all planes are filtered.
7050 Set vertical radius size, if zero it will be same as @code{sizeX}.
7051 Default is @code{0}.
7054 @subsection Commands
7055 This filter supports same commands as options.
7056 The command accepts the same syntax of the corresponding option.
7058 If the specified expression is not valid, it is kept at its current
7063 Compute the bounding box for the non-black pixels in the input frame
7066 This filter computes the bounding box containing all the pixels with a
7067 luminance value greater than the minimum allowed value.
7068 The parameters describing the bounding box are printed on the filter
7071 The filter accepts the following option:
7075 Set the minimal luminance value. Default is @code{16}.
7079 Apply bilateral filter, spatial smoothing while preserving edges.
7081 The filter accepts the following options:
7084 Set sigma of gaussian function to calculate spatial weight.
7085 Allowed range is 0 to 512. Default is 0.1.
7088 Set sigma of gaussian function to calculate range weight.
7089 Allowed range is 0 to 1. Default is 0.1.
7092 Set planes to filter. Default is first only.
7095 @section bitplanenoise
7097 Show and measure bit plane noise.
7099 The filter accepts the following options:
7103 Set which plane to analyze. Default is @code{1}.
7106 Filter out noisy pixels from @code{bitplane} set above.
7107 Default is disabled.
7110 @section blackdetect
7112 Detect video intervals that are (almost) completely black. Can be
7113 useful to detect chapter transitions, commercials, or invalid
7116 The filter outputs its detection analysis to both the log as well as
7117 frame metadata. If a black segment of at least the specified minimum
7118 duration is found, a line with the start and end timestamps as well
7119 as duration is printed to the log with level @code{info}. In addition,
7120 a log line with level @code{debug} is printed per frame showing the
7121 black amount detected for that frame.
7123 The filter also attaches metadata to the first frame of a black
7124 segment with key @code{lavfi.black_start} and to the first frame
7125 after the black segment ends with key @code{lavfi.black_end}. The
7126 value is the frame's timestamp. This metadata is added regardless
7127 of the minimum duration specified.
7129 The filter accepts the following options:
7132 @item black_min_duration, d
7133 Set the minimum detected black duration expressed in seconds. It must
7134 be a non-negative floating point number.
7136 Default value is 2.0.
7138 @item picture_black_ratio_th, pic_th
7139 Set the threshold for considering a picture "black".
7140 Express the minimum value for the ratio:
7142 @var{nb_black_pixels} / @var{nb_pixels}
7145 for which a picture is considered black.
7146 Default value is 0.98.
7148 @item pixel_black_th, pix_th
7149 Set the threshold for considering a pixel "black".
7151 The threshold expresses the maximum pixel luminance value for which a
7152 pixel is considered "black". The provided value is scaled according to
7153 the following equation:
7155 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7158 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7159 the input video format, the range is [0-255] for YUV full-range
7160 formats and [16-235] for YUV non full-range formats.
7162 Default value is 0.10.
7165 The following example sets the maximum pixel threshold to the minimum
7166 value, and detects only black intervals of 2 or more seconds:
7168 blackdetect=d=2:pix_th=0.00
7173 Detect frames that are (almost) completely black. Can be useful to
7174 detect chapter transitions or commercials. Output lines consist of
7175 the frame number of the detected frame, the percentage of blackness,
7176 the position in the file if known or -1 and the timestamp in seconds.
7178 In order to display the output lines, you need to set the loglevel at
7179 least to the AV_LOG_INFO value.
7181 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7182 The value represents the percentage of pixels in the picture that
7183 are below the threshold value.
7185 It accepts the following parameters:
7190 The percentage of the pixels that have to be below the threshold; it defaults to
7193 @item threshold, thresh
7194 The threshold below which a pixel value is considered black; it defaults to
7202 Blend two video frames into each other.
7204 The @code{blend} filter takes two input streams and outputs one
7205 stream, the first input is the "top" layer and second input is
7206 "bottom" layer. By default, the output terminates when the longest input terminates.
7208 The @code{tblend} (time blend) filter takes two consecutive frames
7209 from one single stream, and outputs the result obtained by blending
7210 the new frame on top of the old frame.
7212 A description of the accepted options follows.
7220 Set blend mode for specific pixel component or all pixel components in case
7221 of @var{all_mode}. Default value is @code{normal}.
7223 Available values for component modes are:
7265 Set blend opacity for specific pixel component or all pixel components in case
7266 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7273 Set blend expression for specific pixel component or all pixel components in case
7274 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7276 The expressions can use the following variables:
7280 The sequential number of the filtered frame, starting from @code{0}.
7284 the coordinates of the current sample
7288 the width and height of currently filtered plane
7292 Width and height scale for the plane being filtered. It is the
7293 ratio between the dimensions of the current plane to the luma plane,
7294 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7295 the luma plane and @code{0.5,0.5} for the chroma planes.
7298 Time of the current frame, expressed in seconds.
7301 Value of pixel component at current location for first video frame (top layer).
7304 Value of pixel component at current location for second video frame (bottom layer).
7308 The @code{blend} filter also supports the @ref{framesync} options.
7310 @subsection Examples
7314 Apply transition from bottom layer to top layer in first 10 seconds:
7316 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7320 Apply linear horizontal transition from top layer to bottom layer:
7322 blend=all_expr='A*(X/W)+B*(1-X/W)'
7326 Apply 1x1 checkerboard effect:
7328 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7332 Apply uncover left effect:
7334 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7338 Apply uncover down effect:
7340 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7344 Apply uncover up-left effect:
7346 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7350 Split diagonally video and shows top and bottom layer on each side:
7352 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7356 Display differences between the current and the previous frame:
7358 tblend=all_mode=grainextract
7364 Denoise frames using Block-Matching 3D algorithm.
7366 The filter accepts the following options.
7370 Set denoising strength. Default value is 1.
7371 Allowed range is from 0 to 999.9.
7372 The denoising algorithm is very sensitive to sigma, so adjust it
7373 according to the source.
7376 Set local patch size. This sets dimensions in 2D.
7379 Set sliding step for processing blocks. Default value is 4.
7380 Allowed range is from 1 to 64.
7381 Smaller values allows processing more reference blocks and is slower.
7384 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7385 When set to 1, no block matching is done. Larger values allows more blocks
7387 Allowed range is from 1 to 256.
7390 Set radius for search block matching. Default is 9.
7391 Allowed range is from 1 to INT32_MAX.
7394 Set step between two search locations for block matching. Default is 1.
7395 Allowed range is from 1 to 64. Smaller is slower.
7398 Set threshold of mean square error for block matching. Valid range is 0 to
7402 Set thresholding parameter for hard thresholding in 3D transformed domain.
7403 Larger values results in stronger hard-thresholding filtering in frequency
7407 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7408 Default is @code{basic}.
7411 If enabled, filter will use 2nd stream for block matching.
7412 Default is disabled for @code{basic} value of @var{estim} option,
7413 and always enabled if value of @var{estim} is @code{final}.
7416 Set planes to filter. Default is all available except alpha.
7419 @subsection Examples
7423 Basic filtering with bm3d:
7425 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7429 Same as above, but filtering only luma:
7431 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7435 Same as above, but with both estimation modes:
7437 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
7441 Same as above, but prefilter with @ref{nlmeans} filter instead:
7443 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
7449 Apply a boxblur algorithm to the input video.
7451 It accepts the following parameters:
7455 @item luma_radius, lr
7456 @item luma_power, lp
7457 @item chroma_radius, cr
7458 @item chroma_power, cp
7459 @item alpha_radius, ar
7460 @item alpha_power, ap
7464 A description of the accepted options follows.
7467 @item luma_radius, lr
7468 @item chroma_radius, cr
7469 @item alpha_radius, ar
7470 Set an expression for the box radius in pixels used for blurring the
7471 corresponding input plane.
7473 The radius value must be a non-negative number, and must not be
7474 greater than the value of the expression @code{min(w,h)/2} for the
7475 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7478 Default value for @option{luma_radius} is "2". If not specified,
7479 @option{chroma_radius} and @option{alpha_radius} default to the
7480 corresponding value set for @option{luma_radius}.
7482 The expressions can contain the following constants:
7486 The input width and height in pixels.
7490 The input chroma image width and height in pixels.
7494 The horizontal and vertical chroma subsample values. For example, for the
7495 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7498 @item luma_power, lp
7499 @item chroma_power, cp
7500 @item alpha_power, ap
7501 Specify how many times the boxblur filter is applied to the
7502 corresponding plane.
7504 Default value for @option{luma_power} is 2. If not specified,
7505 @option{chroma_power} and @option{alpha_power} default to the
7506 corresponding value set for @option{luma_power}.
7508 A value of 0 will disable the effect.
7511 @subsection Examples
7515 Apply a boxblur filter with the luma, chroma, and alpha radii
7518 boxblur=luma_radius=2:luma_power=1
7523 Set the luma radius to 2, and alpha and chroma radius to 0:
7525 boxblur=2:1:cr=0:ar=0
7529 Set the luma and chroma radii to a fraction of the video dimension:
7531 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7537 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7538 Deinterlacing Filter").
7540 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7541 interpolation algorithms.
7542 It accepts the following parameters:
7546 The interlacing mode to adopt. It accepts one of the following values:
7550 Output one frame for each frame.
7552 Output one frame for each field.
7555 The default value is @code{send_field}.
7558 The picture field parity assumed for the input interlaced video. It accepts one
7559 of the following values:
7563 Assume the top field is first.
7565 Assume the bottom field is first.
7567 Enable automatic detection of field parity.
7570 The default value is @code{auto}.
7571 If the interlacing is unknown or the decoder does not export this information,
7572 top field first will be assumed.
7575 Specify which frames to deinterlace. Accepts one of the following
7580 Deinterlace all frames.
7582 Only deinterlace frames marked as interlaced.
7585 The default value is @code{all}.
7590 Apply Contrast Adaptive Sharpen filter to video stream.
7592 The filter accepts the following options:
7596 Set the sharpening strength. Default value is 0.
7599 Set planes to filter. Default value is to filter all
7600 planes except alpha plane.
7604 Remove all color information for all colors except for certain one.
7606 The filter accepts the following options:
7610 The color which will not be replaced with neutral chroma.
7613 Similarity percentage with the above color.
7614 0.01 matches only the exact key color, while 1.0 matches everything.
7618 0.0 makes pixels either fully gray, or not gray at all.
7619 Higher values result in more preserved color.
7622 Signals that the color passed is already in YUV instead of RGB.
7624 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7625 This can be used to pass exact YUV values as hexadecimal numbers.
7628 @subsection Commands
7629 This filter supports same @ref{commands} as options.
7630 The command accepts the same syntax of the corresponding option.
7632 If the specified expression is not valid, it is kept at its current
7636 YUV colorspace color/chroma keying.
7638 The filter accepts the following options:
7642 The color which will be replaced with transparency.
7645 Similarity percentage with the key color.
7647 0.01 matches only the exact key color, while 1.0 matches everything.
7652 0.0 makes pixels either fully transparent, or not transparent at all.
7654 Higher values result in semi-transparent pixels, with a higher transparency
7655 the more similar the pixels color is to the key color.
7658 Signals that the color passed is already in YUV instead of RGB.
7660 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7661 This can be used to pass exact YUV values as hexadecimal numbers.
7664 @subsection Commands
7665 This filter supports same @ref{commands} as options.
7666 The command accepts the same syntax of the corresponding option.
7668 If the specified expression is not valid, it is kept at its current
7671 @subsection Examples
7675 Make every green pixel in the input image transparent:
7677 ffmpeg -i input.png -vf chromakey=green out.png
7681 Overlay a greenscreen-video on top of a static black background.
7683 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
7688 Reduce chrominance noise.
7690 The filter accepts the following options:
7694 Set threshold for averaging chrominance values.
7695 Sum of absolute difference of U and V pixel components or current
7696 pixel and neighbour pixels lower than this threshold will be used in
7697 averaging. Luma component is left unchanged and is copied to output.
7698 Default value is 30. Allowed range is from 1 to 200.
7701 Set horizontal radius of rectangle used for averaging.
7702 Allowed range is from 1 to 100. Default value is 5.
7705 Set vertical radius of rectangle used for averaging.
7706 Allowed range is from 1 to 100. Default value is 5.
7709 Set horizontal step when averaging. Default value is 1.
7710 Allowed range is from 1 to 50.
7711 Mostly useful to speed-up filtering.
7714 Set vertical step when averaging. Default value is 1.
7715 Allowed range is from 1 to 50.
7716 Mostly useful to speed-up filtering.
7719 @subsection Commands
7720 This filter supports same @ref{commands} as options.
7721 The command accepts the same syntax of the corresponding option.
7723 @section chromashift
7724 Shift chroma pixels horizontally and/or vertically.
7726 The filter accepts the following options:
7729 Set amount to shift chroma-blue horizontally.
7731 Set amount to shift chroma-blue vertically.
7733 Set amount to shift chroma-red horizontally.
7735 Set amount to shift chroma-red vertically.
7737 Set edge mode, can be @var{smear}, default, or @var{warp}.
7740 @subsection Commands
7742 This filter supports the all above options as @ref{commands}.
7746 Display CIE color diagram with pixels overlaid onto it.
7748 The filter accepts the following options:
7763 @item uhdtv, rec2020
7777 Set what gamuts to draw.
7779 See @code{system} option for available values.
7782 Set ciescope size, by default set to 512.
7785 Set intensity used to map input pixel values to CIE diagram.
7788 Set contrast used to draw tongue colors that are out of active color system gamut.
7791 Correct gamma displayed on scope, by default enabled.
7794 Show white point on CIE diagram, by default disabled.
7797 Set input gamma. Used only with XYZ input color space.
7802 Visualize information exported by some codecs.
7804 Some codecs can export information through frames using side-data or other
7805 means. For example, some MPEG based codecs export motion vectors through the
7806 @var{export_mvs} flag in the codec @option{flags2} option.
7808 The filter accepts the following option:
7812 Set motion vectors to visualize.
7814 Available flags for @var{mv} are:
7818 forward predicted MVs of P-frames
7820 forward predicted MVs of B-frames
7822 backward predicted MVs of B-frames
7826 Display quantization parameters using the chroma planes.
7829 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7831 Available flags for @var{mv_type} are:
7835 forward predicted MVs
7837 backward predicted MVs
7840 @item frame_type, ft
7841 Set frame type to visualize motion vectors of.
7843 Available flags for @var{frame_type} are:
7847 intra-coded frames (I-frames)
7849 predicted frames (P-frames)
7851 bi-directionally predicted frames (B-frames)
7855 @subsection Examples
7859 Visualize forward predicted MVs of all frames using @command{ffplay}:
7861 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7865 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7867 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7871 @section colorbalance
7872 Modify intensity of primary colors (red, green and blue) of input frames.
7874 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7875 regions for the red-cyan, green-magenta or blue-yellow balance.
7877 A positive adjustment value shifts the balance towards the primary color, a negative
7878 value towards the complementary color.
7880 The filter accepts the following options:
7886 Adjust red, green and blue shadows (darkest pixels).
7891 Adjust red, green and blue midtones (medium pixels).
7896 Adjust red, green and blue highlights (brightest pixels).
7898 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7901 Preserve lightness when changing color balance. Default is disabled.
7904 @subsection Examples
7908 Add red color cast to shadows:
7914 @subsection Commands
7916 This filter supports the all above options as @ref{commands}.
7918 @section colorchannelmixer
7920 Adjust video input frames by re-mixing color channels.
7922 This filter modifies a color channel by adding the values associated to
7923 the other channels of the same pixels. For example if the value to
7924 modify is red, the output value will be:
7926 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7929 The filter accepts the following options:
7936 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7937 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7943 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7944 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7950 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7951 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7957 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7958 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7960 Allowed ranges for options are @code{[-2.0, 2.0]}.
7963 @subsection Examples
7967 Convert source to grayscale:
7969 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7972 Simulate sepia tones:
7974 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7978 @subsection Commands
7980 This filter supports the all above options as @ref{commands}.
7983 RGB colorspace color keying.
7985 The filter accepts the following options:
7989 The color which will be replaced with transparency.
7992 Similarity percentage with the key color.
7994 0.01 matches only the exact key color, while 1.0 matches everything.
7999 0.0 makes pixels either fully transparent, or not transparent at all.
8001 Higher values result in semi-transparent pixels, with a higher transparency
8002 the more similar the pixels color is to the key color.
8005 @subsection Examples
8009 Make every green pixel in the input image transparent:
8011 ffmpeg -i input.png -vf colorkey=green out.png
8015 Overlay a greenscreen-video on top of a static background image.
8017 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
8021 @subsection Commands
8022 This filter supports same @ref{commands} as options.
8023 The command accepts the same syntax of the corresponding option.
8025 If the specified expression is not valid, it is kept at its current
8029 Remove all color information for all RGB colors except for certain one.
8031 The filter accepts the following options:
8035 The color which will not be replaced with neutral gray.
8038 Similarity percentage with the above color.
8039 0.01 matches only the exact key color, while 1.0 matches everything.
8042 Blend percentage. 0.0 makes pixels fully gray.
8043 Higher values result in more preserved color.
8046 @subsection Commands
8047 This filter supports same @ref{commands} as options.
8048 The command accepts the same syntax of the corresponding option.
8050 If the specified expression is not valid, it is kept at its current
8053 @section colorlevels
8055 Adjust video input frames using levels.
8057 The filter accepts the following options:
8064 Adjust red, green, blue and alpha input black point.
8065 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8071 Adjust red, green, blue and alpha input white point.
8072 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8074 Input levels are used to lighten highlights (bright tones), darken shadows
8075 (dark tones), change the balance of bright and dark tones.
8081 Adjust red, green, blue and alpha output black point.
8082 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8088 Adjust red, green, blue and alpha output white point.
8089 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8091 Output levels allows manual selection of a constrained output level range.
8094 @subsection Examples
8098 Make video output darker:
8100 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8106 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8110 Make video output lighter:
8112 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8116 Increase brightness:
8118 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8122 @subsection Commands
8124 This filter supports the all above options as @ref{commands}.
8126 @section colormatrix
8128 Convert color matrix.
8130 The filter accepts the following options:
8135 Specify the source and destination color matrix. Both values must be
8138 The accepted values are:
8166 For example to convert from BT.601 to SMPTE-240M, use the command:
8168 colormatrix=bt601:smpte240m
8173 Convert colorspace, transfer characteristics or color primaries.
8174 Input video needs to have an even size.
8176 The filter accepts the following options:
8181 Specify all color properties at once.
8183 The accepted values are:
8213 Specify output colorspace.
8215 The accepted values are:
8224 BT.470BG or BT.601-6 625
8227 SMPTE-170M or BT.601-6 525
8236 BT.2020 with non-constant luminance
8242 Specify output transfer characteristics.
8244 The accepted values are:
8256 Constant gamma of 2.2
8259 Constant gamma of 2.8
8262 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8280 BT.2020 for 10-bits content
8283 BT.2020 for 12-bits content
8289 Specify output color primaries.
8291 The accepted values are:
8300 BT.470BG or BT.601-6 625
8303 SMPTE-170M or BT.601-6 525
8327 Specify output color range.
8329 The accepted values are:
8332 TV (restricted) range
8335 MPEG (restricted) range
8346 Specify output color format.
8348 The accepted values are:
8351 YUV 4:2:0 planar 8-bits
8354 YUV 4:2:0 planar 10-bits
8357 YUV 4:2:0 planar 12-bits
8360 YUV 4:2:2 planar 8-bits
8363 YUV 4:2:2 planar 10-bits
8366 YUV 4:2:2 planar 12-bits
8369 YUV 4:4:4 planar 8-bits
8372 YUV 4:4:4 planar 10-bits
8375 YUV 4:4:4 planar 12-bits
8380 Do a fast conversion, which skips gamma/primary correction. This will take
8381 significantly less CPU, but will be mathematically incorrect. To get output
8382 compatible with that produced by the colormatrix filter, use fast=1.
8385 Specify dithering mode.
8387 The accepted values are:
8393 Floyd-Steinberg dithering
8397 Whitepoint adaptation mode.
8399 The accepted values are:
8402 Bradford whitepoint adaptation
8405 von Kries whitepoint adaptation
8408 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8412 Override all input properties at once. Same accepted values as @ref{all}.
8415 Override input colorspace. Same accepted values as @ref{space}.
8418 Override input color primaries. Same accepted values as @ref{primaries}.
8421 Override input transfer characteristics. Same accepted values as @ref{trc}.
8424 Override input color range. Same accepted values as @ref{range}.
8428 The filter converts the transfer characteristics, color space and color
8429 primaries to the specified user values. The output value, if not specified,
8430 is set to a default value based on the "all" property. If that property is
8431 also not specified, the filter will log an error. The output color range and
8432 format default to the same value as the input color range and format. The
8433 input transfer characteristics, color space, color primaries and color range
8434 should be set on the input data. If any of these are missing, the filter will
8435 log an error and no conversion will take place.
8437 For example to convert the input to SMPTE-240M, use the command:
8439 colorspace=smpte240m
8442 @section convolution
8444 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8446 The filter accepts the following options:
8453 Set matrix for each plane.
8454 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8455 and from 1 to 49 odd number of signed integers in @var{row} mode.
8461 Set multiplier for calculated value for each plane.
8462 If unset or 0, it will be sum of all matrix elements.
8468 Set bias for each plane. This value is added to the result of the multiplication.
8469 Useful for making the overall image brighter or darker. Default is 0.0.
8475 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8476 Default is @var{square}.
8479 @subsection Examples
8485 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"
8491 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"
8497 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"
8503 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"
8507 Apply laplacian edge detector which includes diagonals:
8509 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"
8515 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"
8521 Apply 2D convolution of video stream in frequency domain using second stream
8524 The filter accepts the following options:
8528 Set which planes to process.
8531 Set which impulse video frames will be processed, can be @var{first}
8532 or @var{all}. Default is @var{all}.
8535 The @code{convolve} filter also supports the @ref{framesync} options.
8539 Copy the input video source unchanged to the output. This is mainly useful for
8544 Video filtering on GPU using Apple's CoreImage API on OSX.
8546 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8547 processed by video hardware. However, software-based OpenGL implementations
8548 exist which means there is no guarantee for hardware processing. It depends on
8551 There are many filters and image generators provided by Apple that come with a
8552 large variety of options. The filter has to be referenced by its name along
8555 The coreimage filter accepts the following options:
8558 List all available filters and generators along with all their respective
8559 options as well as possible minimum and maximum values along with the default
8566 Specify all filters by their respective name and options.
8567 Use @var{list_filters} to determine all valid filter names and options.
8568 Numerical options are specified by a float value and are automatically clamped
8569 to their respective value range. Vector and color options have to be specified
8570 by a list of space separated float values. Character escaping has to be done.
8571 A special option name @code{default} is available to use default options for a
8574 It is required to specify either @code{default} or at least one of the filter options.
8575 All omitted options are used with their default values.
8576 The syntax of the filter string is as follows:
8578 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8582 Specify a rectangle where the output of the filter chain is copied into the
8583 input image. It is given by a list of space separated float values:
8585 output_rect=x\ y\ width\ height
8587 If not given, the output rectangle equals the dimensions of the input image.
8588 The output rectangle is automatically cropped at the borders of the input
8589 image. Negative values are valid for each component.
8591 output_rect=25\ 25\ 100\ 100
8595 Several filters can be chained for successive processing without GPU-HOST
8596 transfers allowing for fast processing of complex filter chains.
8597 Currently, only filters with zero (generators) or exactly one (filters) input
8598 image and one output image are supported. Also, transition filters are not yet
8601 Some filters generate output images with additional padding depending on the
8602 respective filter kernel. The padding is automatically removed to ensure the
8603 filter output has the same size as the input image.
8605 For image generators, the size of the output image is determined by the
8606 previous output image of the filter chain or the input image of the whole
8607 filterchain, respectively. The generators do not use the pixel information of
8608 this image to generate their output. However, the generated output is
8609 blended onto this image, resulting in partial or complete coverage of the
8612 The @ref{coreimagesrc} video source can be used for generating input images
8613 which are directly fed into the filter chain. By using it, providing input
8614 images by another video source or an input video is not required.
8616 @subsection Examples
8621 List all filters available:
8623 coreimage=list_filters=true
8627 Use the CIBoxBlur filter with default options to blur an image:
8629 coreimage=filter=CIBoxBlur@@default
8633 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8634 its center at 100x100 and a radius of 50 pixels:
8636 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8640 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8641 given as complete and escaped command-line for Apple's standard bash shell:
8643 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8649 Cover a rectangular object
8651 It accepts the following options:
8655 Filepath of the optional cover image, needs to be in yuv420.
8660 It accepts the following values:
8663 cover it by the supplied image
8665 cover it by interpolating the surrounding pixels
8668 Default value is @var{blur}.
8671 @subsection Examples
8675 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8677 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8683 Crop the input video to given dimensions.
8685 It accepts the following parameters:
8689 The width of the output video. It defaults to @code{iw}.
8690 This expression is evaluated only once during the filter
8691 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8694 The height of the output video. It defaults to @code{ih}.
8695 This expression is evaluated only once during the filter
8696 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8699 The horizontal position, in the input video, of the left edge of the output
8700 video. It defaults to @code{(in_w-out_w)/2}.
8701 This expression is evaluated per-frame.
8704 The vertical position, in the input video, of the top edge of the output video.
8705 It defaults to @code{(in_h-out_h)/2}.
8706 This expression is evaluated per-frame.
8709 If set to 1 will force the output display aspect ratio
8710 to be the same of the input, by changing the output sample aspect
8711 ratio. It defaults to 0.
8714 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8715 width/height/x/y as specified and will not be rounded to nearest smaller value.
8719 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8720 expressions containing the following constants:
8725 The computed values for @var{x} and @var{y}. They are evaluated for
8730 The input width and height.
8734 These are the same as @var{in_w} and @var{in_h}.
8738 The output (cropped) width and height.
8742 These are the same as @var{out_w} and @var{out_h}.
8745 same as @var{iw} / @var{ih}
8748 input sample aspect ratio
8751 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8755 horizontal and vertical chroma subsample values. For example for the
8756 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8759 The number of the input frame, starting from 0.
8762 the position in the file of the input frame, NAN if unknown
8765 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8769 The expression for @var{out_w} may depend on the value of @var{out_h},
8770 and the expression for @var{out_h} may depend on @var{out_w}, but they
8771 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8772 evaluated after @var{out_w} and @var{out_h}.
8774 The @var{x} and @var{y} parameters specify the expressions for the
8775 position of the top-left corner of the output (non-cropped) area. They
8776 are evaluated for each frame. If the evaluated value is not valid, it
8777 is approximated to the nearest valid value.
8779 The expression for @var{x} may depend on @var{y}, and the expression
8780 for @var{y} may depend on @var{x}.
8782 @subsection Examples
8786 Crop area with size 100x100 at position (12,34).
8791 Using named options, the example above becomes:
8793 crop=w=100:h=100:x=12:y=34
8797 Crop the central input area with size 100x100:
8803 Crop the central input area with size 2/3 of the input video:
8805 crop=2/3*in_w:2/3*in_h
8809 Crop the input video central square:
8816 Delimit the rectangle with the top-left corner placed at position
8817 100:100 and the right-bottom corner corresponding to the right-bottom
8818 corner of the input image.
8820 crop=in_w-100:in_h-100:100:100
8824 Crop 10 pixels from the left and right borders, and 20 pixels from
8825 the top and bottom borders
8827 crop=in_w-2*10:in_h-2*20
8831 Keep only the bottom right quarter of the input image:
8833 crop=in_w/2:in_h/2:in_w/2:in_h/2
8837 Crop height for getting Greek harmony:
8839 crop=in_w:1/PHI*in_w
8843 Apply trembling effect:
8845 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)
8849 Apply erratic camera effect depending on timestamp:
8851 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)"
8855 Set x depending on the value of y:
8857 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8861 @subsection Commands
8863 This filter supports the following commands:
8869 Set width/height of the output video and the horizontal/vertical position
8871 The command accepts the same syntax of the corresponding option.
8873 If the specified expression is not valid, it is kept at its current
8879 Auto-detect the crop size.
8881 It calculates the necessary cropping parameters and prints the
8882 recommended parameters via the logging system. The detected dimensions
8883 correspond to the non-black area of the input video.
8885 It accepts the following parameters:
8890 Set higher black value threshold, which can be optionally specified
8891 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8892 value greater to the set value is considered non-black. It defaults to 24.
8893 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8894 on the bitdepth of the pixel format.
8897 The value which the width/height should be divisible by. It defaults to
8898 16. The offset is automatically adjusted to center the video. Use 2 to
8899 get only even dimensions (needed for 4:2:2 video). 16 is best when
8900 encoding to most video codecs.
8903 Set the number of initial frames for which evaluation is skipped.
8904 Default is 2. Range is 0 to INT_MAX.
8906 @item reset_count, reset
8907 Set the counter that determines after how many frames cropdetect will
8908 reset the previously detected largest video area and start over to
8909 detect the current optimal crop area. Default value is 0.
8911 This can be useful when channel logos distort the video area. 0
8912 indicates 'never reset', and returns the largest area encountered during
8919 Delay video filtering until a given wallclock timestamp. The filter first
8920 passes on @option{preroll} amount of frames, then it buffers at most
8921 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8922 it forwards the buffered frames and also any subsequent frames coming in its
8925 The filter can be used synchronize the output of multiple ffmpeg processes for
8926 realtime output devices like decklink. By putting the delay in the filtering
8927 chain and pre-buffering frames the process can pass on data to output almost
8928 immediately after the target wallclock timestamp is reached.
8930 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8936 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8939 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8942 The maximum duration of content to buffer before waiting for the cue expressed
8943 in seconds. Default is 0.
8950 Apply color adjustments using curves.
8952 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8953 component (red, green and blue) has its values defined by @var{N} key points
8954 tied from each other using a smooth curve. The x-axis represents the pixel
8955 values from the input frame, and the y-axis the new pixel values to be set for
8958 By default, a component curve is defined by the two points @var{(0;0)} and
8959 @var{(1;1)}. This creates a straight line where each original pixel value is
8960 "adjusted" to its own value, which means no change to the image.
8962 The filter allows you to redefine these two points and add some more. A new
8963 curve (using a natural cubic spline interpolation) will be define to pass
8964 smoothly through all these new coordinates. The new defined points needs to be
8965 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8966 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8967 the vector spaces, the values will be clipped accordingly.
8969 The filter accepts the following options:
8973 Select one of the available color presets. This option can be used in addition
8974 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8975 options takes priority on the preset values.
8976 Available presets are:
8979 @item color_negative
8982 @item increase_contrast
8984 @item linear_contrast
8985 @item medium_contrast
8987 @item strong_contrast
8990 Default is @code{none}.
8992 Set the master key points. These points will define a second pass mapping. It
8993 is sometimes called a "luminance" or "value" mapping. It can be used with
8994 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8995 post-processing LUT.
8997 Set the key points for the red component.
8999 Set the key points for the green component.
9001 Set the key points for the blue component.
9003 Set the key points for all components (not including master).
9004 Can be used in addition to the other key points component
9005 options. In this case, the unset component(s) will fallback on this
9006 @option{all} setting.
9008 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9010 Save Gnuplot script of the curves in specified file.
9013 To avoid some filtergraph syntax conflicts, each key points list need to be
9014 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9016 @subsection Examples
9020 Increase slightly the middle level of blue:
9022 curves=blue='0/0 0.5/0.58 1/1'
9028 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'
9030 Here we obtain the following coordinates for each components:
9033 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9035 @code{(0;0) (0.50;0.48) (1;1)}
9037 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9041 The previous example can also be achieved with the associated built-in preset:
9043 curves=preset=vintage
9053 Use a Photoshop preset and redefine the points of the green component:
9055 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9059 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9060 and @command{gnuplot}:
9062 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9063 gnuplot -p /tmp/curves.plt
9069 Video data analysis filter.
9071 This filter shows hexadecimal pixel values of part of video.
9073 The filter accepts the following options:
9077 Set output video size.
9080 Set x offset from where to pick pixels.
9083 Set y offset from where to pick pixels.
9086 Set scope mode, can be one of the following:
9089 Draw hexadecimal pixel values with white color on black background.
9092 Draw hexadecimal pixel values with input video pixel color on black
9096 Draw hexadecimal pixel values on color background picked from input video,
9097 the text color is picked in such way so its always visible.
9101 Draw rows and columns numbers on left and top of video.
9104 Set background opacity.
9107 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9111 Apply Directional blur filter.
9113 The filter accepts the following options:
9117 Set angle of directional blur. Default is @code{45}.
9120 Set radius of directional blur. Default is @code{5}.
9123 Set which planes to filter. By default all planes are filtered.
9126 @subsection Commands
9127 This filter supports same @ref{commands} as options.
9128 The command accepts the same syntax of the corresponding option.
9130 If the specified expression is not valid, it is kept at its current
9135 Denoise frames using 2D DCT (frequency domain filtering).
9137 This filter is not designed for real time.
9139 The filter accepts the following options:
9143 Set the noise sigma constant.
9145 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9146 coefficient (absolute value) below this threshold with be dropped.
9148 If you need a more advanced filtering, see @option{expr}.
9150 Default is @code{0}.
9153 Set number overlapping pixels for each block. Since the filter can be slow, you
9154 may want to reduce this value, at the cost of a less effective filter and the
9155 risk of various artefacts.
9157 If the overlapping value doesn't permit processing the whole input width or
9158 height, a warning will be displayed and according borders won't be denoised.
9160 Default value is @var{blocksize}-1, which is the best possible setting.
9163 Set the coefficient factor expression.
9165 For each coefficient of a DCT block, this expression will be evaluated as a
9166 multiplier value for the coefficient.
9168 If this is option is set, the @option{sigma} option will be ignored.
9170 The absolute value of the coefficient can be accessed through the @var{c}
9174 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9175 @var{blocksize}, which is the width and height of the processed blocks.
9177 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9178 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9179 on the speed processing. Also, a larger block size does not necessarily means a
9183 @subsection Examples
9185 Apply a denoise with a @option{sigma} of @code{4.5}:
9190 The same operation can be achieved using the expression system:
9192 dctdnoiz=e='gte(c, 4.5*3)'
9195 Violent denoise using a block size of @code{16x16}:
9202 Remove banding artifacts from input video.
9203 It works by replacing banded pixels with average value of referenced pixels.
9205 The filter accepts the following options:
9212 Set banding detection threshold for each plane. Default is 0.02.
9213 Valid range is 0.00003 to 0.5.
9214 If difference between current pixel and reference pixel is less than threshold,
9215 it will be considered as banded.
9218 Banding detection range in pixels. Default is 16. If positive, random number
9219 in range 0 to set value will be used. If negative, exact absolute value
9221 The range defines square of four pixels around current pixel.
9224 Set direction in radians from which four pixel will be compared. If positive,
9225 random direction from 0 to set direction will be picked. If negative, exact of
9226 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9227 will pick only pixels on same row and -PI/2 will pick only pixels on same
9231 If enabled, current pixel is compared with average value of all four
9232 surrounding pixels. The default is enabled. If disabled current pixel is
9233 compared with all four surrounding pixels. The pixel is considered banded
9234 if only all four differences with surrounding pixels are less than threshold.
9237 If enabled, current pixel is changed if and only if all pixel components are banded,
9238 e.g. banding detection threshold is triggered for all color components.
9239 The default is disabled.
9244 Remove blocking artifacts from input video.
9246 The filter accepts the following options:
9250 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9251 This controls what kind of deblocking is applied.
9254 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9260 Set blocking detection thresholds. Allowed range is 0 to 1.
9261 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9262 Using higher threshold gives more deblocking strength.
9263 Setting @var{alpha} controls threshold detection at exact edge of block.
9264 Remaining options controls threshold detection near the edge. Each one for
9265 below/above or left/right. Setting any of those to @var{0} disables
9269 Set planes to filter. Default is to filter all available planes.
9272 @subsection Examples
9276 Deblock using weak filter and block size of 4 pixels.
9278 deblock=filter=weak:block=4
9282 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9283 deblocking more edges.
9285 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9289 Similar as above, but filter only first plane.
9291 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9295 Similar as above, but filter only second and third plane.
9297 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9304 Drop duplicated frames at regular intervals.
9306 The filter accepts the following options:
9310 Set the number of frames from which one will be dropped. Setting this to
9311 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9312 Default is @code{5}.
9315 Set the threshold for duplicate detection. If the difference metric for a frame
9316 is less than or equal to this value, then it is declared as duplicate. Default
9320 Set scene change threshold. Default is @code{15}.
9324 Set the size of the x and y-axis blocks used during metric calculations.
9325 Larger blocks give better noise suppression, but also give worse detection of
9326 small movements. Must be a power of two. Default is @code{32}.
9329 Mark main input as a pre-processed input and activate clean source input
9330 stream. This allows the input to be pre-processed with various filters to help
9331 the metrics calculation while keeping the frame selection lossless. When set to
9332 @code{1}, the first stream is for the pre-processed input, and the second
9333 stream is the clean source from where the kept frames are chosen. Default is
9337 Set whether or not chroma is considered in the metric calculations. Default is
9343 Apply 2D deconvolution of video stream in frequency domain using second stream
9346 The filter accepts the following options:
9350 Set which planes to process.
9353 Set which impulse video frames will be processed, can be @var{first}
9354 or @var{all}. Default is @var{all}.
9357 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9358 and height are not same and not power of 2 or if stream prior to convolving
9362 The @code{deconvolve} filter also supports the @ref{framesync} options.
9366 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9368 It accepts the following options:
9372 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9373 @var{rainbows} for cross-color reduction.
9376 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9379 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9382 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9385 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9390 Apply deflate effect to the video.
9392 This filter replaces the pixel by the local(3x3) average by taking into account
9393 only values lower than the pixel.
9395 It accepts the following options:
9402 Limit the maximum change for each plane, default is 65535.
9403 If 0, plane will remain unchanged.
9406 @subsection Commands
9408 This filter supports the all above options as @ref{commands}.
9412 Remove temporal frame luminance variations.
9414 It accepts the following options:
9418 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9421 Set averaging mode to smooth temporal luminance variations.
9423 Available values are:
9448 Do not actually modify frame. Useful when one only wants metadata.
9453 Remove judder produced by partially interlaced telecined content.
9455 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9456 source was partially telecined content then the output of @code{pullup,dejudder}
9457 will have a variable frame rate. May change the recorded frame rate of the
9458 container. Aside from that change, this filter will not affect constant frame
9461 The option available in this filter is:
9465 Specify the length of the window over which the judder repeats.
9467 Accepts any integer greater than 1. Useful values are:
9471 If the original was telecined from 24 to 30 fps (Film to NTSC).
9474 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9477 If a mixture of the two.
9480 The default is @samp{4}.
9485 Suppress a TV station logo by a simple interpolation of the surrounding
9486 pixels. Just set a rectangle covering the logo and watch it disappear
9487 (and sometimes something even uglier appear - your mileage may vary).
9489 It accepts the following parameters:
9494 Specify the top left corner coordinates of the logo. They must be
9499 Specify the width and height of the logo to clear. They must be
9503 Specify the thickness of the fuzzy edge of the rectangle (added to
9504 @var{w} and @var{h}). The default value is 1. This option is
9505 deprecated, setting higher values should no longer be necessary and
9509 When set to 1, a green rectangle is drawn on the screen to simplify
9510 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9511 The default value is 0.
9513 The rectangle is drawn on the outermost pixels which will be (partly)
9514 replaced with interpolated values. The values of the next pixels
9515 immediately outside this rectangle in each direction will be used to
9516 compute the interpolated pixel values inside the rectangle.
9520 @subsection Examples
9524 Set a rectangle covering the area with top left corner coordinates 0,0
9525 and size 100x77, and a band of size 10:
9527 delogo=x=0:y=0:w=100:h=77:band=10
9535 Remove the rain in the input image/video by applying the derain methods based on
9536 convolutional neural networks. Supported models:
9540 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9541 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9544 Training as well as model generation scripts are provided in
9545 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9547 Native model files (.model) can be generated from TensorFlow model
9548 files (.pb) by using tools/python/convert.py
9550 The filter accepts the following options:
9554 Specify which filter to use. This option accepts the following values:
9558 Derain filter. To conduct derain filter, you need to use a derain model.
9561 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9563 Default value is @samp{derain}.
9566 Specify which DNN backend to use for model loading and execution. This option accepts
9567 the following values:
9571 Native implementation of DNN loading and execution.
9574 TensorFlow backend. To enable this backend you
9575 need to install the TensorFlow for C library (see
9576 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9577 @code{--enable-libtensorflow}
9579 Default value is @samp{native}.
9582 Set path to model file specifying network architecture and its parameters.
9583 Note that different backends use different file formats. TensorFlow and native
9584 backend can load files for only its format.
9587 It can also be finished with @ref{dnn_processing} filter.
9591 Attempt to fix small changes in horizontal and/or vertical shift. This
9592 filter helps remove camera shake from hand-holding a camera, bumping a
9593 tripod, moving on a vehicle, etc.
9595 The filter accepts the following options:
9603 Specify a rectangular area where to limit the search for motion
9605 If desired the search for motion vectors can be limited to a
9606 rectangular area of the frame defined by its top left corner, width
9607 and height. These parameters have the same meaning as the drawbox
9608 filter which can be used to visualise the position of the bounding
9611 This is useful when simultaneous movement of subjects within the frame
9612 might be confused for camera motion by the motion vector search.
9614 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9615 then the full frame is used. This allows later options to be set
9616 without specifying the bounding box for the motion vector search.
9618 Default - search the whole frame.
9622 Specify the maximum extent of movement in x and y directions in the
9623 range 0-64 pixels. Default 16.
9626 Specify how to generate pixels to fill blanks at the edge of the
9627 frame. Available values are:
9630 Fill zeroes at blank locations
9632 Original image at blank locations
9634 Extruded edge value at blank locations
9636 Mirrored edge at blank locations
9638 Default value is @samp{mirror}.
9641 Specify the blocksize to use for motion search. Range 4-128 pixels,
9645 Specify the contrast threshold for blocks. Only blocks with more than
9646 the specified contrast (difference between darkest and lightest
9647 pixels) will be considered. Range 1-255, default 125.
9650 Specify the search strategy. Available values are:
9653 Set exhaustive search
9655 Set less exhaustive search.
9657 Default value is @samp{exhaustive}.
9660 If set then a detailed log of the motion search is written to the
9667 Remove unwanted contamination of foreground colors, caused by reflected color of
9668 greenscreen or bluescreen.
