1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1617 Same as above but in @code{zp} format:
1619 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3694 (unchanged sound) to 10.0 (maximum effect).
3697 Enable clipping. By default is enabled.
3700 @subsection Commands
3702 This filter supports the all above options as @ref{commands}.
3705 Apply a DC shift to the audio.
3707 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3708 in the recording chain) from the audio. The effect of a DC offset is reduced
3709 headroom and hence volume. The @ref{astats} filter can be used to determine if
3710 a signal has a DC offset.
3714 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3718 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3719 used to prevent clipping.
3724 Apply de-essing to the audio samples.
3728 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3732 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3736 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3740 Set the output mode.
3742 It accepts the following values:
3745 Pass input unchanged.
3748 Pass ess filtered out.
3753 Default value is @var{o}.
3759 Measure audio dynamic range.
3761 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3762 is found in transition material. And anything less that 8 have very poor dynamics
3763 and is very compressed.
3765 The filter accepts the following options:
3769 Set window length in seconds used to split audio into segments of equal length.
3770 Default is 3 seconds.
3774 Dynamic Audio Normalizer.
3776 This filter applies a certain amount of gain to the input audio in order
3777 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3778 contrast to more "simple" normalization algorithms, the Dynamic Audio
3779 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3780 This allows for applying extra gain to the "quiet" sections of the audio
3781 while avoiding distortions or clipping the "loud" sections. In other words:
3782 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3783 sections, in the sense that the volume of each section is brought to the
3784 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3785 this goal *without* applying "dynamic range compressing". It will retain 100%
3786 of the dynamic range *within* each section of the audio file.
3790 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3791 Default is 500 milliseconds.
3792 The Dynamic Audio Normalizer processes the input audio in small chunks,
3793 referred to as frames. This is required, because a peak magnitude has no
3794 meaning for just a single sample value. Instead, we need to determine the
3795 peak magnitude for a contiguous sequence of sample values. While a "standard"
3796 normalizer would simply use the peak magnitude of the complete file, the
3797 Dynamic Audio Normalizer determines the peak magnitude individually for each
3798 frame. The length of a frame is specified in milliseconds. By default, the
3799 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3800 been found to give good results with most files.
3801 Note that the exact frame length, in number of samples, will be determined
3802 automatically, based on the sampling rate of the individual input audio file.
3805 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3806 number. Default is 31.
3807 Probably the most important parameter of the Dynamic Audio Normalizer is the
3808 @code{window size} of the Gaussian smoothing filter. The filter's window size
3809 is specified in frames, centered around the current frame. For the sake of
3810 simplicity, this must be an odd number. Consequently, the default value of 31
3811 takes into account the current frame, as well as the 15 preceding frames and
3812 the 15 subsequent frames. Using a larger window results in a stronger
3813 smoothing effect and thus in less gain variation, i.e. slower gain
3814 adaptation. Conversely, using a smaller window results in a weaker smoothing
3815 effect and thus in more gain variation, i.e. faster gain adaptation.
3816 In other words, the more you increase this value, the more the Dynamic Audio
3817 Normalizer will behave like a "traditional" normalization filter. On the
3818 contrary, the more you decrease this value, the more the Dynamic Audio
3819 Normalizer will behave like a dynamic range compressor.
3822 Set the target peak value. This specifies the highest permissible magnitude
3823 level for the normalized audio input. This filter will try to approach the
3824 target peak magnitude as closely as possible, but at the same time it also
3825 makes sure that the normalized signal will never exceed the peak magnitude.
3826 A frame's maximum local gain factor is imposed directly by the target peak
3827 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3828 It is not recommended to go above this value.