9670 This filter accepts the following options:
9674 Set what type of despill to use.
9677 Set how spillmap will be generated.
9680 Set how much to get rid of still remaining spill.
9683 Controls amount of red in spill area.
9686 Controls amount of green in spill area.
9687 Should be -1 for greenscreen.
9690 Controls amount of blue in spill area.
9691 Should be -1 for bluescreen.
9694 Controls brightness of spill area, preserving colors.
9697 Modify alpha from generated spillmap.
9700 @subsection Commands
9702 This filter supports the all above options as @ref{commands}.
9706 Apply an exact inverse of the telecine operation. It requires a predefined
9707 pattern specified using the pattern option which must be the same as that passed
9708 to the telecine filter.
9710 This filter accepts the following options:
9719 The default value is @code{top}.
9723 A string of numbers representing the pulldown pattern you wish to apply.
9724 The default value is @code{23}.
9727 A number representing position of the first frame with respect to the telecine
9728 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9733 Apply dilation effect to the video.
9735 This filter replaces the pixel by the local(3x3) maximum.
9737 It accepts the following options:
9744 Limit the maximum change for each plane, default is 65535.
9745 If 0, plane will remain unchanged.
9748 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9751 Flags to local 3x3 coordinates maps like this:
9758 @subsection Commands
9760 This filter supports the all above options as @ref{commands}.
9764 Displace pixels as indicated by second and third input stream.
9766 It takes three input streams and outputs one stream, the first input is the
9767 source, and second and third input are displacement maps.
9769 The second input specifies how much to displace pixels along the
9770 x-axis, while the third input specifies how much to displace pixels
9772 If one of displacement map streams terminates, last frame from that
9773 displacement map will be used.
9775 Note that once generated, displacements maps can be reused over and over again.
9777 A description of the accepted options follows.
9781 Set displace behavior for pixels that are out of range.
9783 Available values are:
9786 Missing pixels are replaced by black pixels.
9789 Adjacent pixels will spread out to replace missing pixels.
9792 Out of range pixels are wrapped so they point to pixels of other side.
9795 Out of range pixels will be replaced with mirrored pixels.
9797 Default is @samp{smear}.
9801 @subsection Examples
9805 Add ripple effect to rgb input of video size hd720:
9807 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
9811 Add wave effect to rgb input of video size hd720:
9813 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
9817 @anchor{dnn_processing}
9818 @section dnn_processing
9820 Do image processing with deep neural networks. It works together with another filter
9821 which converts the pixel format of the Frame to what the dnn network requires.
9823 The filter accepts the following options:
9827 Specify which DNN backend to use for model loading and execution. This option accepts
9828 the following values:
9832 Native implementation of DNN loading and execution.
9835 TensorFlow backend. To enable this backend you
9836 need to install the TensorFlow for C library (see
9837 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9838 @code{--enable-libtensorflow}
9841 OpenVINO backend. To enable this backend you
9842 need to build and install the OpenVINO for C library (see
9843 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9844 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9845 be needed if the header files and libraries are not installed into system path)
9849 Default value is @samp{native}.
9852 Set path to model file specifying network architecture and its parameters.
9853 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9854 backend can load files for only its format.
9856 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9859 Set the input name of the dnn network.
9862 Set the output name of the dnn network.
9866 @subsection Examples
9870 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9872 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9876 Halve the pixel value of the frame with format gray32f:
9878 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
9882 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9884 ./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
9888 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9890 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9897 Draw a colored box on the input image.
9899 It accepts the following parameters:
9904 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9908 The expressions which specify the width and height of the box; if 0 they are interpreted as
9909 the input width and height. It defaults to 0.
9912 Specify the color of the box to write. For the general syntax of this option,
9913 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9914 value @code{invert} is used, the box edge color is the same as the
9915 video with inverted luma.
9918 The expression which sets the thickness of the box edge.
9919 A value of @code{fill} will create a filled box. Default value is @code{3}.
9921 See below for the list of accepted constants.
9924 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9925 will overwrite the video's color and alpha pixels.
9926 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9929 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9930 following constants:
9934 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9938 horizontal and vertical chroma subsample values. For example for the
9939 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9943 The input width and height.
9946 The input sample aspect ratio.
9950 The x and y offset coordinates where the box is drawn.
9954 The width and height of the drawn box.
9957 The thickness of the drawn box.
9959 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9960 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9964 @subsection Examples
9968 Draw a black box around the edge of the input image:
9974 Draw a box with color red and an opacity of 50%:
9976 drawbox=10:20:200:60:red@@0.5
9979 The previous example can be specified as:
9981 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9985 Fill the box with pink color:
9987 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9991 Draw a 2-pixel red 2.40:1 mask:
9993 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
9997 @subsection Commands
9998 This filter supports same commands as options.
9999 The command accepts the same syntax of the corresponding option.
10001 If the specified expression is not valid, it is kept at its current
10006 Draw a graph using input video metadata.
10008 It accepts the following parameters:
10012 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10015 Set 1st foreground color expression.
10018 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10021 Set 2nd foreground color expression.
10024 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10027 Set 3rd foreground color expression.
10030 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10033 Set 4th foreground color expression.
10036 Set minimal value of metadata value.
10039 Set maximal value of metadata value.
10042 Set graph background color. Default is white.
10047 Available values for mode is:
10054 Default is @code{line}.
10059 Available values for slide is:
10062 Draw new frame when right border is reached.
10065 Replace old columns with new ones.
10068 Scroll from right to left.
10071 Scroll from left to right.
10074 Draw single picture.
10077 Default is @code{frame}.
10080 Set size of graph video. For the syntax of this option, check the
10081 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10082 The default value is @code{900x256}.
10085 Set the output frame rate. Default value is @code{25}.
10087 The foreground color expressions can use the following variables:
10090 Minimal value of metadata value.
10093 Maximal value of metadata value.
10096 Current metadata key value.
10099 The color is defined as 0xAABBGGRR.
10102 Example using metadata from @ref{signalstats} filter:
10104 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10107 Example using metadata from @ref{ebur128} filter:
10109 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10114 Draw a grid on the input image.
10116 It accepts the following parameters:
10121 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10125 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10126 input width and height, respectively, minus @code{thickness}, so image gets
10127 framed. Default to 0.
10130 Specify the color of the grid. For the general syntax of this option,
10131 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10132 value @code{invert} is used, the grid color is the same as the
10133 video with inverted luma.
10136 The expression which sets the thickness of the grid line. Default value is @code{1}.
10138 See below for the list of accepted constants.
10141 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10142 will overwrite the video's color and alpha pixels.
10143 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10146 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10147 following constants:
10151 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10155 horizontal and vertical chroma subsample values. For example for the
10156 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10160 The input grid cell width and height.
10163 The input sample aspect ratio.
10167 The x and y coordinates of some point of grid intersection (meant to configure offset).
10171 The width and height of the drawn cell.
10174 The thickness of the drawn cell.
10176 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10177 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10181 @subsection Examples
10185 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10187 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10191 Draw a white 3x3 grid with an opacity of 50%:
10193 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10197 @subsection Commands
10198 This filter supports same commands as options.
10199 The command accepts the same syntax of the corresponding option.
10201 If the specified expression is not valid, it is kept at its current
10207 Draw a text string or text from a specified file on top of a video, using the
10208 libfreetype library.
10210 To enable compilation of this filter, you need to configure FFmpeg with
10211 @code{--enable-libfreetype}.
10212 To enable default font fallback and the @var{font} option you need to
10213 configure FFmpeg with @code{--enable-libfontconfig}.
10214 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10215 @code{--enable-libfribidi}.
10219 It accepts the following parameters:
10224 Used to draw a box around text using the background color.
10225 The value must be either 1 (enable) or 0 (disable).
10226 The default value of @var{box} is 0.
10229 Set the width of the border to be drawn around the box using @var{boxcolor}.
10230 The default value of @var{boxborderw} is 0.
10233 The color to be used for drawing box around text. For the syntax of this
10234 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10236 The default value of @var{boxcolor} is "white".
10239 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10240 The default value of @var{line_spacing} is 0.
10243 Set the width of the border to be drawn around the text using @var{bordercolor}.
10244 The default value of @var{borderw} is 0.
10247 Set the color to be used for drawing border around text. For the syntax of this
10248 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10250 The default value of @var{bordercolor} is "black".
10253 Select how the @var{text} is expanded. Can be either @code{none},
10254 @code{strftime} (deprecated) or
10255 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10259 Set a start time for the count. Value is in microseconds. Only applied
10260 in the deprecated strftime expansion mode. To emulate in normal expansion
10261 mode use the @code{pts} function, supplying the start time (in seconds)
10262 as the second argument.
10265 If true, check and fix text coords to avoid clipping.
10268 The color to be used for drawing fonts. For the syntax of this option, check
10269 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10271 The default value of @var{fontcolor} is "black".
10273 @item fontcolor_expr
10274 String which is expanded the same way as @var{text} to obtain dynamic
10275 @var{fontcolor} value. By default this option has empty value and is not
10276 processed. When this option is set, it overrides @var{fontcolor} option.
10279 The font family to be used for drawing text. By default Sans.
10282 The font file to be used for drawing text. The path must be included.
10283 This parameter is mandatory if the fontconfig support is disabled.
10286 Draw the text applying alpha blending. The value can
10287 be a number between 0.0 and 1.0.
10288 The expression accepts the same variables @var{x, y} as well.
10289 The default value is 1.
10290 Please see @var{fontcolor_expr}.
10293 The font size to be used for drawing text.
10294 The default value of @var{fontsize} is 16.
10297 If set to 1, attempt to shape the text (for example, reverse the order of
10298 right-to-left text and join Arabic characters) before drawing it.
10299 Otherwise, just draw the text exactly as given.
10300 By default 1 (if supported).
10302 @item ft_load_flags
10303 The flags to be used for loading the fonts.
10305 The flags map the corresponding flags supported by libfreetype, and are
10306 a combination of the following values:
10313 @item vertical_layout
10314 @item force_autohint
10317 @item ignore_global_advance_width
10319 @item ignore_transform
10321 @item linear_design
10325 Default value is "default".
10327 For more information consult the documentation for the FT_LOAD_*
10331 The color to be used for drawing a shadow behind the drawn text. For the
10332 syntax of this option, check the @ref{color syntax,,"Color" section in the
10333 ffmpeg-utils manual,ffmpeg-utils}.
10335 The default value of @var{shadowcolor} is "black".
10339 The x and y offsets for the text shadow position with respect to the
10340 position of the text. They can be either positive or negative
10341 values. The default value for both is "0".
10344 The starting frame number for the n/frame_num variable. The default value
10348 The size in number of spaces to use for rendering the tab.
10349 Default value is 4.
10352 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10353 format. It can be used with or without text parameter. @var{timecode_rate}
10354 option must be specified.
10356 @item timecode_rate, rate, r
10357 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10358 integer. Minimum value is "1".
10359 Drop-frame timecode is supported for frame rates 30 & 60.
10362 If set to 1, the output of the timecode option will wrap around at 24 hours.
10363 Default is 0 (disabled).
10366 The text string to be drawn. The text must be a sequence of UTF-8
10367 encoded characters.
10368 This parameter is mandatory if no file is specified with the parameter
10372 A text file containing text to be drawn. The text must be a sequence
10373 of UTF-8 encoded characters.
10375 This parameter is mandatory if no text string is specified with the
10376 parameter @var{text}.
10378 If both @var{text} and @var{textfile} are specified, an error is thrown.
10381 If set to 1, the @var{textfile} will be reloaded before each frame.
10382 Be sure to update it atomically, or it may be read partially, or even fail.
10386 The expressions which specify the offsets where text will be drawn
10387 within the video frame. They are relative to the top/left border of the
10390 The default value of @var{x} and @var{y} is "0".
10392 See below for the list of accepted constants and functions.
10395 The parameters for @var{x} and @var{y} are expressions containing the
10396 following constants and functions:
10400 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10404 horizontal and vertical chroma subsample values. For example for the
10405 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10408 the height of each text line
10416 @item max_glyph_a, ascent
10417 the maximum distance from the baseline to the highest/upper grid
10418 coordinate used to place a glyph outline point, for all the rendered
10420 It is a positive value, due to the grid's orientation with the Y axis
10423 @item max_glyph_d, descent
10424 the maximum distance from the baseline to the lowest grid coordinate
10425 used to place a glyph outline point, for all the rendered glyphs.
10426 This is a negative value, due to the grid's orientation, with the Y axis
10430 maximum glyph height, that is the maximum height for all the glyphs
10431 contained in the rendered text, it is equivalent to @var{ascent} -
10435 maximum glyph width, that is the maximum width for all the glyphs
10436 contained in the rendered text
10439 the number of input frame, starting from 0
10441 @item rand(min, max)
10442 return a random number included between @var{min} and @var{max}
10445 The input sample aspect ratio.
10448 timestamp expressed in seconds, NAN if the input timestamp is unknown
10451 the height of the rendered text
10454 the width of the rendered text
10458 the x and y offset coordinates where the text is drawn.
10460 These parameters allow the @var{x} and @var{y} expressions to refer
10461 to each other, so you can for example specify @code{y=x/dar}.
10464 A one character description of the current frame's picture type.
10467 The current packet's position in the input file or stream
10468 (in bytes, from the start of the input). A value of -1 indicates
10469 this info is not available.
10472 The current packet's duration, in seconds.
10475 The current packet's size (in bytes).
10478 @anchor{drawtext_expansion}
10479 @subsection Text expansion
10481 If @option{expansion} is set to @code{strftime},
10482 the filter recognizes strftime() sequences in the provided text and
10483 expands them accordingly. Check the documentation of strftime(). This
10484 feature is deprecated.
10486 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10488 If @option{expansion} is set to @code{normal} (which is the default),
10489 the following expansion mechanism is used.
10491 The backslash character @samp{\}, followed by any character, always expands to
10492 the second character.
10494 Sequences of the form @code{%@{...@}} are expanded. The text between the
10495 braces is a function name, possibly followed by arguments separated by ':'.
10496 If the arguments contain special characters or delimiters (':' or '@}'),
10497 they should be escaped.
10499 Note that they probably must also be escaped as the value for the
10500 @option{text} option in the filter argument string and as the filter
10501 argument in the filtergraph description, and possibly also for the shell,
10502 that makes up to four levels of escaping; using a text file avoids these
10505 The following functions are available:
10510 The expression evaluation result.
10512 It must take one argument specifying the expression to be evaluated,
10513 which accepts the same constants and functions as the @var{x} and
10514 @var{y} values. Note that not all constants should be used, for
10515 example the text size is not known when evaluating the expression, so
10516 the constants @var{text_w} and @var{text_h} will have an undefined
10519 @item expr_int_format, eif
10520 Evaluate the expression's value and output as formatted integer.
10522 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10523 The second argument specifies the output format. Allowed values are @samp{x},
10524 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10525 @code{printf} function.
10526 The third parameter is optional and sets the number of positions taken by the output.
10527 It can be used to add padding with zeros from the left.
10530 The time at which the filter is running, expressed in UTC.
10531 It can accept an argument: a strftime() format string.
10534 The time at which the filter is running, expressed in the local time zone.
10535 It can accept an argument: a strftime() format string.
10538 Frame metadata. Takes one or two arguments.
10540 The first argument is mandatory and specifies the metadata key.
10542 The second argument is optional and specifies a default value, used when the
10543 metadata key is not found or empty.
10545 Available metadata can be identified by inspecting entries
10546 starting with TAG included within each frame section
10547 printed by running @code{ffprobe -show_frames}.
10549 String metadata generated in filters leading to
10550 the drawtext filter are also available.
10553 The frame number, starting from 0.
10556 A one character description of the current picture type.
10559 The timestamp of the current frame.
10560 It can take up to three arguments.
10562 The first argument is the format of the timestamp; it defaults to @code{flt}
10563 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10564 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10565 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10566 @code{localtime} stands for the timestamp of the frame formatted as
10567 local time zone time.
10569 The second argument is an offset added to the timestamp.
10571 If the format is set to @code{hms}, a third argument @code{24HH} may be
10572 supplied to present the hour part of the formatted timestamp in 24h format
10575 If the format is set to @code{localtime} or @code{gmtime},
10576 a third argument may be supplied: a strftime() format string.
10577 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10580 @subsection Commands
10582 This filter supports altering parameters via commands:
10585 Alter existing filter parameters.
10587 Syntax for the argument is the same as for filter invocation, e.g.
10590 fontsize=56:fontcolor=green:text='Hello World'
10593 Full filter invocation with sendcmd would look like this:
10596 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10600 If the entire argument can't be parsed or applied as valid values then the filter will
10601 continue with its existing parameters.
10603 @subsection Examples
10607 Draw "Test Text" with font FreeSerif, using the default values for the
10608 optional parameters.
10611 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10615 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10616 and y=50 (counting from the top-left corner of the screen), text is
10617 yellow with a red box around it. Both the text and the box have an
10621 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10622 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10625 Note that the double quotes are not necessary if spaces are not used
10626 within the parameter list.
10629 Show the text at the center of the video frame:
10631 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10635 Show the text at a random position, switching to a new position every 30 seconds:
10637 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)"
10641 Show a text line sliding from right to left in the last row of the video
10642 frame. The file @file{LONG_LINE} is assumed to contain a single line
10645 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10649 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10651 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10655 Draw a single green letter "g", at the center of the input video.
10656 The glyph baseline is placed at half screen height.
10658 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10662 Show text for 1 second every 3 seconds:
10664 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10668 Use fontconfig to set the font. Note that the colons need to be escaped.
10670 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10674 Draw "Test Text" with font size dependent on height of the video.
10676 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10680 Print the date of a real-time encoding (see strftime(3)):
10682 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10686 Show text fading in and out (appearing/disappearing):
10689 DS=1.0 # display start
10690 DE=10.0 # display end
10691 FID=1.5 # fade in duration
10692 FOD=5 # fade out duration
10693 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 @}"
10697 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10698 and the @option{fontsize} value are included in the @option{y} offset.
10700 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10701 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10705 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10706 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10707 must have option @option{-export_path_metadata 1} for the special metadata fields
10708 to be available for filters.
10710 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10715 For more information about libfreetype, check:
10716 @url{http://www.freetype.org/}.
10718 For more information about fontconfig, check:
10719 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10721 For more information about libfribidi, check:
10722 @url{http://fribidi.org/}.
10724 @section edgedetect
10726 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10728 The filter accepts the following options:
10733 Set low and high threshold values used by the Canny thresholding
10736 The high threshold selects the "strong" edge pixels, which are then
10737 connected through 8-connectivity with the "weak" edge pixels selected
10738 by the low threshold.
10740 @var{low} and @var{high} threshold values must be chosen in the range
10741 [0,1], and @var{low} should be lesser or equal to @var{high}.
10743 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10747 Define the drawing mode.
10751 Draw white/gray wires on black background.
10754 Mix the colors to create a paint/cartoon effect.
10757 Apply Canny edge detector on all selected planes.
10759 Default value is @var{wires}.
10762 Select planes for filtering. By default all available planes are filtered.
10765 @subsection Examples
10769 Standard edge detection with custom values for the hysteresis thresholding:
10771 edgedetect=low=0.1:high=0.4
10775 Painting effect without thresholding:
10777 edgedetect=mode=colormix:high=0
10783 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10785 For each input image, the filter will compute the optimal mapping from
10786 the input to the output given the codebook length, that is the number
10787 of distinct output colors.
10789 This filter accepts the following options.
10792 @item codebook_length, l
10793 Set codebook length. The value must be a positive integer, and
10794 represents the number of distinct output colors. Default value is 256.
10797 Set the maximum number of iterations to apply for computing the optimal
10798 mapping. The higher the value the better the result and the higher the
10799 computation time. Default value is 1.
10802 Set a random seed, must be an integer included between 0 and
10803 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10804 will try to use a good random seed on a best effort basis.
10807 Set pal8 output pixel format. This option does not work with codebook
10808 length greater than 256.
10813 Measure graylevel entropy in histogram of color channels of video frames.
10815 It accepts the following parameters:
10819 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10821 @var{diff} mode measures entropy of histogram delta values, absolute differences
10822 between neighbour histogram values.
10826 Set brightness, contrast, saturation and approximate gamma adjustment.
10828 The filter accepts the following options:
10832 Set the contrast expression. The value must be a float value in range
10833 @code{-1000.0} to @code{1000.0}. The default value is "1".
10836 Set the brightness expression. The value must be a float value in
10837 range @code{-1.0} to @code{1.0}. The default value is "0".
10840 Set the saturation expression. The value must be a float in
10841 range @code{0.0} to @code{3.0}. The default value is "1".
10844 Set the gamma expression. The value must be a float in range
10845 @code{0.1} to @code{10.0}. The default value is "1".
10848 Set the gamma expression for red. The value must be a float in
10849 range @code{0.1} to @code{10.0}. The default value is "1".
10852 Set the gamma expression for green. The value must be a float in range
10853 @code{0.1} to @code{10.0}. The default value is "1".
10856 Set the gamma expression for blue. The value must be a float in range
10857 @code{0.1} to @code{10.0}. The default value is "1".
10860 Set the gamma weight expression. It can be used to reduce the effect
10861 of a high gamma value on bright image areas, e.g. keep them from
10862 getting overamplified and just plain white. The value must be a float
10863 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10864 gamma correction all the way down while @code{1.0} leaves it at its
10865 full strength. Default is "1".
10868 Set when the expressions for brightness, contrast, saturation and
10869 gamma expressions are evaluated.
10871 It accepts the following values:
10874 only evaluate expressions once during the filter initialization or
10875 when a command is processed
10878 evaluate expressions for each incoming frame
10881 Default value is @samp{init}.
10884 The expressions accept the following parameters:
10887 frame count of the input frame starting from 0
10890 byte position of the corresponding packet in the input file, NAN if
10894 frame rate of the input video, NAN if the input frame rate is unknown
10897 timestamp expressed in seconds, NAN if the input timestamp is unknown
10900 @subsection Commands
10901 The filter supports the following commands:
10905 Set the contrast expression.
10908 Set the brightness expression.
10911 Set the saturation expression.
10914 Set the gamma expression.
10917 Set the gamma_r expression.
10920 Set gamma_g expression.
10923 Set gamma_b expression.
10926 Set gamma_weight expression.
10928 The command accepts the same syntax of the corresponding option.
10930 If the specified expression is not valid, it is kept at its current
10937 Apply erosion effect to the video.
10939 This filter replaces the pixel by the local(3x3) minimum.
10941 It accepts the following options:
10948 Limit the maximum change for each plane, default is 65535.
10949 If 0, plane will remain unchanged.
10952 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10955 Flags to local 3x3 coordinates maps like this:
10962 @subsection Commands
10964 This filter supports the all above options as @ref{commands}.
10966 @section extractplanes
10968 Extract color channel components from input video stream into
10969 separate grayscale video streams.
10971 The filter accepts the following option:
10975 Set plane(s) to extract.
10977 Available values for planes are:
10988 Choosing planes not available in the input will result in an error.
10989 That means you cannot select @code{r}, @code{g}, @code{b} planes
10990 with @code{y}, @code{u}, @code{v} planes at same time.
10993 @subsection Examples
10997 Extract luma, u and v color channel component from input video frame
10998 into 3 grayscale outputs:
11000 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
11006 Apply a fade-in/out effect to the input video.
11008 It accepts the following parameters:
11012 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11014 Default is @code{in}.
11016 @item start_frame, s
11017 Specify the number of the frame to start applying the fade
11018 effect at. Default is 0.
11021 The number of frames that the fade effect lasts. At the end of the
11022 fade-in effect, the output video will have the same intensity as the input video.
11023 At the end of the fade-out transition, the output video will be filled with the
11024 selected @option{color}.
11028 If set to 1, fade only alpha channel, if one exists on the input.
11029 Default value is 0.
11031 @item start_time, st
11032 Specify the timestamp (in seconds) of the frame to start to apply the fade
11033 effect. If both start_frame and start_time are specified, the fade will start at
11034 whichever comes last. Default is 0.
11037 The number of seconds for which the fade effect has to last. At the end of the
11038 fade-in effect the output video will have the same intensity as the input video,
11039 at the end of the fade-out transition the output video will be filled with the
11040 selected @option{color}.
11041 If both duration and nb_frames are specified, duration is used. Default is 0
11042 (nb_frames is used by default).
11045 Specify the color of the fade. Default is "black".
11048 @subsection Examples
11052 Fade in the first 30 frames of video:
11057 The command above is equivalent to:
11063 Fade out the last 45 frames of a 200-frame video:
11066 fade=type=out:start_frame=155:nb_frames=45
11070 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11072 fade=in:0:25, fade=out:975:25
11076 Make the first 5 frames yellow, then fade in from frame 5-24:
11078 fade=in:5:20:color=yellow
11082 Fade in alpha over first 25 frames of video:
11084 fade=in:0:25:alpha=1
11088 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11090 fade=t=in:st=5.5:d=0.5
11096 Denoise frames using 3D FFT (frequency domain filtering).
11098 The filter accepts the following options:
11102 Set the noise sigma constant. This sets denoising strength.
11103 Default value is 1. Allowed range is from 0 to 30.
11104 Using very high sigma with low overlap may give blocking artifacts.
11107 Set amount of denoising. By default all detected noise is reduced.
11108 Default value is 1. Allowed range is from 0 to 1.
11111 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11112 Actual size of block in pixels is 2 to power of @var{block}, so by default
11113 block size in pixels is 2^4 which is 16.
11116 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11119 Set number of previous frames to use for denoising. By default is set to 0.
11122 Set number of next frames to to use for denoising. By default is set to 0.
11125 Set planes which will be filtered, by default are all available filtered
11130 Apply arbitrary expressions to samples in frequency domain
11134 Adjust the dc value (gain) of the luma plane of the image. The filter
11135 accepts an integer value in range @code{0} to @code{1000}. The default
11136 value is set to @code{0}.
11139 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11140 filter accepts an integer value in range @code{0} to @code{1000}. The
11141 default value is set to @code{0}.
11144 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11145 filter accepts an integer value in range @code{0} to @code{1000}. The
11146 default value is set to @code{0}.
11149 Set the frequency domain weight expression for the luma plane.
11152 Set the frequency domain weight expression for the 1st chroma plane.
11155 Set the frequency domain weight expression for the 2nd chroma plane.
11158 Set when the expressions are evaluated.
11160 It accepts the following values:
11163 Only evaluate expressions once during the filter initialization.
11166 Evaluate expressions for each incoming frame.
11169 Default value is @samp{init}.
11171 The filter accepts the following variables:
11174 The coordinates of the current sample.
11178 The width and height of the image.
11181 The number of input frame, starting from 0.
11184 @subsection Examples
11190 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11196 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11202 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11208 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11215 Extract a single field from an interlaced image using stride
11216 arithmetic to avoid wasting CPU time. The output frames are marked as
11219 The filter accepts the following options:
11223 Specify whether to extract the top (if the value is @code{0} or
11224 @code{top}) or the bottom field (if the value is @code{1} or
11230 Create new frames by copying the top and bottom fields from surrounding frames
11231 supplied as numbers by the hint file.
11235 Set file containing hints: absolute/relative frame numbers.
11237 There must be one line for each frame in a clip. Each line must contain two
11238 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11239 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11240 is current frame number for @code{absolute} mode or out of [-1, 1] range
11241 for @code{relative} mode. First number tells from which frame to pick up top
11242 field and second number tells from which frame to pick up bottom field.
11244 If optionally followed by @code{+} output frame will be marked as interlaced,
11245 else if followed by @code{-} output frame will be marked as progressive, else
11246 it will be marked same as input frame.
11247 If optionally followed by @code{t} output frame will use only top field, or in
11248 case of @code{b} it will use only bottom field.
11249 If line starts with @code{#} or @code{;} that line is skipped.
11252 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11255 Example of first several lines of @code{hint} file for @code{relative} mode:
11257 0,0 - # first frame
11258 1,0 - # second frame, use third's frame top field and second's frame bottom field
11259 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11274 @section fieldmatch
11276 Field matching filter for inverse telecine. It is meant to reconstruct the
11277 progressive frames from a telecined stream. The filter does not drop duplicated
11278 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11279 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11281 The separation of the field matching and the decimation is notably motivated by
11282 the possibility of inserting a de-interlacing filter fallback between the two.
11283 If the source has mixed telecined and real interlaced content,
11284 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11285 But these remaining combed frames will be marked as interlaced, and thus can be
11286 de-interlaced by a later filter such as @ref{yadif} before decimation.
11288 In addition to the various configuration options, @code{fieldmatch} can take an
11289 optional second stream, activated through the @option{ppsrc} option. If
11290 enabled, the frames reconstruction will be based on the fields and frames from
11291 this second stream. This allows the first input to be pre-processed in order to
11292 help the various algorithms of the filter, while keeping the output lossless
11293 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11294 or brightness/contrast adjustments can help.
11296 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11297 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11298 which @code{fieldmatch} is based on. While the semantic and usage are very
11299 close, some behaviour and options names can differ.
11301 The @ref{decimate} filter currently only works for constant frame rate input.
11302 If your input has mixed telecined (30fps) and progressive content with a lower
11303 framerate like 24fps use the following filterchain to produce the necessary cfr
11304 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11306 The filter accepts the following options:
11310 Specify the assumed field order of the input stream. Available values are:
11314 Auto detect parity (use FFmpeg's internal parity value).
11316 Assume bottom field first.
11318 Assume top field first.
11321 Note that it is sometimes recommended not to trust the parity announced by the
11324 Default value is @var{auto}.
11327 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11328 sense that it won't risk creating jerkiness due to duplicate frames when
11329 possible, but if there are bad edits or blended fields it will end up
11330 outputting combed frames when a good match might actually exist. On the other
11331 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11332 but will almost always find a good frame if there is one. The other values are
11333 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11334 jerkiness and creating duplicate frames versus finding good matches in sections
11335 with bad edits, orphaned fields, blended fields, etc.
11337 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11339 Available values are:
11343 2-way matching (p/c)
11345 2-way matching, and trying 3rd match if still combed (p/c + n)
11347 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11349 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11350 still combed (p/c + n + u/b)
11352 3-way matching (p/c/n)
11354 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11355 detected as combed (p/c/n + u/b)
11358 The parenthesis at the end indicate the matches that would be used for that
11359 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11362 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11365 Default value is @var{pc_n}.
11368 Mark the main input stream as a pre-processed input, and enable the secondary
11369 input stream as the clean source to pick the fields from. See the filter
11370 introduction for more details. It is similar to the @option{clip2} feature from
11373 Default value is @code{0} (disabled).
11376 Set the field to match from. It is recommended to set this to the same value as
11377 @option{order} unless you experience matching failures with that setting. In
11378 certain circumstances changing the field that is used to match from can have a
11379 large impact on matching performance. Available values are:
11383 Automatic (same value as @option{order}).
11385 Match from the bottom field.
11387 Match from the top field.
11390 Default value is @var{auto}.
11393 Set whether or not chroma is included during the match comparisons. In most
11394 cases it is recommended to leave this enabled. You should set this to @code{0}
11395 only if your clip has bad chroma problems such as heavy rainbowing or other
11396 artifacts. Setting this to @code{0} could also be used to speed things up at
11397 the cost of some accuracy.
11399 Default value is @code{1}.
11403 These define an exclusion band which excludes the lines between @option{y0} and
11404 @option{y1} from being included in the field matching decision. An exclusion
11405 band can be used to ignore subtitles, a logo, or other things that may
11406 interfere with the matching. @option{y0} sets the starting scan line and
11407 @option{y1} sets the ending line; all lines in between @option{y0} and
11408 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11409 @option{y0} and @option{y1} to the same value will disable the feature.
11410 @option{y0} and @option{y1} defaults to @code{0}.
11413 Set the scene change detection threshold as a percentage of maximum change on
11414 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11415 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11416 @option{scthresh} is @code{[0.0, 100.0]}.
11418 Default value is @code{12.0}.
11421 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11422 account the combed scores of matches when deciding what match to use as the
11423 final match. Available values are:
11427 No final matching based on combed scores.
11429 Combed scores are only used when a scene change is detected.
11431 Use combed scores all the time.
11434 Default is @var{sc}.
11437 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11438 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11439 Available values are:
11443 No forced calculation.
11445 Force p/c/n calculations.
11447 Force p/c/n/u/b calculations.
11450 Default value is @var{none}.
11453 This is the area combing threshold used for combed frame detection. This
11454 essentially controls how "strong" or "visible" combing must be to be detected.
11455 Larger values mean combing must be more visible and smaller values mean combing
11456 can be less visible or strong and still be detected. Valid settings are from
11457 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11458 be detected as combed). This is basically a pixel difference value. A good
11459 range is @code{[8, 12]}.
11461 Default value is @code{9}.
11464 Sets whether or not chroma is considered in the combed frame decision. Only
11465 disable this if your source has chroma problems (rainbowing, etc.) that are
11466 causing problems for the combed frame detection with chroma enabled. Actually,
11467 using @option{chroma}=@var{0} is usually more reliable, except for the case
11468 where there is chroma only combing in the source.
11470 Default value is @code{0}.
11474 Respectively set the x-axis and y-axis size of the window used during combed
11475 frame detection. This has to do with the size of the area in which
11476 @option{combpel} pixels are required to be detected as combed for a frame to be
11477 declared combed. See the @option{combpel} parameter description for more info.
11478 Possible values are any number that is a power of 2 starting at 4 and going up
11481 Default value is @code{16}.
11484 The number of combed pixels inside any of the @option{blocky} by
11485 @option{blockx} size blocks on the frame for the frame to be detected as
11486 combed. While @option{cthresh} controls how "visible" the combing must be, this
11487 setting controls "how much" combing there must be in any localized area (a
11488 window defined by the @option{blockx} and @option{blocky} settings) on the
11489 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11490 which point no frames will ever be detected as combed). This setting is known
11491 as @option{MI} in TFM/VFM vocabulary.
11493 Default value is @code{80}.
11496 @anchor{p/c/n/u/b meaning}
11497 @subsection p/c/n/u/b meaning
11499 @subsubsection p/c/n
11501 We assume the following telecined stream:
11504 Top fields: 1 2 2 3 4
11505 Bottom fields: 1 2 3 4 4
11508 The numbers correspond to the progressive frame the fields relate to. Here, the
11509 first two frames are progressive, the 3rd and 4th are combed, and so on.
11511 When @code{fieldmatch} is configured to run a matching from bottom
11512 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11517 B 1 2 3 4 4 <-- matching reference
11526 As a result of the field matching, we can see that some frames get duplicated.
11527 To perform a complete inverse telecine, you need to rely on a decimation filter
11528 after this operation. See for instance the @ref{decimate} filter.
11530 The same operation now matching from top fields (@option{field}=@var{top})
11535 T 1 2 2 3 4 <-- matching reference
11545 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11546 basically, they refer to the frame and field of the opposite parity:
11549 @item @var{p} matches the field of the opposite parity in the previous frame
11550 @item @var{c} matches the field of the opposite parity in the current frame
11551 @item @var{n} matches the field of the opposite parity in the next frame
11556 The @var{u} and @var{b} matching are a bit special in the sense that they match
11557 from the opposite parity flag. In the following examples, we assume that we are
11558 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11559 'x' is placed above and below each matched fields.
11561 With bottom matching (@option{field}=@var{bottom}):
11566 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11567 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11575 With top matching (@option{field}=@var{top}):
11580 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11581 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11589 @subsection Examples
11591 Simple IVTC of a top field first telecined stream:
11593 fieldmatch=order=tff:combmatch=none, decimate
11596 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11598 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11601 @section fieldorder
11603 Transform the field order of the input video.
11605 It accepts the following parameters:
11610 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11611 for bottom field first.
11614 The default value is @samp{tff}.
11616 The transformation is done by shifting the picture content up or down
11617 by one line, and filling the remaining line with appropriate picture content.
11618 This method is consistent with most broadcast field order converters.
11620 If the input video is not flagged as being interlaced, or it is already
11621 flagged as being of the required output field order, then this filter does
11622 not alter the incoming video.
11624 It is very useful when converting to or from PAL DV material,
11625 which is bottom field first.
11629 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11632 @section fifo, afifo
11634 Buffer input images and send them when they are requested.
11636 It is mainly useful when auto-inserted by the libavfilter
11639 It does not take parameters.
11641 @section fillborders
11643 Fill borders of the input video, without changing video stream dimensions.
11644 Sometimes video can have garbage at the four edges and you may not want to
11645 crop video input to keep size multiple of some number.
11647 This filter accepts the following options:
11651 Number of pixels to fill from left border.
11654 Number of pixels to fill from right border.
11657 Number of pixels to fill from top border.
11660 Number of pixels to fill from bottom border.
11665 It accepts the following values:
11668 fill pixels using outermost pixels
11671 fill pixels using mirroring
11674 fill pixels with constant value
11677 Default is @var{smear}.
11680 Set color for pixels in fixed mode. Default is @var{black}.
11683 @subsection Commands
11684 This filter supports same @ref{commands} as options.
11685 The command accepts the same syntax of the corresponding option.
11687 If the specified expression is not valid, it is kept at its current
11692 Find a rectangular object
11694 It accepts the following options:
11698 Filepath of the object image, needs to be in gray8.
11701 Detection threshold, default is 0.5.
11704 Number of mipmaps, default is 3.
11706 @item xmin, ymin, xmax, ymax
11707 Specifies the rectangle in which to search.
11710 @subsection Examples
11714 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11716 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11722 Flood area with values of same pixel components with another values.
11724 It accepts the following options:
11727 Set pixel x coordinate.
11730 Set pixel y coordinate.
11733 Set source #0 component value.
11736 Set source #1 component value.
11739 Set source #2 component value.
11742 Set source #3 component value.
11745 Set destination #0 component value.
11748 Set destination #1 component value.
11751 Set destination #2 component value.
11754 Set destination #3 component value.
11760 Convert the input video to one of the specified pixel formats.
11761 Libavfilter will try to pick one that is suitable as input to
11764 It accepts the following parameters:
11768 A '|'-separated list of pixel format names, such as
11769 "pix_fmts=yuv420p|monow|rgb24".
11773 @subsection Examples
11777 Convert the input video to the @var{yuv420p} format
11779 format=pix_fmts=yuv420p
11782 Convert the input video to any of the formats in the list
11784 format=pix_fmts=yuv420p|yuv444p|yuv410p
11791 Convert the video to specified constant frame rate by duplicating or dropping
11792 frames as necessary.
11794 It accepts the following parameters:
11798 The desired output frame rate. The default is @code{25}.
11801 Assume the first PTS should be the given value, in seconds. This allows for
11802 padding/trimming at the start of stream. By default, no assumption is made
11803 about the first frame's expected PTS, so no padding or trimming is done.
11804 For example, this could be set to 0 to pad the beginning with duplicates of
11805 the first frame if a video stream starts after the audio stream or to trim any
11806 frames with a negative PTS.
11809 Timestamp (PTS) rounding method.
11811 Possible values are:
11818 round towards -infinity
11820 round towards +infinity
11824 The default is @code{near}.
11827 Action performed when reading the last frame.
11829 Possible values are:
11832 Use same timestamp rounding method as used for other frames.
11834 Pass through last frame if input duration has not been reached yet.
11836 The default is @code{round}.
11840 Alternatively, the options can be specified as a flat string:
11841 @var{fps}[:@var{start_time}[:@var{round}]].
11843 See also the @ref{setpts} filter.
11845 @subsection Examples
11849 A typical usage in order to set the fps to 25:
11855 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11857 fps=fps=film:round=near
11863 Pack two different video streams into a stereoscopic video, setting proper
11864 metadata on supported codecs. The two views should have the same size and
11865 framerate and processing will stop when the shorter video ends. Please note
11866 that you may conveniently adjust view properties with the @ref{scale} and
11869 It accepts the following parameters:
11873 The desired packing format. Supported values are:
11878 The views are next to each other (default).
11881 The views are on top of each other.
11884 The views are packed by line.
11887 The views are packed by column.
11890 The views are temporally interleaved.
11899 # Convert left and right views into a frame-sequential video
11900 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11902 # Convert views into a side-by-side video with the same output resolution as the input
11903 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
11908 Change the frame rate by interpolating new video output frames from the source
11911 This filter is not designed to function correctly with interlaced media. If
11912 you wish to change the frame rate of interlaced media then you are required
11913 to deinterlace before this filter and re-interlace after this filter.
11915 A description of the accepted options follows.
11919 Specify the output frames per second. This option can also be specified
11920 as a value alone. The default is @code{50}.
11923 Specify the start of a range where the output frame will be created as a
11924 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11925 the default is @code{15}.
11928 Specify the end of a range where the output frame will be created as a
11929 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11930 the default is @code{240}.
11933 Specify the level at which a scene change is detected as a value between
11934 0 and 100 to indicate a new scene; a low value reflects a low
11935 probability for the current frame to introduce a new scene, while a higher
11936 value means the current frame is more likely to be one.
11937 The default is @code{8.2}.
11940 Specify flags influencing the filter process.
11942 Available value for @var{flags} is:
11945 @item scene_change_detect, scd
11946 Enable scene change detection using the value of the option @var{scene}.
11947 This flag is enabled by default.
11953 Select one frame every N-th frame.
11955 This filter accepts the following option:
11958 Select frame after every @code{step} frames.
11959 Allowed values are positive integers higher than 0. Default value is @code{1}.
11962 @section freezedetect
11964 Detect frozen video.
11966 This filter logs a message and sets frame metadata when it detects that the
11967 input video has no significant change in content during a specified duration.
11968 Video freeze detection calculates the mean average absolute difference of all
11969 the components of video frames and compares it to a noise floor.
11971 The printed times and duration are expressed in seconds. The
11972 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11973 whose timestamp equals or exceeds the detection duration and it contains the
11974 timestamp of the first frame of the freeze. The
11975 @code{lavfi.freezedetect.freeze_duration} and
11976 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11979 The filter accepts the following options:
11983 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11984 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11988 Set freeze duration until notification (default is 2 seconds).
11991 @section freezeframes
11993 Freeze video frames.
11995 This filter freezes video frames using frame from 2nd input.
11997 The filter accepts the following options:
12001 Set number of first frame from which to start freeze.
12004 Set number of last frame from which to end freeze.
12007 Set number of frame from 2nd input which will be used instead of replaced frames.
12013 Apply a frei0r effect to the input video.
12015 To enable the compilation of this filter, you need to install the frei0r
12016 header and configure FFmpeg with @code{--enable-frei0r}.
12018 It accepts the following parameters:
12023 The name of the frei0r effect to load. If the environment variable
12024 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12025 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12026 Otherwise, the standard frei0r paths are searched, in this order:
12027 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12028 @file{/usr/lib/frei0r-1/}.
12030 @item filter_params
12031 A '|'-separated list of parameters to pass to the frei0r effect.
12035 A frei0r effect parameter can be a boolean (its value is either
12036 "y" or "n"), a double, a color (specified as
12037 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12038 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12039 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12040 a position (specified as @var{X}/@var{Y}, where
12041 @var{X} and @var{Y} are floating point numbers) and/or a string.
12043 The number and types of parameters depend on the loaded effect. If an
12044 effect parameter is not specified, the default value is set.
12046 @subsection Examples
12050 Apply the distort0r effect, setting the first two double parameters:
12052 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12056 Apply the colordistance effect, taking a color as the first parameter:
12058 frei0r=colordistance:0.2/0.3/0.4
12059 frei0r=colordistance:violet
12060 frei0r=colordistance:0x112233
12064 Apply the perspective effect, specifying the top left and top right image
12067 frei0r=perspective:0.2/0.2|0.8/0.2
12071 For more information, see
12072 @url{http://frei0r.dyne.org}
12074 @subsection Commands
12076 This filter supports the @option{filter_params} option as @ref{commands}.
12080 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12082 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12083 processing filter, one of them is performed once per block, not per pixel.
12084 This allows for much higher speed.
12086 The filter accepts the following options:
12090 Set quality. This option defines the number of levels for averaging. It accepts
12091 an integer in the range 4-5. Default value is @code{4}.
12094 Force a constant quantization parameter. It accepts an integer in range 0-63.
12095 If not set, the filter will use the QP from the video stream (if available).
12098 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12099 more details but also more artifacts, while higher values make the image smoother
12100 but also blurrier. Default value is @code{0} − PSNR optimal.
12102 @item use_bframe_qp
12103 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12104 option may cause flicker since the B-Frames have often larger QP. Default is
12105 @code{0} (not enabled).
12111 Apply Gaussian blur filter.
12113 The filter accepts the following options:
12117 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12120 Set number of steps for Gaussian approximation. Default is @code{1}.
12123 Set which planes to filter. By default all planes are filtered.
12126 Set vertical sigma, if negative it will be same as @code{sigma}.
12127 Default is @code{-1}.
12130 @subsection Commands
12131 This filter supports same commands as options.
12132 The command accepts the same syntax of the corresponding option.
12134 If the specified expression is not valid, it is kept at its current
12139 Apply generic equation to each pixel.
12141 The filter accepts the following options:
12144 @item lum_expr, lum
12145 Set the luminance expression.
12147 Set the chrominance blue expression.
12149 Set the chrominance red expression.
12150 @item alpha_expr, a
12151 Set the alpha expression.
12153 Set the red expression.
12154 @item green_expr, g
12155 Set the green expression.
12157 Set the blue expression.
12160 The colorspace is selected according to the specified options. If one
12161 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12162 options is specified, the filter will automatically select a YCbCr
12163 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12164 @option{blue_expr} options is specified, it will select an RGB
12167 If one of the chrominance expression is not defined, it falls back on the other
12168 one. If no alpha expression is specified it will evaluate to opaque value.
12169 If none of chrominance expressions are specified, they will evaluate
12170 to the luminance expression.
12172 The expressions can use the following variables and functions:
12176 The sequential number of the filtered frame, starting from @code{0}.
12180 The coordinates of the current sample.
12184 The width and height of the image.
12188 Width and height scale depending on the currently filtered plane. It is the
12189 ratio between the corresponding luma plane number of pixels and the current
12190 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12191 @code{0.5,0.5} for chroma planes.
12194 Time of the current frame, expressed in seconds.
12197 Return the value of the pixel at location (@var{x},@var{y}) of the current
12201 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12205 Return the value of the pixel at location (@var{x},@var{y}) of the
12206 blue-difference chroma plane. Return 0 if there is no such plane.
12209 Return the value of the pixel at location (@var{x},@var{y}) of the
12210 red-difference chroma plane. Return 0 if there is no such plane.
12215 Return the value of the pixel at location (@var{x},@var{y}) of the
12216 red/green/blue component. Return 0 if there is no such component.
12219 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12220 plane. Return 0 if there is no such plane.
12222 @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)
12223 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12224 sums of samples within a rectangle. See the functions without the sum postfix.
12226 @item interpolation
12227 Set one of interpolation methods:
12232 Default is bilinear.
12235 For functions, if @var{x} and @var{y} are outside the area, the value will be
12236 automatically clipped to the closer edge.
12238 Please note that this filter can use multiple threads in which case each slice
12239 will have its own expression state. If you want to use only a single expression
12240 state because your expressions depend on previous state then you should limit
12241 the number of filter threads to 1.
12243 @subsection Examples
12247 Flip the image horizontally:
12253 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12254 wavelength of 100 pixels:
12256 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12260 Generate a fancy enigmatic moving light:
12262 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
12266 Generate a quick emboss effect:
12268 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12272 Modify RGB components depending on pixel position:
12274 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12278 Create a radial gradient that is the same size as the input (also see
12279 the @ref{vignette} filter):
12281 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12287 Fix the banding artifacts that are sometimes introduced into nearly flat
12288 regions by truncation to 8-bit color depth.
12289 Interpolate the gradients that should go where the bands are, and
12292 It is designed for playback only. Do not use it prior to
12293 lossy compression, because compression tends to lose the dither and
12294 bring back the bands.
12296 It accepts the following parameters:
12301 The maximum amount by which the filter will change any one pixel. This is also
12302 the threshold for detecting nearly flat regions. Acceptable values range from
12303 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12307 The neighborhood to fit the gradient to. A larger radius makes for smoother
12308 gradients, but also prevents the filter from modifying the pixels near detailed
12309 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12310 values will be clipped to the valid range.
12314 Alternatively, the options can be specified as a flat string:
12315 @var{strength}[:@var{radius}]
12317 @subsection Examples
12321 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12327 Specify radius, omitting the strength (which will fall-back to the default
12335 @anchor{graphmonitor}
12336 @section graphmonitor
12337 Show various filtergraph stats.
12339 With this filter one can debug complete filtergraph.
12340 Especially issues with links filling with queued frames.
12342 The filter accepts the following options:
12346 Set video output size. Default is @var{hd720}.
12349 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12352 Set output mode, can be @var{fulll} or @var{compact}.
12353 In @var{compact} mode only filters with some queued frames have displayed stats.
12356 Set flags which enable which stats are shown in video.
12358 Available values for flags are:
12361 Display number of queued frames in each link.
12363 @item frame_count_in
12364 Display number of frames taken from filter.
12366 @item frame_count_out
12367 Display number of frames given out from filter.
12370 Display current filtered frame pts.
12373 Display current filtered frame time.
12376 Display time base for filter link.
12379 Display used format for filter link.
12382 Display video size or number of audio channels in case of audio used by filter link.
12385 Display video frame rate or sample rate in case of audio used by filter link.
12388 Display link output status.
12392 Set upper limit for video rate of output stream, Default value is @var{25}.
12393 This guarantee that output video frame rate will not be higher than this value.
12397 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12398 and corrects the scene colors accordingly.
12400 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12402 The filter accepts the following options:
12406 The order of differentiation to be applied on the scene. Must be chosen in the range
12407 [0,2] and default value is 1.
12410 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12411 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12412 max value instead of calculating Minkowski distance.
12415 The standard deviation of Gaussian blur to be applied on the scene. Must be
12416 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12417 can't be equal to 0 if @var{difford} is greater than 0.
12420 @subsection Examples
12426 greyedge=difford=1:minknorm=5:sigma=2
12432 greyedge=difford=1:minknorm=0:sigma=2
12440 Apply a Hald CLUT to a video stream.
12442 First input is the video stream to process, and second one is the Hald CLUT.
12443 The Hald CLUT input can be a simple picture or a complete video stream.
12445 The filter accepts the following options:
12449 Force termination when the shortest input terminates. Default is @code{0}.
12451 Continue applying the last CLUT after the end of the stream. A value of
12452 @code{0} disable the filter after the last frame of the CLUT is reached.
12453 Default is @code{1}.
12456 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12457 filters share the same internals).
12459 This filter also supports the @ref{framesync} options.
12461 More information about the Hald CLUT can be found on Eskil Steenberg's website
12462 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12464 @subsection Workflow examples
12466 @subsubsection Hald CLUT video stream
12468 Generate an identity Hald CLUT stream altered with various effects:
12470 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
12473 Note: make sure you use a lossless codec.
12475 Then use it with @code{haldclut} to apply it on some random stream:
12477 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12480 The Hald CLUT will be applied to the 10 first seconds (duration of
12481 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12482 to the remaining frames of the @code{mandelbrot} stream.
12484 @subsubsection Hald CLUT with preview
12486 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12487 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12488 biggest possible square starting at the top left of the picture. The remaining
12489 padding pixels (bottom or right) will be ignored. This area can be used to add
12490 a preview of the Hald CLUT.
12492 Typically, the following generated Hald CLUT will be supported by the
12493 @code{haldclut} filter:
12496 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12497 pad=iw+320 [padded_clut];
12498 smptebars=s=320x256, split [a][b];
12499 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12500 [main][b] overlay=W-320" -frames:v 1 clut.png
12503 It contains the original and a preview of the effect of the CLUT: SMPTE color
12504 bars are displayed on the right-top, and below the same color bars processed by
12507 Then, the effect of this Hald CLUT can be visualized with:
12509 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12514 Flip the input video horizontally.
12516 For example, to horizontally flip the input video with @command{ffmpeg}:
12518 ffmpeg -i in.avi -vf "hflip" out.avi
12522 This filter applies a global color histogram equalization on a
12525 It can be used to correct video that has a compressed range of pixel
12526 intensities. The filter redistributes the pixel intensities to
12527 equalize their distribution across the intensity range. It may be
12528 viewed as an "automatically adjusting contrast filter". This filter is
12529 useful only for correcting degraded or poorly captured source
12532 The filter accepts the following options:
12536 Determine the amount of equalization to be applied. As the strength
12537 is reduced, the distribution of pixel intensities more-and-more
12538 approaches that of the input frame. The value must be a float number
12539 in the range [0,1] and defaults to 0.200.
12542 Set the maximum intensity that can generated and scale the output
12543 values appropriately. The strength should be set as desired and then
12544 the intensity can be limited if needed to avoid washing-out. The value
12545 must be a float number in the range [0,1] and defaults to 0.210.
12548 Set the antibanding level. If enabled the filter will randomly vary
12549 the luminance of output pixels by a small amount to avoid banding of
12550 the histogram. Possible values are @code{none}, @code{weak} or
12551 @code{strong}. It defaults to @code{none}.
12557 Compute and draw a color distribution histogram for the input video.
12559 The computed histogram is a representation of the color component
12560 distribution in an image.
12562 Standard histogram displays the color components distribution in an image.
12563 Displays color graph for each color component. Shows distribution of
12564 the Y, U, V, A or R, G, B components, depending on input format, in the
12565 current frame. Below each graph a color component scale meter is shown.
12567 The filter accepts the following options:
12571 Set height of level. Default value is @code{200}.
12572 Allowed range is [50, 2048].
12575 Set height of color scale. Default value is @code{12}.
12576 Allowed range is [0, 40].
12580 It accepts the following values:
12583 Per color component graphs are placed below each other.
12586 Per color component graphs are placed side by side.
12589 Presents information identical to that in the @code{parade}, except
12590 that the graphs representing color components are superimposed directly
12593 Default is @code{stack}.
12596 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12597 Default is @code{linear}.
12600 Set what color components to display.
12601 Default is @code{7}.
12604 Set foreground opacity. Default is @code{0.7}.
12607 Set background opacity. Default is @code{0.5}.
12610 @subsection Examples
12615 Calculate and draw histogram:
12617 ffplay -i input -vf histogram
12625 This is a high precision/quality 3d denoise filter. It aims to reduce
12626 image noise, producing smooth images and making still images really
12627 still. It should enhance compressibility.
12629 It accepts the following optional parameters:
12633 A non-negative floating point number which specifies spatial luma strength.
12634 It defaults to 4.0.
12636 @item chroma_spatial
12637 A non-negative floating point number which specifies spatial chroma strength.
12638 It defaults to 3.0*@var{luma_spatial}/4.0.
12641 A floating point number which specifies luma temporal strength. It defaults to
12642 6.0*@var{luma_spatial}/4.0.
12645 A floating point number which specifies chroma temporal strength. It defaults to
12646 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12649 @subsection Commands
12650 This filter supports same @ref{commands} as options.
12651 The command accepts the same syntax of the corresponding option.
12653 If the specified expression is not valid, it is kept at its current
12656 @anchor{hwdownload}
12657 @section hwdownload
12659 Download hardware frames to system memory.
12661 The input must be in hardware frames, and the output a non-hardware format.
12662 Not all formats will be supported on the output - it may be necessary to insert
12663 an additional @option{format} filter immediately following in the graph to get
12664 the output in a supported format.
12668 Map hardware frames to system memory or to another device.
12670 This filter has several different modes of operation; which one is used depends
12671 on the input and output formats:
12674 Hardware frame input, normal frame output
12676 Map the input frames to system memory and pass them to the output. If the
12677 original hardware frame is later required (for example, after overlaying
12678 something else on part of it), the @option{hwmap} filter can be used again
12679 in the next mode to retrieve it.
12681 Normal frame input, hardware frame output
12683 If the input is actually a software-mapped hardware frame, then unmap it -
12684 that is, return the original hardware frame.
12686 Otherwise, a device must be provided. Create new hardware surfaces on that
12687 device for the output, then map them back to the software format at the input
12688 and give those frames to the preceding filter. This will then act like the
12689 @option{hwupload} filter, but may be able to avoid an additional copy when
12690 the input is already in a compatible format.
12692 Hardware frame input and output
12694 A device must be supplied for the output, either directly or with the
12695 @option{derive_device} option. The input and output devices must be of
12696 different types and compatible - the exact meaning of this is
12697 system-dependent, but typically it means that they must refer to the same
12698 underlying hardware context (for example, refer to the same graphics card).
12700 If the input frames were originally created on the output device, then unmap
12701 to retrieve the original frames.
12703 Otherwise, map the frames to the output device - create new hardware frames
12704 on the output corresponding to the frames on the input.
12707 The following additional parameters are accepted:
12711 Set the frame mapping mode. Some combination of:
12714 The mapped frame should be readable.
12716 The mapped frame should be writeable.
12718 The mapping will always overwrite the entire frame.
12720 This may improve performance in some cases, as the original contents of the
12721 frame need not be loaded.
12723 The mapping must not involve any copying.
12725 Indirect mappings to copies of frames are created in some cases where either
12726 direct mapping is not possible or it would have unexpected properties.
12727 Setting this flag ensures that the mapping is direct and will fail if that is
12730 Defaults to @var{read+write} if not specified.
12732 @item derive_device @var{type}
12733 Rather than using the device supplied at initialisation, instead derive a new
12734 device of type @var{type} from the device the input frames exist on.
12737 In a hardware to hardware mapping, map in reverse - create frames in the sink
12738 and map them back to the source. This may be necessary in some cases where
12739 a mapping in one direction is required but only the opposite direction is
12740 supported by the devices being used.
12742 This option is dangerous - it may break the preceding filter in undefined
12743 ways if there are any additional constraints on that filter's output.
12744 Do not use it without fully understanding the implications of its use.
12750 Upload system memory frames to hardware surfaces.
12752 The device to upload to must be supplied when the filter is initialised. If
12753 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12754 option or with the @option{derive_device} option. The input and output devices
12755 must be of different types and compatible - the exact meaning of this is
12756 system-dependent, but typically it means that they must refer to the same
12757 underlying hardware context (for example, refer to the same graphics card).
12759 The following additional parameters are accepted:
12762 @item derive_device @var{type}
12763 Rather than using the device supplied at initialisation, instead derive a new
12764 device of type @var{type} from the device the input frames exist on.
12767 @anchor{hwupload_cuda}
12768 @section hwupload_cuda
12770 Upload system memory frames to a CUDA device.
12772 It accepts the following optional parameters:
12776 The number of the CUDA device to use
12781 Apply a high-quality magnification filter designed for pixel art. This filter
12782 was originally created by Maxim Stepin.
12784 It accepts the following option:
12788 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12789 @code{hq3x} and @code{4} for @code{hq4x}.
12790 Default is @code{3}.
12794 Stack input videos horizontally.
12796 All streams must be of same pixel format and of same height.
12798 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12799 to create same output.
12801 The filter accepts the following option:
12805 Set number of input streams. Default is 2.
12808 If set to 1, force the output to terminate when the shortest input
12809 terminates. Default value is 0.
12814 Modify the hue and/or the saturation of the input.
12816 It accepts the following parameters:
12820 Specify the hue angle as a number of degrees. It accepts an expression,
12821 and defaults to "0".
12824 Specify the saturation in the [-10,10] range. It accepts an expression and
12828 Specify the hue angle as a number of radians. It accepts an
12829 expression, and defaults to "0".
12832 Specify the brightness in the [-10,10] range. It accepts an expression and
12836 @option{h} and @option{H} are mutually exclusive, and can't be
12837 specified at the same time.
12839 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12840 expressions containing the following constants:
12844 frame count of the input frame starting from 0
12847 presentation timestamp of the input frame expressed in time base units
12850 frame rate of the input video, NAN if the input frame rate is unknown
12853 timestamp expressed in seconds, NAN if the input timestamp is unknown
12856 time base of the input video
12859 @subsection Examples
12863 Set the hue to 90 degrees and the saturation to 1.0:
12869 Same command but expressing the hue in radians:
12875 Rotate hue and make the saturation swing between 0
12876 and 2 over a period of 1 second:
12878 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12882 Apply a 3 seconds saturation fade-in effect starting at 0:
12884 hue="s=min(t/3\,1)"
12887 The general fade-in expression can be written as:
12889 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12893 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12895 hue="s=max(0\, min(1\, (8-t)/3))"
12898 The general fade-out expression can be written as:
12900 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12905 @subsection Commands
12907 This filter supports the following commands:
12913 Modify the hue and/or the saturation and/or brightness of the input video.
12914 The command accepts the same syntax of the corresponding option.
12916 If the specified expression is not valid, it is kept at its current
12920 @section hysteresis
12922 Grow first stream into second stream by connecting components.
12923 This makes it possible to build more robust edge masks.
12925 This filter accepts the following options:
12929 Set which planes will be processed as bitmap, unprocessed planes will be
12930 copied from first stream.
12931 By default value 0xf, all planes will be processed.
12934 Set threshold which is used in filtering. If pixel component value is higher than
12935 this value filter algorithm for connecting components is activated.
12936 By default value is 0.
12939 The @code{hysteresis} filter also supports the @ref{framesync} options.
12943 Detect video interlacing type.
12945 This filter tries to detect if the input frames are interlaced, progressive,
12946 top or bottom field first. It will also try to detect fields that are
12947 repeated between adjacent frames (a sign of telecine).
12949 Single frame detection considers only immediately adjacent frames when classifying each frame.
12950 Multiple frame detection incorporates the classification history of previous frames.
12952 The filter will log these metadata values:
12955 @item single.current_frame
12956 Detected type of current frame using single-frame detection. One of:
12957 ``tff'' (top field first), ``bff'' (bottom field first),
12958 ``progressive'', or ``undetermined''
12961 Cumulative number of frames detected as top field first using single-frame detection.
12964 Cumulative number of frames detected as top field first using multiple-frame detection.
12967 Cumulative number of frames detected as bottom field first using single-frame detection.
12969 @item multiple.current_frame
12970 Detected type of current frame using multiple-frame detection. One of:
12971 ``tff'' (top field first), ``bff'' (bottom field first),
12972 ``progressive'', or ``undetermined''
12975 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12977 @item single.progressive
12978 Cumulative number of frames detected as progressive using single-frame detection.
12980 @item multiple.progressive
12981 Cumulative number of frames detected as progressive using multiple-frame detection.
12983 @item single.undetermined
12984 Cumulative number of frames that could not be classified using single-frame detection.
12986 @item multiple.undetermined
12987 Cumulative number of frames that could not be classified using multiple-frame detection.
12989 @item repeated.current_frame
12990 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12992 @item repeated.neither
12993 Cumulative number of frames with no repeated field.
12996 Cumulative number of frames with the top field repeated from the previous frame's top field.
12998 @item repeated.bottom
12999 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13002 The filter accepts the following options:
13006 Set interlacing threshold.
13008 Set progressive threshold.
13010 Threshold for repeated field detection.
13012 Number of frames after which a given frame's contribution to the
13013 statistics is halved (i.e., it contributes only 0.5 to its
13014 classification). The default of 0 means that all frames seen are given
13015 full weight of 1.0 forever.
13016 @item analyze_interlaced_flag
13017 When this is not 0 then idet will use the specified number of frames to determine
13018 if the interlaced flag is accurate, it will not count undetermined frames.
13019 If the flag is found to be accurate it will be used without any further
13020 computations, if it is found to be inaccurate it will be cleared without any
13021 further computations. This allows inserting the idet filter as a low computational
13022 method to clean up the interlaced flag
13027 Deinterleave or interleave fields.
13029 This filter allows one to process interlaced images fields without
13030 deinterlacing them. Deinterleaving splits the input frame into 2
13031 fields (so called half pictures). Odd lines are moved to the top
13032 half of the output image, even lines to the bottom half.
13033 You can process (filter) them independently and then re-interleave them.
13035 The filter accepts the following options:
13039 @item chroma_mode, c
13040 @item alpha_mode, a
13041 Available values for @var{luma_mode}, @var{chroma_mode} and
13042 @var{alpha_mode} are:
13048 @item deinterleave, d
13049 Deinterleave fields, placing one above the other.
13051 @item interleave, i
13052 Interleave fields. Reverse the effect of deinterleaving.
13054 Default value is @code{none}.
13056 @item luma_swap, ls
13057 @item chroma_swap, cs
13058 @item alpha_swap, as
13059 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13062 @subsection Commands
13064 This filter supports the all above options as @ref{commands}.
13068 Apply inflate effect to the video.
13070 This filter replaces the pixel by the local(3x3) average by taking into account
13071 only values higher than the pixel.
13073 It accepts the following options:
13080 Limit the maximum change for each plane, default is 65535.
13081 If 0, plane will remain unchanged.
13084 @subsection Commands
13086 This filter supports the all above options as @ref{commands}.
13090 Simple interlacing filter from progressive contents. This interleaves upper (or
13091 lower) lines from odd frames with lower (or upper) lines from even frames,
13092 halving the frame rate and preserving image height.
13095 Original Original New Frame
13096 Frame 'j' Frame 'j+1' (tff)
13097 ========== =========== ==================
13098 Line 0 --------------------> Frame 'j' Line 0
13099 Line 1 Line 1 ----> Frame 'j+1' Line 1
13100 Line 2 ---------------------> Frame 'j' Line 2
13101 Line 3 Line 3 ----> Frame 'j+1' Line 3
13103 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13106 It accepts the following optional parameters:
13110 This determines whether the interlaced frame is taken from the even
13111 (tff - default) or odd (bff) lines of the progressive frame.
13114 Vertical lowpass filter to avoid twitter interlacing and
13115 reduce moire patterns.
13119 Disable vertical lowpass filter
13122 Enable linear filter (default)
13125 Enable complex filter. This will slightly less reduce twitter and moire
13126 but better retain detail and subjective sharpness impression.
13133 Deinterlace input video by applying Donald Graft's adaptive kernel
13134 deinterling. Work on interlaced parts of a video to produce
13135 progressive frames.
13137 The description of the accepted parameters follows.
13141 Set the threshold which affects the filter's tolerance when
13142 determining if a pixel line must be processed. It must be an integer
13143 in the range [0,255] and defaults to 10. A value of 0 will result in
13144 applying the process on every pixels.
13147 Paint pixels exceeding the threshold value to white if set to 1.
13151 Set the fields order. Swap fields if set to 1, leave fields alone if
13155 Enable additional sharpening if set to 1. Default is 0.
13158 Enable twoway sharpening if set to 1. Default is 0.
13161 @subsection Examples
13165 Apply default values:
13167 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13171 Enable additional sharpening:
13177 Paint processed pixels in white:
13185 Slowly update darker pixels.
13187 This filter makes short flashes of light appear longer.
13188 This filter accepts the following options:
13192 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13195 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13198 @section lenscorrection
13200 Correct radial lens distortion
13202 This filter can be used to correct for radial distortion as can result from the use
13203 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13204 one can use tools available for example as part of opencv or simply trial-and-error.
13205 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13206 and extract the k1 and k2 coefficients from the resulting matrix.
13208 Note that effectively the same filter is available in the open-source tools Krita and
13209 Digikam from the KDE project.
13211 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13212 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13213 brightness distribution, so you may want to use both filters together in certain
13214 cases, though you will have to take care of ordering, i.e. whether vignetting should
13215 be applied before or after lens correction.
13217 @subsection Options
13219 The filter accepts the following options:
13223 Relative x-coordinate of the focal point of the image, and thereby the center of the
13224 distortion. This value has a range [0,1] and is expressed as fractions of the image
13225 width. Default is 0.5.
13227 Relative y-coordinate of the focal point of the image, and thereby the center of the
13228 distortion. This value has a range [0,1] and is expressed as fractions of the image
13229 height. Default is 0.5.
13231 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13232 no correction. Default is 0.
13234 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13235 0 means no correction. Default is 0.
13238 The formula that generates the correction is:
13240 @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)
13242 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13243 distances from the focal point in the source and target images, respectively.
13247 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13249 The @code{lensfun} filter requires the camera make, camera model, and lens model
13250 to apply the lens correction. The filter will load the lensfun database and
13251 query it to find the corresponding camera and lens entries in the database. As
13252 long as these entries can be found with the given options, the filter can
13253 perform corrections on frames. Note that incomplete strings will result in the
13254 filter choosing the best match with the given options, and the filter will
13255 output the chosen camera and lens models (logged with level "info"). You must
13256 provide the make, camera model, and lens model as they are required.
13258 The filter accepts the following options:
13262 The make of the camera (for example, "Canon"). This option is required.
13265 The model of the camera (for example, "Canon EOS 100D"). This option is
13269 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13270 option is required.
13273 The type of correction to apply. The following values are valid options:
13277 Enables fixing lens vignetting.
13280 Enables fixing lens geometry. This is the default.
13283 Enables fixing chromatic aberrations.
13286 Enables fixing lens vignetting and lens geometry.
13289 Enables fixing lens vignetting and chromatic aberrations.
13292 Enables fixing both lens geometry and chromatic aberrations.
13295 Enables all possible corrections.
13299 The focal length of the image/video (zoom; expected constant for video). For
13300 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13301 range should be chosen when using that lens. Default 18.
13304 The aperture of the image/video (expected constant for video). Note that
13305 aperture is only used for vignetting correction. Default 3.5.
13307 @item focus_distance
13308 The focus distance of the image/video (expected constant for video). Note that
13309 focus distance is only used for vignetting and only slightly affects the
13310 vignetting correction process. If unknown, leave it at the default value (which
13314 The scale factor which is applied after transformation. After correction the
13315 video is no longer necessarily rectangular. This parameter controls how much of
13316 the resulting image is visible. The value 0 means that a value will be chosen
13317 automatically such that there is little or no unmapped area in the output
13318 image. 1.0 means that no additional scaling is done. Lower values may result
13319 in more of the corrected image being visible, while higher values may avoid
13320 unmapped areas in the output.
13322 @item target_geometry
13323 The target geometry of the output image/video. The following values are valid
13327 @item rectilinear (default)
13330 @item equirectangular
13331 @item fisheye_orthographic
13332 @item fisheye_stereographic
13333 @item fisheye_equisolid
13334 @item fisheye_thoby
13337 Apply the reverse of image correction (instead of correcting distortion, apply
13340 @item interpolation
13341 The type of interpolation used when correcting distortion. The following values
13346 @item linear (default)
13351 @subsection Examples
13355 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13356 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13360 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
13364 Apply the same as before, but only for the first 5 seconds of video.
13367 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
13374 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13375 score between two input videos.
13377 The obtained VMAF score is printed through the logging system.
13379 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13380 After installing the library it can be enabled using:
13381 @code{./configure --enable-libvmaf}.
13382 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13384 The filter has following options:
13388 Set the model path which is to be used for SVM.
13389 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13392 Set the file path to be used to store logs.
13395 Set the format of the log file (csv, json or xml).
13397 @item enable_transform
13398 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13399 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13400 Default value: @code{false}
13403 Invokes the phone model which will generate VMAF scores higher than in the
13404 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13405 Default value: @code{false}
13408 Enables computing psnr along with vmaf.
13409 Default value: @code{false}
13412 Enables computing ssim along with vmaf.
13413 Default value: @code{false}
13416 Enables computing ms_ssim along with vmaf.
13417 Default value: @code{false}
13420 Set the pool method to be used for computing vmaf.
13421 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13424 Set number of threads to be used when computing vmaf.
13425 Default value: @code{0}, which makes use of all available logical processors.
13428 Set interval for frame subsampling used when computing vmaf.
13429 Default value: @code{1}
13431 @item enable_conf_interval
13432 Enables confidence interval.
13433 Default value: @code{false}
13436 This filter also supports the @ref{framesync} options.
13438 @subsection Examples
13441 On the below examples the input file @file{main.mpg} being processed is
13442 compared with the reference file @file{ref.mpg}.
13445 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13449 Example with options:
13451 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13455 Example with options and different containers:
13457 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 -
13463 Limits the pixel components values to the specified range [min, max].
13465 The filter accepts the following options:
13469 Lower bound. Defaults to the lowest allowed value for the input.
13472 Upper bound. Defaults to the highest allowed value for the input.
13475 Specify which planes will be processed. Defaults to all available.
13482 The filter accepts the following options:
13486 Set the number of loops. Setting this value to -1 will result in infinite loops.
13490 Set maximal size in number of frames. Default is 0.
13493 Set first frame of loop. Default is 0.
13496 @subsection Examples
13500 Loop single first frame infinitely:
13502 loop=loop=-1:size=1:start=0
13506 Loop single first frame 10 times:
13508 loop=loop=10:size=1:start=0
13512 Loop 10 first frames 5 times:
13514 loop=loop=5:size=10:start=0
13520 Apply a 1D LUT to an input video.
13522 The filter accepts the following options:
13526 Set the 1D LUT file name.
13528 Currently supported formats:
13537 Select interpolation mode.
13539 Available values are:
13543 Use values from the nearest defined point.
13545 Interpolate values using the linear interpolation.
13547 Interpolate values using the cosine interpolation.
13549 Interpolate values using the cubic interpolation.
13551 Interpolate values using the spline interpolation.
13558 Apply a 3D LUT to an input video.
13560 The filter accepts the following options:
13564 Set the 3D LUT file name.
13566 Currently supported formats:
13580 Select interpolation mode.
13582 Available values are:
13586 Use values from the nearest defined point.
13588 Interpolate values using the 8 points defining a cube.
13590 Interpolate values using a tetrahedron.
13596 Turn certain luma values into transparency.
13598 The filter accepts the following options:
13602 Set the luma which will be used as base for transparency.
13603 Default value is @code{0}.
13606 Set the range of luma values to be keyed out.
13607 Default value is @code{0.01}.
13610 Set the range of softness. Default value is @code{0}.
13611 Use this to control gradual transition from zero to full transparency.
13614 @subsection Commands
13615 This filter supports same @ref{commands} as options.
13616 The command accepts the same syntax of the corresponding option.
13618 If the specified expression is not valid, it is kept at its current
13621 @section lut, lutrgb, lutyuv
13623 Compute a look-up table for binding each pixel component input value
13624 to an output value, and apply it to the input video.
13626 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13627 to an RGB input video.
13629 These filters accept the following parameters:
13632 set first pixel component expression
13634 set second pixel component expression
13636 set third pixel component expression
13638 set fourth pixel component expression, corresponds to the alpha component
13641 set red component expression
13643 set green component expression
13645 set blue component expression
13647 alpha component expression
13650 set Y/luminance component expression
13652 set U/Cb component expression
13654 set V/Cr component expression
13657 Each of them specifies the expression to use for computing the lookup table for
13658 the corresponding pixel component values.
13660 The exact component associated to each of the @var{c*} options depends on the
13663 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13664 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13666 The expressions can contain the following constants and functions:
13671 The input width and height.
13674 The input value for the pixel component.
13677 The input value, clipped to the @var{minval}-@var{maxval} range.
13680 The maximum value for the pixel component.
13683 The minimum value for the pixel component.
13686 The negated value for the pixel component value, clipped to the
13687 @var{minval}-@var{maxval} range; it corresponds to the expression
13688 "maxval-clipval+minval".
13691 The computed value in @var{val}, clipped to the
13692 @var{minval}-@var{maxval} range.
13694 @item gammaval(gamma)
13695 The computed gamma correction value of the pixel component value,
13696 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13698 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13702 All expressions default to "val".
13704 @subsection Examples
13708 Negate input video:
13710 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13711 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13714 The above is the same as:
13716 lutrgb="r=negval:g=negval:b=negval"
13717 lutyuv="y=negval:u=negval:v=negval"
13727 Remove chroma components, turning the video into a graytone image:
13729 lutyuv="u=128:v=128"
13733 Apply a luma burning effect:
13739 Remove green and blue components:
13745 Set a constant alpha channel value on input:
13747 format=rgba,lutrgb=a="maxval-minval/2"
13751 Correct luminance gamma by a factor of 0.5:
13753 lutyuv=y=gammaval(0.5)
13757 Discard least significant bits of luma:
13759 lutyuv=y='bitand(val, 128+64+32)'
13763 Technicolor like effect:
13765 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13769 @section lut2, tlut2
13771 The @code{lut2} filter takes two input streams and outputs one
13774 The @code{tlut2} (time lut2) filter takes two consecutive frames
13775 from one single stream.
13777 This filter accepts the following parameters:
13780 set first pixel component expression
13782 set second pixel component expression
13784 set third pixel component expression
13786 set fourth pixel component expression, corresponds to the alpha component
13789 set output bit depth, only available for @code{lut2} filter. By default is 0,
13790 which means bit depth is automatically picked from first input format.
13793 The @code{lut2} filter also supports the @ref{framesync} options.
13795 Each of them specifies the expression to use for computing the lookup table for
13796 the corresponding pixel component values.
13798 The exact component associated to each of the @var{c*} options depends on the
13801 The expressions can contain the following constants:
13806 The input width and height.
13809 The first input value for the pixel component.
13812 The second input value for the pixel component.
13815 The first input video bit depth.
13818 The second input video bit depth.
13821 All expressions default to "x".
13823 @subsection Examples
13827 Highlight differences between two RGB video streams:
13829 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)'
13833 Highlight differences between two YUV video streams:
13835 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)'
13839 Show max difference between two video streams:
13841 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)))'
13845 @section maskedclamp
13847 Clamp the first input stream with the second input and third input stream.
13849 Returns the value of first stream to be between second input
13850 stream - @code{undershoot} and third input stream + @code{overshoot}.
13852 This filter accepts the following options:
13855 Default value is @code{0}.
13858 Default value is @code{0}.
13861 Set which planes will be processed as bitmap, unprocessed planes will be
13862 copied from first stream.
13863 By default value 0xf, all planes will be processed.
13868 Merge the second and third input stream into output stream using absolute differences
13869 between second input stream and first input stream and absolute difference between
13870 third input stream and first input stream. The picked value will be from second input
13871 stream if second absolute difference is greater than first one or from third input stream
13874 This filter accepts the following options:
13877 Set which planes will be processed as bitmap, unprocessed planes will be
13878 copied from first stream.
13879 By default value 0xf, all planes will be processed.
13882 @section maskedmerge
13884 Merge the first input stream with the second input stream using per pixel
13885 weights in the third input stream.
13887 A value of 0 in the third stream pixel component means that pixel component
13888 from first stream is returned unchanged, while maximum value (eg. 255 for
13889 8-bit videos) means that pixel component from second stream is returned
13890 unchanged. Intermediate values define the amount of merging between both
13891 input stream's pixel components.
13893 This filter accepts the following options:
13896 Set which planes will be processed as bitmap, unprocessed planes will be
13897 copied from first stream.
13898 By default value 0xf, all planes will be processed.
13903 Merge the second and third input stream into output stream using absolute differences
13904 between second input stream and first input stream and absolute difference between
13905 third input stream and first input stream. The picked value will be from second input
13906 stream if second absolute difference is less than first one or from third input stream
13909 This filter accepts the following options:
13912 Set which planes will be processed as bitmap, unprocessed planes will be
13913 copied from first stream.
13914 By default value 0xf, all planes will be processed.
13917 @section maskedthreshold
13918 Pick pixels comparing absolute difference of two video streams with fixed
13921 If absolute difference between pixel component of first and second video
13922 stream is equal or lower than user supplied threshold than pixel component
13923 from first video stream is picked, otherwise pixel component from second
13924 video stream is picked.
13926 This filter accepts the following options:
13929 Set threshold used when picking pixels from absolute difference from two input
13933 Set which planes will be processed as bitmap, unprocessed planes will be
13934 copied from second stream.
13935 By default value 0xf, all planes will be processed.
13939 Create mask from input video.
13941 For example it is useful to create motion masks after @code{tblend} filter.
13943 This filter accepts the following options:
13947 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13950 Set high threshold. Any pixel component higher than this value will be set to max value
13951 allowed for current pixel format.
13954 Set planes to filter, by default all available planes are filtered.
13957 Fill all frame pixels with this value.
13960 Set max average pixel value for frame. If sum of all pixel components is higher that this
13961 average, output frame will be completely filled with value set by @var{fill} option.
13962 Typically useful for scene changes when used in combination with @code{tblend} filter.
13967 Apply motion-compensation deinterlacing.
13969 It needs one field per frame as input and must thus be used together
13970 with yadif=1/3 or equivalent.
13972 This filter accepts the following options:
13975 Set the deinterlacing mode.
13977 It accepts one of the following values:
13982 use iterative motion estimation
13984 like @samp{slow}, but use multiple reference frames.
13986 Default value is @samp{fast}.
13989 Set the picture field parity assumed for the input video. It must be
13990 one of the following values:
13994 assume top field first
13996 assume bottom field first
13999 Default value is @samp{bff}.
14002 Set per-block quantization parameter (QP) used by the internal
14005 Higher values should result in a smoother motion vector field but less
14006 optimal individual vectors. Default value is 1.
14011 Pick median pixel from certain rectangle defined by radius.
14013 This filter accepts the following options:
14017 Set horizontal radius size. Default value is @code{1}.
14018 Allowed range is integer from 1 to 127.
14021 Set which planes to process. Default is @code{15}, which is all available planes.
14024 Set vertical radius size. Default value is @code{0}.
14025 Allowed range is integer from 0 to 127.
14026 If it is 0, value will be picked from horizontal @code{radius} option.
14029 Set median percentile. Default value is @code{0.5}.
14030 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14031 minimum values, and @code{1} maximum values.
14034 @subsection Commands
14035 This filter supports same @ref{commands} as options.
14036 The command accepts the same syntax of the corresponding option.
14038 If the specified expression is not valid, it is kept at its current
14041 @section mergeplanes
14043 Merge color channel components from several video streams.
14045 The filter accepts up to 4 input streams, and merge selected input
14046 planes to the output video.
14048 This filter accepts the following options:
14051 Set input to output plane mapping. Default is @code{0}.
14053 The mappings is specified as a bitmap. It should be specified as a
14054 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14055 mapping for the first plane of the output stream. 'A' sets the number of
14056 the input stream to use (from 0 to 3), and 'a' the plane number of the
14057 corresponding input to use (from 0 to 3). The rest of the mappings is
14058 similar, 'Bb' describes the mapping for the output stream second
14059 plane, 'Cc' describes the mapping for the output stream third plane and
14060 'Dd' describes the mapping for the output stream fourth plane.
14063 Set output pixel format. Default is @code{yuva444p}.
14066 @subsection Examples
14070 Merge three gray video streams of same width and height into single video stream:
14072 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14076 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14078 [a0][a1]mergeplanes=0x00010210:yuva444p
14082 Swap Y and A plane in yuva444p stream:
14084 format=yuva444p,mergeplanes=0x03010200:yuva444p
14088 Swap U and V plane in yuv420p stream:
14090 format=yuv420p,mergeplanes=0x000201:yuv420p
14094 Cast a rgb24 clip to yuv444p:
14096 format=rgb24,mergeplanes=0x000102:yuv444p
14102 Estimate and export motion vectors using block matching algorithms.
14103 Motion vectors are stored in frame side data to be used by other filters.
14105 This filter accepts the following options:
14108 Specify the motion estimation method. Accepts one of the following values:
14112 Exhaustive search algorithm.
14114 Three step search algorithm.
14116 Two dimensional logarithmic search algorithm.
14118 New three step search algorithm.
14120 Four step search algorithm.
14122 Diamond search algorithm.
14124 Hexagon-based search algorithm.
14126 Enhanced predictive zonal search algorithm.
14128 Uneven multi-hexagon search algorithm.
14130 Default value is @samp{esa}.
14133 Macroblock size. Default @code{16}.
14136 Search parameter. Default @code{7}.
14139 @section midequalizer
14141 Apply Midway Image Equalization effect using two video streams.
14143 Midway Image Equalization adjusts a pair of images to have the same
14144 histogram, while maintaining their dynamics as much as possible. It's
14145 useful for e.g. matching exposures from a pair of stereo cameras.
14147 This filter has two inputs and one output, which must be of same pixel format, but
14148 may be of different sizes. The output of filter is first input adjusted with
14149 midway histogram of both inputs.
14151 This filter accepts the following option:
14155 Set which planes to process. Default is @code{15}, which is all available planes.
14158 @section minterpolate
14160 Convert the video to specified frame rate using motion interpolation.
14162 This filter accepts the following options:
14165 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}.
14168 Motion interpolation mode. Following values are accepted:
14171 Duplicate previous or next frame for interpolating new ones.
14173 Blend source frames. Interpolated frame is mean of previous and next frames.
14175 Motion compensated interpolation. Following options are effective when this mode is selected:
14179 Motion compensation mode. Following values are accepted:
14182 Overlapped block motion compensation.
14184 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14186 Default mode is @samp{obmc}.
14189 Motion estimation mode. Following values are accepted:
14192 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14194 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14196 Default mode is @samp{bilat}.
14199 The algorithm to be used for motion estimation. Following values are accepted:
14202 Exhaustive search algorithm.
14204 Three step search algorithm.
14206 Two dimensional logarithmic search algorithm.
14208 New three step search algorithm.
14210 Four step search algorithm.
14212 Diamond search algorithm.
14214 Hexagon-based search algorithm.
14216 Enhanced predictive zonal search algorithm.
14218 Uneven multi-hexagon search algorithm.
14220 Default algorithm is @samp{epzs}.
14223 Macroblock size. Default @code{16}.
14226 Motion estimation search parameter. Default @code{32}.
14229 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).