3831 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3832 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3833 factor for each input frame, i.e. the maximum gain factor that does not
3834 result in clipping or distortion. The maximum gain factor is determined by
3835 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3836 additionally bounds the frame's maximum gain factor by a predetermined
3837 (global) maximum gain factor. This is done in order to avoid excessive gain
3838 factors in "silent" or almost silent frames. By default, the maximum gain
3839 factor is 10.0, For most inputs the default value should be sufficient and
3840 it usually is not recommended to increase this value. Though, for input
3841 with an extremely low overall volume level, it may be necessary to allow even
3842 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3843 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3844 Instead, a "sigmoid" threshold function will be applied. This way, the
3845 gain factors will smoothly approach the threshold value, but never exceed that
3849 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3850 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3851 This means that the maximum local gain factor for each frame is defined
3852 (only) by the frame's highest magnitude sample. This way, the samples can
3853 be amplified as much as possible without exceeding the maximum signal
3854 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3855 Normalizer can also take into account the frame's root mean square,
3856 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3857 determine the power of a time-varying signal. It is therefore considered
3858 that the RMS is a better approximation of the "perceived loudness" than
3859 just looking at the signal's peak magnitude. Consequently, by adjusting all
3860 frames to a constant RMS value, a uniform "perceived loudness" can be
3861 established. If a target RMS value has been specified, a frame's local gain
3862 factor is defined as the factor that would result in exactly that RMS value.
3863 Note, however, that the maximum local gain factor is still restricted by the
3864 frame's highest magnitude sample, in order to prevent clipping.
3867 Enable channels coupling. By default is enabled.
3868 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3869 amount. This means the same gain factor will be applied to all channels, i.e.
3870 the maximum possible gain factor is determined by the "loudest" channel.
3871 However, in some recordings, it may happen that the volume of the different
3872 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3873 In this case, this option can be used to disable the channel coupling. This way,
3874 the gain factor will be determined independently for each channel, depending
3875 only on the individual channel's highest magnitude sample. This allows for
3876 harmonizing the volume of the different channels.
3879 Enable DC bias correction. By default is disabled.
3880 An audio signal (in the time domain) is a sequence of sample values.
3881 In the Dynamic Audio Normalizer these sample values are represented in the
3882 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3883 audio signal, or "waveform", should be centered around the zero point.
3884 That means if we calculate the mean value of all samples in a file, or in a
3885 single frame, then the result should be 0.0 or at least very close to that
3886 value. If, however, there is a significant deviation of the mean value from
3887 0.0, in either positive or negative direction, this is referred to as a
3888 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3889 Audio Normalizer provides optional DC bias correction.
3890 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3891 the mean value, or "DC correction" offset, of each input frame and subtract
3892 that value from all of the frame's sample values which ensures those samples
3893 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3894 boundaries, the DC correction offset values will be interpolated smoothly
3895 between neighbouring frames.
3897 @item altboundary, b
3898 Enable alternative boundary mode. By default is disabled.
3899 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3900 around each frame. This includes the preceding frames as well as the
3901 subsequent frames. However, for the "boundary" frames, located at the very
3902 beginning and at the very end of the audio file, not all neighbouring
3903 frames are available. In particular, for the first few frames in the audio
3904 file, the preceding frames are not known. And, similarly, for the last few
3905 frames in the audio file, the subsequent frames are not known. Thus, the
3906 question arises which gain factors should be assumed for the missing frames
3907 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3908 to deal with this situation. The default boundary mode assumes a gain factor
3909 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3910 "fade out" at the beginning and at the end of the input, respectively.
3913 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3914 By default, the Dynamic Audio Normalizer does not apply "traditional"
3915 compression. This means that signal peaks will not be pruned and thus the
3916 full dynamic range will be retained within each local neighbourhood. However,
3917 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3918 normalization algorithm with a more "traditional" compression.
3919 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3920 (thresholding) function. If (and only if) the compression feature is enabled,
3921 all input frames will be processed by a soft knee thresholding function prior
3922 to the actual normalization process. Put simply, the thresholding function is
3923 going to prune all samples whose magnitude exceeds a certain threshold value.
3924 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3925 value. Instead, the threshold value will be adjusted for each individual
3927 In general, smaller parameters result in stronger compression, and vice versa.
3928 Values below 3.0 are not recommended, because audible distortion may appear.
3931 Set the target threshold value. This specifies the lowest permissible
3932 magnitude level for the audio input which will be normalized.
3933 If input frame volume is above this value frame will be normalized.
3934 Otherwise frame may not be normalized at all. The default value is set
3935 to 0, which means all input frames will be normalized.