14234 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:
14237 Disable scene change detection.
14239 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14241 Default method is @samp{fdiff}.
14243 @item scd_threshold
14244 Scene change detection threshold. Default is @code{10.}.
14249 Mix several video input streams into one video stream.
14251 A description of the accepted options follows.
14255 The number of inputs. If unspecified, it defaults to 2.
14258 Specify weight of each input video stream as sequence.
14259 Each weight is separated by space. If number of weights
14260 is smaller than number of @var{frames} last specified
14261 weight will be used for all remaining unset weights.
14264 Specify scale, if it is set it will be multiplied with sum
14265 of each weight multiplied with pixel values to give final destination
14266 pixel value. By default @var{scale} is auto scaled to sum of weights.
14269 Specify how end of stream is determined.
14272 The duration of the longest input. (default)
14275 The duration of the shortest input.
14278 The duration of the first input.
14282 @section mpdecimate
14284 Drop frames that do not differ greatly from the previous frame in
14285 order to reduce frame rate.
14287 The main use of this filter is for very-low-bitrate encoding
14288 (e.g. streaming over dialup modem), but it could in theory be used for
14289 fixing movies that were inverse-telecined incorrectly.
14291 A description of the accepted options follows.
14295 Set the maximum number of consecutive frames which can be dropped (if
14296 positive), or the minimum interval between dropped frames (if
14297 negative). If the value is 0, the frame is dropped disregarding the
14298 number of previous sequentially dropped frames.
14300 Default value is 0.
14305 Set the dropping threshold values.
14307 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14308 represent actual pixel value differences, so a threshold of 64
14309 corresponds to 1 unit of difference for each pixel, or the same spread
14310 out differently over the block.
14312 A frame is a candidate for dropping if no 8x8 blocks differ by more
14313 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14314 meaning the whole image) differ by more than a threshold of @option{lo}.
14316 Default value for @option{hi} is 64*12, default value for @option{lo} is
14317 64*5, and default value for @option{frac} is 0.33.
14323 Negate (invert) the input video.
14325 It accepts the following option:
14330 With value 1, it negates the alpha component, if present. Default value is 0.
14336 Denoise frames using Non-Local Means algorithm.
14338 Each pixel is adjusted by looking for other pixels with similar contexts. This
14339 context similarity is defined by comparing their surrounding patches of size
14340 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14343 Note that the research area defines centers for patches, which means some
14344 patches will be made of pixels outside that research area.
14346 The filter accepts the following options.
14350 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14353 Set patch size. Default is 7. Must be odd number in range [0, 99].
14356 Same as @option{p} but for chroma planes.
14358 The default value is @var{0} and means automatic.
14361 Set research size. Default is 15. Must be odd number in range [0, 99].
14364 Same as @option{r} but for chroma planes.
14366 The default value is @var{0} and means automatic.
14371 Deinterlace video using neural network edge directed interpolation.
14373 This filter accepts the following options:
14377 Mandatory option, without binary file filter can not work.
14378 Currently file can be found here:
14379 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14382 Set which frames to deinterlace, by default it is @code{all}.
14383 Can be @code{all} or @code{interlaced}.
14386 Set mode of operation.
14388 Can be one of the following:
14392 Use frame flags, both fields.
14394 Use frame flags, single field.
14396 Use top field only.
14398 Use bottom field only.
14400 Use both fields, top first.
14402 Use both fields, bottom first.
14406 Set which planes to process, by default filter process all frames.
14409 Set size of local neighborhood around each pixel, used by the predictor neural
14412 Can be one of the following:
14425 Set the number of neurons in predictor neural network.
14426 Can be one of the following:
14437 Controls the number of different neural network predictions that are blended
14438 together to compute the final output value. Can be @code{fast}, default or
14442 Set which set of weights to use in the predictor.
14443 Can be one of the following:
14447 weights trained to minimize absolute error
14449 weights trained to minimize squared error
14453 Controls whether or not the prescreener neural network is used to decide
14454 which pixels should be processed by the predictor neural network and which
14455 can be handled by simple cubic interpolation.
14456 The prescreener is trained to know whether cubic interpolation will be
14457 sufficient for a pixel or whether it should be predicted by the predictor nn.
14458 The computational complexity of the prescreener nn is much less than that of
14459 the predictor nn. Since most pixels can be handled by cubic interpolation,
14460 using the prescreener generally results in much faster processing.
14461 The prescreener is pretty accurate, so the difference between using it and not
14462 using it is almost always unnoticeable.
14464 Can be one of the following:
14472 Default is @code{new}.
14475 Set various debugging flags.
14480 Force libavfilter not to use any of the specified pixel formats for the
14481 input to the next filter.
14483 It accepts the following parameters:
14487 A '|'-separated list of pixel format names, such as
14488 pix_fmts=yuv420p|monow|rgb24".
14492 @subsection Examples
14496 Force libavfilter to use a format different from @var{yuv420p} for the
14497 input to the vflip filter:
14499 noformat=pix_fmts=yuv420p,vflip
14503 Convert the input video to any of the formats not contained in the list:
14505 noformat=yuv420p|yuv444p|yuv410p
14511 Add noise on video input frame.
14513 The filter accepts the following options:
14521 Set noise seed for specific pixel component or all pixel components in case
14522 of @var{all_seed}. Default value is @code{123457}.
14524 @item all_strength, alls
14525 @item c0_strength, c0s
14526 @item c1_strength, c1s
14527 @item c2_strength, c2s
14528 @item c3_strength, c3s
14529 Set noise strength for specific pixel component or all pixel components in case
14530 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14532 @item all_flags, allf
14533 @item c0_flags, c0f
14534 @item c1_flags, c1f
14535 @item c2_flags, c2f
14536 @item c3_flags, c3f
14537 Set pixel component flags or set flags for all components if @var{all_flags}.
14538 Available values for component flags are:
14541 averaged temporal noise (smoother)
14543 mix random noise with a (semi)regular pattern
14545 temporal noise (noise pattern changes between frames)
14547 uniform noise (gaussian otherwise)
14551 @subsection Examples
14553 Add temporal and uniform noise to input video:
14555 noise=alls=20:allf=t+u
14560 Normalize RGB video (aka histogram stretching, contrast stretching).
14561 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14563 For each channel of each frame, the filter computes the input range and maps
14564 it linearly to the user-specified output range. The output range defaults
14565 to the full dynamic range from pure black to pure white.
14567 Temporal smoothing can be used on the input range to reduce flickering (rapid
14568 changes in brightness) caused when small dark or bright objects enter or leave
14569 the scene. This is similar to the auto-exposure (automatic gain control) on a
14570 video camera, and, like a video camera, it may cause a period of over- or
14571 under-exposure of the video.
14573 The R,G,B channels can be normalized independently, which may cause some
14574 color shifting, or linked together as a single channel, which prevents
14575 color shifting. Linked normalization preserves hue. Independent normalization
14576 does not, so it can be used to remove some color casts. Independent and linked
14577 normalization can be combined in any ratio.
14579 The normalize filter accepts the following options:
14584 Colors which define the output range. The minimum input value is mapped to
14585 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14586 The defaults are black and white respectively. Specifying white for
14587 @var{blackpt} and black for @var{whitept} will give color-inverted,
14588 normalized video. Shades of grey can be used to reduce the dynamic range
14589 (contrast). Specifying saturated colors here can create some interesting
14593 The number of previous frames to use for temporal smoothing. The input range
14594 of each channel is smoothed using a rolling average over the current frame
14595 and the @var{smoothing} previous frames. The default is 0 (no temporal
14599 Controls the ratio of independent (color shifting) channel normalization to
14600 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14601 independent. Defaults to 1.0 (fully independent).
14604 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14605 expensive no-op. Defaults to 1.0 (full strength).
14609 @subsection Commands
14610 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14611 The command accepts the same syntax of the corresponding option.
14613 If the specified expression is not valid, it is kept at its current
14616 @subsection Examples
14618 Stretch video contrast to use the full dynamic range, with no temporal
14619 smoothing; may flicker depending on the source content:
14621 normalize=blackpt=black:whitept=white:smoothing=0
14624 As above, but with 50 frames of temporal smoothing; flicker should be
14625 reduced, depending on the source content:
14627 normalize=blackpt=black:whitept=white:smoothing=50
14630 As above, but with hue-preserving linked channel normalization:
14632 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14635 As above, but with half strength:
14637 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14640 Map the darkest input color to red, the brightest input color to cyan:
14642 normalize=blackpt=red:whitept=cyan
14647 Pass the video source unchanged to the output.
14650 Optical Character Recognition
14652 This filter uses Tesseract for optical character recognition. To enable
14653 compilation of this filter, you need to configure FFmpeg with
14654 @code{--enable-libtesseract}.
14656 It accepts the following options:
14660 Set datapath to tesseract data. Default is to use whatever was
14661 set at installation.
14664 Set language, default is "eng".
14667 Set character whitelist.
14670 Set character blacklist.
14673 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14674 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14678 Apply a video transform using libopencv.
14680 To enable this filter, install the libopencv library and headers and
14681 configure FFmpeg with @code{--enable-libopencv}.
14683 It accepts the following parameters:
14688 The name of the libopencv filter to apply.
14690 @item filter_params
14691 The parameters to pass to the libopencv filter. If not specified, the default
14692 values are assumed.
14696 Refer to the official libopencv documentation for more precise
14698 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14700 Several libopencv filters are supported; see the following subsections.
14705 Dilate an image by using a specific structuring element.
14706 It corresponds to the libopencv function @code{cvDilate}.
14708 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14710 @var{struct_el} represents a structuring element, and has the syntax:
14711 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14713 @var{cols} and @var{rows} represent the number of columns and rows of
14714 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14715 point, and @var{shape} the shape for the structuring element. @var{shape}
14716 must be "rect", "cross", "ellipse", or "custom".
14718 If the value for @var{shape} is "custom", it must be followed by a
14719 string of the form "=@var{filename}". The file with name
14720 @var{filename} is assumed to represent a binary image, with each
14721 printable character corresponding to a bright pixel. When a custom
14722 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14723 or columns and rows of the read file are assumed instead.
14725 The default value for @var{struct_el} is "3x3+0x0/rect".
14727 @var{nb_iterations} specifies the number of times the transform is
14728 applied to the image, and defaults to 1.
14732 # Use the default values
14735 # Dilate using a structuring element with a 5x5 cross, iterating two times
14736 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14738 # Read the shape from the file diamond.shape, iterating two times.
14739 # The file diamond.shape may contain a pattern of characters like this
14745 # The specified columns and rows are ignored
14746 # but the anchor point coordinates are not
14747 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14752 Erode an image by using a specific structuring element.
14753 It corresponds to the libopencv function @code{cvErode}.
14755 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14756 with the same syntax and semantics as the @ref{dilate} filter.
14760 Smooth the input video.
14762 The filter takes the following parameters:
14763 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14765 @var{type} is the type of smooth filter to apply, and must be one of
14766 the following values: "blur", "blur_no_scale", "median", "gaussian",
14767 or "bilateral". The default value is "gaussian".
14769 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14770 depends on the smooth type. @var{param1} and
14771 @var{param2} accept integer positive values or 0. @var{param3} and
14772 @var{param4} accept floating point values.
14774 The default value for @var{param1} is 3. The default value for the
14775 other parameters is 0.
14777 These parameters correspond to the parameters assigned to the
14778 libopencv function @code{cvSmooth}.
14780 @section oscilloscope
14782 2D Video Oscilloscope.
14784 Useful to measure spatial impulse, step responses, chroma delays, etc.
14786 It accepts the following parameters:
14790 Set scope center x position.
14793 Set scope center y position.
14796 Set scope size, relative to frame diagonal.
14799 Set scope tilt/rotation.
14805 Set trace center x position.
14808 Set trace center y position.
14811 Set trace width, relative to width of frame.
14814 Set trace height, relative to height of frame.
14817 Set which components to trace. By default it traces first three components.
14820 Draw trace grid. By default is enabled.
14823 Draw some statistics. By default is enabled.
14826 Draw scope. By default is enabled.
14829 @subsection Commands
14830 This filter supports same @ref{commands} as options.
14831 The command accepts the same syntax of the corresponding option.
14833 If the specified expression is not valid, it is kept at its current
14836 @subsection Examples
14840 Inspect full first row of video frame.
14842 oscilloscope=x=0.5:y=0:s=1
14846 Inspect full last row of video frame.
14848 oscilloscope=x=0.5:y=1:s=1
14852 Inspect full 5th line of video frame of height 1080.
14854 oscilloscope=x=0.5:y=5/1080:s=1
14858 Inspect full last column of video frame.
14860 oscilloscope=x=1:y=0.5:s=1:t=1
14868 Overlay one video on top of another.
14870 It takes two inputs and has one output. The first input is the "main"
14871 video on which the second input is overlaid.
14873 It accepts the following parameters:
14875 A description of the accepted options follows.
14880 Set the expression for the x and y coordinates of the overlaid video
14881 on the main video. Default value is "0" for both expressions. In case
14882 the expression is invalid, it is set to a huge value (meaning that the
14883 overlay will not be displayed within the output visible area).
14886 See @ref{framesync}.
14889 Set when the expressions for @option{x}, and @option{y} are evaluated.
14891 It accepts the following values:
14894 only evaluate expressions once during the filter initialization or
14895 when a command is processed
14898 evaluate expressions for each incoming frame
14901 Default value is @samp{frame}.
14904 See @ref{framesync}.
14907 Set the format for the output video.
14909 It accepts the following values:
14912 force YUV420 output
14915 force YUV420p10 output
14918 force YUV422 output
14921 force YUV422p10 output
14924 force YUV444 output
14927 force packed RGB output
14930 force planar RGB output
14933 automatically pick format
14936 Default value is @samp{yuv420}.
14939 See @ref{framesync}.
14942 Set format of alpha of the overlaid video, it can be @var{straight} or
14943 @var{premultiplied}. Default is @var{straight}.
14946 The @option{x}, and @option{y} expressions can contain the following
14952 The main input width and height.
14956 The overlay input width and height.
14960 The computed values for @var{x} and @var{y}. They are evaluated for
14965 horizontal and vertical chroma subsample values of the output
14966 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14970 the number of input frame, starting from 0
14973 the position in the file of the input frame, NAN if unknown
14976 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14980 This filter also supports the @ref{framesync} options.
14982 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14983 when evaluation is done @emph{per frame}, and will evaluate to NAN
14984 when @option{eval} is set to @samp{init}.
14986 Be aware that frames are taken from each input video in timestamp
14987 order, hence, if their initial timestamps differ, it is a good idea
14988 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14989 have them begin in the same zero timestamp, as the example for
14990 the @var{movie} filter does.
14992 You can chain together more overlays but you should test the
14993 efficiency of such approach.
14995 @subsection Commands
14997 This filter supports the following commands:
15001 Modify the x and y of the overlay input.
15002 The command accepts the same syntax of the corresponding option.
15004 If the specified expression is not valid, it is kept at its current
15008 @subsection Examples
15012 Draw the overlay at 10 pixels from the bottom right corner of the main
15015 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15018 Using named options the example above becomes:
15020 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15024 Insert a transparent PNG logo in the bottom left corner of the input,
15025 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15027 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15031 Insert 2 different transparent PNG logos (second logo on bottom
15032 right corner) using the @command{ffmpeg} tool:
15034 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
15038 Add a transparent color layer on top of the main video; @code{WxH}
15039 must specify the size of the main input to the overlay filter:
15041 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15045 Play an original video and a filtered version (here with the deshake
15046 filter) side by side using the @command{ffplay} tool:
15048 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15051 The above command is the same as:
15053 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15057 Make a sliding overlay appearing from the left to the right top part of the
15058 screen starting since time 2:
15060 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15064 Compose output by putting two input videos side to side:
15066 ffmpeg -i left.avi -i right.avi -filter_complex "
15067 nullsrc=size=200x100 [background];
15068 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15069 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15070 [background][left] overlay=shortest=1 [background+left];
15071 [background+left][right] overlay=shortest=1:x=100 [left+right]
15076 Mask 10-20 seconds of a video by applying the delogo filter to a section
15078 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15079 -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]'
15084 Chain several overlays in cascade:
15086 nullsrc=s=200x200 [bg];
15087 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15088 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15089 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15090 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15091 [in3] null, [mid2] overlay=100:100 [out0]
15096 @anchor{overlay_cuda}
15097 @section overlay_cuda
15099 Overlay one video on top of another.
15101 This is the CUDA variant of the @ref{overlay} filter.
15102 It only accepts CUDA frames. The underlying input pixel formats have to match.
15104 It takes two inputs and has one output. The first input is the "main"
15105 video on which the second input is overlaid.
15107 It accepts the following parameters:
15112 Set the x and y coordinates of the overlaid video on the main video.
15113 Default value is "0" for both expressions.
15116 See @ref{framesync}.
15119 See @ref{framesync}.
15122 See @ref{framesync}.
15126 This filter also supports the @ref{framesync} options.
15130 Apply Overcomplete Wavelet denoiser.
15132 The filter accepts the following options:
15138 Larger depth values will denoise lower frequency components more, but
15139 slow down filtering.
15141 Must be an int in the range 8-16, default is @code{8}.
15143 @item luma_strength, ls
15146 Must be a double value in the range 0-1000, default is @code{1.0}.
15148 @item chroma_strength, cs
15149 Set chroma strength.
15151 Must be a double value in the range 0-1000, default is @code{1.0}.
15157 Add paddings to the input image, and place the original input at the
15158 provided @var{x}, @var{y} coordinates.
15160 It accepts the following parameters:
15165 Specify an expression for the size of the output image with the
15166 paddings added. If the value for @var{width} or @var{height} is 0, the
15167 corresponding input size is used for the output.
15169 The @var{width} expression can reference the value set by the
15170 @var{height} expression, and vice versa.
15172 The default value of @var{width} and @var{height} is 0.
15176 Specify the offsets to place the input image at within the padded area,
15177 with respect to the top/left border of the output image.
15179 The @var{x} expression can reference the value set by the @var{y}
15180 expression, and vice versa.
15182 The default value of @var{x} and @var{y} is 0.
15184 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15185 so the input image is centered on the padded area.
15188 Specify the color of the padded area. For the syntax of this option,
15189 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15190 manual,ffmpeg-utils}.
15192 The default value of @var{color} is "black".
15195 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15197 It accepts the following values:
15201 Only evaluate expressions once during the filter initialization or when
15202 a command is processed.
15205 Evaluate expressions for each incoming frame.
15209 Default value is @samp{init}.
15212 Pad to aspect instead to a resolution.
15216 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15217 options are expressions containing the following constants:
15222 The input video width and height.
15226 These are the same as @var{in_w} and @var{in_h}.
15230 The output width and height (the size of the padded area), as
15231 specified by the @var{width} and @var{height} expressions.
15235 These are the same as @var{out_w} and @var{out_h}.
15239 The x and y offsets as specified by the @var{x} and @var{y}
15240 expressions, or NAN if not yet specified.
15243 same as @var{iw} / @var{ih}
15246 input sample aspect ratio
15249 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15253 The horizontal and vertical chroma subsample values. For example for the
15254 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15257 @subsection Examples
15261 Add paddings with the color "violet" to the input video. The output video
15262 size is 640x480, and the top-left corner of the input video is placed at
15265 pad=640:480:0:40:violet
15268 The example above is equivalent to the following command:
15270 pad=width=640:height=480:x=0:y=40:color=violet
15274 Pad the input to get an output with dimensions increased by 3/2,
15275 and put the input video at the center of the padded area:
15277 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15281 Pad the input to get a squared output with size equal to the maximum
15282 value between the input width and height, and put the input video at
15283 the center of the padded area:
15285 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15289 Pad the input to get a final w/h ratio of 16:9:
15291 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15295 In case of anamorphic video, in order to set the output display aspect
15296 correctly, it is necessary to use @var{sar} in the expression,
15297 according to the relation:
15299 (ih * X / ih) * sar = output_dar
15300 X = output_dar / sar
15303 Thus the previous example needs to be modified to:
15305 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15309 Double the output size and put the input video in the bottom-right
15310 corner of the output padded area:
15312 pad="2*iw:2*ih:ow-iw:oh-ih"
15316 @anchor{palettegen}
15317 @section palettegen
15319 Generate one palette for a whole video stream.
15321 It accepts the following options:
15325 Set the maximum number of colors to quantize in the palette.
15326 Note: the palette will still contain 256 colors; the unused palette entries
15329 @item reserve_transparent
15330 Create a palette of 255 colors maximum and reserve the last one for
15331 transparency. Reserving the transparency color is useful for GIF optimization.
15332 If not set, the maximum of colors in the palette will be 256. You probably want
15333 to disable this option for a standalone image.
15336 @item transparency_color
15337 Set the color that will be used as background for transparency.
15340 Set statistics mode.
15342 It accepts the following values:
15345 Compute full frame histograms.
15347 Compute histograms only for the part that differs from previous frame. This
15348 might be relevant to give more importance to the moving part of your input if
15349 the background is static.
15351 Compute new histogram for each frame.
15354 Default value is @var{full}.
15357 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15358 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15359 color quantization of the palette. This information is also visible at
15360 @var{info} logging level.
15362 @subsection Examples
15366 Generate a representative palette of a given video using @command{ffmpeg}:
15368 ffmpeg -i input.mkv -vf palettegen palette.png
15372 @section paletteuse
15374 Use a palette to downsample an input video stream.
15376 The filter takes two inputs: one video stream and a palette. The palette must
15377 be a 256 pixels image.
15379 It accepts the following options:
15383 Select dithering mode. Available algorithms are:
15386 Ordered 8x8 bayer dithering (deterministic)
15388 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15389 Note: this dithering is sometimes considered "wrong" and is included as a
15391 @item floyd_steinberg
15392 Floyd and Steingberg dithering (error diffusion)
15394 Frankie Sierra dithering v2 (error diffusion)
15396 Frankie Sierra dithering v2 "Lite" (error diffusion)
15399 Default is @var{sierra2_4a}.
15402 When @var{bayer} dithering is selected, this option defines the scale of the
15403 pattern (how much the crosshatch pattern is visible). A low value means more
15404 visible pattern for less banding, and higher value means less visible pattern
15405 at the cost of more banding.
15407 The option must be an integer value in the range [0,5]. Default is @var{2}.
15410 If set, define the zone to process
15414 Only the changing rectangle will be reprocessed. This is similar to GIF
15415 cropping/offsetting compression mechanism. This option can be useful for speed
15416 if only a part of the image is changing, and has use cases such as limiting the
15417 scope of the error diffusal @option{dither} to the rectangle that bounds the
15418 moving scene (it leads to more deterministic output if the scene doesn't change
15419 much, and as a result less moving noise and better GIF compression).
15422 Default is @var{none}.
15425 Take new palette for each output frame.
15427 @item alpha_threshold
15428 Sets the alpha threshold for transparency. Alpha values above this threshold
15429 will be treated as completely opaque, and values below this threshold will be
15430 treated as completely transparent.
15432 The option must be an integer value in the range [0,255]. Default is @var{128}.
15435 @subsection Examples
15439 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15440 using @command{ffmpeg}:
15442 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15446 @section perspective
15448 Correct perspective of video not recorded perpendicular to the screen.
15450 A description of the accepted parameters follows.
15461 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15462 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15463 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15464 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15465 then the corners of the source will be sent to the specified coordinates.
15467 The expressions can use the following variables:
15472 the width and height of video frame.
15476 Output frame count.
15479 @item interpolation
15480 Set interpolation for perspective correction.
15482 It accepts the following values:
15488 Default value is @samp{linear}.
15491 Set interpretation of coordinate options.
15493 It accepts the following values:
15497 Send point in the source specified by the given coordinates to
15498 the corners of the destination.
15500 @item 1, destination
15502 Send the corners of the source to the point in the destination specified
15503 by the given coordinates.
15505 Default value is @samp{source}.
15509 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15511 It accepts the following values:
15514 only evaluate expressions once during the filter initialization or
15515 when a command is processed
15518 evaluate expressions for each incoming frame
15521 Default value is @samp{init}.
15526 Delay interlaced video by one field time so that the field order changes.
15528 The intended use is to fix PAL movies that have been captured with the
15529 opposite field order to the film-to-video transfer.
15531 A description of the accepted parameters follows.
15537 It accepts the following values:
15540 Capture field order top-first, transfer bottom-first.
15541 Filter will delay the bottom field.
15544 Capture field order bottom-first, transfer top-first.
15545 Filter will delay the top field.
15548 Capture and transfer with the same field order. This mode only exists
15549 for the documentation of the other options to refer to, but if you
15550 actually select it, the filter will faithfully do nothing.
15553 Capture field order determined automatically by field flags, transfer
15555 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15556 basis using field flags. If no field information is available,
15557 then this works just like @samp{u}.
15560 Capture unknown or varying, transfer opposite.
15561 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15562 analyzing the images and selecting the alternative that produces best
15563 match between the fields.
15566 Capture top-first, transfer unknown or varying.
15567 Filter selects among @samp{t} and @samp{p} using image analysis.
15570 Capture bottom-first, transfer unknown or varying.
15571 Filter selects among @samp{b} and @samp{p} using image analysis.
15574 Capture determined by field flags, transfer unknown or varying.
15575 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15576 image analysis. If no field information is available, then this works just
15577 like @samp{U}. This is the default mode.
15580 Both capture and transfer unknown or varying.
15581 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15585 @section photosensitivity
15586 Reduce various flashes in video, so to help users with epilepsy.
15588 It accepts the following options:
15591 Set how many frames to use when filtering. Default is 30.
15594 Set detection threshold factor. Default is 1.
15598 Set how many pixels to skip when sampling frames. Default is 1.
15599 Allowed range is from 1 to 1024.
15602 Leave frames unchanged. Default is disabled.
15605 @section pixdesctest
15607 Pixel format descriptor test filter, mainly useful for internal
15608 testing. The output video should be equal to the input video.
15612 format=monow, pixdesctest
15615 can be used to test the monowhite pixel format descriptor definition.
15619 Display sample values of color channels. Mainly useful for checking color
15620 and levels. Minimum supported resolution is 640x480.
15622 The filters accept the following options:
15626 Set scope X position, relative offset on X axis.
15629 Set scope Y position, relative offset on Y axis.
15638 Set window opacity. This window also holds statistics about pixel area.
15641 Set window X position, relative offset on X axis.
15644 Set window Y position, relative offset on Y axis.
15649 Enable the specified chain of postprocessing subfilters using libpostproc. This
15650 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15651 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15652 Each subfilter and some options have a short and a long name that can be used
15653 interchangeably, i.e. dr/dering are the same.
15655 The filters accept the following options:
15659 Set postprocessing subfilters string.
15662 All subfilters share common options to determine their scope:
15666 Honor the quality commands for this subfilter.
15669 Do chrominance filtering, too (default).
15672 Do luminance filtering only (no chrominance).
15675 Do chrominance filtering only (no luminance).
15678 These options can be appended after the subfilter name, separated by a '|'.
15680 Available subfilters are:
15683 @item hb/hdeblock[|difference[|flatness]]
15684 Horizontal deblocking filter
15687 Difference factor where higher values mean more deblocking (default: @code{32}).
15689 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15692 @item vb/vdeblock[|difference[|flatness]]
15693 Vertical deblocking filter
15696 Difference factor where higher values mean more deblocking (default: @code{32}).
15698 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15701 @item ha/hadeblock[|difference[|flatness]]
15702 Accurate horizontal deblocking filter
15705 Difference factor where higher values mean more deblocking (default: @code{32}).
15707 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15710 @item va/vadeblock[|difference[|flatness]]
15711 Accurate vertical deblocking filter
15714 Difference factor where higher values mean more deblocking (default: @code{32}).
15716 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15720 The horizontal and vertical deblocking filters share the difference and
15721 flatness values so you cannot set different horizontal and vertical
15725 @item h1/x1hdeblock
15726 Experimental horizontal deblocking filter
15728 @item v1/x1vdeblock
15729 Experimental vertical deblocking filter
15734 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15737 larger -> stronger filtering
15739 larger -> stronger filtering
15741 larger -> stronger filtering
15744 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15747 Stretch luminance to @code{0-255}.
15750 @item lb/linblenddeint
15751 Linear blend deinterlacing filter that deinterlaces the given block by
15752 filtering all lines with a @code{(1 2 1)} filter.
15754 @item li/linipoldeint
15755 Linear interpolating deinterlacing filter that deinterlaces the given block by
15756 linearly interpolating every second line.
15758 @item ci/cubicipoldeint
15759 Cubic interpolating deinterlacing filter deinterlaces the given block by
15760 cubically interpolating every second line.
15762 @item md/mediandeint
15763 Median deinterlacing filter that deinterlaces the given block by applying a
15764 median filter to every second line.
15766 @item fd/ffmpegdeint
15767 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15768 second line with a @code{(-1 4 2 4 -1)} filter.
15771 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15772 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15774 @item fq/forceQuant[|quantizer]
15775 Overrides the quantizer table from the input with the constant quantizer you
15783 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15786 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15789 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15792 @subsection Examples
15796 Apply horizontal and vertical deblocking, deringing and automatic
15797 brightness/contrast:
15803 Apply default filters without brightness/contrast correction:
15809 Apply default filters and temporal denoiser:
15811 pp=default/tmpnoise|1|2|3
15815 Apply deblocking on luminance only, and switch vertical deblocking on or off
15816 automatically depending on available CPU time:
15823 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15824 similar to spp = 6 with 7 point DCT, where only the center sample is
15827 The filter accepts the following options:
15831 Force a constant quantization parameter. It accepts an integer in range
15832 0 to 63. If not set, the filter will use the QP from the video stream
15836 Set thresholding mode. Available modes are:
15840 Set hard thresholding.
15842 Set soft thresholding (better de-ringing effect, but likely blurrier).
15844 Set medium thresholding (good results, default).
15848 @section premultiply
15849 Apply alpha premultiply effect to input video stream using first plane
15850 of second stream as alpha.
15852 Both streams must have same dimensions and same pixel format.
15854 The filter accepts the following option:
15858 Set which planes will be processed, unprocessed planes will be copied.
15859 By default value 0xf, all planes will be processed.
15862 Do not require 2nd input for processing, instead use alpha plane from input stream.
15866 Apply prewitt operator to input video stream.
15868 The filter accepts the following option:
15872 Set which planes will be processed, unprocessed planes will be copied.
15873 By default value 0xf, all planes will be processed.
15876 Set value which will be multiplied with filtered result.
15879 Set value which will be added to filtered result.
15882 @section pseudocolor
15884 Alter frame colors in video with pseudocolors.
15886 This filter accepts the following options:
15890 set pixel first component expression
15893 set pixel second component expression
15896 set pixel third component expression
15899 set pixel fourth component expression, corresponds to the alpha component
15902 set component to use as base for altering colors
15905 Each of them specifies the expression to use for computing the lookup table for
15906 the corresponding pixel component values.
15908 The expressions can contain the following constants and functions:
15913 The input width and height.
15916 The input value for the pixel component.
15918 @item ymin, umin, vmin, amin
15919 The minimum allowed component value.
15921 @item ymax, umax, vmax, amax
15922 The maximum allowed component value.
15925 All expressions default to "val".
15927 @subsection Examples
15931 Change too high luma values to gradient:
15933 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'"
15939 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15940 Ratio) between two input videos.
15942 This filter takes in input two input videos, the first input is
15943 considered the "main" source and is passed unchanged to the
15944 output. The second input is used as a "reference" video for computing
15947 Both video inputs must have the same resolution and pixel format for
15948 this filter to work correctly. Also it assumes that both inputs
15949 have the same number of frames, which are compared one by one.
15951 The obtained average PSNR is printed through the logging system.
15953 The filter stores the accumulated MSE (mean squared error) of each
15954 frame, and at the end of the processing it is averaged across all frames
15955 equally, and the following formula is applied to obtain the PSNR:
15958 PSNR = 10*log10(MAX^2/MSE)
15961 Where MAX is the average of the maximum values of each component of the
15964 The description of the accepted parameters follows.
15967 @item stats_file, f
15968 If specified the filter will use the named file to save the PSNR of
15969 each individual frame. When filename equals "-" the data is sent to
15972 @item stats_version
15973 Specifies which version of the stats file format to use. Details of
15974 each format are written below.
15975 Default value is 1.
15977 @item stats_add_max
15978 Determines whether the max value is output to the stats log.
15979 Default value is 0.
15980 Requires stats_version >= 2. If this is set and stats_version < 2,
15981 the filter will return an error.
15984 This filter also supports the @ref{framesync} options.
15986 The file printed if @var{stats_file} is selected, contains a sequence of
15987 key/value pairs of the form @var{key}:@var{value} for each compared
15990 If a @var{stats_version} greater than 1 is specified, a header line precedes
15991 the list of per-frame-pair stats, with key value pairs following the frame
15992 format with the following parameters:
15995 @item psnr_log_version
15996 The version of the log file format. Will match @var{stats_version}.
15999 A comma separated list of the per-frame-pair parameters included in
16003 A description of each shown per-frame-pair parameter follows:
16007 sequential number of the input frame, starting from 1
16010 Mean Square Error pixel-by-pixel average difference of the compared
16011 frames, averaged over all the image components.
16013 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16014 Mean Square Error pixel-by-pixel average difference of the compared
16015 frames for the component specified by the suffix.
16017 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16018 Peak Signal to Noise ratio of the compared frames for the component
16019 specified by the suffix.
16021 @item max_avg, max_y, max_u, max_v
16022 Maximum allowed value for each channel, and average over all
16026 @subsection Examples
16031 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16032 [main][ref] psnr="stats_file=stats.log" [out]
16035 On this example the input file being processed is compared with the
16036 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16037 is stored in @file{stats.log}.
16040 Another example with different containers:
16042 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 -
16049 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16050 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16053 The pullup filter is designed to take advantage of future context in making
16054 its decisions. This filter is stateless in the sense that it does not lock
16055 onto a pattern to follow, but it instead looks forward to the following
16056 fields in order to identify matches and rebuild progressive frames.
16058 To produce content with an even framerate, insert the fps filter after
16059 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16060 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16062 The filter accepts the following options:
16069 These options set the amount of "junk" to ignore at the left, right, top, and
16070 bottom of the image, respectively. Left and right are in units of 8 pixels,
16071 while top and bottom are in units of 2 lines.
16072 The default is 8 pixels on each side.
16075 Set the strict breaks. Setting this option to 1 will reduce the chances of
16076 filter generating an occasional mismatched frame, but it may also cause an
16077 excessive number of frames to be dropped during high motion sequences.
16078 Conversely, setting it to -1 will make filter match fields more easily.
16079 This may help processing of video where there is slight blurring between
16080 the fields, but may also cause there to be interlaced frames in the output.
16081 Default value is @code{0}.
16084 Set the metric plane to use. It accepts the following values:
16090 Use chroma blue plane.
16093 Use chroma red plane.
16096 This option may be set to use chroma plane instead of the default luma plane
16097 for doing filter's computations. This may improve accuracy on very clean
16098 source material, but more likely will decrease accuracy, especially if there
16099 is chroma noise (rainbow effect) or any grayscale video.
16100 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16101 load and make pullup usable in realtime on slow machines.
16104 For best results (without duplicated frames in the output file) it is
16105 necessary to change the output frame rate. For example, to inverse
16106 telecine NTSC input:
16108 ffmpeg -i input -vf pullup -r 24000/1001 ...
16113 Change video quantization parameters (QP).
16115 The filter accepts the following option:
16119 Set expression for quantization parameter.
16122 The expression is evaluated through the eval API and can contain, among others,
16123 the following constants:
16127 1 if index is not 129, 0 otherwise.
16130 Sequential index starting from -129 to 128.
16133 @subsection Examples
16137 Some equation like:
16145 Flush video frames from internal cache of frames into a random order.
16146 No frame is discarded.
16147 Inspired by @ref{frei0r} nervous filter.
16151 Set size in number of frames of internal cache, in range from @code{2} to
16152 @code{512}. Default is @code{30}.
16155 Set seed for random number generator, must be an integer included between
16156 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16157 less than @code{0}, the filter will try to use a good random seed on a
16161 @section readeia608
16163 Read closed captioning (EIA-608) information from the top lines of a video frame.
16165 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16166 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16167 with EIA-608 data (starting from 0). A description of each metadata value follows:
16170 @item lavfi.readeia608.X.cc
16171 The two bytes stored as EIA-608 data (printed in hexadecimal).
16173 @item lavfi.readeia608.X.line
16174 The number of the line on which the EIA-608 data was identified and read.
16177 This filter accepts the following options:
16181 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16184 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16187 Set the ratio of width reserved for sync code detection.
16188 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16191 Enable checking the parity bit. In the event of a parity error, the filter will output
16192 @code{0x00} for that character. Default is false.
16195 Lowpass lines prior to further processing. Default is enabled.
16198 @subsection Commands
16200 This filter supports the all above options as @ref{commands}.
16202 @subsection Examples
16206 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16208 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
16214 Read vertical interval timecode (VITC) information from the top lines of a
16217 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16218 timecode value, if a valid timecode has been detected. Further metadata key
16219 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16220 timecode data has been found or not.
16222 This filter accepts the following options:
16226 Set the maximum number of lines to scan for VITC data. If the value is set to
16227 @code{-1} the full video frame is scanned. Default is @code{45}.
16230 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16231 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16234 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16235 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16238 @subsection Examples
16242 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16243 draw @code{--:--:--:--} as a placeholder:
16245 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16251 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16253 Destination pixel at position (X, Y) will be picked from source (x, y) position
16254 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16255 value for pixel will be used for destination pixel.
16257 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16258 will have Xmap/Ymap video stream dimensions.
16259 Xmap and Ymap input video streams are 16bit depth, single channel.
16263 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16264 Default is @code{color}.
16267 Specify the color of the unmapped pixels. For the syntax of this option,
16268 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16269 manual,ffmpeg-utils}. Default color is @code{black}.
16272 @section removegrain
16274 The removegrain filter is a spatial denoiser for progressive video.
16278 Set mode for the first plane.
16281 Set mode for the second plane.
16284 Set mode for the third plane.
16287 Set mode for the fourth plane.
16290 Range of mode is from 0 to 24. Description of each mode follows:
16294 Leave input plane unchanged. Default.
16297 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16300 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16303 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16306 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16307 This is equivalent to a median filter.
16310 Line-sensitive clipping giving the minimal change.
16313 Line-sensitive clipping, intermediate.
16316 Line-sensitive clipping, intermediate.
16319 Line-sensitive clipping, intermediate.
16322 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16325 Replaces the target pixel with the closest neighbour.
16328 [1 2 1] horizontal and vertical kernel blur.
16334 Bob mode, interpolates top field from the line where the neighbours
16335 pixels are the closest.
16338 Bob mode, interpolates bottom field from the line where the neighbours
16339 pixels are the closest.
16342 Bob mode, interpolates top field. Same as 13 but with a more complicated
16343 interpolation formula.
16346 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16347 interpolation formula.
16350 Clips the pixel with the minimum and maximum of respectively the maximum and
16351 minimum of each pair of opposite neighbour pixels.
16354 Line-sensitive clipping using opposite neighbours whose greatest distance from
16355 the current pixel is minimal.
16358 Replaces the pixel with the average of its 8 neighbours.
16361 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16364 Clips pixels using the averages of opposite neighbour.
16367 Same as mode 21 but simpler and faster.
16370 Small edge and halo removal, but reputed useless.
16376 @section removelogo
16378 Suppress a TV station logo, using an image file to determine which
16379 pixels comprise the logo. It works by filling in the pixels that
16380 comprise the logo with neighboring pixels.
16382 The filter accepts the following options:
16386 Set the filter bitmap file, which can be any image format supported by
16387 libavformat. The width and height of the image file must match those of the
16388 video stream being processed.
16391 Pixels in the provided bitmap image with a value of zero are not
16392 considered part of the logo, non-zero pixels are considered part of
16393 the logo. If you use white (255) for the logo and black (0) for the
16394 rest, you will be safe. For making the filter bitmap, it is
16395 recommended to take a screen capture of a black frame with the logo
16396 visible, and then using a threshold filter followed by the erode
16397 filter once or twice.
16399 If needed, little splotches can be fixed manually. Remember that if
16400 logo pixels are not covered, the filter quality will be much
16401 reduced. Marking too many pixels as part of the logo does not hurt as
16402 much, but it will increase the amount of blurring needed to cover over
16403 the image and will destroy more information than necessary, and extra
16404 pixels will slow things down on a large logo.
16406 @section repeatfields
16408 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16409 fields based on its value.
16413 Reverse a video clip.
16415 Warning: This filter requires memory to buffer the entire clip, so trimming
16418 @subsection Examples
16422 Take the first 5 seconds of a clip, and reverse it.
16429 Shift R/G/B/A pixels horizontally and/or vertically.
16431 The filter accepts the following options:
16434 Set amount to shift red horizontally.
16436 Set amount to shift red vertically.
16438 Set amount to shift green horizontally.
16440 Set amount to shift green vertically.
16442 Set amount to shift blue horizontally.
16444 Set amount to shift blue vertically.
16446 Set amount to shift alpha horizontally.
16448 Set amount to shift alpha vertically.
16450 Set edge mode, can be @var{smear}, default, or @var{warp}.
16453 @subsection Commands
16455 This filter supports the all above options as @ref{commands}.
16458 Apply roberts cross operator to input video stream.
16460 The filter accepts the following option:
16464 Set which planes will be processed, unprocessed planes will be copied.
16465 By default value 0xf, all planes will be processed.
16468 Set value which will be multiplied with filtered result.
16471 Set value which will be added to filtered result.
16476 Rotate video by an arbitrary angle expressed in radians.
16478 The filter accepts the following options:
16480 A description of the optional parameters follows.
16483 Set an expression for the angle by which to rotate the input video
16484 clockwise, expressed as a number of radians. A negative value will
16485 result in a counter-clockwise rotation. By default it is set to "0".
16487 This expression is evaluated for each frame.
16490 Set the output width expression, default value is "iw".
16491 This expression is evaluated just once during configuration.
16494 Set the output height expression, default value is "ih".
16495 This expression is evaluated just once during configuration.
16498 Enable bilinear interpolation if set to 1, a value of 0 disables
16499 it. Default value is 1.
16502 Set the color used to fill the output area not covered by the rotated
16503 image. For the general syntax of this option, check the
16504 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16505 If the special value "none" is selected then no
16506 background is printed (useful for example if the background is never shown).
16508 Default value is "black".
16511 The expressions for the angle and the output size can contain the
16512 following constants and functions:
16516 sequential number of the input frame, starting from 0. It is always NAN
16517 before the first frame is filtered.
16520 time in seconds of the input frame, it is set to 0 when the filter is
16521 configured. It is always NAN before the first frame is filtered.
16525 horizontal and vertical chroma subsample values. For example for the
16526 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16530 the input video width and height
16534 the output width and height, that is the size of the padded area as
16535 specified by the @var{width} and @var{height} expressions
16539 the minimal width/height required for completely containing the input
16540 video rotated by @var{a} radians.
16542 These are only available when computing the @option{out_w} and
16543 @option{out_h} expressions.
16546 @subsection Examples
16550 Rotate the input by PI/6 radians clockwise:
16556 Rotate the input by PI/6 radians counter-clockwise:
16562 Rotate the input by 45 degrees clockwise:
16568 Apply a constant rotation with period T, starting from an angle of PI/3:
16570 rotate=PI/3+2*PI*t/T
16574 Make the input video rotation oscillating with a period of T
16575 seconds and an amplitude of A radians:
16577 rotate=A*sin(2*PI/T*t)
16581 Rotate the video, output size is chosen so that the whole rotating
16582 input video is always completely contained in the output:
16584 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16588 Rotate the video, reduce the output size so that no background is ever
16591 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16595 @subsection Commands
16597 The filter supports the following commands:
16601 Set the angle expression.
16602 The command accepts the same syntax of the corresponding option.
16604 If the specified expression is not valid, it is kept at its current
16610 Apply Shape Adaptive Blur.
16612 The filter accepts the following options:
16615 @item luma_radius, lr
16616 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16617 value is 1.0. A greater value will result in a more blurred image, and
16618 in slower processing.
16620 @item luma_pre_filter_radius, lpfr
16621 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16624 @item luma_strength, ls
16625 Set luma maximum difference between pixels to still be considered, must
16626 be a value in the 0.1-100.0 range, default value is 1.0.
16628 @item chroma_radius, cr
16629 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16630 greater value will result in a more blurred image, and in slower
16633 @item chroma_pre_filter_radius, cpfr
16634 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16636 @item chroma_strength, cs
16637 Set chroma maximum difference between pixels to still be considered,
16638 must be a value in the -0.9-100.0 range.
16641 Each chroma option value, if not explicitly specified, is set to the
16642 corresponding luma option value.
16647 Scale (resize) the input video, using the libswscale library.
16649 The scale filter forces the output display aspect ratio to be the same
16650 of the input, by changing the output sample aspect ratio.
16652 If the input image format is different from the format requested by
16653 the next filter, the scale filter will convert the input to the
16656 @subsection Options
16657 The filter accepts the following options, or any of the options
16658 supported by the libswscale scaler.
16660 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16661 the complete list of scaler options.
16666 Set the output video dimension expression. Default value is the input
16669 If the @var{width} or @var{w} value is 0, the input width is used for
16670 the output. If the @var{height} or @var{h} value is 0, the input height
16671 is used for the output.
16673 If one and only one of the values is -n with n >= 1, the scale filter
16674 will use a value that maintains the aspect ratio of the input image,
16675 calculated from the other specified dimension. After that it will,
16676 however, make sure that the calculated dimension is divisible by n and
16677 adjust the value if necessary.
16679 If both values are -n with n >= 1, the behavior will be identical to
16680 both values being set to 0 as previously detailed.
16682 See below for the list of accepted constants for use in the dimension
16686 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16690 Only evaluate expressions once during the filter initialization or when a command is processed.
16693 Evaluate expressions for each incoming frame.
16697 Default value is @samp{init}.
16701 Set the interlacing mode. It accepts the following values:
16705 Force interlaced aware scaling.
16708 Do not apply interlaced scaling.
16711 Select interlaced aware scaling depending on whether the source frames
16712 are flagged as interlaced or not.
16715 Default value is @samp{0}.
16718 Set libswscale scaling flags. See
16719 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16720 complete list of values. If not explicitly specified the filter applies
16724 @item param0, param1
16725 Set libswscale input parameters for scaling algorithms that need them. See
16726 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16727 complete documentation. If not explicitly specified the filter applies
16733 Set the video size. For the syntax of this option, check the
16734 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16736 @item in_color_matrix
16737 @item out_color_matrix
16738 Set in/output YCbCr color space type.
16740 This allows the autodetected value to be overridden as well as allows forcing
16741 a specific value used for the output and encoder.
16743 If not specified, the color space type depends on the pixel format.
16749 Choose automatically.
16752 Format conforming to International Telecommunication Union (ITU)
16753 Recommendation BT.709.
16756 Set color space conforming to the United States Federal Communications
16757 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16762 Set color space conforming to:
16766 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16769 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16772 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16777 Set color space conforming to SMPTE ST 240:1999.
16780 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16785 Set in/output YCbCr sample range.
16787 This allows the autodetected value to be overridden as well as allows forcing
16788 a specific value used for the output and encoder. If not specified, the
16789 range depends on the pixel format. Possible values:
16793 Choose automatically.
16796 Set full range (0-255 in case of 8-bit luma).
16798 @item mpeg/limited/tv
16799 Set "MPEG" range (16-235 in case of 8-bit luma).
16802 @item force_original_aspect_ratio
16803 Enable decreasing or increasing output video width or height if necessary to
16804 keep the original aspect ratio. Possible values:
16808 Scale the video as specified and disable this feature.
16811 The output video dimensions will automatically be decreased if needed.
16814 The output video dimensions will automatically be increased if needed.
16818 One useful instance of this option is that when you know a specific device's
16819 maximum allowed resolution, you can use this to limit the output video to
16820 that, while retaining the aspect ratio. For example, device A allows
16821 1280x720 playback, and your video is 1920x800. Using this option (set it to
16822 decrease) and specifying 1280x720 to the command line makes the output
16825 Please note that this is a different thing than specifying -1 for @option{w}
16826 or @option{h}, you still need to specify the output resolution for this option
16829 @item force_divisible_by
16830 Ensures that both the output dimensions, width and height, are divisible by the
16831 given integer when used together with @option{force_original_aspect_ratio}. This
16832 works similar to using @code{-n} in the @option{w} and @option{h} options.
16834 This option respects the value set for @option{force_original_aspect_ratio},
16835 increasing or decreasing the resolution accordingly. The video's aspect ratio
16836 may be slightly modified.
16838 This option can be handy if you need to have a video fit within or exceed
16839 a defined resolution using @option{force_original_aspect_ratio} but also have
16840 encoder restrictions on width or height divisibility.
16844 The values of the @option{w} and @option{h} options are expressions
16845 containing the following constants:
16850 The input width and height
16854 These are the same as @var{in_w} and @var{in_h}.
16858 The output (scaled) width and height
16862 These are the same as @var{out_w} and @var{out_h}
16865 The same as @var{iw} / @var{ih}
16868 input sample aspect ratio
16871 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16875 horizontal and vertical input chroma subsample values. For example for the
16876 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16880 horizontal and vertical output chroma subsample values. For example for the
16881 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16884 The (sequential) number of the input frame, starting from 0.
16885 Only available with @code{eval=frame}.
16888 The presentation timestamp of the input frame, expressed as a number of
16889 seconds. Only available with @code{eval=frame}.
16892 The position (byte offset) of the frame in the input stream, or NaN if
16893 this information is unavailable and/or meaningless (for example in case of synthetic video).
16894 Only available with @code{eval=frame}.
16897 @subsection Examples
16901 Scale the input video to a size of 200x100
16906 This is equivalent to:
16917 Specify a size abbreviation for the output size:
16922 which can also be written as:
16928 Scale the input to 2x:
16930 scale=w=2*iw:h=2*ih
16934 The above is the same as:
16936 scale=2*in_w:2*in_h
16940 Scale the input to 2x with forced interlaced scaling:
16942 scale=2*iw:2*ih:interl=1
16946 Scale the input to half size:
16948 scale=w=iw/2:h=ih/2
16952 Increase the width, and set the height to the same size:
16958 Seek Greek harmony:
16965 Increase the height, and set the width to 3/2 of the height:
16967 scale=w=3/2*oh:h=3/5*ih
16971 Increase the size, making the size a multiple of the chroma
16974 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16978 Increase the width to a maximum of 500 pixels,
16979 keeping the same aspect ratio as the input:
16981 scale=w='min(500\, iw*3/2):h=-1'
16985 Make pixels square by combining scale and setsar:
16987 scale='trunc(ih*dar):ih',setsar=1/1
16991 Make pixels square by combining scale and setsar,
16992 making sure the resulting resolution is even (required by some codecs):
16994 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16998 @subsection Commands
17000 This filter supports the following commands:
17004 Set the output video dimension expression.
17005 The command accepts the same syntax of the corresponding option.
17007 If the specified expression is not valid, it is kept at its current
17013 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17014 format conversion on CUDA video frames. Setting the output width and height
17015 works in the same way as for the @var{scale} filter.
17017 The following additional options are accepted:
17020 The pixel format of the output CUDA frames. If set to the string "same" (the
17021 default), the input format will be kept. Note that automatic format negotiation
17022 and conversion is not yet supported for hardware frames
17025 The interpolation algorithm used for resizing. One of the following:
17032 @item cubic2p_bspline
17033 2-parameter cubic (B=1, C=0)
17035 @item cubic2p_catmullrom
17036 2-parameter cubic (B=0, C=1/2)
17038 @item cubic2p_b05c03
17039 2-parameter cubic (B=1/2, C=3/10)
17047 @item force_original_aspect_ratio
17048 Enable decreasing or increasing output video width or height if necessary to
17049 keep the original aspect ratio. Possible values:
17053 Scale the video as specified and disable this feature.
17056 The output video dimensions will automatically be decreased if needed.
17059 The output video dimensions will automatically be increased if needed.
17063 One useful instance of this option is that when you know a specific device's
17064 maximum allowed resolution, you can use this to limit the output video to
17065 that, while retaining the aspect ratio. For example, device A allows
17066 1280x720 playback, and your video is 1920x800. Using this option (set it to
17067 decrease) and specifying 1280x720 to the command line makes the output
17070 Please note that this is a different thing than specifying -1 for @option{w}
17071 or @option{h}, you still need to specify the output resolution for this option
17074 @item force_divisible_by
17075 Ensures that both the output dimensions, width and height, are divisible by the
17076 given integer when used together with @option{force_original_aspect_ratio}. This
17077 works similar to using @code{-n} in the @option{w} and @option{h} options.
17079 This option respects the value set for @option{force_original_aspect_ratio},
17080 increasing or decreasing the resolution accordingly. The video's aspect ratio
17081 may be slightly modified.
17083 This option can be handy if you need to have a video fit within or exceed
17084 a defined resolution using @option{force_original_aspect_ratio} but also have
17085 encoder restrictions on width or height divisibility.
17091 Scale (resize) the input video, based on a reference video.
17093 See the scale filter for available options, scale2ref supports the same but
17094 uses the reference video instead of the main input as basis. scale2ref also
17095 supports the following additional constants for the @option{w} and
17096 @option{h} options:
17101 The main input video's width and height
17104 The same as @var{main_w} / @var{main_h}
17107 The main input video's sample aspect ratio
17109 @item main_dar, mdar
17110 The main input video's display aspect ratio. Calculated from
17111 @code{(main_w / main_h) * main_sar}.
17115 The main input video's horizontal and vertical chroma subsample values.
17116 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17120 The (sequential) number of the main input frame, starting from 0.
17121 Only available with @code{eval=frame}.
17124 The presentation timestamp of the main input frame, expressed as a number of
17125 seconds. Only available with @code{eval=frame}.
17128 The position (byte offset) of the frame in the main input stream, or NaN if
17129 this information is unavailable and/or meaningless (for example in case of synthetic video).
17130 Only available with @code{eval=frame}.
17133 @subsection Examples
17137 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17139 'scale2ref[b][a];[a][b]overlay'
17143 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17145 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17149 @subsection Commands
17151 This filter supports the following commands:
17155 Set the output video dimension expression.
17156 The command accepts the same syntax of the corresponding option.
17158 If the specified expression is not valid, it is kept at its current
17163 Scroll input video horizontally and/or vertically by constant speed.
17165 The filter accepts the following options:
17167 @item horizontal, h
17168 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17169 Negative values changes scrolling direction.
17172 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17173 Negative values changes scrolling direction.
17176 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17179 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17182 @subsection Commands
17184 This filter supports the following @ref{commands}:
17186 @item horizontal, h
17187 Set the horizontal scrolling speed.
17189 Set the vertical scrolling speed.
17195 Detect video scene change.
17197 This filter sets frame metadata with mafd between frame, the scene score, and
17198 forward the frame to the next filter, so they can use these metadata to detect
17199 scene change or others.
17201 In addition, this filter logs a message and sets frame metadata when it detects
17202 a scene change by @option{threshold}.
17204 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17206 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17207 to detect scene change.
17209 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17210 detect scene change with @option{threshold}.
17212 The filter accepts the following options:
17216 Set the scene change detection threshold as a percentage of maximum change. Good
17217 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17220 Default value is @code{10.}.
17223 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17224 You can enable it if you want to get snapshot of scene change frames only.
17227 @anchor{selectivecolor}
17228 @section selectivecolor
17230 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17231 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17232 by the "purity" of the color (that is, how saturated it already is).
17234 This filter is similar to the Adobe Photoshop Selective Color tool.
17236 The filter accepts the following options:
17239 @item correction_method
17240 Select color correction method.
17242 Available values are:
17245 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17248 Specified adjustments are relative to the original component value.
17250 Default is @code{absolute}.
17252 Adjustments for red pixels (pixels where the red component is the maximum)
17254 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17256 Adjustments for green pixels (pixels where the green component is the maximum)
17258 Adjustments for cyan pixels (pixels where the red component is the minimum)
17260 Adjustments for blue pixels (pixels where the blue component is the maximum)
17262 Adjustments for magenta pixels (pixels where the green component is the minimum)
17264 Adjustments for white pixels (pixels where all components are greater than 128)
17266 Adjustments for all pixels except pure black and pure white
17268 Adjustments for black pixels (pixels where all components are lesser than 128)
17270 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17273 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17274 4 space separated floating point adjustment values in the [-1,1] range,
17275 respectively to adjust the amount of cyan, magenta, yellow and black for the
17276 pixels of its range.
17278 @subsection Examples
17282 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17283 increase magenta by 27% in blue areas:
17285 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17289 Use a Photoshop selective color preset:
17291 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17295 @anchor{separatefields}
17296 @section separatefields
17298 The @code{separatefields} takes a frame-based video input and splits
17299 each frame into its components fields, producing a new half height clip
17300 with twice the frame rate and twice the frame count.
17302 This filter use field-dominance information in frame to decide which
17303 of each pair of fields to place first in the output.
17304 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17306 @section setdar, setsar
17308 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17311 This is done by changing the specified Sample (aka Pixel) Aspect
17312 Ratio, according to the following equation:
17314 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17317 Keep in mind that the @code{setdar} filter does not modify the pixel
17318 dimensions of the video frame. Also, the display aspect ratio set by
17319 this filter may be changed by later filters in the filterchain,
17320 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17323 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17324 the filter output video.
17326 Note that as a consequence of the application of this filter, the
17327 output display aspect ratio will change according to the equation
17330 Keep in mind that the sample aspect ratio set by the @code{setsar}
17331 filter may be changed by later filters in the filterchain, e.g. if
17332 another "setsar" or a "setdar" filter is applied.
17334 It accepts the following parameters:
17337 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17338 Set the aspect ratio used by the filter.
17340 The parameter can be a floating point number string, an expression, or
17341 a string of the form @var{num}:@var{den}, where @var{num} and
17342 @var{den} are the numerator and denominator of the aspect ratio. If
17343 the parameter is not specified, it is assumed the value "0".
17344 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17348 Set the maximum integer value to use for expressing numerator and
17349 denominator when reducing the expressed aspect ratio to a rational.
17350 Default value is @code{100}.
17354 The parameter @var{sar} is an expression containing
17355 the following constants:
17359 These are approximated values for the mathematical constants e
17360 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17363 The input width and height.
17366 These are the same as @var{w} / @var{h}.
17369 The input sample aspect ratio.
17372 The input display aspect ratio. It is the same as
17373 (@var{w} / @var{h}) * @var{sar}.
17376 Horizontal and vertical chroma subsample values. For example, for the
17377 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17380 @subsection Examples
17385 To change the display aspect ratio to 16:9, specify one of the following:
17392 To change the sample aspect ratio to 10:11, specify:
17398 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17399 1000 in the aspect ratio reduction, use the command:
17401 setdar=ratio=16/9:max=1000
17409 Force field for the output video frame.
17411 The @code{setfield} filter marks the interlace type field for the
17412 output frames. It does not change the input frame, but only sets the
17413 corresponding property, which affects how the frame is treated by
17414 following filters (e.g. @code{fieldorder} or @code{yadif}).
17416 The filter accepts the following options:
17421 Available values are:
17425 Keep the same field property.
17428 Mark the frame as bottom-field-first.
17431 Mark the frame as top-field-first.
17434 Mark the frame as progressive.
17441 Force frame parameter for the output video frame.
17443 The @code{setparams} filter marks interlace and color range for the
17444 output frames. It does not change the input frame, but only sets the
17445 corresponding property, which affects how the frame is treated by
17450 Available values are:
17454 Keep the same field property (default).
17457 Mark the frame as bottom-field-first.
17460 Mark the frame as top-field-first.
17463 Mark the frame as progressive.
17467 Available values are:
17471 Keep the same color range property (default).
17473 @item unspecified, unknown
17474 Mark the frame as unspecified color range.
17476 @item limited, tv, mpeg
17477 Mark the frame as limited range.
17479 @item full, pc, jpeg
17480 Mark the frame as full range.
17483 @item color_primaries
17484 Set the color primaries.
17485 Available values are:
17489 Keep the same color primaries property (default).
17506 Set the color transfer.
17507 Available values are:
17511 Keep the same color trc property (default).
17533 Set the colorspace.
17534 Available values are:
17538 Keep the same colorspace property (default).
17551 @item chroma-derived-nc
17552 @item chroma-derived-c
17559 Show a line containing various information for each input video frame.
17560 The input video is not modified.
17562 This filter supports the following options:
17566 Calculate checksums of each plane. By default enabled.
17569 The shown line contains a sequence of key/value pairs of the form
17570 @var{key}:@var{value}.
17572 The following values are shown in the output:
17576 The (sequential) number of the input frame, starting from 0.
17579 The Presentation TimeStamp of the input frame, expressed as a number of
17580 time base units. The time base unit depends on the filter input pad.
17583 The Presentation TimeStamp of the input frame, expressed as a number of
17587 The position of the frame in the input stream, or -1 if this information is
17588 unavailable and/or meaningless (for example in case of synthetic video).
17591 The pixel format name.
17594 The sample aspect ratio of the input frame, expressed in the form
17595 @var{num}/@var{den}.
17598 The size of the input frame. For the syntax of this option, check the
17599 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17602 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17603 for bottom field first).
17606 This is 1 if the frame is a key frame, 0 otherwise.
17609 The picture type of the input frame ("I" for an I-frame, "P" for a
17610 P-frame, "B" for a B-frame, or "?" for an unknown type).
17611 Also refer to the documentation of the @code{AVPictureType} enum and of
17612 the @code{av_get_picture_type_char} function defined in
17613 @file{libavutil/avutil.h}.
17616 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17618 @item plane_checksum
17619 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17620 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17623 The mean value of pixels in each plane of the input frame, expressed in the form
17624 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17627 The standard deviation of pixel values in each plane of the input frame, expressed
17628 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17632 @section showpalette
17634 Displays the 256 colors palette of each frame. This filter is only relevant for
17635 @var{pal8} pixel format frames.
17637 It accepts the following option:
17641 Set the size of the box used to represent one palette color entry. Default is
17642 @code{30} (for a @code{30x30} pixel box).
17645 @section shuffleframes
17647 Reorder and/or duplicate and/or drop video frames.
17649 It accepts the following parameters:
17653 Set the destination indexes of input frames.
17654 This is space or '|' separated list of indexes that maps input frames to output
17655 frames. Number of indexes also sets maximal value that each index may have.
17656 '-1' index have special meaning and that is to drop frame.
17659 The first frame has the index 0. The default is to keep the input unchanged.
17661 @subsection Examples
17665 Swap second and third frame of every three frames of the input:
17667 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17671 Swap 10th and 1st frame of every ten frames of the input:
17673 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17677 @section shuffleplanes
17679 Reorder and/or duplicate video planes.
17681 It accepts the following parameters:
17686 The index of the input plane to be used as the first output plane.
17689 The index of the input plane to be used as the second output plane.
17692 The index of the input plane to be used as the third output plane.
17695 The index of the input plane to be used as the fourth output plane.
17699 The first plane has the index 0. The default is to keep the input unchanged.
17701 @subsection Examples
17705 Swap the second and third planes of the input:
17707 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17711 @anchor{signalstats}
17712 @section signalstats
17713 Evaluate various visual metrics that assist in determining issues associated
17714 with the digitization of analog video media.
17716 By default the filter will log these metadata values:
17720 Display the minimal Y value contained within the input frame. Expressed in
17724 Display the Y value at the 10% percentile within the input frame. Expressed in
17728 Display the average Y value within the input frame. Expressed in range of
17732 Display the Y value at the 90% percentile within the input frame. Expressed in
17736 Display the maximum Y value contained within the input frame. Expressed in
17740 Display the minimal U value contained within the input frame. Expressed in
17744 Display the U value at the 10% percentile within the input frame. Expressed in
17748 Display the average U value within the input frame. Expressed in range of
17752 Display the U value at the 90% percentile within the input frame. Expressed in
17756 Display the maximum U value contained within the input frame. Expressed in
17760 Display the minimal V value contained within the input frame. Expressed in
17764 Display the V value at the 10% percentile within the input frame. Expressed in
17768 Display the average V value within the input frame. Expressed in range of
17772 Display the V value at the 90% percentile within the input frame. Expressed in
17776 Display the maximum V value contained within the input frame. Expressed in
17780 Display the minimal saturation value contained within the input frame.
17781 Expressed in range of [0-~181.02].
17784 Display the saturation value at the 10% percentile within the input frame.
17785 Expressed in range of [0-~181.02].
17788 Display the average saturation value within the input frame. Expressed in range
17792 Display the saturation value at the 90% percentile within the input frame.
17793 Expressed in range of [0-~181.02].
17796 Display the maximum saturation value contained within the input frame.
17797 Expressed in range of [0-~181.02].
17800 Display the median value for hue within the input frame. Expressed in range of
17804 Display the average value for hue within the input frame. Expressed in range of
17808 Display the average of sample value difference between all values of the Y
17809 plane in the current frame and corresponding values of the previous input frame.
17810 Expressed in range of [0-255].
17813 Display the average of sample value difference between all values of the U
17814 plane in the current frame and corresponding values of the previous input frame.
17815 Expressed in range of [0-255].
17818 Display the average of sample value difference between all values of the V
17819 plane in the current frame and corresponding values of the previous input frame.
17820 Expressed in range of [0-255].
17823 Display bit depth of Y plane in current frame.
17824 Expressed in range of [0-16].
17827 Display bit depth of U plane in current frame.
17828 Expressed in range of [0-16].
17831 Display bit depth of V plane in current frame.
17832 Expressed in range of [0-16].
17835 The filter accepts the following options:
17841 @option{stat} specify an additional form of image analysis.
17842 @option{out} output video with the specified type of pixel highlighted.
17844 Both options accept the following values:
17848 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17849 unlike the neighboring pixels of the same field. Examples of temporal outliers
17850 include the results of video dropouts, head clogs, or tape tracking issues.
17853 Identify @var{vertical line repetition}. Vertical line repetition includes
17854 similar rows of pixels within a frame. In born-digital video vertical line
17855 repetition is common, but this pattern is uncommon in video digitized from an
17856 analog source. When it occurs in video that results from the digitization of an
17857 analog source it can indicate concealment from a dropout compensator.
17860 Identify pixels that fall outside of legal broadcast range.
17864 Set the highlight color for the @option{out} option. The default color is
17868 @subsection Examples
17872 Output data of various video metrics:
17874 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17878 Output specific data about the minimum and maximum values of the Y plane per frame:
17880 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17884 Playback video while highlighting pixels that are outside of broadcast range in red.
17886 ffplay example.mov -vf signalstats="out=brng:color=red"
17890 Playback video with signalstats metadata drawn over the frame.
17892 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17895 The contents of signalstat_drawtext.txt used in the command are:
17898 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17899 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17900 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17901 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17909 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17910 input. In this case the matching between the inputs can be calculated additionally.
17911 The filter always passes through the first input. The signature of each stream can
17912 be written into a file.
17914 It accepts the following options:
17918 Enable or disable the matching process.
17920 Available values are:
17924 Disable the calculation of a matching (default).
17926 Calculate the matching for the whole video and output whether the whole video
17927 matches or only parts.
17929 Calculate only until a matching is found or the video ends. Should be faster in
17934 Set the number of inputs. The option value must be a non negative integer.
17935 Default value is 1.
17938 Set the path to which the output is written. If there is more than one input,
17939 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17940 integer), that will be replaced with the input number. If no filename is
17941 specified, no output will be written. This is the default.
17944 Choose the output format.
17946 Available values are:
17950 Use the specified binary representation (default).
17952 Use the specified xml representation.
17956 Set threshold to detect one word as similar. The option value must be an integer
17957 greater than zero. The default value is 9000.
17960 Set threshold to detect all words as similar. The option value must be an integer
17961 greater than zero. The default value is 60000.
17964 Set threshold to detect frames as similar. The option value must be an integer
17965 greater than zero. The default value is 116.
17968 Set the minimum length of a sequence in frames to recognize it as matching
17969 sequence. The option value must be a non negative integer value.
17970 The default value is 0.
17973 Set the minimum relation, that matching frames to all frames must have.
17974 The option value must be a double value between 0 and 1. The default value is 0.5.
17977 @subsection Examples
17981 To calculate the signature of an input video and store it in signature.bin:
17983 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17987 To detect whether two videos match and store the signatures in XML format in
17988 signature0.xml and signature1.xml:
17990 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 -
17998 Blur the input video without impacting the outlines.
18000 It accepts the following options:
18003 @item luma_radius, lr
18004 Set the luma radius. The option value must be a float number in
18005 the range [0.1,5.0] that specifies the variance of the gaussian filter
18006 used to blur the image (slower if larger). Default value is 1.0.
18008 @item luma_strength, ls
18009 Set the luma strength. The option value must be a float number
18010 in the range [-1.0,1.0] that configures the blurring. A value included
18011 in [0.0,1.0] will blur the image whereas a value included in
18012 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18014 @item luma_threshold, lt
18015 Set the luma threshold used as a coefficient to determine
18016 whether a pixel should be blurred or not. The option value must be an
18017 integer in the range [-30,30]. A value of 0 will filter all the image,
18018 a value included in [0,30] will filter flat areas and a value included
18019 in [-30,0] will filter edges. Default value is 0.
18021 @item chroma_radius, cr
18022 Set the chroma radius. The option value must be a float number in
18023 the range [0.1,5.0] that specifies the variance of the gaussian filter
18024 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18026 @item chroma_strength, cs
18027 Set the chroma strength. The option value must be a float number
18028 in the range [-1.0,1.0] that configures the blurring. A value included
18029 in [0.0,1.0] will blur the image whereas a value included in
18030 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18032 @item chroma_threshold, ct
18033 Set the chroma threshold used as a coefficient to determine
18034 whether a pixel should be blurred or not. The option value must be an
18035 integer in the range [-30,30]. A value of 0 will filter all the image,
18036 a value included in [0,30] will filter flat areas and a value included
18037 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18040 If a chroma option is not explicitly set, the corresponding luma value
18044 Apply sobel operator to input video stream.
18046 The filter accepts the following option:
18050 Set which planes will be processed, unprocessed planes will be copied.
18051 By default value 0xf, all planes will be processed.
18054 Set value which will be multiplied with filtered result.
18057 Set value which will be added to filtered result.
18063 Apply a simple postprocessing filter that compresses and decompresses the image
18064 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18065 and average the results.
18067 The filter accepts the following options:
18071 Set quality. This option defines the number of levels for averaging. It accepts
18072 an integer in the range 0-6. If set to @code{0}, the filter will have no
18073 effect. A value of @code{6} means the higher quality. For each increment of
18074 that value the speed drops by a factor of approximately 2. Default value is
18078 Force a constant quantization parameter. If not set, the filter will use the QP
18079 from the video stream (if available).
18082 Set thresholding mode. Available modes are:
18086 Set hard thresholding (default).
18088 Set soft thresholding (better de-ringing effect, but likely blurrier).
18091 @item use_bframe_qp
18092 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18093 option may cause flicker since the B-Frames have often larger QP. Default is
18094 @code{0} (not enabled).
18097 @subsection Commands
18099 This filter supports the following commands:
18101 @item quality, level
18102 Set quality level. The value @code{max} can be used to set the maximum level,
18103 currently @code{6}.
18109 Scale the input by applying one of the super-resolution methods based on
18110 convolutional neural networks. Supported models:
18114 Super-Resolution Convolutional Neural Network model (SRCNN).
18115 See @url{https://arxiv.org/abs/1501.00092}.
18118 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18119 See @url{https://arxiv.org/abs/1609.05158}.
18122 Training scripts as well as scripts for model file (.pb) saving can be found at
18123 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18124 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18126 Native model files (.model) can be generated from TensorFlow model
18127 files (.pb) by using tools/python/convert.py
18129 The filter accepts the following options:
18133 Specify which DNN backend to use for model loading and execution. This option accepts
18134 the following values:
18138 Native implementation of DNN loading and execution.
18141 TensorFlow backend. To enable this backend you
18142 need to install the TensorFlow for C library (see
18143 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18144 @code{--enable-libtensorflow}
18147 Default value is @samp{native}.
18150 Set path to model file specifying network architecture and its parameters.
18151 Note that different backends use different file formats. TensorFlow backend
18152 can load files for both formats, while native backend can load files for only
18156 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18157 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18158 input upscaled using bicubic upscaling with proper scale factor.
18161 This feature can also be finished with @ref{dnn_processing} filter.
18165 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18167 This filter takes in input two input videos, the first input is
18168 considered the "main" source and is passed unchanged to the
18169 output. The second input is used as a "reference" video for computing
18172 Both video inputs must have the same resolution and pixel format for
18173 this filter to work correctly. Also it assumes that both inputs
18174 have the same number of frames, which are compared one by one.
18176 The filter stores the calculated SSIM of each frame.
18178 The description of the accepted parameters follows.
18181 @item stats_file, f
18182 If specified the filter will use the named file to save the SSIM of
18183 each individual frame. When filename equals "-" the data is sent to
18187 The file printed if @var{stats_file} is selected, contains a sequence of
18188 key/value pairs of the form @var{key}:@var{value} for each compared
18191 A description of each shown parameter follows:
18195 sequential number of the input frame, starting from 1
18197 @item Y, U, V, R, G, B
18198 SSIM of the compared frames for the component specified by the suffix.
18201 SSIM of the compared frames for the whole frame.
18204 Same as above but in dB representation.
18207 This filter also supports the @ref{framesync} options.
18209 @subsection Examples
18214 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18215 [main][ref] ssim="stats_file=stats.log" [out]
18218 On this example the input file being processed is compared with the
18219 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18220 is stored in @file{stats.log}.
18223 Another example with both psnr and ssim at same time:
18225 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18229 Another example with different containers:
18231 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 -
18237 Convert between different stereoscopic image formats.
18239 The filters accept the following options:
18243 Set stereoscopic image format of input.
18245 Available values for input image formats are:
18248 side by side parallel (left eye left, right eye right)
18251 side by side crosseye (right eye left, left eye right)
18254 side by side parallel with half width resolution
18255 (left eye left, right eye right)
18258 side by side crosseye with half width resolution
18259 (right eye left, left eye right)
18263 above-below (left eye above, right eye below)
18267 above-below (right eye above, left eye below)
18271 above-below with half height resolution
18272 (left eye above, right eye below)
18276 above-below with half height resolution
18277 (right eye above, left eye below)
18280 alternating frames (left eye first, right eye second)
18283 alternating frames (right eye first, left eye second)
18286 interleaved rows (left eye has top row, right eye starts on next row)
18289 interleaved rows (right eye has top row, left eye starts on next row)
18292 interleaved columns, left eye first
18295 interleaved columns, right eye first
18297 Default value is @samp{sbsl}.
18301 Set stereoscopic image format of output.
18305 side by side parallel (left eye left, right eye right)
18308 side by side crosseye (right eye left, left eye right)
18311 side by side parallel with half width resolution
18312 (left eye left, right eye right)
18315 side by side crosseye with half width resolution
18316 (right eye left, left eye right)
18320 above-below (left eye above, right eye below)
18324 above-below (right eye above, left eye below)
18328 above-below with half height resolution
18329 (left eye above, right eye below)
18333 above-below with half height resolution
18334 (right eye above, left eye below)
18337 alternating frames (left eye first, right eye second)
18340 alternating frames (right eye first, left eye second)
18343 interleaved rows (left eye has top row, right eye starts on next row)
18346 interleaved rows (right eye has top row, left eye starts on next row)
18349 anaglyph red/blue gray
18350 (red filter on left eye, blue filter on right eye)
18353 anaglyph red/green gray
18354 (red filter on left eye, green filter on right eye)
18357 anaglyph red/cyan gray
18358 (red filter on left eye, cyan filter on right eye)
18361 anaglyph red/cyan half colored
18362 (red filter on left eye, cyan filter on right eye)
18365 anaglyph red/cyan color
18366 (red filter on left eye, cyan filter on right eye)
18369 anaglyph red/cyan color optimized with the least squares projection of dubois
18370 (red filter on left eye, cyan filter on right eye)
18373 anaglyph green/magenta gray
18374 (green filter on left eye, magenta filter on right eye)
18377 anaglyph green/magenta half colored
18378 (green filter on left eye, magenta filter on right eye)
18381 anaglyph green/magenta colored
18382 (green filter on left eye, magenta filter on right eye)
18385 anaglyph green/magenta color optimized with the least squares projection of dubois
18386 (green filter on left eye, magenta filter on right eye)
18389 anaglyph yellow/blue gray
18390 (yellow filter on left eye, blue filter on right eye)
18393 anaglyph yellow/blue half colored
18394 (yellow filter on left eye, blue filter on right eye)
18397 anaglyph yellow/blue colored
18398 (yellow filter on left eye, blue filter on right eye)
18401 anaglyph yellow/blue color optimized with the least squares projection of dubois
18402 (yellow filter on left eye, blue filter on right eye)
18405 mono output (left eye only)
18408 mono output (right eye only)
18411 checkerboard, left eye first
18414 checkerboard, right eye first
18417 interleaved columns, left eye first
18420 interleaved columns, right eye first
18426 Default value is @samp{arcd}.
18429 @subsection Examples
18433 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18439 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18445 @section streamselect, astreamselect
18446 Select video or audio streams.
18448 The filter accepts the following options:
18452 Set number of inputs. Default is 2.
18455 Set input indexes to remap to outputs.
18458 @subsection Commands
18460 The @code{streamselect} and @code{astreamselect} filter supports the following
18465 Set input indexes to remap to outputs.
18468 @subsection Examples
18472 Select first 5 seconds 1st stream and rest of time 2nd stream:
18474 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18478 Same as above, but for audio:
18480 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18487 Draw subtitles on top of input video using the libass library.
18489 To enable compilation of this filter you need to configure FFmpeg with
18490 @code{--enable-libass}. This filter also requires a build with libavcodec and
18491 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18492 Alpha) subtitles format.
18494 The filter accepts the following options:
18498 Set the filename of the subtitle file to read. It must be specified.
18500 @item original_size
18501 Specify the size of the original video, the video for which the ASS file
18502 was composed. For the syntax of this option, check the
18503 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18504 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18505 correctly scale the fonts if the aspect ratio has been changed.
18508 Set a directory path containing fonts that can be used by the filter.
18509 These fonts will be used in addition to whatever the font provider uses.
18512 Process alpha channel, by default alpha channel is untouched.
18515 Set subtitles input character encoding. @code{subtitles} filter only. Only
18516 useful if not UTF-8.
18518 @item stream_index, si
18519 Set subtitles stream index. @code{subtitles} filter only.
18522 Override default style or script info parameters of the subtitles. It accepts a
18523 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18526 If the first key is not specified, it is assumed that the first value
18527 specifies the @option{filename}.
18529 For example, to render the file @file{sub.srt} on top of the input
18530 video, use the command:
18535 which is equivalent to:
18537 subtitles=filename=sub.srt
18540 To render the default subtitles stream from file @file{video.mkv}, use:
18542 subtitles=video.mkv
18545 To render the second subtitles stream from that file, use:
18547 subtitles=video.mkv:si=1
18550 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18551 @code{DejaVu Serif}, use:
18553 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18556 @section super2xsai
18558 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18559 Interpolate) pixel art scaling algorithm.
18561 Useful for enlarging pixel art images without reducing sharpness.
18565 Swap two rectangular objects in video.
18567 This filter accepts the following options:
18577 Set 1st rect x coordinate.
18580 Set 1st rect y coordinate.
18583 Set 2nd rect x coordinate.
18586 Set 2nd rect y coordinate.
18588 All expressions are evaluated once for each frame.
18591 The all options are expressions containing the following constants:
18596 The input width and height.
18599 same as @var{w} / @var{h}
18602 input sample aspect ratio
18605 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18608 The number of the input frame, starting from 0.
18611 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18614 the position in the file of the input frame, NAN if unknown
18621 Blend successive video frames.
18627 Apply telecine process to the video.
18629 This filter accepts the following options:
18638 The default value is @code{top}.
18642 A string of numbers representing the pulldown pattern you wish to apply.
18643 The default value is @code{23}.
18647 Some typical patterns:
18652 24p: 2332 (preferred)
18659 24p: 222222222223 ("Euro pulldown")
18664 @section thistogram
18666 Compute and draw a color distribution histogram for the input video across time.
18668 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18669 at certain time, this filter shows also past histograms of number of frames defined
18670 by @code{width} option.
18672 The computed histogram is a representation of the color component
18673 distribution in an image.
18675 The filter accepts the following options:
18679 Set width of single color component output. Default value is @code{0}.
18680 Value of @code{0} means width will be picked from input video.
18681 This also set number of passed histograms to keep.
18682 Allowed range is [0, 8192].
18684 @item display_mode, d
18686 It accepts the following values:
18689 Per color component graphs are placed below each other.
18692 Per color component graphs are placed side by side.
18695 Presents information identical to that in the @code{parade}, except
18696 that the graphs representing color components are superimposed directly
18699 Default is @code{stack}.
18701 @item levels_mode, m
18702 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18703 Default is @code{linear}.
18705 @item components, c
18706 Set what color components to display.
18707 Default is @code{7}.
18710 Set background opacity. Default is @code{0.9}.
18713 Show envelope. Default is disabled.
18716 Set envelope color. Default is @code{gold}.
18721 Available values for slide is:
18724 Draw new frame when right border is reached.
18727 Replace old columns with new ones.
18730 Scroll from right to left.
18733 Scroll from left to right.
18736 Draw single picture.
18739 Default is @code{replace}.
18744 Apply threshold effect to video stream.
18746 This filter needs four video streams to perform thresholding.
18747 First stream is stream we are filtering.
18748 Second stream is holding threshold values, third stream is holding min values,
18749 and last, fourth stream is holding max values.
18751 The filter accepts the following option:
18755 Set which planes will be processed, unprocessed planes will be copied.
18756 By default value 0xf, all planes will be processed.
18759 For example if first stream pixel's component value is less then threshold value
18760 of pixel component from 2nd threshold stream, third stream value will picked,
18761 otherwise fourth stream pixel component value will be picked.
18763 Using color source filter one can perform various types of thresholding:
18765 @subsection Examples
18769 Binary threshold, using gray color as threshold:
18771 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18775 Inverted binary threshold, using gray color as threshold:
18777 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18781 Truncate binary threshold, using gray color as threshold:
18783 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18787 Threshold to zero, using gray color as threshold:
18789 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18793 Inverted threshold to zero, using gray color as threshold:
18795 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18800 Select the most representative frame in a given sequence of consecutive frames.
18802 The filter accepts the following options:
18806 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18807 will pick one of them, and then handle the next batch of @var{n} frames until
18808 the end. Default is @code{100}.
18811 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18812 value will result in a higher memory usage, so a high value is not recommended.
18814 @subsection Examples
18818 Extract one picture each 50 frames:
18824 Complete example of a thumbnail creation with @command{ffmpeg}:
18826 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18833 Tile several successive frames together.
18835 The @ref{untile} filter can do the reverse.
18837 The filter accepts the following options:
18842 Set the grid size (i.e. the number of lines and columns). For the syntax of
18843 this option, check the
18844 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18847 Set the maximum number of frames to render in the given area. It must be less
18848 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18849 the area will be used.
18852 Set the outer border margin in pixels.
18855 Set the inner border thickness (i.e. the number of pixels between frames). For
18856 more advanced padding options (such as having different values for the edges),
18857 refer to the pad video filter.
18860 Specify the color of the unused area. For the syntax of this option, check the
18861 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18862 The default value of @var{color} is "black".
18865 Set the number of frames to overlap when tiling several successive frames together.
18866 The value must be between @code{0} and @var{nb_frames - 1}.
18869 Set the number of frames to initially be empty before displaying first output frame.
18870 This controls how soon will one get first output frame.
18871 The value must be between @code{0} and @var{nb_frames - 1}.
18874 @subsection Examples
18878 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18880 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18882 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18883 duplicating each output frame to accommodate the originally detected frame
18887 Display @code{5} pictures in an area of @code{3x2} frames,
18888 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18889 mixed flat and named options:
18891 tile=3x2:nb_frames=5:padding=7:margin=2
18895 @section tinterlace
18897 Perform various types of temporal field interlacing.
18899 Frames are counted starting from 1, so the first input frame is
18902 The filter accepts the following options:
18907 Specify the mode of the interlacing. This option can also be specified
18908 as a value alone. See below for a list of values for this option.
18910 Available values are:
18914 Move odd frames into the upper field, even into the lower field,
18915 generating a double height frame at half frame rate.
18919 Frame 1 Frame 2 Frame 3 Frame 4
18921 11111 22222 33333 44444
18922 11111 22222 33333 44444
18923 11111 22222 33333 44444
18924 11111 22222 33333 44444
18938 Only output odd frames, even frames are dropped, generating a frame with
18939 unchanged height at half frame rate.
18944 Frame 1 Frame 2 Frame 3 Frame 4
18946 11111 22222 33333 44444
18947 11111 22222 33333 44444
18948 11111 22222 33333 44444
18949 11111 22222 33333 44444
18959 Only output even frames, odd frames are dropped, generating a frame with
18960 unchanged height at half frame rate.