3936 This option is mostly useful if digital noise is not wanted to be amplified.
3939 @subsection Commands
3941 This filter supports the all above options as @ref{commands}.
3945 Make audio easier to listen to on headphones.
3947 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3948 so that when listened to on headphones the stereo image is moved from
3949 inside your head (standard for headphones) to outside and in front of
3950 the listener (standard for speakers).
3956 Apply a two-pole peaking equalisation (EQ) filter. With this
3957 filter, the signal-level at and around a selected frequency can
3958 be increased or decreased, whilst (unlike bandpass and bandreject
3959 filters) that at all other frequencies is unchanged.
3961 In order to produce complex equalisation curves, this filter can
3962 be given several times, each with a different central frequency.
3964 The filter accepts the following options:
3968 Set the filter's central frequency in Hz.
3971 Set method to specify band-width of filter.
3986 Specify the band-width of a filter in width_type units.
3989 Set the required gain or attenuation in dB.
3990 Beware of clipping when using a positive gain.
3993 How much to use filtered signal in output. Default is 1.
3994 Range is between 0 and 1.
3997 Specify which channels to filter, by default all available are filtered.
4000 Normalize biquad coefficients, by default is disabled.
4001 Enabling it will normalize magnitude response at DC to 0dB.
4004 Set transform type of IIR filter.
4013 Set precison of filtering.
4016 Pick automatic sample format depending on surround filters.
4018 Always use signed 16-bit.
4020 Always use signed 32-bit.
4022 Always use float 32-bit.
4024 Always use float 64-bit.
4028 @subsection Examples
4031 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4033 equalizer=f=1000:t=h:width=200:g=-10
4037 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4039 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4043 @subsection Commands
4045 This filter supports the following commands:
4048 Change equalizer frequency.
4049 Syntax for the command is : "@var{frequency}"
4052 Change equalizer width_type.
4053 Syntax for the command is : "@var{width_type}"
4056 Change equalizer width.
4057 Syntax for the command is : "@var{width}"
4060 Change equalizer gain.
4061 Syntax for the command is : "@var{gain}"
4064 Change equalizer mix.
4065 Syntax for the command is : "@var{mix}"
4068 @section extrastereo
4070 Linearly increases the difference between left and right channels which
4071 adds some sort of "live" effect to playback.
4073 The filter accepts the following options:
4077 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4078 (average of both channels), with 1.0 sound will be unchanged, with
4079 -1.0 left and right channels will be swapped.
4082 Enable clipping. By default is enabled.
4085 @subsection Commands
4087 This filter supports the all above options as @ref{commands}.
4089 @section firequalizer
4090 Apply FIR Equalization using arbitrary frequency response.
4092 The filter accepts the following option:
4096 Set gain curve equation (in dB). The expression can contain variables:
4099 the evaluated frequency
4103 channel number, set to 0 when multichannels evaluation is disabled
4105 channel id, see libavutil/channel_layout.h, set to the first channel id when
4106 multichannels evaluation is disabled
4110 channel_layout, see libavutil/channel_layout.h
4115 @item gain_interpolate(f)
4116 interpolate gain on frequency f based on gain_entry
4117 @item cubic_interpolate(f)
4118 same as gain_interpolate, but smoother
4120 This option is also available as command. Default is @code{gain_interpolate(f)}.
4123 Set gain entry for gain_interpolate function. The expression can
4127 store gain entry at frequency f with value g
4129 This option is also available as command.
4132 Set filter delay in seconds. Higher value means more accurate.
4133 Default is @code{0.01}.
4136 Set filter accuracy in Hz. Lower value means more accurate.
4137 Default is @code{5}.
4140 Set window function. Acceptable values are:
4143 rectangular window, useful when gain curve is already smooth
4145 hann window (default)
4151 3-terms continuous 1st derivative nuttall window
4153 minimum 3-terms discontinuous nuttall window
4155 4-terms continuous 1st derivative nuttall window
4157 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4159 blackman-harris window
4165 If enabled, use fixed number of audio samples. This improves speed when
4166 filtering with large delay. Default is disabled.