18965 Frame 1 Frame 2 Frame 3 Frame 4
18967 11111 22222 33333 44444
18968 11111 22222 33333 44444
18969 11111 22222 33333 44444
18970 11111 22222 33333 44444
18980 Expand each frame to full height, but pad alternate lines with black,
18981 generating a frame with double height at the same input frame rate.
18986 Frame 1 Frame 2 Frame 3 Frame 4
18988 11111 22222 33333 44444
18989 11111 22222 33333 44444
18990 11111 22222 33333 44444
18991 11111 22222 33333 44444
18994 11111 ..... 33333 .....
18995 ..... 22222 ..... 44444
18996 11111 ..... 33333 .....
18997 ..... 22222 ..... 44444
18998 11111 ..... 33333 .....
18999 ..... 22222 ..... 44444
19000 11111 ..... 33333 .....
19001 ..... 22222 ..... 44444
19005 @item interleave_top, 4
19006 Interleave the upper field from odd frames with the lower field from
19007 even frames, generating a frame with unchanged height at half frame rate.
19012 Frame 1 Frame 2 Frame 3 Frame 4
19014 11111<- 22222 33333<- 44444
19015 11111 22222<- 33333 44444<-
19016 11111<- 22222 33333<- 44444
19017 11111 22222<- 33333 44444<-
19027 @item interleave_bottom, 5
19028 Interleave the lower field from odd frames with the upper field from
19029 even frames, generating a frame with unchanged height at half frame rate.
19034 Frame 1 Frame 2 Frame 3 Frame 4
19036 11111 22222<- 33333 44444<-
19037 11111<- 22222 33333<- 44444
19038 11111 22222<- 33333 44444<-
19039 11111<- 22222 33333<- 44444
19049 @item interlacex2, 6
19050 Double frame rate with unchanged height. Frames are inserted each
19051 containing the second temporal field from the previous input frame and
19052 the first temporal field from the next input frame. This mode relies on
19053 the top_field_first flag. Useful for interlaced video displays with no
19054 field synchronisation.
19059 Frame 1 Frame 2 Frame 3 Frame 4
19061 11111 22222 33333 44444
19062 11111 22222 33333 44444
19063 11111 22222 33333 44444
19064 11111 22222 33333 44444
19067 11111 22222 22222 33333 33333 44444 44444
19068 11111 11111 22222 22222 33333 33333 44444
19069 11111 22222 22222 33333 33333 44444 44444
19070 11111 11111 22222 22222 33333 33333 44444
19075 Move odd frames into the upper field, even into the lower field,
19076 generating a double height frame at same frame rate.
19081 Frame 1 Frame 2 Frame 3 Frame 4
19083 11111 22222 33333 44444
19084 11111 22222 33333 44444
19085 11111 22222 33333 44444
19086 11111 22222 33333 44444
19089 11111 33333 33333 55555
19090 22222 22222 44444 44444
19091 11111 33333 33333 55555
19092 22222 22222 44444 44444
19093 11111 33333 33333 55555
19094 22222 22222 44444 44444
19095 11111 33333 33333 55555
19096 22222 22222 44444 44444
19101 Numeric values are deprecated but are accepted for backward
19102 compatibility reasons.
19104 Default mode is @code{merge}.
19107 Specify flags influencing the filter process.
19109 Available value for @var{flags} is:
19112 @item low_pass_filter, vlpf
19113 Enable linear vertical low-pass filtering in the filter.
19114 Vertical low-pass filtering is required when creating an interlaced
19115 destination from a progressive source which contains high-frequency
19116 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19119 @item complex_filter, cvlpf
19120 Enable complex vertical low-pass filtering.
19121 This will slightly less reduce interlace 'twitter' and Moire
19122 patterning but better retain detail and subjective sharpness impression.
19125 Bypass already interlaced frames, only adjust the frame rate.
19128 Vertical low-pass filtering and bypassing already interlaced frames can only be
19129 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19134 Pick median pixels from several successive input video frames.
19136 The filter accepts the following options:
19140 Set radius of median filter.
19141 Default is 1. Allowed range is from 1 to 127.
19144 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19147 Set median percentile. Default value is @code{0.5}.
19148 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19149 minimum values, and @code{1} maximum values.
19154 Mix successive video frames.
19156 A description of the accepted options follows.
19160 The number of successive frames to mix. If unspecified, it defaults to 3.
19163 Specify weight of each input video frame.
19164 Each weight is separated by space. If number of weights is smaller than
19165 number of @var{frames} last specified weight will be used for all remaining
19169 Specify scale, if it is set it will be multiplied with sum
19170 of each weight multiplied with pixel values to give final destination
19171 pixel value. By default @var{scale} is auto scaled to sum of weights.
19174 @subsection Examples
19178 Average 7 successive frames:
19180 tmix=frames=7:weights="1 1 1 1 1 1 1"
19184 Apply simple temporal convolution:
19186 tmix=frames=3:weights="-1 3 -1"
19190 Similar as above but only showing temporal differences:
19192 tmix=frames=3:weights="-1 2 -1":scale=1
19198 Tone map colors from different dynamic ranges.
19200 This filter expects data in single precision floating point, as it needs to
19201 operate on (and can output) out-of-range values. Another filter, such as
19202 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19204 The tonemapping algorithms implemented only work on linear light, so input
19205 data should be linearized beforehand (and possibly correctly tagged).
19208 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19211 @subsection Options
19212 The filter accepts the following options.
19216 Set the tone map algorithm to use.
19218 Possible values are:
19221 Do not apply any tone map, only desaturate overbright pixels.
19224 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19225 in-range values, while distorting out-of-range values.
19228 Stretch the entire reference gamut to a linear multiple of the display.
19231 Fit a logarithmic transfer between the tone curves.
19234 Preserve overall image brightness with a simple curve, using nonlinear
19235 contrast, which results in flattening details and degrading color accuracy.
19238 Preserve both dark and bright details better than @var{reinhard}, at the cost
19239 of slightly darkening everything. Use it when detail preservation is more
19240 important than color and brightness accuracy.
19243 Smoothly map out-of-range values, while retaining contrast and colors for
19244 in-range material as much as possible. Use it when color accuracy is more
19245 important than detail preservation.
19251 Tune the tone mapping algorithm.
19253 This affects the following algorithms:
19259 Specifies the scale factor to use while stretching.
19263 Specifies the exponent of the function.
19267 Specify an extra linear coefficient to multiply into the signal before clipping.
19271 Specify the local contrast coefficient at the display peak.
19272 Default to 0.5, which means that in-gamut values will be about half as bright
19279 Specify the transition point from linear to mobius transform. Every value
19280 below this point is guaranteed to be mapped 1:1. The higher the value, the
19281 more accurate the result will be, at the cost of losing bright details.
19282 Default to 0.3, which due to the steep initial slope still preserves in-range
19283 colors fairly accurately.
19287 Apply desaturation for highlights that exceed this level of brightness. The
19288 higher the parameter, the more color information will be preserved. This
19289 setting helps prevent unnaturally blown-out colors for super-highlights, by
19290 (smoothly) turning into white instead. This makes images feel more natural,
19291 at the cost of reducing information about out-of-range colors.
19293 The default of 2.0 is somewhat conservative and will mostly just apply to
19294 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19296 This option works only if the input frame has a supported color tag.
19299 Override signal/nominal/reference peak with this value. Useful when the
19300 embedded peak information in display metadata is not reliable or when tone
19301 mapping from a lower range to a higher range.
19306 Temporarily pad video frames.
19308 The filter accepts the following options:
19312 Specify number of delay frames before input video stream. Default is 0.
19315 Specify number of padding frames after input video stream.
19316 Set to -1 to pad indefinitely. Default is 0.
19319 Set kind of frames added to beginning of stream.
19320 Can be either @var{add} or @var{clone}.
19321 With @var{add} frames of solid-color are added.
19322 With @var{clone} frames are clones of first frame.
19323 Default is @var{add}.
19326 Set kind of frames added to end of stream.
19327 Can be either @var{add} or @var{clone}.
19328 With @var{add} frames of solid-color are added.
19329 With @var{clone} frames are clones of last frame.
19330 Default is @var{add}.
19332 @item start_duration, stop_duration
19333 Specify the duration of the start/stop delay. See
19334 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19335 for the accepted syntax.
19336 These options override @var{start} and @var{stop}. Default is 0.
19339 Specify the color of the padded area. For the syntax of this option,
19340 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19341 manual,ffmpeg-utils}.
19343 The default value of @var{color} is "black".
19349 Transpose rows with columns in the input video and optionally flip it.
19351 It accepts the following parameters:
19356 Specify the transposition direction.
19358 Can assume the following values:
19360 @item 0, 4, cclock_flip
19361 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19369 Rotate by 90 degrees clockwise, that is:
19377 Rotate by 90 degrees counterclockwise, that is:
19384 @item 3, 7, clock_flip
19385 Rotate by 90 degrees clockwise and vertically flip, that is:
19393 For values between 4-7, the transposition is only done if the input
19394 video geometry is portrait and not landscape. These values are
19395 deprecated, the @code{passthrough} option should be used instead.
19397 Numerical values are deprecated, and should be dropped in favor of
19398 symbolic constants.
19401 Do not apply the transposition if the input geometry matches the one
19402 specified by the specified value. It accepts the following values:
19405 Always apply transposition.
19407 Preserve portrait geometry (when @var{height} >= @var{width}).
19409 Preserve landscape geometry (when @var{width} >= @var{height}).
19412 Default value is @code{none}.
19415 For example to rotate by 90 degrees clockwise and preserve portrait
19418 transpose=dir=1:passthrough=portrait
19421 The command above can also be specified as:
19423 transpose=1:portrait
19426 @section transpose_npp
19428 Transpose rows with columns in the input video and optionally flip it.
19429 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19431 It accepts the following parameters:
19436 Specify the transposition direction.
19438 Can assume the following values:
19441 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19444 Rotate by 90 degrees clockwise.
19447 Rotate by 90 degrees counterclockwise.
19450 Rotate by 90 degrees clockwise and vertically flip.
19454 Do not apply the transposition if the input geometry matches the one
19455 specified by the specified value. It accepts the following values:
19458 Always apply transposition. (default)
19460 Preserve portrait geometry (when @var{height} >= @var{width}).
19462 Preserve landscape geometry (when @var{width} >= @var{height}).
19468 Trim the input so that the output contains one continuous subpart of the input.
19470 It accepts the following parameters:
19473 Specify the time of the start of the kept section, i.e. the frame with the
19474 timestamp @var{start} will be the first frame in the output.
19477 Specify the time of the first frame that will be dropped, i.e. the frame
19478 immediately preceding the one with the timestamp @var{end} will be the last
19479 frame in the output.
19482 This is the same as @var{start}, except this option sets the start timestamp
19483 in timebase units instead of seconds.
19486 This is the same as @var{end}, except this option sets the end timestamp
19487 in timebase units instead of seconds.
19490 The maximum duration of the output in seconds.
19493 The number of the first frame that should be passed to the output.
19496 The number of the first frame that should be dropped.
19499 @option{start}, @option{end}, and @option{duration} are expressed as time
19500 duration specifications; see
19501 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19502 for the accepted syntax.
19504 Note that the first two sets of the start/end options and the @option{duration}
19505 option look at the frame timestamp, while the _frame variants simply count the
19506 frames that pass through the filter. Also note that this filter does not modify
19507 the timestamps. If you wish for the output timestamps to start at zero, insert a
19508 setpts filter after the trim filter.
19510 If multiple start or end options are set, this filter tries to be greedy and
19511 keep all the frames that match at least one of the specified constraints. To keep
19512 only the part that matches all the constraints at once, chain multiple trim
19515 The defaults are such that all the input is kept. So it is possible to set e.g.
19516 just the end values to keep everything before the specified time.
19521 Drop everything except the second minute of input:
19523 ffmpeg -i INPUT -vf trim=60:120
19527 Keep only the first second:
19529 ffmpeg -i INPUT -vf trim=duration=1
19534 @section unpremultiply
19535 Apply alpha unpremultiply effect to input video stream using first plane
19536 of second stream as alpha.
19538 Both streams must have same dimensions and same pixel format.
19540 The filter accepts the following option:
19544 Set which planes will be processed, unprocessed planes will be copied.
19545 By default value 0xf, all planes will be processed.
19547 If the format has 1 or 2 components, then luma is bit 0.
19548 If the format has 3 or 4 components:
19549 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19550 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19551 If present, the alpha channel is always the last bit.
19554 Do not require 2nd input for processing, instead use alpha plane from input stream.
19560 Sharpen or blur the input video.
19562 It accepts the following parameters:
19565 @item luma_msize_x, lx
19566 Set the luma matrix horizontal size. It must be an odd integer between
19567 3 and 23. The default value is 5.
19569 @item luma_msize_y, ly
19570 Set the luma matrix vertical size. It must be an odd integer between 3
19571 and 23. The default value is 5.
19573 @item luma_amount, la
19574 Set the luma effect strength. It must be a floating point number, reasonable
19575 values lay between -1.5 and 1.5.
19577 Negative values will blur the input video, while positive values will
19578 sharpen it, a value of zero will disable the effect.
19580 Default value is 1.0.
19582 @item chroma_msize_x, cx
19583 Set the chroma matrix horizontal size. It must be an odd integer
19584 between 3 and 23. The default value is 5.
19586 @item chroma_msize_y, cy
19587 Set the chroma matrix vertical size. It must be an odd integer
19588 between 3 and 23. The default value is 5.
19590 @item chroma_amount, ca
19591 Set the chroma effect strength. It must be a floating point number, reasonable
19592 values lay between -1.5 and 1.5.
19594 Negative values will blur the input video, while positive values will
19595 sharpen it, a value of zero will disable the effect.
19597 Default value is 0.0.
19601 All parameters are optional and default to the equivalent of the
19602 string '5:5:1.0:5:5:0.0'.
19604 @subsection Examples
19608 Apply strong luma sharpen effect:
19610 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19614 Apply a strong blur of both luma and chroma parameters:
19616 unsharp=7:7:-2:7:7:-2
19623 Decompose a video made of tiled images into the individual images.
19625 The frame rate of the output video is the frame rate of the input video
19626 multiplied by the number of tiles.
19628 This filter does the reverse of @ref{tile}.
19630 The filter accepts the following options:
19635 Set the grid size (i.e. the number of lines and columns). For the syntax of
19636 this option, check the
19637 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19640 @subsection Examples
19644 Produce a 1-second video from a still image file made of 25 frames stacked
19645 vertically, like an analogic film reel:
19647 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19653 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19654 the image at several (or - in the case of @option{quality} level @code{8} - all)
19655 shifts and average the results.
19657 The way this differs from the behavior of spp is that uspp actually encodes &
19658 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19659 DCT similar to MJPEG.
19661 The filter accepts the following options:
19665 Set quality. This option defines the number of levels for averaging. It accepts
19666 an integer in the range 0-8. If set to @code{0}, the filter will have no
19667 effect. A value of @code{8} means the higher quality. For each increment of
19668 that value the speed drops by a factor of approximately 2. Default value is
19672 Force a constant quantization parameter. If not set, the filter will use the QP
19673 from the video stream (if available).
19678 Convert 360 videos between various formats.
19680 The filter accepts the following options:
19686 Set format of the input/output video.
19694 Equirectangular projection.
19699 Cubemap with 3x2/6x1/1x6 layout.
19701 Format specific options:
19706 Set padding proportion for the input/output cubemap. Values in decimals.
19713 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)
19716 Default value is @b{@samp{0}}.
19717 Maximum value is @b{@samp{0.1}}.
19721 Set fixed padding for the input/output cubemap. Values in pixels.
19723 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19727 Set order of faces for the input/output cubemap. Choose one direction for each position.
19729 Designation of directions:
19745 Default value is @b{@samp{rludfb}}.
19749 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19751 Designation of angles:
19754 0 degrees clockwise
19756 90 degrees clockwise
19758 180 degrees clockwise
19760 270 degrees clockwise
19763 Default value is @b{@samp{000000}}.
19767 Equi-Angular Cubemap.
19774 Format specific options:
19779 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19781 If diagonal field of view is set it overrides horizontal and vertical field of view.
19786 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19788 If diagonal field of view is set it overrides horizontal and vertical field of view.
19794 Format specific options:
19799 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19801 If diagonal field of view is set it overrides horizontal and vertical field of view.
19806 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19808 If diagonal field of view is set it overrides horizontal and vertical field of view.
19814 Facebook's 360 formats.
19817 Stereographic format.
19819 Format specific options:
19824 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19826 If diagonal field of view is set it overrides horizontal and vertical field of view.
19831 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19833 If diagonal field of view is set it overrides horizontal and vertical field of view.
19840 Ball format, gives significant distortion toward the back.
19843 Hammer-Aitoff map projection format.
19846 Sinusoidal map projection format.
19849 Fisheye projection.
19851 Format specific options:
19856 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19858 If diagonal field of view is set it overrides horizontal and vertical field of view.
19863 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19865 If diagonal field of view is set it overrides horizontal and vertical field of view.
19869 Pannini projection.
19871 Format specific options:
19874 Set output pannini parameter.
19877 Set input pannini parameter.
19881 Cylindrical projection.
19883 Format specific options:
19888 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19890 If diagonal field of view is set it overrides horizontal and vertical field of view.
19895 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19897 If diagonal field of view is set it overrides horizontal and vertical field of view.
19901 Perspective projection. @i{(output only)}
19903 Format specific options:
19906 Set perspective parameter.
19910 Tetrahedron projection.
19913 Truncated square pyramid projection.
19917 Half equirectangular projection.
19922 Format specific options:
19927 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19929 If diagonal field of view is set it overrides horizontal and vertical field of view.
19934 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19936 If diagonal field of view is set it overrides horizontal and vertical field of view.
19940 Orthographic format.
19942 Format specific options:
19947 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19949 If diagonal field of view is set it overrides horizontal and vertical field of view.
19954 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19956 If diagonal field of view is set it overrides horizontal and vertical field of view.
19960 Octahedron projection.
19964 Set interpolation method.@*
19965 @i{Note: more complex interpolation methods require much more memory to run.}
19975 Bilinear interpolation.
19977 Lagrange9 interpolation.
19980 Bicubic interpolation.
19983 Lanczos interpolation.
19986 Spline16 interpolation.
19989 Gaussian interpolation.
19991 Mitchell interpolation.
19994 Default value is @b{@samp{line}}.
19998 Set the output video resolution.
20000 Default resolution depends on formats.
20004 Set the input/output stereo format.
20015 Default value is @b{@samp{2d}} for input and output format.
20020 Set rotation for the output video. Values in degrees.
20023 Set rotation order for the output video. Choose one item for each position.
20034 Default value is @b{@samp{ypr}}.
20039 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20043 Set if input video is flipped horizontally/vertically. Boolean values.
20046 Set if input video is transposed. Boolean value, by default disabled.
20049 Set if output video needs to be transposed. Boolean value, by default disabled.
20052 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20055 @subsection Examples
20059 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20061 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20064 Extract back view of Equi-Angular Cubemap:
20066 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20069 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20071 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20075 @subsection Commands
20077 This filter supports subset of above options as @ref{commands}.
20079 @section vaguedenoiser
20081 Apply a wavelet based denoiser.
20083 It transforms each frame from the video input into the wavelet domain,
20084 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20085 the obtained coefficients. It does an inverse wavelet transform after.
20086 Due to wavelet properties, it should give a nice smoothed result, and
20087 reduced noise, without blurring picture features.
20089 This filter accepts the following options:
20093 The filtering strength. The higher, the more filtered the video will be.
20094 Hard thresholding can use a higher threshold than soft thresholding
20095 before the video looks overfiltered. Default value is 2.
20098 The filtering method the filter will use.
20100 It accepts the following values:
20103 All values under the threshold will be zeroed.
20106 All values under the threshold will be zeroed. All values above will be
20107 reduced by the threshold.
20110 Scales or nullifies coefficients - intermediary between (more) soft and
20111 (less) hard thresholding.
20114 Default is garrote.
20117 Number of times, the wavelet will decompose the picture. Picture can't
20118 be decomposed beyond a particular point (typically, 8 for a 640x480
20119 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20122 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20125 A list of the planes to process. By default all planes are processed.
20128 The threshold type the filter will use.
20130 It accepts the following values:
20133 Threshold used is same for all decompositions.
20136 Threshold used depends also on each decomposition coefficients.
20139 Default is universal.
20142 @section vectorscope
20144 Display 2 color component values in the two dimensional graph (which is called
20147 This filter accepts the following options:
20151 Set vectorscope mode.
20153 It accepts the following values:
20157 Gray values are displayed on graph, higher brightness means more pixels have
20158 same component color value on location in graph. This is the default mode.
20161 Gray values are displayed on graph. Surrounding pixels values which are not
20162 present in video frame are drawn in gradient of 2 color components which are
20163 set by option @code{x} and @code{y}. The 3rd color component is static.
20166 Actual color components values present in video frame are displayed on graph.
20169 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20170 on graph increases value of another color component, which is luminance by
20171 default values of @code{x} and @code{y}.
20174 Actual colors present in video frame are displayed on graph. If two different
20175 colors map to same position on graph then color with higher value of component
20176 not present in graph is picked.
20179 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20180 component picked from radial gradient.
20184 Set which color component will be represented on X-axis. Default is @code{1}.
20187 Set which color component will be represented on Y-axis. Default is @code{2}.
20190 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20191 of color component which represents frequency of (X, Y) location in graph.
20196 No envelope, this is default.
20199 Instant envelope, even darkest single pixel will be clearly highlighted.
20202 Hold maximum and minimum values presented in graph over time. This way you
20203 can still spot out of range values without constantly looking at vectorscope.
20206 Peak and instant envelope combined together.
20210 Set what kind of graticule to draw.
20219 Set graticule opacity.
20222 Set graticule flags.
20226 Draw graticule for white point.
20229 Draw graticule for black point.
20232 Draw color points short names.
20236 Set background opacity.
20238 @item lthreshold, l
20239 Set low threshold for color component not represented on X or Y axis.
20240 Values lower than this value will be ignored. Default is 0.
20241 Note this value is multiplied with actual max possible value one pixel component
20242 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20245 @item hthreshold, h
20246 Set high threshold for color component not represented on X or Y axis.
20247 Values higher than this value will be ignored. Default is 1.
20248 Note this value is multiplied with actual max possible value one pixel component
20249 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20250 is 0.9 * 255 = 230.
20252 @item colorspace, c
20253 Set what kind of colorspace to use when drawing graticule.
20263 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20264 This means no tint, and output will remain gray.
20267 @anchor{vidstabdetect}
20268 @section vidstabdetect
20270 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20271 @ref{vidstabtransform} for pass 2.
20273 This filter generates a file with relative translation and rotation
20274 transform information about subsequent frames, which is then used by
20275 the @ref{vidstabtransform} filter.
20277 To enable compilation of this filter you need to configure FFmpeg with
20278 @code{--enable-libvidstab}.
20280 This filter accepts the following options:
20284 Set the path to the file used to write the transforms information.
20285 Default value is @file{transforms.trf}.
20288 Set how shaky the video is and how quick the camera is. It accepts an
20289 integer in the range 1-10, a value of 1 means little shakiness, a
20290 value of 10 means strong shakiness. Default value is 5.
20293 Set the accuracy of the detection process. It must be a value in the
20294 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20295 accuracy. Default value is 15.
20298 Set stepsize of the search process. The region around minimum is
20299 scanned with 1 pixel resolution. Default value is 6.
20302 Set minimum contrast. Below this value a local measurement field is
20303 discarded. Must be a floating point value in the range 0-1. Default
20307 Set reference frame number for tripod mode.
20309 If enabled, the motion of the frames is compared to a reference frame
20310 in the filtered stream, identified by the specified number. The idea
20311 is to compensate all movements in a more-or-less static scene and keep
20312 the camera view absolutely still.
20314 If set to 0, it is disabled. The frames are counted starting from 1.
20317 Show fields and transforms in the resulting frames. It accepts an
20318 integer in the range 0-2. Default value is 0, which disables any
20322 @subsection Examples
20326 Use default values:
20332 Analyze strongly shaky movie and put the results in file
20333 @file{mytransforms.trf}:
20335 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20339 Visualize the result of internal transformations in the resulting
20342 vidstabdetect=show=1
20346 Analyze a video with medium shakiness using @command{ffmpeg}:
20348 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20352 @anchor{vidstabtransform}
20353 @section vidstabtransform
20355 Video stabilization/deshaking: pass 2 of 2,
20356 see @ref{vidstabdetect} for pass 1.
20358 Read a file with transform information for each frame and
20359 apply/compensate them. Together with the @ref{vidstabdetect}
20360 filter this can be used to deshake videos. See also
20361 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20362 the @ref{unsharp} filter, see below.
20364 To enable compilation of this filter you need to configure FFmpeg with
20365 @code{--enable-libvidstab}.
20367 @subsection Options
20371 Set path to the file used to read the transforms. Default value is
20372 @file{transforms.trf}.
20375 Set the number of frames (value*2 + 1) used for lowpass filtering the
20376 camera movements. Default value is 10.
20378 For example a number of 10 means that 21 frames are used (10 in the
20379 past and 10 in the future) to smoothen the motion in the video. A
20380 larger value leads to a smoother video, but limits the acceleration of
20381 the camera (pan/tilt movements). 0 is a special case where a static
20382 camera is simulated.
20385 Set the camera path optimization algorithm.
20387 Accepted values are:
20390 gaussian kernel low-pass filter on camera motion (default)
20392 averaging on transformations
20396 Set maximal number of pixels to translate frames. Default value is -1,
20400 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20401 value is -1, meaning no limit.
20404 Specify how to deal with borders that may be visible due to movement
20407 Available values are:
20410 keep image information from previous frame (default)
20412 fill the border black
20416 Invert transforms if set to 1. Default value is 0.
20419 Consider transforms as relative to previous frame if set to 1,
20420 absolute if set to 0. Default value is 0.
20423 Set percentage to zoom. A positive value will result in a zoom-in
20424 effect, a negative value in a zoom-out effect. Default value is 0 (no
20428 Set optimal zooming to avoid borders.
20430 Accepted values are:
20435 optimal static zoom value is determined (only very strong movements
20436 will lead to visible borders) (default)
20438 optimal adaptive zoom value is determined (no borders will be
20439 visible), see @option{zoomspeed}
20442 Note that the value given at zoom is added to the one calculated here.
20445 Set percent to zoom maximally each frame (enabled when
20446 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20450 Specify type of interpolation.
20452 Available values are:
20457 linear only horizontal
20459 linear in both directions (default)
20461 cubic in both directions (slow)
20465 Enable virtual tripod mode if set to 1, which is equivalent to
20466 @code{relative=0:smoothing=0}. Default value is 0.
20468 Use also @code{tripod} option of @ref{vidstabdetect}.
20471 Increase log verbosity if set to 1. Also the detected global motions
20472 are written to the temporary file @file{global_motions.trf}. Default
20476 @subsection Examples
20480 Use @command{ffmpeg} for a typical stabilization with default values:
20482 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20485 Note the use of the @ref{unsharp} filter which is always recommended.
20488 Zoom in a bit more and load transform data from a given file:
20490 vidstabtransform=zoom=5:input="mytransforms.trf"
20494 Smoothen the video even more:
20496 vidstabtransform=smoothing=30
20502 Flip the input video vertically.
20504 For example, to vertically flip a video with @command{ffmpeg}:
20506 ffmpeg -i in.avi -vf "vflip" out.avi
20511 Detect variable frame rate video.
20513 This filter tries to detect if the input is variable or constant frame rate.
20515 At end it will output number of frames detected as having variable delta pts,
20516 and ones with constant delta pts.
20517 If there was frames with variable delta, than it will also show min, max and
20518 average delta encountered.
20522 Boost or alter saturation.
20524 The filter accepts the following options:
20527 Set strength of boost if positive value or strength of alter if negative value.
20528 Default is 0. Allowed range is from -2 to 2.
20531 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20534 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20537 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20540 Set the red luma coefficient.
20543 Set the green luma coefficient.
20546 Set the blue luma coefficient.
20549 If @code{intensity} is negative and this is set to 1, colors will change,
20550 otherwise colors will be less saturated, more towards gray.
20553 @subsection Commands
20555 This filter supports the all above options as @ref{commands}.
20560 Make or reverse a natural vignetting effect.
20562 The filter accepts the following options:
20566 Set lens angle expression as a number of radians.
20568 The value is clipped in the @code{[0,PI/2]} range.
20570 Default value: @code{"PI/5"}
20574 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20578 Set forward/backward mode.
20580 Available modes are:
20583 The larger the distance from the central point, the darker the image becomes.
20586 The larger the distance from the central point, the brighter the image becomes.
20587 This can be used to reverse a vignette effect, though there is no automatic
20588 detection to extract the lens @option{angle} and other settings (yet). It can
20589 also be used to create a burning effect.
20592 Default value is @samp{forward}.
20595 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20597 It accepts the following values:
20600 Evaluate expressions only once during the filter initialization.
20603 Evaluate expressions for each incoming frame. This is way slower than the
20604 @samp{init} mode since it requires all the scalers to be re-computed, but it
20605 allows advanced dynamic expressions.
20608 Default value is @samp{init}.
20611 Set dithering to reduce the circular banding effects. Default is @code{1}
20615 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20616 Setting this value to the SAR of the input will make a rectangular vignetting
20617 following the dimensions of the video.
20619 Default is @code{1/1}.
20622 @subsection Expressions
20624 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20625 following parameters.
20630 input width and height
20633 the number of input frame, starting from 0
20636 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20637 @var{TB} units, NAN if undefined
20640 frame rate of the input video, NAN if the input frame rate is unknown
20643 the PTS (Presentation TimeStamp) of the filtered video frame,
20644 expressed in seconds, NAN if undefined
20647 time base of the input video
20651 @subsection Examples
20655 Apply simple strong vignetting effect:
20661 Make a flickering vignetting:
20663 vignette='PI/4+random(1)*PI/50':eval=frame
20668 @section vmafmotion
20670 Obtain the average VMAF motion score of a video.
20671 It is one of the component metrics of VMAF.
20673 The obtained average motion score is printed through the logging system.
20675 The filter accepts the following options:
20679 If specified, the filter will use the named file to save the motion score of
20680 each frame with respect to the previous frame.
20681 When filename equals "-" the data is sent to standard output.
20686 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20690 Stack input videos vertically.
20692 All streams must be of same pixel format and of same width.
20694 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20695 to create same output.
20697 The filter accepts the following options:
20701 Set number of input streams. Default is 2.
20704 If set to 1, force the output to terminate when the shortest input
20705 terminates. Default value is 0.
20710 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20711 Deinterlacing Filter").
20713 Based on the process described by Martin Weston for BBC R&D, and
20714 implemented based on the de-interlace algorithm written by Jim
20715 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20716 uses filter coefficients calculated by BBC R&D.
20718 This filter uses field-dominance information in frame to decide which
20719 of each pair of fields to place first in the output.
20720 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20722 There are two sets of filter coefficients, so called "simple"
20723 and "complex". Which set of filter coefficients is used can
20724 be set by passing an optional parameter:
20728 Set the interlacing filter coefficients. Accepts one of the following values:
20732 Simple filter coefficient set.
20734 More-complex filter coefficient set.
20736 Default value is @samp{complex}.
20739 Specify which frames to deinterlace. Accepts one of the following values:
20743 Deinterlace all frames,
20745 Only deinterlace frames marked as interlaced.
20748 Default value is @samp{all}.
20752 Video waveform monitor.
20754 The waveform monitor plots color component intensity. By default luminance
20755 only. Each column of the waveform corresponds to a column of pixels in the
20758 It accepts the following options:
20762 Can be either @code{row}, or @code{column}. Default is @code{column}.
20763 In row mode, the graph on the left side represents color component value 0 and
20764 the right side represents value = 255. In column mode, the top side represents
20765 color component value = 0 and bottom side represents value = 255.
20768 Set intensity. Smaller values are useful to find out how many values of the same
20769 luminance are distributed across input rows/columns.
20770 Default value is @code{0.04}. Allowed range is [0, 1].
20773 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20774 In mirrored mode, higher values will be represented on the left
20775 side for @code{row} mode and at the top for @code{column} mode. Default is
20776 @code{1} (mirrored).
20780 It accepts the following values:
20783 Presents information identical to that in the @code{parade}, except
20784 that the graphs representing color components are superimposed directly
20787 This display mode makes it easier to spot relative differences or similarities
20788 in overlapping areas of the color components that are supposed to be identical,
20789 such as neutral whites, grays, or blacks.
20792 Display separate graph for the color components side by side in
20793 @code{row} mode or one below the other in @code{column} mode.
20796 Display separate graph for the color components side by side in
20797 @code{column} mode or one below the other in @code{row} mode.
20799 Using this display mode makes it easy to spot color casts in the highlights
20800 and shadows of an image, by comparing the contours of the top and the bottom
20801 graphs of each waveform. Since whites, grays, and blacks are characterized
20802 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20803 should display three waveforms of roughly equal width/height. If not, the
20804 correction is easy to perform by making level adjustments the three waveforms.
20806 Default is @code{stack}.
20808 @item components, c
20809 Set which color components to display. Default is 1, which means only luminance
20810 or red color component if input is in RGB colorspace. If is set for example to
20811 7 it will display all 3 (if) available color components.
20816 No envelope, this is default.
20819 Instant envelope, minimum and maximum values presented in graph will be easily
20820 visible even with small @code{step} value.
20823 Hold minimum and maximum values presented in graph across time. This way you
20824 can still spot out of range values without constantly looking at waveforms.
20827 Peak and instant envelope combined together.
20833 No filtering, this is default.
20836 Luma and chroma combined together.
20839 Similar as above, but shows difference between blue and red chroma.
20842 Similar as above, but use different colors.
20845 Similar as above, but again with different colors.
20848 Displays only chroma.
20851 Displays actual color value on waveform.
20854 Similar as above, but with luma showing frequency of chroma values.
20858 Set which graticule to display.
20862 Do not display graticule.
20865 Display green graticule showing legal broadcast ranges.
20868 Display orange graticule showing legal broadcast ranges.
20871 Display invert graticule showing legal broadcast ranges.
20875 Set graticule opacity.
20878 Set graticule flags.
20882 Draw numbers above lines. By default enabled.
20885 Draw dots instead of lines.
20889 Set scale used for displaying graticule.
20896 Default is digital.
20899 Set background opacity.
20903 Set tint for output.
20904 Only used with lowpass filter and when display is not overlay and input
20905 pixel formats are not RGB.
20908 @section weave, doubleweave
20910 The @code{weave} takes a field-based video input and join
20911 each two sequential fields into single frame, producing a new double
20912 height clip with half the frame rate and half the frame count.
20914 The @code{doubleweave} works same as @code{weave} but without
20915 halving frame rate and frame count.
20917 It accepts the following option:
20921 Set first field. Available values are:
20925 Set the frame as top-field-first.
20928 Set the frame as bottom-field-first.
20932 @subsection Examples
20936 Interlace video using @ref{select} and @ref{separatefields} filter:
20938 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20943 Apply the xBR high-quality magnification filter which is designed for pixel
20944 art. It follows a set of edge-detection rules, see
20945 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20947 It accepts the following option:
20951 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20952 @code{3xBR} and @code{4} for @code{4xBR}.
20953 Default is @code{3}.
20958 Apply cross fade from one input video stream to another input video stream.
20959 The cross fade is applied for specified duration.
20961 The filter accepts the following options:
20965 Set one of available transition effects:
21013 Default transition effect is fade.
21016 Set cross fade duration in seconds.
21017 Default duration is 1 second.
21020 Set cross fade start relative to first input stream in seconds.
21021 Default offset is 0.
21024 Set expression for custom transition effect.
21026 The expressions can use the following variables and functions:
21031 The coordinates of the current sample.
21035 The width and height of the image.
21038 Progress of transition effect.
21041 Currently processed plane.
21044 Return value of first input at current location and plane.
21047 Return value of second input at current location and plane.
21053 Return the value of the pixel at location (@var{x},@var{y}) of the
21054 first/second/third/fourth component of first input.
21060 Return the value of the pixel at location (@var{x},@var{y}) of the
21061 first/second/third/fourth component of second input.
21065 @subsection Examples
21069 Cross fade from one input video to another input video, with fade transition and duration of transition
21070 of 2 seconds starting at offset of 5 seconds:
21072 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21077 Pick median pixels from several input videos.
21079 The filter accepts the following options:
21083 Set number of inputs.
21084 Default is 3. Allowed range is from 3 to 255.
21085 If number of inputs is even number, than result will be mean value between two median values.
21088 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21091 Set median percentile. Default value is @code{0.5}.
21092 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21093 minimum values, and @code{1} maximum values.
21097 Stack video inputs into custom layout.
21099 All streams must be of same pixel format.
21101 The filter accepts the following options:
21105 Set number of input streams. Default is 2.
21108 Specify layout of inputs.
21109 This option requires the desired layout configuration to be explicitly set by the user.
21110 This sets position of each video input in output. Each input
21111 is separated by '|'.
21112 The first number represents the column, and the second number represents the row.
21113 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21114 where X is video input from which to take width or height.
21115 Multiple values can be used when separated by '+'. In such
21116 case values are summed together.
21118 Note that if inputs are of different sizes gaps may appear, as not all of
21119 the output video frame will be filled. Similarly, videos can overlap each
21120 other if their position doesn't leave enough space for the full frame of
21123 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21124 a layout must be set by the user.
21127 If set to 1, force the output to terminate when the shortest input
21128 terminates. Default value is 0.
21131 If set to valid color, all unused pixels will be filled with that color.
21132 By default fill is set to none, so it is disabled.
21135 @subsection Examples
21139 Display 4 inputs into 2x2 grid.
21143 input1(0, 0) | input3(w0, 0)
21144 input2(0, h0) | input4(w0, h0)
21148 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21151 Note that if inputs are of different sizes, gaps or overlaps may occur.
21154 Display 4 inputs into 1x4 grid.
21161 input4(0, h0+h1+h2)
21165 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21168 Note that if inputs are of different widths, unused space will appear.
21171 Display 9 inputs into 3x3 grid.
21175 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21176 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21177 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21181 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
21184 Note that if inputs are of different sizes, gaps or overlaps may occur.
21187 Display 16 inputs into 4x4 grid.
21191 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21192 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21193 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21194 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21198 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|
21199 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
21202 Note that if inputs are of different sizes, gaps or overlaps may occur.
21209 Deinterlace the input video ("yadif" means "yet another deinterlacing
21212 It accepts the following parameters:
21218 The interlacing mode to adopt. It accepts one of the following values:
21221 @item 0, send_frame
21222 Output one frame for each frame.
21223 @item 1, send_field
21224 Output one frame for each field.
21225 @item 2, send_frame_nospatial
21226 Like @code{send_frame}, but it skips the spatial interlacing check.
21227 @item 3, send_field_nospatial
21228 Like @code{send_field}, but it skips the spatial interlacing check.
21231 The default value is @code{send_frame}.
21234 The picture field parity assumed for the input interlaced video. It accepts one
21235 of the following values:
21239 Assume the top field is first.
21241 Assume the bottom field is first.
21243 Enable automatic detection of field parity.
21246 The default value is @code{auto}.
21247 If the interlacing is unknown or the decoder does not export this information,
21248 top field first will be assumed.
21251 Specify which frames to deinterlace. Accepts one of the following
21256 Deinterlace all frames.
21257 @item 1, interlaced
21258 Only deinterlace frames marked as interlaced.
21261 The default value is @code{all}.
21264 @section yadif_cuda
21266 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21267 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21270 It accepts the following parameters:
21276 The interlacing mode to adopt. It accepts one of the following values:
21279 @item 0, send_frame
21280 Output one frame for each frame.
21281 @item 1, send_field
21282 Output one frame for each field.
21283 @item 2, send_frame_nospatial
21284 Like @code{send_frame}, but it skips the spatial interlacing check.
21285 @item 3, send_field_nospatial
21286 Like @code{send_field}, but it skips the spatial interlacing check.
21289 The default value is @code{send_frame}.
21292 The picture field parity assumed for the input interlaced video. It accepts one
21293 of the following values:
21297 Assume the top field is first.
21299 Assume the bottom field is first.
21301 Enable automatic detection of field parity.
21304 The default value is @code{auto}.
21305 If the interlacing is unknown or the decoder does not export this information,
21306 top field first will be assumed.
21309 Specify which frames to deinterlace. Accepts one of the following
21314 Deinterlace all frames.
21315 @item 1, interlaced
21316 Only deinterlace frames marked as interlaced.
21319 The default value is @code{all}.
21324 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21325 The algorithm is described in
21326 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21328 It accepts the following parameters:
21332 Set the window radius. Default value is 3.
21335 Set which planes to filter. Default is only the first plane.
21338 Set blur strength. Default value is 128.
21341 @subsection Commands
21342 This filter supports same @ref{commands} as options.
21346 Apply Zoom & Pan effect.
21348 This filter accepts the following options:
21352 Set the zoom expression. Range is 1-10. Default is 1.
21356 Set the x and y expression. Default is 0.
21359 Set the duration expression in number of frames.
21360 This sets for how many number of frames effect will last for
21361 single input image.
21364 Set the output image size, default is 'hd720'.
21367 Set the output frame rate, default is '25'.
21370 Each expression can contain the following constants:
21389 Output frame count.
21392 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21394 @item out_time, time, ot
21395 The output timestamp expressed in seconds.
21399 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21400 for current input frame.
21404 'x' and 'y' of last output frame of previous input frame or 0 when there was
21405 not yet such frame (first input frame).
21408 Last calculated zoom from 'z' expression for current input frame.
21411 Last calculated zoom of last output frame of previous input frame.
21414 Number of output frames for current input frame. Calculated from 'd' expression
21415 for each input frame.
21418 number of output frames created for previous input frame
21421 Rational number: input width / input height
21424 sample aspect ratio
21427 display aspect ratio
21431 @subsection Examples
21435 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21437 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
21441 Zoom in up to 1.5x and pan always at center of picture:
21443 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21447 Same as above but without pausing:
21449 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21453 Zoom in 2x into center of picture only for the first second of the input video:
21455 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21462 Scale (resize) the input video, using the z.lib library:
21463 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21464 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21466 The zscale filter forces the output display aspect ratio to be the same
21467 as the input, by changing the output sample aspect ratio.
21469 If the input image format is different from the format requested by
21470 the next filter, the zscale filter will convert the input to the
21473 @subsection Options
21474 The filter accepts the following options.
21479 Set the output video dimension expression. Default value is the input
21482 If the @var{width} or @var{w} value is 0, the input width is used for
21483 the output. If the @var{height} or @var{h} value is 0, the input height
21484 is used for the output.