4169 Enable multichannels evaluation on gain. Default is disabled.
4172 Enable zero phase mode by subtracting timestamp to compensate delay.
4173 Default is disabled.
4176 Set scale used by gain. Acceptable values are:
4179 linear frequency, linear gain
4181 linear frequency, logarithmic (in dB) gain (default)
4183 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4185 logarithmic frequency, logarithmic gain
4189 Set file for dumping, suitable for gnuplot.
4192 Set scale for dumpfile. Acceptable values are same with scale option.
4196 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4197 Default is disabled.
4200 Enable minimum phase impulse response. Default is disabled.
4203 @subsection Examples
4208 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4211 lowpass at 1000 Hz with gain_entry:
4213 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4216 custom equalization:
4218 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4221 higher delay with zero phase to compensate delay:
4223 firequalizer=delay=0.1:fixed=on:zero_phase=on
4226 lowpass on left channel, highpass on right channel:
4228 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4229 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4234 Apply a flanging effect to the audio.
4236 The filter accepts the following options:
4240 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4243 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4246 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4250 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4251 Default value is 71.
4254 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4257 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4258 Default value is @var{sinusoidal}.
4261 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4262 Default value is 25.
4265 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4266 Default is @var{linear}.
4270 Apply Haas effect to audio.
4272 Note that this makes most sense to apply on mono signals.
4273 With this filter applied to mono signals it give some directionality and
4274 stretches its stereo image.
4276 The filter accepts the following options:
4280 Set input level. By default is @var{1}, or 0dB
4283 Set output level. By default is @var{1}, or 0dB.
4286 Set gain applied to side part of signal. By default is @var{1}.
4289 Set kind of middle source. Can be one of the following:
4299 Pick middle part signal of stereo image.
4302 Pick side part signal of stereo image.
4306 Change middle phase. By default is disabled.
4309 Set left channel delay. By default is @var{2.05} milliseconds.
4312 Set left channel balance. By default is @var{-1}.
4315 Set left channel gain. By default is @var{1}.
4318 Change left phase. By default is disabled.
4321 Set right channel delay. By defaults is @var{2.12} milliseconds.
4324 Set right channel balance. By default is @var{1}.
4327 Set right channel gain. By default is @var{1}.
4330 Change right phase. By default is enabled.
4335 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4336 embedded HDCD codes is expanded into a 20-bit PCM stream.
4338 The filter supports the Peak Extend and Low-level Gain Adjustment features
4339 of HDCD, and detects the Transient Filter flag.
4342 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4345 When using the filter with wav, note the default encoding for wav is 16-bit,
4346 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4347 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4349 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4350 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4353 The filter accepts the following options:
4356 @item disable_autoconvert
4357 Disable any automatic format conversion or resampling in the filter graph.
4359 @item process_stereo
4360 Process the stereo channels together. If target_gain does not match between
4361 channels, consider it invalid and use the last valid target_gain.
4364 Set the code detect timer period in ms.
4367 Always extend peaks above -3dBFS even if PE isn't signaled.
4370 Replace audio with a solid tone and adjust the amplitude to signal some
4371 specific aspect of the decoding process. The output file can be loaded in
4372 an audio editor alongside the original to aid analysis.
4374 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4381 Gain adjustment level at each sample
4383 Samples where peak extend occurs
4385 Samples where the code detect timer is active
4387 Samples where the target gain does not match between channels
4393 Apply head-related transfer functions (HRTFs) to create virtual
4394 loudspeakers around the user for binaural listening via headphones.
4395 The HRIRs are provided via additional streams, for each channel
4396 one stereo input stream is needed.
4398 The filter accepts the following options:
4402 Set mapping of input streams for convolution.
4403 The argument is a '|'-separated list of channel names in order as they
4404 are given as additional stream inputs for filter.
4405 This also specify number of input streams. Number of input streams
4406 must be not less than number of channels in first stream plus one.
4409 Set gain applied to audio. Value is in dB. Default is 0.
4412 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4413 processing audio in time domain which is slow.
4414 @var{freq} is processing audio in frequency domain which is fast.
4415 Default is @var{freq}.
4418 Set custom gain for LFE channels. Value is in dB. Default is 0.
4421 Set size of frame in number of samples which will be processed at once.
4422 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4425 Set format of hrir stream.