21486 If one and only one of the values is -n with n >= 1, the zscale filter
21487 will use a value that maintains the aspect ratio of the input image,
21488 calculated from the other specified dimension. After that it will,
21489 however, make sure that the calculated dimension is divisible by n and
21490 adjust the value if necessary.
21492 If both values are -n with n >= 1, the behavior will be identical to
21493 both values being set to 0 as previously detailed.
21495 See below for the list of accepted constants for use in the dimension
21499 Set the video size. For the syntax of this option, check the
21500 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21503 Set the dither type.
21505 Possible values are:
21510 @item error_diffusion
21516 Set the resize filter type.
21518 Possible values are:
21528 Default is bilinear.
21531 Set the color range.
21533 Possible values are:
21540 Default is same as input.
21543 Set the color primaries.
21545 Possible values are:
21555 Default is same as input.
21558 Set the transfer characteristics.
21560 Possible values are:
21574 Default is same as input.
21577 Set the colorspace matrix.
21579 Possible value are:
21590 Default is same as input.
21593 Set the input color range.
21595 Possible values are:
21602 Default is same as input.
21604 @item primariesin, pin
21605 Set the input color primaries.
21607 Possible values are:
21617 Default is same as input.
21619 @item transferin, tin
21620 Set the input transfer characteristics.
21622 Possible values are:
21633 Default is same as input.
21635 @item matrixin, min
21636 Set the input colorspace matrix.
21638 Possible value are:
21650 Set the output chroma location.
21652 Possible values are:
21663 @item chromalin, cin
21664 Set the input chroma location.
21666 Possible values are:
21678 Set the nominal peak luminance.
21681 The values of the @option{w} and @option{h} options are expressions
21682 containing the following constants:
21687 The input width and height
21691 These are the same as @var{in_w} and @var{in_h}.
21695 The output (scaled) width and height
21699 These are the same as @var{out_w} and @var{out_h}
21702 The same as @var{iw} / @var{ih}
21705 input sample aspect ratio
21708 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21712 horizontal and vertical input chroma subsample values. For example for the
21713 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21717 horizontal and vertical output chroma subsample values. For example for the
21718 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21721 @subsection Commands
21723 This filter supports the following commands:
21727 Set the output video dimension expression.
21728 The command accepts the same syntax of the corresponding option.
21730 If the specified expression is not valid, it is kept at its current
21734 @c man end VIDEO FILTERS
21736 @chapter OpenCL Video Filters
21737 @c man begin OPENCL VIDEO FILTERS
21739 Below is a description of the currently available OpenCL video filters.
21741 To enable compilation of these filters you need to configure FFmpeg with
21742 @code{--enable-opencl}.
21744 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21747 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21748 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21749 given device parameters.
21751 @item -filter_hw_device @var{name}
21752 Pass the hardware device called @var{name} to all filters in any filter graph.
21756 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21760 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21762 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21766 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.
21768 @section avgblur_opencl
21770 Apply average blur filter.
21772 The filter accepts the following options:
21776 Set horizontal radius size.
21777 Range is @code{[1, 1024]} and default value is @code{1}.
21780 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21783 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21786 @subsection Example
21790 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.
21792 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21796 @section boxblur_opencl
21798 Apply a boxblur algorithm to the input video.
21800 It accepts the following parameters:
21804 @item luma_radius, lr
21805 @item luma_power, lp
21806 @item chroma_radius, cr
21807 @item chroma_power, cp
21808 @item alpha_radius, ar
21809 @item alpha_power, ap
21813 A description of the accepted options follows.
21816 @item luma_radius, lr
21817 @item chroma_radius, cr
21818 @item alpha_radius, ar
21819 Set an expression for the box radius in pixels used for blurring the
21820 corresponding input plane.
21822 The radius value must be a non-negative number, and must not be
21823 greater than the value of the expression @code{min(w,h)/2} for the
21824 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21827 Default value for @option{luma_radius} is "2". If not specified,
21828 @option{chroma_radius} and @option{alpha_radius} default to the
21829 corresponding value set for @option{luma_radius}.
21831 The expressions can contain the following constants:
21835 The input width and height in pixels.
21839 The input chroma image width and height in pixels.
21843 The horizontal and vertical chroma subsample values. For example, for the
21844 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21847 @item luma_power, lp
21848 @item chroma_power, cp
21849 @item alpha_power, ap
21850 Specify how many times the boxblur filter is applied to the
21851 corresponding plane.
21853 Default value for @option{luma_power} is 2. If not specified,
21854 @option{chroma_power} and @option{alpha_power} default to the
21855 corresponding value set for @option{luma_power}.
21857 A value of 0 will disable the effect.
21860 @subsection Examples
21862 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.
21866 Apply a boxblur filter with the luma, chroma, and alpha radius
21867 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.
21869 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21870 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21874 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.
21876 For the luma plane, a 2x2 box radius will be run once.
21878 For the chroma plane, a 4x4 box radius will be run 5 times.
21880 For the alpha plane, a 3x3 box radius will be run 7 times.
21882 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21886 @section colorkey_opencl
21887 RGB colorspace color keying.
21889 The filter accepts the following options:
21893 The color which will be replaced with transparency.
21896 Similarity percentage with the key color.
21898 0.01 matches only the exact key color, while 1.0 matches everything.
21903 0.0 makes pixels either fully transparent, or not transparent at all.
21905 Higher values result in semi-transparent pixels, with a higher transparency
21906 the more similar the pixels color is to the key color.
21909 @subsection Examples
21913 Make every semi-green pixel in the input transparent with some slight blending:
21915 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
21919 @section convolution_opencl
21921 Apply convolution of 3x3, 5x5, 7x7 matrix.
21923 The filter accepts the following options:
21930 Set matrix for each plane.
21931 Matrix is sequence of 9, 25 or 49 signed numbers.
21932 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
21938 Set multiplier for calculated value for each plane.
21939 If unset or 0, it will be sum of all matrix elements.
21940 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
21946 Set bias for each plane. This value is added to the result of the multiplication.
21947 Useful for making the overall image brighter or darker.
21948 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
21952 @subsection Examples
21958 -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
21964 -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
21968 Apply edge enhance:
21970 -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
21976 -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
21980 Apply laplacian edge detector which includes diagonals:
21982 -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
21988 -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
21992 @section erosion_opencl
21994 Apply erosion effect to the video.
21996 This filter replaces the pixel by the local(3x3) minimum.
21998 It accepts the following options:
22005 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22006 If @code{0}, plane will remain unchanged.
22009 Flag which specifies the pixel to refer to.
22010 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22012 Flags to local 3x3 coordinates region centered on @code{x}:
22021 @subsection Example
22025 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.
22027 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22031 @section deshake_opencl
22032 Feature-point based video stabilization filter.
22034 The filter accepts the following options:
22038 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22041 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22043 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22045 Viewing point matches in the output video is only supported for RGB input.
22047 Defaults to @code{0}.
22049 @item adaptive_crop
22050 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22052 Defaults to @code{1}.
22054 @item refine_features
22055 Whether or not feature points should be refined at a sub-pixel level.
22057 This can be turned off for a slight performance gain at the cost of precision.
22059 Defaults to @code{1}.
22061 @item smooth_strength
22062 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22064 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22066 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22068 Defaults to @code{0.0}.
22070 @item smooth_window_multiplier
22071 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22073 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22075 Acceptable values range from @code{0.1} to @code{10.0}.
22077 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22078 potentially improving smoothness, but also increase latency and memory usage.
22080 Defaults to @code{2.0}.
22084 @subsection Examples
22088 Stabilize a video with a fixed, medium smoothing strength:
22090 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22094 Stabilize a video with debugging (both in console and in rendered video):
22096 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22100 @section dilation_opencl
22102 Apply dilation effect to the video.
22104 This filter replaces the pixel by the local(3x3) maximum.
22106 It accepts the following options:
22113 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22114 If @code{0}, plane will remain unchanged.
22117 Flag which specifies the pixel to refer to.
22118 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22120 Flags to local 3x3 coordinates region centered on @code{x}:
22129 @subsection Example
22133 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.
22135 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22139 @section nlmeans_opencl
22141 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22143 @section overlay_opencl
22145 Overlay one video on top of another.
22147 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22148 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22150 The filter accepts the following options:
22155 Set the x coordinate of the overlaid video on the main video.
22156 Default value is @code{0}.
22159 Set the y coordinate of the overlaid video on the main video.
22160 Default value is @code{0}.
22164 @subsection Examples
22168 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22170 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22173 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22175 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22180 @section pad_opencl
22182 Add paddings to the input image, and place the original input at the
22183 provided @var{x}, @var{y} coordinates.
22185 It accepts the following options:
22190 Specify an expression for the size of the output image with the
22191 paddings added. If the value for @var{width} or @var{height} is 0, the
22192 corresponding input size is used for the output.
22194 The @var{width} expression can reference the value set by the
22195 @var{height} expression, and vice versa.
22197 The default value of @var{width} and @var{height} is 0.
22201 Specify the offsets to place the input image at within the padded area,
22202 with respect to the top/left border of the output image.
22204 The @var{x} expression can reference the value set by the @var{y}
22205 expression, and vice versa.
22207 The default value of @var{x} and @var{y} is 0.
22209 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22210 so the input image is centered on the padded area.
22213 Specify the color of the padded area. For the syntax of this option,
22214 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22215 manual,ffmpeg-utils}.
22218 Pad to an aspect instead to a resolution.
22221 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22222 options are expressions containing the following constants:
22227 The input video width and height.
22231 These are the same as @var{in_w} and @var{in_h}.
22235 The output width and height (the size of the padded area), as
22236 specified by the @var{width} and @var{height} expressions.
22240 These are the same as @var{out_w} and @var{out_h}.
22244 The x and y offsets as specified by the @var{x} and @var{y}
22245 expressions, or NAN if not yet specified.
22248 same as @var{iw} / @var{ih}
22251 input sample aspect ratio
22254 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22257 @section prewitt_opencl
22259 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22261 The filter accepts the following option:
22265 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22268 Set value which will be multiplied with filtered result.
22269 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22272 Set value which will be added to filtered result.
22273 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22276 @subsection Example
22280 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22282 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22286 @anchor{program_opencl}
22287 @section program_opencl
22289 Filter video using an OpenCL program.
22294 OpenCL program source file.
22297 Kernel name in program.
22300 Number of inputs to the filter. Defaults to 1.
22303 Size of output frames. Defaults to the same as the first input.
22307 The @code{program_opencl} filter also supports the @ref{framesync} options.
22309 The program source file must contain a kernel function with the given name,
22310 which will be run once for each plane of the output. Each run on a plane
22311 gets enqueued as a separate 2D global NDRange with one work-item for each
22312 pixel to be generated. The global ID offset for each work-item is therefore
22313 the coordinates of a pixel in the destination image.
22315 The kernel function needs to take the following arguments:
22318 Destination image, @var{__write_only image2d_t}.
22320 This image will become the output; the kernel should write all of it.
22322 Frame index, @var{unsigned int}.
22324 This is a counter starting from zero and increasing by one for each frame.
22326 Source images, @var{__read_only image2d_t}.
22328 These are the most recent images on each input. The kernel may read from
22329 them to generate the output, but they can't be written to.
22336 Copy the input to the output (output must be the same size as the input).
22338 __kernel void copy(__write_only image2d_t destination,
22339 unsigned int index,
22340 __read_only image2d_t source)
22342 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22344 int2 location = (int2)(get_global_id(0), get_global_id(1));
22346 float4 value = read_imagef(source, sampler, location);
22348 write_imagef(destination, location, value);
22353 Apply a simple transformation, rotating the input by an amount increasing
22354 with the index counter. Pixel values are linearly interpolated by the
22355 sampler, and the output need not have the same dimensions as the input.
22357 __kernel void rotate_image(__write_only image2d_t dst,
22358 unsigned int index,
22359 __read_only image2d_t src)
22361 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22362 CLK_FILTER_LINEAR);
22364 float angle = (float)index / 100.0f;
22366 float2 dst_dim = convert_float2(get_image_dim(dst));
22367 float2 src_dim = convert_float2(get_image_dim(src));
22369 float2 dst_cen = dst_dim / 2.0f;
22370 float2 src_cen = src_dim / 2.0f;
22372 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22374 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22376 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22377 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22379 src_pos = src_pos * src_dim / dst_dim;
22381 float2 src_loc = src_pos + src_cen;
22383 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22384 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22385 write_imagef(dst, dst_loc, 0.5f);
22387 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22392 Blend two inputs together, with the amount of each input used varying
22393 with the index counter.
22395 __kernel void blend_images(__write_only image2d_t dst,
22396 unsigned int index,
22397 __read_only image2d_t src1,
22398 __read_only image2d_t src2)
22400 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22401 CLK_FILTER_LINEAR);
22403 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22405 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22406 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22407 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22409 float4 val1 = read_imagef(src1, sampler, src1_loc);
22410 float4 val2 = read_imagef(src2, sampler, src2_loc);
22412 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22418 @section roberts_opencl
22419 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22421 The filter accepts the following option:
22425 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22428 Set value which will be multiplied with filtered result.
22429 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22432 Set value which will be added to filtered result.
22433 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22436 @subsection Example
22440 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22442 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22446 @section sobel_opencl
22448 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22450 The filter accepts the following option:
22454 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22457 Set value which will be multiplied with filtered result.
22458 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22461 Set value which will be added to filtered result.
22462 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22465 @subsection Example
22469 Apply sobel operator with scale set to 2 and delta set to 10
22471 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22475 @section tonemap_opencl
22477 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22479 It accepts the following parameters:
22483 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22486 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22489 Apply desaturation for highlights that exceed this level of brightness. The
22490 higher the parameter, the more color information will be preserved. This
22491 setting helps prevent unnaturally blown-out colors for super-highlights, by
22492 (smoothly) turning into white instead. This makes images feel more natural,
22493 at the cost of reducing information about out-of-range colors.
22495 The default value is 0.5, and the algorithm here is a little different from
22496 the cpu version tonemap currently. A setting of 0.0 disables this option.
22499 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22500 is used to detect whether the scene has changed or not. If the distance between
22501 the current frame average brightness and the current running average exceeds
22502 a threshold value, we would re-calculate scene average and peak brightness.
22503 The default value is 0.2.
22506 Specify the output pixel format.
22508 Currently supported formats are:
22515 Set the output color range.
22517 Possible values are:
22523 Default is same as input.
22526 Set the output color primaries.
22528 Possible values are:
22534 Default is same as input.
22537 Set the output transfer characteristics.
22539 Possible values are:
22548 Set the output colorspace matrix.
22550 Possible value are:
22556 Default is same as input.
22560 @subsection Example
22564 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22566 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22570 @section unsharp_opencl
22572 Sharpen or blur the input video.
22574 It accepts the following parameters:
22577 @item luma_msize_x, lx
22578 Set the luma matrix horizontal size.
22579 Range is @code{[1, 23]} and default value is @code{5}.
22581 @item luma_msize_y, ly
22582 Set the luma matrix vertical size.
22583 Range is @code{[1, 23]} and default value is @code{5}.
22585 @item luma_amount, la
22586 Set the luma effect strength.
22587 Range is @code{[-10, 10]} and default value is @code{1.0}.
22589 Negative values will blur the input video, while positive values will
22590 sharpen it, a value of zero will disable the effect.
22592 @item chroma_msize_x, cx
22593 Set the chroma matrix horizontal size.
22594 Range is @code{[1, 23]} and default value is @code{5}.
22596 @item chroma_msize_y, cy
22597 Set the chroma matrix vertical size.
22598 Range is @code{[1, 23]} and default value is @code{5}.
22600 @item chroma_amount, ca
22601 Set the chroma effect strength.
22602 Range is @code{[-10, 10]} and default value is @code{0.0}.
22604 Negative values will blur the input video, while positive values will
22605 sharpen it, a value of zero will disable the effect.
22609 All parameters are optional and default to the equivalent of the
22610 string '5:5:1.0:5:5:0.0'.
22612 @subsection Examples
22616 Apply strong luma sharpen effect:
22618 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22622 Apply a strong blur of both luma and chroma parameters:
22624 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22628 @section xfade_opencl
22630 Cross fade two videos with custom transition effect by using OpenCL.
22632 It accepts the following options:
22636 Set one of possible transition effects.
22640 Select custom transition effect, the actual transition description
22641 will be picked from source and kernel options.
22653 Default transition is fade.
22657 OpenCL program source file for custom transition.
22660 Set name of kernel to use for custom transition from program source file.
22663 Set duration of video transition.
22666 Set time of start of transition relative to first video.
22669 The program source file must contain a kernel function with the given name,
22670 which will be run once for each plane of the output. Each run on a plane
22671 gets enqueued as a separate 2D global NDRange with one work-item for each
22672 pixel to be generated. The global ID offset for each work-item is therefore
22673 the coordinates of a pixel in the destination image.
22675 The kernel function needs to take the following arguments:
22678 Destination image, @var{__write_only image2d_t}.
22680 This image will become the output; the kernel should write all of it.
22683 First Source image, @var{__read_only image2d_t}.
22684 Second Source image, @var{__read_only image2d_t}.
22686 These are the most recent images on each input. The kernel may read from
22687 them to generate the output, but they can't be written to.
22690 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22697 Apply dots curtain transition effect:
22699 __kernel void blend_images(__write_only image2d_t dst,
22700 __read_only image2d_t src1,
22701 __read_only image2d_t src2,
22704 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22705 CLK_FILTER_LINEAR);
22706 int2 p = (int2)(get_global_id(0), get_global_id(1));
22707 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22708 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22711 float2 dots = (float2)(20.0, 20.0);
22712 float2 center = (float2)(0,0);
22715 float4 val1 = read_imagef(src1, sampler, p);
22716 float4 val2 = read_imagef(src2, sampler, p);
22717 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22719 write_imagef(dst, p, next ? val1 : val2);
22725 @c man end OPENCL VIDEO FILTERS
22727 @chapter VAAPI Video Filters
22728 @c man begin VAAPI VIDEO FILTERS
22730 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22732 To enable compilation of these filters you need to configure FFmpeg with
22733 @code{--enable-vaapi}.
22735 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}
22737 @section tonemap_vaapi
22739 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22740 It maps the dynamic range of HDR10 content to the SDR content.
22741 It currently only accepts HDR10 as input.
22743 It accepts the following parameters:
22747 Specify the output pixel format.
22749 Currently supported formats are:
22758 Set the output color primaries.
22760 Default is same as input.
22763 Set the output transfer characteristics.
22768 Set the output colorspace matrix.
22770 Default is same as input.
22774 @subsection Example
22778 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22780 tonemap_vaapi=format=p010:t=bt2020-10
22784 @c man end VAAPI VIDEO FILTERS
22786 @chapter Video Sources
22787 @c man begin VIDEO SOURCES
22789 Below is a description of the currently available video sources.
22793 Buffer video frames, and make them available to the filter chain.
22795 This source is mainly intended for a programmatic use, in particular
22796 through the interface defined in @file{libavfilter/buffersrc.h}.
22798 It accepts the following parameters:
22803 Specify the size (width and height) of the buffered video frames. For the
22804 syntax of this option, check the
22805 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22808 The input video width.
22811 The input video height.
22814 A string representing the pixel format of the buffered video frames.
22815 It may be a number corresponding to a pixel format, or a pixel format
22819 Specify the timebase assumed by the timestamps of the buffered frames.
22822 Specify the frame rate expected for the video stream.
22824 @item pixel_aspect, sar
22825 The sample (pixel) aspect ratio of the input video.
22828 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22829 to the filtergraph description to specify swscale flags for automatically
22830 inserted scalers. See @ref{Filtergraph syntax}.
22832 @item hw_frames_ctx
22833 When using a hardware pixel format, this should be a reference to an
22834 AVHWFramesContext describing input frames.
22839 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22842 will instruct the source to accept video frames with size 320x240 and
22843 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22844 square pixels (1:1 sample aspect ratio).
22845 Since the pixel format with name "yuv410p" corresponds to the number 6
22846 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22847 this example corresponds to:
22849 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22852 Alternatively, the options can be specified as a flat string, but this
22853 syntax is deprecated:
22855 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22859 Create a pattern generated by an elementary cellular automaton.
22861 The initial state of the cellular automaton can be defined through the
22862 @option{filename} and @option{pattern} options. If such options are
22863 not specified an initial state is created randomly.
22865 At each new frame a new row in the video is filled with the result of
22866 the cellular automaton next generation. The behavior when the whole
22867 frame is filled is defined by the @option{scroll} option.
22869 This source accepts the following options:
22873 Read the initial cellular automaton state, i.e. the starting row, from
22874 the specified file.
22875 In the file, each non-whitespace character is considered an alive
22876 cell, a newline will terminate the row, and further characters in the
22877 file will be ignored.
22880 Read the initial cellular automaton state, i.e. the starting row, from
22881 the specified string.
22883 Each non-whitespace character in the string is considered an alive
22884 cell, a newline will terminate the row, and further characters in the
22885 string will be ignored.
22888 Set the video rate, that is the number of frames generated per second.
22891 @item random_fill_ratio, ratio
22892 Set the random fill ratio for the initial cellular automaton row. It
22893 is a floating point number value ranging from 0 to 1, defaults to
22896 This option is ignored when a file or a pattern is specified.
22898 @item random_seed, seed
22899 Set the seed for filling randomly the initial row, must be an integer
22900 included between 0 and UINT32_MAX. If not specified, or if explicitly
22901 set to -1, the filter will try to use a good random seed on a best
22905 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22906 Default value is 110.
22909 Set the size of the output video. For the syntax of this option, check the
22910 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22912 If @option{filename} or @option{pattern} is specified, the size is set
22913 by default to the width of the specified initial state row, and the
22914 height is set to @var{width} * PHI.
22916 If @option{size} is set, it must contain the width of the specified
22917 pattern string, and the specified pattern will be centered in the
22920 If a filename or a pattern string is not specified, the size value
22921 defaults to "320x518" (used for a randomly generated initial state).
22924 If set to 1, scroll the output upward when all the rows in the output
22925 have been already filled. If set to 0, the new generated row will be
22926 written over the top row just after the bottom row is filled.
22929 @item start_full, full
22930 If set to 1, completely fill the output with generated rows before
22931 outputting the first frame.
22932 This is the default behavior, for disabling set the value to 0.
22935 If set to 1, stitch the left and right row edges together.
22936 This is the default behavior, for disabling set the value to 0.
22939 @subsection Examples
22943 Read the initial state from @file{pattern}, and specify an output of
22946 cellauto=f=pattern:s=200x400
22950 Generate a random initial row with a width of 200 cells, with a fill
22953 cellauto=ratio=2/3:s=200x200
22957 Create a pattern generated by rule 18 starting by a single alive cell
22958 centered on an initial row with width 100:
22960 cellauto=p=@@:s=100x400:full=0:rule=18
22964 Specify a more elaborated initial pattern:
22966 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
22971 @anchor{coreimagesrc}
22972 @section coreimagesrc
22973 Video source generated on GPU using Apple's CoreImage API on OSX.
22975 This video source is a specialized version of the @ref{coreimage} video filter.
22976 Use a core image generator at the beginning of the applied filterchain to
22977 generate the content.
22979 The coreimagesrc video source accepts the following options:
22981 @item list_generators
22982 List all available generators along with all their respective options as well as
22983 possible minimum and maximum values along with the default values.
22985 list_generators=true
22989 Specify the size of the sourced video. For the syntax of this option, check the
22990 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22991 The default value is @code{320x240}.
22994 Specify the frame rate of the sourced video, as the number of frames
22995 generated per second. It has to be a string in the format
22996 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22997 number or a valid video frame rate abbreviation. The default value is
23001 Set the sample aspect ratio of the sourced video.
23004 Set the duration of the sourced video. See
23005 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23006 for the accepted syntax.
23008 If not specified, or the expressed duration is negative, the video is
23009 supposed to be generated forever.
23012 Additionally, all options of the @ref{coreimage} video filter are accepted.
23013 A complete filterchain can be used for further processing of the
23014 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23015 and examples for details.
23017 @subsection Examples
23022 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23023 given as complete and escaped command-line for Apple's standard bash shell:
23025 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23027 This example is equivalent to the QRCode example of @ref{coreimage} without the
23028 need for a nullsrc video source.
23033 Generate several gradients.
23037 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23038 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23041 Set frame rate, expressed as number of frames per second. Default
23044 @item c0, c1, c2, c3, c4, c5, c6, c7
23045 Set 8 colors. Default values for colors is to pick random one.
23047 @item x0, y0, y0, y1
23048 Set gradient line source and destination points. If negative or out of range, random ones
23052 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23055 Set seed for picking gradient line points.
23058 Set the duration of the sourced video. See
23059 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23060 for the accepted syntax.
23062 If not specified, or the expressed duration is negative, the video is
23063 supposed to be generated forever.
23066 Set speed of gradients rotation.
23070 @section mandelbrot
23072 Generate a Mandelbrot set fractal, and progressively zoom towards the
23073 point specified with @var{start_x} and @var{start_y}.
23075 This source accepts the following options:
23080 Set the terminal pts value. Default value is 400.
23083 Set the terminal scale value.
23084 Must be a floating point value. Default value is 0.3.
23087 Set the inner coloring mode, that is the algorithm used to draw the
23088 Mandelbrot fractal internal region.
23090 It shall assume one of the following values:
23095 Show time until convergence.
23097 Set color based on point closest to the origin of the iterations.
23102 Default value is @var{mincol}.
23105 Set the bailout value. Default value is 10.0.
23108 Set the maximum of iterations performed by the rendering
23109 algorithm. Default value is 7189.
23112 Set outer coloring mode.
23113 It shall assume one of following values:
23115 @item iteration_count
23116 Set iteration count mode.
23117 @item normalized_iteration_count
23118 set normalized iteration count mode.
23120 Default value is @var{normalized_iteration_count}.
23123 Set frame rate, expressed as number of frames per second. Default
23127 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23128 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23131 Set the initial scale value. Default value is 3.0.
23134 Set the initial x position. Must be a floating point value between
23135 -100 and 100. Default value is -0.743643887037158704752191506114774.
23138 Set the initial y position. Must be a floating point value between
23139 -100 and 100. Default value is -0.131825904205311970493132056385139.
23144 Generate various test patterns, as generated by the MPlayer test filter.
23146 The size of the generated video is fixed, and is 256x256.
23147 This source is useful in particular for testing encoding features.
23149 This source accepts the following options:
23154 Specify the frame rate of the sourced video, as the number of frames
23155 generated per second. It has to be a string in the format
23156 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23157 number or a valid video frame rate abbreviation. The default value is
23161 Set the duration of the sourced video. See
23162 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23163 for the accepted syntax.
23165 If not specified, or the expressed duration is negative, the video is
23166 supposed to be generated forever.
23170 Set the number or the name of the test to perform. Supported tests are:
23184 @item max_frames, m
23185 Set the maximum number of frames generated for each test, default value is 30.
23189 Default value is "all", which will cycle through the list of all tests.
23194 mptestsrc=t=dc_luma
23197 will generate a "dc_luma" test pattern.
23199 @section frei0r_src
23201 Provide a frei0r source.
23203 To enable compilation of this filter you need to install the frei0r
23204 header and configure FFmpeg with @code{--enable-frei0r}.
23206 This source accepts the following parameters:
23211 The size of the video to generate. For the syntax of this option, check the
23212 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23215 The framerate of the generated video. It may be a string of the form
23216 @var{num}/@var{den} or a frame rate abbreviation.
23219 The name to the frei0r source to load. For more information regarding frei0r and
23220 how to set the parameters, read the @ref{frei0r} section in the video filters
23223 @item filter_params
23224 A '|'-separated list of parameters to pass to the frei0r source.
23228 For example, to generate a frei0r partik0l source with size 200x200
23229 and frame rate 10 which is overlaid on the overlay filter main input:
23231 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23236 Generate a life pattern.
23238 This source is based on a generalization of John Conway's life game.
23240 The sourced input represents a life grid, each pixel represents a cell
23241 which can be in one of two possible states, alive or dead. Every cell
23242 interacts with its eight neighbours, which are the cells that are
23243 horizontally, vertically, or diagonally adjacent.
23245 At each interaction the grid evolves according to the adopted rule,
23246 which specifies the number of neighbor alive cells which will make a
23247 cell stay alive or born. The @option{rule} option allows one to specify
23250 This source accepts the following options:
23254 Set the file from which to read the initial grid state. In the file,
23255 each non-whitespace character is considered an alive cell, and newline
23256 is used to delimit the end of each row.
23258 If this option is not specified, the initial grid is generated
23262 Set the video rate, that is the number of frames generated per second.
23265 @item random_fill_ratio, ratio
23266 Set the random fill ratio for the initial random grid. It is a
23267 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23268 It is ignored when a file is specified.
23270 @item random_seed, seed
23271 Set the seed for filling the initial random grid, must be an integer
23272 included between 0 and UINT32_MAX. If not specified, or if explicitly
23273 set to -1, the filter will try to use a good random seed on a best
23279 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23280 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23281 @var{NS} specifies the number of alive neighbor cells which make a
23282 live cell stay alive, and @var{NB} the number of alive neighbor cells
23283 which make a dead cell to become alive (i.e. to "born").
23284 "s" and "b" can be used in place of "S" and "B", respectively.
23286 Alternatively a rule can be specified by an 18-bits integer. The 9
23287 high order bits are used to encode the next cell state if it is alive
23288 for each number of neighbor alive cells, the low order bits specify
23289 the rule for "borning" new cells. Higher order bits encode for an
23290 higher number of neighbor cells.
23291 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23292 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23294 Default value is "S23/B3", which is the original Conway's game of life
23295 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23296 cells, and will born a new cell if there are three alive cells around
23300 Set the size of the output video. For the syntax of this option, check the
23301 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23303 If @option{filename} is specified, the size is set by default to the
23304 same size of the input file. If @option{size} is set, it must contain
23305 the size specified in the input file, and the initial grid defined in
23306 that file is centered in the larger resulting area.
23308 If a filename is not specified, the size value defaults to "320x240"
23309 (used for a randomly generated initial grid).
23312 If set to 1, stitch the left and right grid edges together, and the
23313 top and bottom edges also. Defaults to 1.
23316 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23317 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23318 value from 0 to 255.
23321 Set the color of living (or new born) cells.
23324 Set the color of dead cells. If @option{mold} is set, this is the first color
23325 used to represent a dead cell.
23328 Set mold color, for definitely dead and moldy cells.
23330 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23331 ffmpeg-utils manual,ffmpeg-utils}.
23334 @subsection Examples
23338 Read a grid from @file{pattern}, and center it on a grid of size
23341 life=f=pattern:s=300x300
23345 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23347 life=ratio=2/3:s=200x200
23351 Specify a custom rule for evolving a randomly generated grid:
23357 Full example with slow death effect (mold) using @command{ffplay}:
23359 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23366 @anchor{haldclutsrc}
23369 @anchor{pal100bars}
23370 @anchor{rgbtestsrc}
23372 @anchor{smptehdbars}
23375 @anchor{yuvtestsrc}
23376 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23378 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23380 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23382 The @code{color} source provides an uniformly colored input.
23384 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23385 @ref{haldclut} filter.
23387 The @code{nullsrc} source returns unprocessed video frames. It is
23388 mainly useful to be employed in analysis / debugging tools, or as the
23389 source for filters which ignore the input data.
23391 The @code{pal75bars} source generates a color bars pattern, based on
23392 EBU PAL recommendations with 75% color levels.
23394 The @code{pal100bars} source generates a color bars pattern, based on
23395 EBU PAL recommendations with 100% color levels.
23397 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23398 detecting RGB vs BGR issues. You should see a red, green and blue
23399 stripe from top to bottom.
23401 The @code{smptebars} source generates a color bars pattern, based on
23402 the SMPTE Engineering Guideline EG 1-1990.
23404 The @code{smptehdbars} source generates a color bars pattern, based on
23405 the SMPTE RP 219-2002.
23407 The @code{testsrc} source generates a test video pattern, showing a
23408 color pattern, a scrolling gradient and a timestamp. This is mainly
23409 intended for testing purposes.
23411 The @code{testsrc2} source is similar to testsrc, but supports more
23412 pixel formats instead of just @code{rgb24}. This allows using it as an
23413 input for other tests without requiring a format conversion.
23415 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23416 see a y, cb and cr stripe from top to bottom.
23418 The sources accept the following parameters:
23423 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23424 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23425 pixels to be used as identity matrix for 3D lookup tables. Each component is
23426 coded on a @code{1/(N*N)} scale.
23429 Specify the color of the source, only available in the @code{color}
23430 source. For the syntax of this option, check the
23431 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23434 Specify the size of the sourced video. For the syntax of this option, check the
23435 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23436 The default value is @code{320x240}.
23438 This option is not available with the @code{allrgb}, @code{allyuv}, and
23439 @code{haldclutsrc} filters.
23442 Specify the frame rate of the sourced video, as the number of frames
23443 generated per second. It has to be a string in the format
23444 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23445 number or a valid video frame rate abbreviation. The default value is
23449 Set the duration of the sourced video. See
23450 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23451 for the accepted syntax.
23453 If not specified, or the expressed duration is negative, the video is
23454 supposed to be generated forever.
23456 Since the frame rate is used as time base, all frames including the last one
23457 will have their full duration. If the specified duration is not a multiple
23458 of the frame duration, it will be rounded up.
23461 Set the sample aspect ratio of the sourced video.
23464 Specify the alpha (opacity) of the background, only available in the
23465 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23466 255 (fully opaque, the default).
23469 Set the number of decimals to show in the timestamp, only available in the
23470 @code{testsrc} source.
23472 The displayed timestamp value will correspond to the original
23473 timestamp value multiplied by the power of 10 of the specified
23474 value. Default value is 0.
23477 @subsection Examples
23481 Generate a video with a duration of 5.3 seconds, with size
23482 176x144 and a frame rate of 10 frames per second:
23484 testsrc=duration=5.3:size=qcif:rate=10
23488 The following graph description will generate a red source
23489 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23492 color=c=red@@0.2:s=qcif:r=10
23496 If the input content is to be ignored, @code{nullsrc} can be used. The
23497 following command generates noise in the luminance plane by employing
23498 the @code{geq} filter:
23500 nullsrc=s=256x256, geq=random(1)*255:128:128
23504 @subsection Commands
23506 The @code{color} source supports the following commands:
23510 Set the color of the created image. Accepts the same syntax of the
23511 corresponding @option{color} option.
23516 Generate video using an OpenCL program.
23521 OpenCL program source file.
23524 Kernel name in program.
23527 Size of frames to generate. This must be set.
23530 Pixel format to use for the generated frames. This must be set.
23533 Number of frames generated every second. Default value is '25'.
23537 For details of how the program loading works, see the @ref{program_opencl}
23544 Generate a colour ramp by setting pixel values from the position of the pixel
23545 in the output image. (Note that this will work with all pixel formats, but
23546 the generated output will not be the same.)
23548 __kernel void ramp(__write_only image2d_t dst,
23549 unsigned int index)
23551 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23554 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23556 write_imagef(dst, loc, val);
23561 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23563 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23564 unsigned int index)
23566 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23568 float4 value = 0.0f;
23569 int x = loc.x + index;
23570 int y = loc.y + index;
23571 while (x > 0 || y > 0) {
23572 if (x % 3 == 1 && y % 3 == 1) {
23580 write_imagef(dst, loc, value);
23586 @section sierpinski
23588 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23590 This source accepts the following options:
23594 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23595 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23598 Set frame rate, expressed as number of frames per second. Default
23602 Set seed which is used for random panning.
23605 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23608 Set fractal type, can be default @code{carpet} or @code{triangle}.
23611 @c man end VIDEO SOURCES
23613 @chapter Video Sinks
23614 @c man begin VIDEO SINKS
23616 Below is a description of the currently available video sinks.
23618 @section buffersink
23620 Buffer video frames, and make them available to the end of the filter
23623 This sink is mainly intended for programmatic use, in particular
23624 through the interface defined in @file{libavfilter/buffersink.h}
23625 or the options system.
23627 It accepts a pointer to an AVBufferSinkContext structure, which
23628 defines the incoming buffers' formats, to be passed as the opaque
23629 parameter to @code{avfilter_init_filter} for initialization.
23633 Null video sink: do absolutely nothing with the input video. It is
23634 mainly useful as a template and for use in analysis / debugging
23637 @c man end VIDEO SINKS
23639 @chapter Multimedia Filters
23640 @c man begin MULTIMEDIA FILTERS
23642 Below is a description of the currently available multimedia filters.
23646 Convert input audio to a video output, displaying the audio bit scope.
23648 The filter accepts the following options:
23652 Set frame rate, expressed as number of frames per second. Default
23656 Specify the video size for the output. For the syntax of this option, check the
23657 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23658 Default value is @code{1024x256}.
23661 Specify list of colors separated by space or by '|' which will be used to
23662 draw channels. Unrecognized or missing colors will be replaced
23666 @section adrawgraph
23667 Draw a graph using input audio metadata.
23669 See @ref{drawgraph}
23671 @section agraphmonitor
23673 See @ref{graphmonitor}.
23675 @section ahistogram
23677 Convert input audio to a video output, displaying the volume histogram.
23679 The filter accepts the following options:
23683 Specify how histogram is calculated.
23685 It accepts the following values:
23688 Use single histogram for all channels.
23690 Use separate histogram for each channel.
23692 Default is @code{single}.
23695 Set frame rate, expressed as number of frames per second. Default
23699 Specify the video size for the output. For the syntax of this option, check the
23700 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23701 Default value is @code{hd720}.
23706 It accepts the following values:
23717 reverse logarithmic
23719 Default is @code{log}.
23722 Set amplitude scale.
23724 It accepts the following values:
23731 Default is @code{log}.
23734 Set how much frames to accumulate in histogram.
23735 Default is 1. Setting this to -1 accumulates all frames.
23738 Set histogram ratio of window height.
23741 Set sonogram sliding.
23743 It accepts the following values:
23746 replace old rows with new ones.
23748 scroll from top to bottom.
23750 Default is @code{replace}.
23753 @section aphasemeter
23755 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23756 representing mean phase of current audio frame. A video output can also be produced and is
23757 enabled by default. The audio is passed through as first output.
23759 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23760 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23761 and @code{1} means channels are in phase.
23763 The filter accepts the following options, all related to its video output:
23767 Set the output frame rate. Default value is @code{25}.
23770 Set the video size for the output. For the syntax of this option, check the
23771 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23772 Default value is @code{800x400}.
23777 Specify the red, green, blue contrast. Default values are @code{2},
23778 @code{7} and @code{1}.
23779 Allowed range is @code{[0, 255]}.
23782 Set color which will be used for drawing median phase. If color is
23783 @code{none} which is default, no median phase value will be drawn.
23786 Enable video output. Default is enabled.
23789 @subsection phasing detection
23791 The filter also detects out of phase and mono sequences in stereo streams.
23792 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23794 The filter accepts the following options for this detection:
23798 Enable mono and out of phase detection. Default is disabled.
23801 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23802 Allowed range is @code{[0, 1]}.
23805 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23806 Allowed range is @code{[90, 180]}.
23809 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23812 @subsection Examples
23816 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23818 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23822 @section avectorscope
23824 Convert input audio to a video output, representing the audio vector
23827 The filter is used to measure the difference between channels of stereo
23828 audio stream. A monaural signal, consisting of identical left and right
23829 signal, results in straight vertical line. Any stereo separation is visible
23830 as a deviation from this line, creating a Lissajous figure.
23831 If the straight (or deviation from it) but horizontal line appears this
23832 indicates that the left and right channels are out of phase.
23834 The filter accepts the following options:
23838 Set the vectorscope mode.
23840 Available values are:
23843 Lissajous rotated by 45 degrees.
23846 Same as above but not rotated.
23849 Shape resembling half of circle.
23852 Default value is @samp{lissajous}.
23855 Set the video size for the output. For the syntax of this option, check the
23856 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23857 Default value is @code{400x400}.
23860 Set the output frame rate. Default value is @code{25}.
23866 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23867 @code{160}, @code{80} and @code{255}.
23868 Allowed range is @code{[0, 255]}.
23874 Specify the red, green, blue and alpha fade. Default values are @code{15},
23875 @code{10}, @code{5} and @code{5}.
23876 Allowed range is @code{[0, 255]}.
23879 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23880 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23883 Set the vectorscope drawing mode.
23885 Available values are:
23888 Draw dot for each sample.
23891 Draw line between previous and current sample.
23894 Default value is @samp{dot}.
23897 Specify amplitude scale of audio samples.
23899 Available values are:
23915 Swap left channel axis with right channel axis.
23925 Mirror only x axis.
23928 Mirror only y axis.
23936 @subsection Examples
23940 Complete example using @command{ffplay}:
23942 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23943 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
23947 @section bench, abench
23949 Benchmark part of a filtergraph.
23951 The filter accepts the following options:
23955 Start or stop a timer.
23957 Available values are:
23960 Get the current time, set it as frame metadata (using the key
23961 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
23964 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
23965 the input frame metadata to get the time difference. Time difference, average,
23966 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
23967 @code{min}) are then printed. The timestamps are expressed in seconds.
23971 @subsection Examples
23975 Benchmark @ref{selectivecolor} filter:
23977 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
23983 Concatenate audio and video streams, joining them together one after the
23986 The filter works on segments of synchronized video and audio streams. All
23987 segments must have the same number of streams of each type, and that will
23988 also be the number of streams at output.
23990 The filter accepts the following options:
23995 Set the number of segments. Default is 2.
23998 Set the number of output video streams, that is also the number of video
23999 streams in each segment. Default is 1.
24002 Set the number of output audio streams, that is also the number of audio
24003 streams in each segment. Default is 0.
24006 Activate unsafe mode: do not fail if segments have a different format.
24010 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24011 @var{a} audio outputs.
24013 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24014 segment, in the same order as the outputs, then the inputs for the second
24017 Related streams do not always have exactly the same duration, for various
24018 reasons including codec frame size or sloppy authoring. For that reason,
24019 related synchronized streams (e.g. a video and its audio track) should be
24020 concatenated at once. The concat filter will use the duration of the longest
24021 stream in each segment (except the last one), and if necessary pad shorter
24022 audio streams with silence.
24024 For this filter to work correctly, all segments must start at timestamp 0.
24026 All corresponding streams must have the same parameters in all segments; the
24027 filtering system will automatically select a common pixel format for video
24028 streams, and a common sample format, sample rate and channel layout for
24029 audio streams, but other settings, such as resolution, must be converted
24030 explicitly by the user.
24032 Different frame rates are acceptable but will result in variable frame rate
24033 at output; be sure to configure the output file to handle it.
24035 @subsection Examples
24039 Concatenate an opening, an episode and an ending, all in bilingual version
24040 (video in stream 0, audio in streams 1 and 2):
24042 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24043 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24044 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24045 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24049 Concatenate two parts, handling audio and video separately, using the
24050 (a)movie sources, and adjusting the resolution:
24052 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24053 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24054 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24056 Note that a desync will happen at the stitch if the audio and video streams
24057 do not have exactly the same duration in the first file.
24061 @subsection Commands
24063 This filter supports the following commands:
24066 Close the current segment and step to the next one
24072 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24073 level. By default, it logs a message at a frequency of 10Hz with the
24074 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24075 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24077 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24078 sample format is double-precision floating point. The input stream will be converted to
24079 this specification, if needed. Users may need to insert aformat and/or aresample filters
24080 after this filter to obtain the original parameters.
24082 The filter also has a video output (see the @var{video} option) with a real
24083 time graph to observe the loudness evolution. The graphic contains the logged
24084 message mentioned above, so it is not printed anymore when this option is set,
24085 unless the verbose logging is set. The main graphing area contains the
24086 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24087 the momentary loudness (400 milliseconds), but can optionally be configured
24088 to instead display short-term loudness (see @var{gauge}).
24090 The green area marks a +/- 1LU target range around the target loudness
24091 (-23LUFS by default, unless modified through @var{target}).
24093 More information about the Loudness Recommendation EBU R128 on
24094 @url{http://tech.ebu.ch/loudness}.
24096 The filter accepts the following options:
24101 Activate the video output. The audio stream is passed unchanged whether this
24102 option is set or no. The video stream will be the first output stream if
24103 activated. Default is @code{0}.
24106 Set the video size. This option is for video only. For the syntax of this
24108 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24109 Default and minimum resolution is @code{640x480}.
24112 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24113 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24114 other integer value between this range is allowed.
24117 Set metadata injection. If set to @code{1}, the audio input will be segmented
24118 into 100ms output frames, each of them containing various loudness information
24119 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24121 Default is @code{0}.
24124 Force the frame logging level.
24126 Available values are:
24129 information logging level
24131 verbose logging level
24134 By default, the logging level is set to @var{info}. If the @option{video} or
24135 the @option{metadata} options are set, it switches to @var{verbose}.
24140 Available modes can be cumulated (the option is a @code{flag} type). Possible
24144 Disable any peak mode (default).
24146 Enable sample-peak mode.
24148 Simple peak mode looking for the higher sample value. It logs a message
24149 for sample-peak (identified by @code{SPK}).
24151 Enable true-peak mode.
24153 If enabled, the peak lookup is done on an over-sampled version of the input
24154 stream for better peak accuracy. It logs a message for true-peak.
24155 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24156 This mode requires a build with @code{libswresample}.
24160 Treat mono input files as "dual mono". If a mono file is intended for playback
24161 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24162 If set to @code{true}, this option will compensate for this effect.
24163 Multi-channel input files are not affected by this option.
24166 Set a specific pan law to be used for the measurement of dual mono files.
24167 This parameter is optional, and has a default value of -3.01dB.
24170 Set a specific target level (in LUFS) used as relative zero in the visualization.
24171 This parameter is optional and has a default value of -23LUFS as specified
24172 by EBU R128. However, material published online may prefer a level of -16LUFS
24173 (e.g. for use with podcasts or video platforms).
24176 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24177 @code{shortterm}. By default the momentary value will be used, but in certain
24178 scenarios it may be more useful to observe the short term value instead (e.g.
24182 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24183 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24184 video output, not the summary or continuous log output.
24187 @subsection Examples
24191 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24193 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24197 Run an analysis with @command{ffmpeg}:
24199 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24203 @section interleave, ainterleave
24205 Temporally interleave frames from several inputs.
24207 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24209 These filters read frames from several inputs and send the oldest
24210 queued frame to the output.
24212 Input streams must have well defined, monotonically increasing frame
24215 In order to submit one frame to output, these filters need to enqueue
24216 at least one frame for each input, so they cannot work in case one
24217 input is not yet terminated and will not receive incoming frames.
24219 For example consider the case when one input is a @code{select} filter
24220 which always drops input frames. The @code{interleave} filter will keep
24221 reading from that input, but it will never be able to send new frames
24222 to output until the input sends an end-of-stream signal.
24224 Also, depending on inputs synchronization, the filters will drop
24225 frames in case one input receives more frames than the other ones, and
24226 the queue is already filled.
24228 These filters accept the following options:
24232 Set the number of different inputs, it is 2 by default.
24235 How to determine the end-of-stream.
24239 The duration of the longest input. (default)
24242 The duration of the shortest input.
24245 The duration of the first input.
24250 @subsection Examples
24254 Interleave frames belonging to different streams using @command{ffmpeg}:
24256 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24260 Add flickering blur effect:
24262 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24266 @section metadata, ametadata
24268 Manipulate frame metadata.
24270 This filter accepts the following options:
24274 Set mode of operation of the filter.
24276 Can be one of the following:
24280 If both @code{value} and @code{key} is set, select frames
24281 which have such metadata. If only @code{key} is set, select
24282 every frame that has such key in metadata.
24285 Add new metadata @code{key} and @code{value}. If key is already available
24289 Modify value of already present key.
24292 If @code{value} is set, delete only keys that have such value.
24293 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24297 Print key and its value if metadata was found. If @code{key} is not set print all
24298 metadata values available in frame.
24302 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24305 Set metadata value which will be used. This option is mandatory for
24306 @code{modify} and @code{add} mode.
24309 Which function to use when comparing metadata value and @code{value}.
24311 Can be one of following:
24315 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24318 Values are interpreted as strings, returns true if metadata value starts with
24319 the @code{value} option string.
24322 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24325 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24328 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24331 Values are interpreted as floats, returns true if expression from option @code{expr}
24335 Values are interpreted as strings, returns true if metadata value ends with
24336 the @code{value} option string.
24340 Set expression which is used when @code{function} is set to @code{expr}.
24341 The expression is evaluated through the eval API and can contain the following
24346 Float representation of @code{value} from metadata key.
24349 Float representation of @code{value} as supplied by user in @code{value} option.
24353 If specified in @code{print} mode, output is written to the named file. Instead of
24354 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24355 for standard output. If @code{file} option is not set, output is written to the log
24356 with AV_LOG_INFO loglevel.
24359 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24363 @subsection Examples
24367 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24370 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24373 Print silencedetect output to file @file{metadata.txt}.
24375 silencedetect,ametadata=mode=print:file=metadata.txt
24378 Direct all metadata to a pipe with file descriptor 4.
24380 metadata=mode=print:file='pipe\:4'
24384 @section perms, aperms
24386 Set read/write permissions for the output frames.
24388 These filters are mainly aimed at developers to test direct path in the
24389 following filter in the filtergraph.
24391 The filters accept the following options:
24395 Select the permissions mode.
24397 It accepts the following values:
24400 Do nothing. This is the default.
24402 Set all the output frames read-only.
24404 Set all the output frames directly writable.
24406 Make the frame read-only if writable, and writable if read-only.
24408 Set each output frame read-only or writable randomly.
24412 Set the seed for the @var{random} mode, must be an integer included between
24413 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24414 @code{-1}, the filter will try to use a good random seed on a best effort
24418 Note: in case of auto-inserted filter between the permission filter and the
24419 following one, the permission might not be received as expected in that
24420 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24421 perms/aperms filter can avoid this problem.
24423 @section realtime, arealtime
24425 Slow down filtering to match real time approximately.
24427 These filters will pause the filtering for a variable amount of time to
24428 match the output rate with the input timestamps.
24429 They are similar to the @option{re} option to @code{ffmpeg}.
24431 They accept the following options:
24435 Time limit for the pauses. Any pause longer than that will be considered
24436 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24438 Speed factor for processing. The value must be a float larger than zero.
24439 Values larger than 1.0 will result in faster than realtime processing,
24440 smaller will slow processing down. The @var{limit} is automatically adapted
24441 accordingly. Default is 1.0.
24443 A processing speed faster than what is possible without these filters cannot
24448 @section select, aselect
24450 Select frames to pass in output.
24452 This filter accepts the following options:
24457 Set expression, which is evaluated for each input frame.
24459 If the expression is evaluated to zero, the frame is discarded.
24461 If the evaluation result is negative or NaN, the frame is sent to the
24462 first output; otherwise it is sent to the output with index
24463 @code{ceil(val)-1}, assuming that the input index starts from 0.
24465 For example a value of @code{1.2} corresponds to the output with index
24466 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24469 Set the number of outputs. The output to which to send the selected
24470 frame is based on the result of the evaluation. Default value is 1.
24473 The expression can contain the following constants:
24477 The (sequential) number of the filtered frame, starting from 0.
24480 The (sequential) number of the selected frame, starting from 0.
24482 @item prev_selected_n
24483 The sequential number of the last selected frame. It's NAN if undefined.
24486 The timebase of the input timestamps.
24489 The PTS (Presentation TimeStamp) of the filtered video frame,
24490 expressed in @var{TB} units. It's NAN if undefined.
24493 The PTS of the filtered video frame,
24494 expressed in seconds. It's NAN if undefined.
24497 The PTS of the previously filtered video frame. It's NAN if undefined.
24499 @item prev_selected_pts
24500 The PTS of the last previously filtered video frame. It's NAN if undefined.
24502 @item prev_selected_t
24503 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24506 The PTS of the first video frame in the video. It's NAN if undefined.
24509 The time of the first video frame in the video. It's NAN if undefined.
24511 @item pict_type @emph{(video only)}
24512 The type of the filtered frame. It can assume one of the following
24524 @item interlace_type @emph{(video only)}
24525 The frame interlace type. It can assume one of the following values:
24528 The frame is progressive (not interlaced).
24530 The frame is top-field-first.
24532 The frame is bottom-field-first.
24535 @item consumed_sample_n @emph{(audio only)}
24536 the number of selected samples before the current frame
24538 @item samples_n @emph{(audio only)}
24539 the number of samples in the current frame
24541 @item sample_rate @emph{(audio only)}
24542 the input sample rate
24545 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24548 the position in the file of the filtered frame, -1 if the information
24549 is not available (e.g. for synthetic video)
24551 @item scene @emph{(video only)}
24552 value between 0 and 1 to indicate a new scene; a low value reflects a low
24553 probability for the current frame to introduce a new scene, while a higher
24554 value means the current frame is more likely to be one (see the example below)
24556 @item concatdec_select
24557 The concat demuxer can select only part of a concat input file by setting an
24558 inpoint and an outpoint, but the output packets may not be entirely contained
24559 in the selected interval. By using this variable, it is possible to skip frames
24560 generated by the concat demuxer which are not exactly contained in the selected
24563 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24564 and the @var{lavf.concat.duration} packet metadata values which are also
24565 present in the decoded frames.
24567 The @var{concatdec_select} variable is -1 if the frame pts is at least
24568 start_time and either the duration metadata is missing or the frame pts is less
24569 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24572 That basically means that an input frame is selected if its pts is within the
24573 interval set by the concat demuxer.
24577 The default value of the select expression is "1".
24579 @subsection Examples
24583 Select all frames in input:
24588 The example above is the same as:
24600 Select only I-frames:
24602 select='eq(pict_type\,I)'
24606 Select one frame every 100:
24608 select='not(mod(n\,100))'
24612 Select only frames contained in the 10-20 time interval:
24614 select=between(t\,10\,20)
24618 Select only I-frames contained in the 10-20 time interval:
24620 select=between(t\,10\,20)*eq(pict_type\,I)
24624 Select frames with a minimum distance of 10 seconds:
24626 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24630 Use aselect to select only audio frames with samples number > 100:
24632 aselect='gt(samples_n\,100)'
24636 Create a mosaic of the first scenes:
24638 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24641 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24645 Send even and odd frames to separate outputs, and compose them:
24647 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24651 Select useful frames from an ffconcat file which is using inpoints and
24652 outpoints but where the source files are not intra frame only.
24654 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24658 @section sendcmd, asendcmd
24660 Send commands to filters in the filtergraph.
24662 These filters read commands to be sent to other filters in the
24665 @code{sendcmd} must be inserted between two video filters,
24666 @code{asendcmd} must be inserted between two audio filters, but apart
24667 from that they act the same way.
24669 The specification of commands can be provided in the filter arguments
24670 with the @var{commands} option, or in a file specified by the
24671 @var{filename} option.
24673 These filters accept the following options:
24676 Set the commands to be read and sent to the other filters.
24678 Set the filename of the commands to be read and sent to the other
24682 @subsection Commands syntax
24684 A commands description consists of a sequence of interval
24685 specifications, comprising a list of commands to be executed when a
24686 particular event related to that interval occurs. The occurring event
24687 is typically the current frame time entering or leaving a given time
24690 An interval is specified by the following syntax:
24692 @var{START}[-@var{END}] @var{COMMANDS};
24695 The time interval is specified by the @var{START} and @var{END} times.
24696 @var{END} is optional and defaults to the maximum time.
24698 The current frame time is considered within the specified interval if
24699 it is included in the interval [@var{START}, @var{END}), that is when
24700 the time is greater or equal to @var{START} and is lesser than
24703 @var{COMMANDS} consists of a sequence of one or more command
24704 specifications, separated by ",", relating to that interval. The
24705 syntax of a command specification is given by:
24707 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24710 @var{FLAGS} is optional and specifies the type of events relating to
24711 the time interval which enable sending the specified command, and must
24712 be a non-null sequence of identifier flags separated by "+" or "|" and
24713 enclosed between "[" and "]".
24715 The following flags are recognized:
24718 The command is sent when the current frame timestamp enters the
24719 specified interval. In other words, the command is sent when the
24720 previous frame timestamp was not in the given interval, and the
24724 The command is sent when the current frame timestamp leaves the
24725 specified interval. In other words, the command is sent when the
24726 previous frame timestamp was in the given interval, and the
24730 The command @var{ARG} is interpreted as expression and result of
24731 expression is passed as @var{ARG}.
24733 The expression is evaluated through the eval API and can contain the following
24738 Original position in the file of the frame, or undefined if undefined
24739 for the current frame.
24742 The presentation timestamp in input.
24745 The count of the input frame for video or audio, starting from 0.
24748 The time in seconds of the current frame.
24751 The start time in seconds of the current command interval.
24754 The end time in seconds of the current command interval.
24757 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24762 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24765 @var{TARGET} specifies the target of the command, usually the name of
24766 the filter class or a specific filter instance name.
24768 @var{COMMAND} specifies the name of the command for the target filter.
24770 @var{ARG} is optional and specifies the optional list of argument for
24771 the given @var{COMMAND}.
24773 Between one interval specification and another, whitespaces, or
24774 sequences of characters starting with @code{#} until the end of line,
24775 are ignored and can be used to annotate comments.
24777 A simplified BNF description of the commands specification syntax
24780 @var{COMMAND_FLAG} ::= "enter" | "leave"
24781 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24782 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24783 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24784 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24785 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24788 @subsection Examples
24792 Specify audio tempo change at second 4:
24794 asendcmd=c='4.0 atempo tempo 1.5',atempo
24798 Target a specific filter instance:
24800 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24804 Specify a list of drawtext and hue commands in a file.
24806 # show text in the interval 5-10
24807 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24808 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24810 # desaturate the image in the interval 15-20
24811 15.0-20.0 [enter] hue s 0,
24812 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24814 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24816 # apply an exponential saturation fade-out effect, starting from time 25
24817 25 [enter] hue s exp(25-t)
24820 A filtergraph allowing to read and process the above command list
24821 stored in a file @file{test.cmd}, can be specified with:
24823 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24828 @section setpts, asetpts
24830 Change the PTS (presentation timestamp) of the input frames.
24832 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24834 This filter accepts the following options:
24839 The expression which is evaluated for each frame to construct its timestamp.
24843 The expression is evaluated through the eval API and can contain the following
24847 @item FRAME_RATE, FR
24848 frame rate, only defined for constant frame-rate video
24851 The presentation timestamp in input
24854 The count of the input frame for video or the number of consumed samples,
24855 not including the current frame for audio, starting from 0.
24857 @item NB_CONSUMED_SAMPLES
24858 The number of consumed samples, not including the current frame (only
24861 @item NB_SAMPLES, S
24862 The number of samples in the current frame (only audio)
24864 @item SAMPLE_RATE, SR
24865 The audio sample rate.
24868 The PTS of the first frame.
24871 the time in seconds of the first frame
24874 State whether the current frame is interlaced.
24877 the time in seconds of the current frame
24880 original position in the file of the frame, or undefined if undefined
24881 for the current frame
24884 The previous input PTS.
24887 previous input time in seconds
24890 The previous output PTS.
24893 previous output time in seconds
24896 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
24900 The wallclock (RTC) time at the start of the movie in microseconds.
24903 The timebase of the input timestamps.
24907 @subsection Examples
24911 Start counting PTS from zero
24913 setpts=PTS-STARTPTS
24917 Apply fast motion effect:
24923 Apply slow motion effect:
24929 Set fixed rate of 25 frames per second:
24935 Set fixed rate 25 fps with some jitter:
24937 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
24941 Apply an offset of 10 seconds to the input PTS:
24947 Generate timestamps from a "live source" and rebase onto the current timebase:
24949 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
24953 Generate timestamps by counting samples:
24962 Force color range for the output video frame.
24964 The @code{setrange} filter marks the color range property for the
24965 output frames. It does not change the input frame, but only sets the
24966 corresponding property, which affects how the frame is treated by
24969 The filter accepts the following options:
24974 Available values are:
24978 Keep the same color range property.
24980 @item unspecified, unknown
24981 Set the color range as unspecified.
24983 @item limited, tv, mpeg
24984 Set the color range as limited.
24986 @item full, pc, jpeg
24987 Set the color range as full.
24991 @section settb, asettb
24993 Set the timebase to use for the output frames timestamps.
24994 It is mainly useful for testing timebase configuration.
24996 It accepts the following parameters:
25001 The expression which is evaluated into the output timebase.
25005 The value for @option{tb} is an arithmetic expression representing a
25006 rational. The expression can contain the constants "AVTB" (the default
25007 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25008 audio only). Default value is "intb".
25010 @subsection Examples
25014 Set the timebase to 1/25:
25020 Set the timebase to 1/10:
25026 Set the timebase to 1001/1000:
25032 Set the timebase to 2*intb:
25038 Set the default timebase value:
25045 Convert input audio to a video output representing frequency spectrum
25046 logarithmically using Brown-Puckette constant Q transform algorithm with
25047 direct frequency domain coefficient calculation (but the transform itself
25048 is not really constant Q, instead the Q factor is actually variable/clamped),
25049 with musical tone scale, from E0 to D#10.
25051 The filter accepts the following options:
25055 Specify the video size for the output. It must be even. For the syntax of this option,
25056 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25057 Default value is @code{1920x1080}.
25060 Set the output frame rate. Default value is @code{25}.
25063 Set the bargraph height. It must be even. Default value is @code{-1} which
25064 computes the bargraph height automatically.
25067 Set the axis height. It must be even. Default value is @code{-1} which computes
25068 the axis height automatically.
25071 Set the sonogram height. It must be even. Default value is @code{-1} which
25072 computes the sonogram height automatically.
25075 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25076 instead. Default value is @code{1}.
25078 @item sono_v, volume
25079 Specify the sonogram volume expression. It can contain variables:
25082 the @var{bar_v} evaluated expression
25083 @item frequency, freq, f
25084 the frequency where it is evaluated
25085 @item timeclamp, tc
25086 the value of @var{timeclamp} option
25090 @item a_weighting(f)
25091 A-weighting of equal loudness
25092 @item b_weighting(f)
25093 B-weighting of equal loudness
25094 @item c_weighting(f)
25095 C-weighting of equal loudness.
25097 Default value is @code{16}.
25099 @item bar_v, volume2
25100 Specify the bargraph volume expression. It can contain variables:
25103 the @var{sono_v} evaluated expression
25104 @item frequency, freq, f
25105 the frequency where it is evaluated
25106 @item timeclamp, tc
25107 the value of @var{timeclamp} option
25111 @item a_weighting(f)
25112 A-weighting of equal loudness
25113 @item b_weighting(f)
25114 B-weighting of equal loudness
25115 @item c_weighting(f)
25116 C-weighting of equal loudness.
25118 Default value is @code{sono_v}.
25120 @item sono_g, gamma
25121 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25122 higher gamma makes the spectrum having more range. Default value is @code{3}.
25123 Acceptable range is @code{[1, 7]}.
25125 @item bar_g, gamma2
25126 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25130 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25131 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25133 @item timeclamp, tc
25134 Specify the transform timeclamp. At low frequency, there is trade-off between
25135 accuracy in time domain and frequency domain. If timeclamp is lower,
25136 event in time domain is represented more accurately (such as fast bass drum),
25137 otherwise event in frequency domain is represented more accurately
25138 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25141 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25142 limits future samples by applying asymmetric windowing in time domain, useful
25143 when low latency is required. Accepted range is @code{[0, 1]}.
25146 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25147 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25150 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25151 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25154 This option is deprecated and ignored.
25157 Specify the transform length in time domain. Use this option to control accuracy
25158 trade-off between time domain and frequency domain at every frequency sample.
25159 It can contain variables:
25161 @item frequency, freq, f
25162 the frequency where it is evaluated
25163 @item timeclamp, tc
25164 the value of @var{timeclamp} option.
25166 Default value is @code{384*tc/(384+tc*f)}.
25169 Specify the transform count for every video frame. Default value is @code{6}.
25170 Acceptable range is @code{[1, 30]}.
25173 Specify the transform count for every single pixel. Default value is @code{0},
25174 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25177 Specify font file for use with freetype to draw the axis. If not specified,
25178 use embedded font. Note that drawing with font file or embedded font is not
25179 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25183 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25184 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25188 Specify font color expression. This is arithmetic expression that should return
25189 integer value 0xRRGGBB. It can contain variables:
25191 @item frequency, freq, f
25192 the frequency where it is evaluated
25193 @item timeclamp, tc
25194 the value of @var{timeclamp} option
25199 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25200 @item r(x), g(x), b(x)
25201 red, green, and blue value of intensity x.
25203 Default value is @code{st(0, (midi(f)-59.5)/12);
25204 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25205 r(1-ld(1)) + b(ld(1))}.
25208 Specify image file to draw the axis. This option override @var{fontfile} and
25209 @var{fontcolor} option.
25212 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25213 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25214 Default value is @code{1}.
25217 Set colorspace. The accepted values are:
25220 Unspecified (default)
25229 BT.470BG or BT.601-6 625
25232 SMPTE-170M or BT.601-6 525
25238 BT.2020 with non-constant luminance
25243 Set spectrogram color scheme. This is list of floating point values with format
25244 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25245 The default is @code{1|0.5|0|0|0.5|1}.
25249 @subsection Examples
25253 Playing audio while showing the spectrum:
25255 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25259 Same as above, but with frame rate 30 fps:
25261 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25265 Playing at 1280x720:
25267 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25271 Disable sonogram display:
25277 A1 and its harmonics: A1, A2, (near)E3, A3:
25279 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),
25280 asplit[a][out1]; [a] showcqt [out0]'
25284 Same as above, but with more accuracy in frequency domain:
25286 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),
25287 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25293 bar_v=10:sono_v=bar_v*a_weighting(f)
25297 Custom gamma, now spectrum is linear to the amplitude.
25303 Custom tlength equation:
25305 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)))'
25309 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25311 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25315 Custom font using fontconfig:
25317 font='Courier New,Monospace,mono|bold'
25321 Custom frequency range with custom axis using image file:
25323 axisfile=myaxis.png:basefreq=40:endfreq=10000
25329 Convert input audio to video output representing the audio power spectrum.
25330 Audio amplitude is on Y-axis while frequency is on X-axis.
25332 The filter accepts the following options:
25336 Specify size of video. For the syntax of this option, check the
25337 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25338 Default is @code{1024x512}.
25342 This set how each frequency bin will be represented.
25344 It accepts the following values:
25350 Default is @code{bar}.
25353 Set amplitude scale.
25355 It accepts the following values:
25369 Default is @code{log}.
25372 Set frequency scale.
25374 It accepts the following values:
25383 Reverse logarithmic scale.
25385 Default is @code{lin}.
25388 Set window size. Allowed range is from 16 to 65536.
25390 Default is @code{2048}
25393 Set windowing function.
25395 It accepts the following values:
25418 Default is @code{hanning}.
25421 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25422 which means optimal overlap for selected window function will be picked.
25425 Set time averaging. Setting this to 0 will display current maximal peaks.
25426 Default is @code{1}, which means time averaging is disabled.
25429 Specify list of colors separated by space or by '|' which will be used to
25430 draw channel frequencies. Unrecognized or missing colors will be replaced
25434 Set channel display mode.
25436 It accepts the following values:
25441 Default is @code{combined}.
25444 Set minimum amplitude used in @code{log} amplitude scaler.
25447 Set data display mode.
25449 It accepts the following values:
25455 Default is @code{magnitude}.
25458 @section showspatial
25460 Convert stereo input audio to a video output, representing the spatial relationship
25461 between two channels.
25463 The filter accepts the following options:
25467 Specify the video size for the output. For the syntax of this option, check the
25468 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25469 Default value is @code{512x512}.
25472 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25475 Set window function.
25477 It accepts the following values:
25502 Default value is @code{hann}.
25505 Set ratio of overlap window. Default value is @code{0.5}.
25506 When value is @code{1} overlap is set to recommended size for specific
25507 window function currently used.
25510 @anchor{showspectrum}
25511 @section showspectrum
25513 Convert input audio to a video output, representing the audio frequency
25516 The filter accepts the following options:
25520 Specify the video size for the output. For the syntax of this option, check the
25521 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25522 Default value is @code{640x512}.
25525 Specify how the spectrum should slide along the window.
25527 It accepts the following values:
25530 the samples start again on the left when they reach the right
25532 the samples scroll from right to left
25534 frames are only produced when the samples reach the right
25536 the samples scroll from left to right
25539 Default value is @code{replace}.
25542 Specify display mode.
25544 It accepts the following values:
25547 all channels are displayed in the same row
25549 all channels are displayed in separate rows
25552 Default value is @samp{combined}.
25555 Specify display color mode.
25557 It accepts the following values:
25560 each channel is displayed in a separate color
25562 each channel is displayed using the same color scheme
25564 each channel is displayed using the rainbow color scheme
25566 each channel is displayed using the moreland color scheme
25568 each channel is displayed using the nebulae color scheme
25570 each channel is displayed using the fire color scheme
25572 each channel is displayed using the fiery color scheme
25574 each channel is displayed using the fruit color scheme
25576 each channel is displayed using the cool color scheme
25578 each channel is displayed using the magma color scheme
25580 each channel is displayed using the green color scheme
25582 each channel is displayed using the viridis color scheme
25584 each channel is displayed using the plasma color scheme
25586 each channel is displayed using the cividis color scheme
25588 each channel is displayed using the terrain color scheme
25591 Default value is @samp{channel}.
25594 Specify scale used for calculating intensity color values.
25596 It accepts the following values:
25601 square root, default
25612 Default value is @samp{sqrt}.
25615 Specify frequency scale.
25617 It accepts the following values:
25625 Default value is @samp{lin}.
25628 Set saturation modifier for displayed colors. Negative values provide
25629 alternative color scheme. @code{0} is no saturation at all.
25630 Saturation must be in [-10.0, 10.0] range.
25631 Default value is @code{1}.
25634 Set window function.
25636 It accepts the following values:
25661 Default value is @code{hann}.
25664 Set orientation of time vs frequency axis. Can be @code{vertical} or
25665 @code{horizontal}. Default is @code{vertical}.
25668 Set ratio of overlap window. Default value is @code{0}.
25669 When value is @code{1} overlap is set to recommended size for specific
25670 window function currently used.
25673 Set scale gain for calculating intensity color values.
25674 Default value is @code{1}.
25677 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25680 Set color rotation, must be in [-1.0, 1.0] range.
25681 Default value is @code{0}.
25684 Set start frequency from which to display spectrogram. Default is @code{0}.
25687 Set stop frequency to which to display spectrogram. Default is @code{0}.
25690 Set upper frame rate limit. Default is @code{auto}, unlimited.
25693 Draw time and frequency axes and legends. Default is disabled.
25696 The usage is very similar to the showwaves filter; see the examples in that
25699 @subsection Examples
25703 Large window with logarithmic color scaling:
25705 showspectrum=s=1280x480:scale=log
25709 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25711 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25712 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25716 @section showspectrumpic
25718 Convert input audio to a single video frame, representing the audio frequency
25721 The filter accepts the following options:
25725 Specify the video size for the output. For the syntax of this option, check the
25726 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25727 Default value is @code{4096x2048}.
25730 Specify display mode.
25732 It accepts the following values:
25735 all channels are displayed in the same row
25737 all channels are displayed in separate rows
25739 Default value is @samp{combined}.
25742 Specify display color mode.
25744 It accepts the following values:
25747 each channel is displayed in a separate color
25749 each channel is displayed using the same color scheme
25751 each channel is displayed using the rainbow color scheme
25753 each channel is displayed using the moreland color scheme
25755 each channel is displayed using the nebulae color scheme
25757 each channel is displayed using the fire color scheme
25759 each channel is displayed using the fiery color scheme
25761 each channel is displayed using the fruit color scheme
25763 each channel is displayed using the cool color scheme
25765 each channel is displayed using the magma color scheme
25767 each channel is displayed using the green color scheme
25769 each channel is displayed using the viridis color scheme
25771 each channel is displayed using the plasma color scheme
25773 each channel is displayed using the cividis color scheme
25775 each channel is displayed using the terrain color scheme
25777 Default value is @samp{intensity}.
25780 Specify scale used for calculating intensity color values.
25782 It accepts the following values:
25787 square root, default
25797 Default value is @samp{log}.
25800 Specify frequency scale.
25802 It accepts the following values:
25810 Default value is @samp{lin}.
25813 Set saturation modifier for displayed colors. Negative values provide
25814 alternative color scheme. @code{0} is no saturation at all.
25815 Saturation must be in [-10.0, 10.0] range.
25816 Default value is @code{1}.
25819 Set window function.
25821 It accepts the following values:
25845 Default value is @code{hann}.
25848 Set orientation of time vs frequency axis. Can be @code{vertical} or
25849 @code{horizontal}. Default is @code{vertical}.
25852 Set scale gain for calculating intensity color values.
25853 Default value is @code{1}.
25856 Draw time and frequency axes and legends. Default is enabled.
25859 Set color rotation, must be in [-1.0, 1.0] range.
25860 Default value is @code{0}.
25863 Set start frequency from which to display spectrogram. Default is @code{0}.
25866 Set stop frequency to which to display spectrogram. Default is @code{0}.
25869 @subsection Examples
25873 Extract an audio spectrogram of a whole audio track
25874 in a 1024x1024 picture using @command{ffmpeg}:
25876 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25880 @section showvolume
25882 Convert input audio volume to a video output.
25884 The filter accepts the following options:
25891 Set border width, allowed range is [0, 5]. Default is 1.
25894 Set channel width, allowed range is [80, 8192]. Default is 400.
25897 Set channel height, allowed range is [1, 900]. Default is 20.
25900 Set fade, allowed range is [0, 1]. Default is 0.95.
25903 Set volume color expression.
25905 The expression can use the following variables:
25909 Current max volume of channel in dB.
25915 Current channel number, starting from 0.
25919 If set, displays channel names. Default is enabled.
25922 If set, displays volume values. Default is enabled.
25925 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
25926 default is @code{h}.
25929 Set step size, allowed range is [0, 5]. Default is 0, which means
25933 Set background opacity, allowed range is [0, 1]. Default is 0.
25936 Set metering mode, can be peak: @code{p} or rms: @code{r},
25937 default is @code{p}.
25940 Set display scale, can be linear: @code{lin} or log: @code{log},
25941 default is @code{lin}.
25945 If set to > 0., display a line for the max level
25946 in the previous seconds.
25947 default is disabled: @code{0.}
25950 The color of the max line. Use when @code{dm} option is set to > 0.
25951 default is: @code{orange}
25956 Convert input audio to a video output, representing the samples waves.
25958 The filter accepts the following options:
25962 Specify the video size for the output. For the syntax of this option, check the
25963 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25964 Default value is @code{600x240}.
25969 Available values are:
25972 Draw a point for each sample.
25975 Draw a vertical line for each sample.
25978 Draw a point for each sample and a line between them.
25981 Draw a centered vertical line for each sample.
25984 Default value is @code{point}.
25987 Set the number of samples which are printed on the same column. A
25988 larger value will decrease the frame rate. Must be a positive
25989 integer. This option can be set only if the value for @var{rate}
25990 is not explicitly specified.
25993 Set the (approximate) output frame rate. This is done by setting the
25994 option @var{n}. Default value is "25".
25996 @item split_channels
25997 Set if channels should be drawn separately or overlap. Default value is 0.
26000 Set colors separated by '|' which are going to be used for drawing of each channel.
26003 Set amplitude scale.
26005 Available values are:
26023 Set the draw mode. This is mostly useful to set for high @var{n}.
26025 Available values are:
26028 Scale pixel values for each drawn sample.
26031 Draw every sample directly.
26034 Default value is @code{scale}.
26037 @subsection Examples
26041 Output the input file audio and the corresponding video representation
26044 amovie=a.mp3,asplit[out0],showwaves[out1]
26048 Create a synthetic signal and show it with showwaves, forcing a
26049 frame rate of 30 frames per second:
26051 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26055 @section showwavespic
26057 Convert input audio to a single video frame, representing the samples waves.
26059 The filter accepts the following options:
26063 Specify the video size for the output. For the syntax of this option, check the
26064 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26065 Default value is @code{600x240}.
26067 @item split_channels
26068 Set if channels should be drawn separately or overlap. Default value is 0.
26071 Set colors separated by '|' which are going to be used for drawing of each channel.
26074 Set amplitude scale.
26076 Available values are:
26096 Available values are:
26099 Scale pixel values for each drawn sample.
26102 Draw every sample directly.
26105 Default value is @code{scale}.
26108 Set the filter mode.
26110 Available values are:
26113 Use average samples values for each drawn sample.
26116 Use peak samples values for each drawn sample.
26119 Default value is @code{average}.
26122 @subsection Examples
26126 Extract a channel split representation of the wave form of a whole audio track
26127 in a 1024x800 picture using @command{ffmpeg}:
26129 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26133 @section sidedata, asidedata
26135 Delete frame side data, or select frames based on it.
26137 This filter accepts the following options:
26141 Set mode of operation of the filter.
26143 Can be one of the following:
26147 Select every frame with side data of @code{type}.
26150 Delete side data of @code{type}. If @code{type} is not set, delete all side
26156 Set side data type used with all modes. Must be set for @code{select} mode. For
26157 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26158 in @file{libavutil/frame.h}. For example, to choose
26159 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26163 @section spectrumsynth
26165 Synthesize audio from 2 input video spectrums, first input stream represents
26166 magnitude across time and second represents phase across time.
26167 The filter will transform from frequency domain as displayed in videos back
26168 to time domain as presented in audio output.
26170 This filter is primarily created for reversing processed @ref{showspectrum}
26171 filter outputs, but can synthesize sound from other spectrograms too.
26172 But in such case results are going to be poor if the phase data is not
26173 available, because in such cases phase data need to be recreated, usually
26174 it's just recreated from random noise.
26175 For best results use gray only output (@code{channel} color mode in
26176 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26177 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26178 @code{data} option. Inputs videos should generally use @code{fullframe}
26179 slide mode as that saves resources needed for decoding video.
26181 The filter accepts the following options:
26185 Specify sample rate of output audio, the sample rate of audio from which
26186 spectrum was generated may differ.
26189 Set number of channels represented in input video spectrums.
26192 Set scale which was used when generating magnitude input spectrum.
26193 Can be @code{lin} or @code{log}. Default is @code{log}.
26196 Set slide which was used when generating inputs spectrums.
26197 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26198 Default is @code{fullframe}.
26201 Set window function used for resynthesis.
26204 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26205 which means optimal overlap for selected window function will be picked.
26208 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26209 Default is @code{vertical}.
26212 @subsection Examples
26216 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26217 then resynthesize videos back to audio with spectrumsynth:
26219 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
26220 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
26221 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26225 @section split, asplit
26227 Split input into several identical outputs.
26229 @code{asplit} works with audio input, @code{split} with video.
26231 The filter accepts a single parameter which specifies the number of outputs. If
26232 unspecified, it defaults to 2.
26234 @subsection Examples
26238 Create two separate outputs from the same input:
26240 [in] split [out0][out1]
26244 To create 3 or more outputs, you need to specify the number of
26247 [in] asplit=3 [out0][out1][out2]
26251 Create two separate outputs from the same input, one cropped and
26254 [in] split [splitout1][splitout2];
26255 [splitout1] crop=100:100:0:0 [cropout];
26256 [splitout2] pad=200:200:100:100 [padout];
26260 Create 5 copies of the input audio with @command{ffmpeg}:
26262 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26268 Receive commands sent through a libzmq client, and forward them to
26269 filters in the filtergraph.
26271 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26272 must be inserted between two video filters, @code{azmq} between two
26273 audio filters. Both are capable to send messages to any filter type.
26275 To enable these filters you need to install the libzmq library and
26276 headers and configure FFmpeg with @code{--enable-libzmq}.
26278 For more information about libzmq see:
26279 @url{http://www.zeromq.org/}
26281 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26282 receives messages sent through a network interface defined by the
26283 @option{bind_address} (or the abbreviation "@option{b}") option.
26284 Default value of this option is @file{tcp://localhost:5555}. You may
26285 want to alter this value to your needs, but do not forget to escape any
26286 ':' signs (see @ref{filtergraph escaping}).
26288 The received message must be in the form:
26290 @var{TARGET} @var{COMMAND} [@var{ARG}]
26293 @var{TARGET} specifies the target of the command, usually the name of
26294 the filter class or a specific filter instance name. The default
26295 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26296 but you can override this by using the @samp{filter_name@@id} syntax
26297 (see @ref{Filtergraph syntax}).
26299 @var{COMMAND} specifies the name of the command for the target filter.
26301 @var{ARG} is optional and specifies the optional argument list for the
26302 given @var{COMMAND}.
26304 Upon reception, the message is processed and the corresponding command
26305 is injected into the filtergraph. Depending on the result, the filter
26306 will send a reply to the client, adopting the format:
26308 @var{ERROR_CODE} @var{ERROR_REASON}
26312 @var{MESSAGE} is optional.
26314 @subsection Examples
26316 Look at @file{tools/zmqsend} for an example of a zmq client which can
26317 be used to send commands processed by these filters.
26319 Consider the following filtergraph generated by @command{ffplay}.
26320 In this example the last overlay filter has an instance name. All other
26321 filters will have default instance names.
26324 ffplay -dumpgraph 1 -f lavfi "
26325 color=s=100x100:c=red [l];
26326 color=s=100x100:c=blue [r];
26327 nullsrc=s=200x100, zmq [bg];
26328 [bg][l] overlay [bg+l];
26329 [bg+l][r] overlay@@my=x=100 "
26332 To change the color of the left side of the video, the following
26333 command can be used:
26335 echo Parsed_color_0 c yellow | tools/zmqsend
26338 To change the right side:
26340 echo Parsed_color_1 c pink | tools/zmqsend
26343 To change the position of the right side:
26345 echo overlay@@my x 150 | tools/zmqsend
26349 @c man end MULTIMEDIA FILTERS
26351 @chapter Multimedia Sources
26352 @c man begin MULTIMEDIA SOURCES
26354 Below is a description of the currently available multimedia sources.
26358 This is the same as @ref{movie} source, except it selects an audio
26364 Read audio and/or video stream(s) from a movie container.
26366 It accepts the following parameters:
26370 The name of the resource to read (not necessarily a file; it can also be a
26371 device or a stream accessed through some protocol).
26373 @item format_name, f
26374 Specifies the format assumed for the movie to read, and can be either
26375 the name of a container or an input device. If not specified, the
26376 format is guessed from @var{movie_name} or by probing.
26378 @item seek_point, sp
26379 Specifies the seek point in seconds. The frames will be output
26380 starting from this seek point. The parameter is evaluated with
26381 @code{av_strtod}, so the numerical value may be suffixed by an IS
26382 postfix. The default value is "0".
26385 Specifies the streams to read. Several streams can be specified,
26386 separated by "+". The source will then have as many outputs, in the
26387 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26388 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26389 respectively the default (best suited) video and audio stream. Default
26390 is "dv", or "da" if the filter is called as "amovie".
26392 @item stream_index, si
26393 Specifies the index of the video stream to read. If the value is -1,
26394 the most suitable video stream will be automatically selected. The default
26395 value is "-1". Deprecated. If the filter is called "amovie", it will select
26396 audio instead of video.
26399 Specifies how many times to read the stream in sequence.
26400 If the value is 0, the stream will be looped infinitely.
26401 Default value is "1".
26403 Note that when the movie is looped the source timestamps are not
26404 changed, so it will generate non monotonically increasing timestamps.
26406 @item discontinuity
26407 Specifies the time difference between frames above which the point is
26408 considered a timestamp discontinuity which is removed by adjusting the later
26412 It allows overlaying a second video on top of the main input of
26413 a filtergraph, as shown in this graph:
26415 input -----------> deltapts0 --> overlay --> output
26418 movie --> scale--> deltapts1 -------+
26420 @subsection Examples
26424 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26425 on top of the input labelled "in":
26427 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26428 [in] setpts=PTS-STARTPTS [main];
26429 [main][over] overlay=16:16 [out]
26433 Read from a video4linux2 device, and overlay it on top of the input
26436 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26437 [in] setpts=PTS-STARTPTS [main];
26438 [main][over] overlay=16:16 [out]
26442 Read the first video stream and the audio stream with id 0x81 from
26443 dvd.vob; the video is connected to the pad named "video" and the audio is
26444 connected to the pad named "audio":
26446 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26450 @subsection Commands
26452 Both movie and amovie support the following commands:
26455 Perform seek using "av_seek_frame".
26456 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26459 @var{stream_index}: If stream_index is -1, a default
26460 stream is selected, and @var{timestamp} is automatically converted
26461 from AV_TIME_BASE units to the stream specific time_base.
26463 @var{timestamp}: Timestamp in AVStream.time_base units
26464 or, if no stream is specified, in AV_TIME_BASE units.
26466 @var{flags}: Flags which select direction and seeking mode.
26470 Get movie duration in AV_TIME_BASE units.
26474 @c man end MULTIMEDIA SOURCES