4426 Default value is @var{stereo}. Alternative value is @var{multich}.
4427 If value is set to @var{stereo}, number of additional streams should
4428 be greater or equal to number of input channels in first input stream.
4429 Also each additional stream should have stereo number of channels.
4430 If value is set to @var{multich}, number of additional streams should
4431 be exactly one. Also number of input channels of additional stream
4432 should be equal or greater than twice number of channels of first input
4436 @subsection Examples
4440 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4441 each amovie filter use stereo file with IR coefficients as input.
4442 The files give coefficients for each position of virtual loudspeaker:
4445 -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"
4450 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4451 but now in @var{multich} @var{hrir} format.
4453 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"
4460 Apply a high-pass filter with 3dB point frequency.
4461 The filter can be either single-pole, or double-pole (the default).
4462 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4464 The filter accepts the following options:
4468 Set frequency in Hz. Default is 3000.
4471 Set number of poles. Default is 2.
4474 Set method to specify band-width of filter.
4489 Specify the band-width of a filter in width_type units.
4490 Applies only to double-pole filter.
4491 The default is 0.707q and gives a Butterworth response.
4494 How much to use filtered signal in output. Default is 1.
4495 Range is between 0 and 1.
4498 Specify which channels to filter, by default all available are filtered.
4501 Normalize biquad coefficients, by default is disabled.
4502 Enabling it will normalize magnitude response at DC to 0dB.
4505 Set transform type of IIR filter.
4514 Set precison of filtering.
4517 Pick automatic sample format depending on surround filters.
4519 Always use signed 16-bit.
4521 Always use signed 32-bit.
4523 Always use float 32-bit.
4525 Always use float 64-bit.
4529 @subsection Commands
4531 This filter supports the following commands:
4534 Change highpass frequency.
4535 Syntax for the command is : "@var{frequency}"
4538 Change highpass width_type.
4539 Syntax for the command is : "@var{width_type}"
4542 Change highpass width.
4543 Syntax for the command is : "@var{width}"
4546 Change highpass mix.
4547 Syntax for the command is : "@var{mix}"
4552 Join multiple input streams into one multi-channel stream.
4554 It accepts the following parameters:
4558 The number of input streams. It defaults to 2.
4560 @item channel_layout
4561 The desired output channel layout. It defaults to stereo.
4564 Map channels from inputs to output. The argument is a '|'-separated list of
4565 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4566 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4567 can be either the name of the input channel (e.g. FL for front left) or its
4568 index in the specified input stream. @var{out_channel} is the name of the output
4572 The filter will attempt to guess the mappings when they are not specified
4573 explicitly. It does so by first trying to find an unused matching input channel
4574 and if that fails it picks the first unused input channel.
4576 Join 3 inputs (with properly set channel layouts):
4578 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4581 Build a 5.1 output from 6 single-channel streams:
4583 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4584 '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'
4590 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4592 To enable compilation of this filter you need to configure FFmpeg with
4593 @code{--enable-ladspa}.
4597 Specifies the name of LADSPA plugin library to load. If the environment
4598 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4599 each one of the directories specified by the colon separated list in
4600 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4601 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4602 @file{/usr/lib/ladspa/}.
4605 Specifies the plugin within the library. Some libraries contain only
4606 one plugin, but others contain many of them. If this is not set filter
4607 will list all available plugins within the specified library.
4610 Set the '|' separated list of controls which are zero or more floating point
4611 values that determine the behavior of the loaded plugin (for example delay,
4613 Controls need to be defined using the following syntax:
4614 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4615 @var{valuei} is the value set on the @var{i}-th control.
4616 Alternatively they can be also defined using the following syntax:
4617 @var{value0}|@var{value1}|@var{value2}|..., where
4618 @var{valuei} is the value set on the @var{i}-th control.
4619 If @option{controls} is set to @code{help}, all available controls and
4620 their valid ranges are printed.
4622 @item sample_rate, s
4623 Specify the sample rate, default to 44100. Only used if plugin have
4627 Set the number of samples per channel per each output frame, default
4628 is 1024. Only used if plugin have zero inputs.
4631 Set the minimum duration of the sourced audio. See
4632 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4633 for the accepted syntax.
4634 Note that the resulting duration may be greater than the specified duration,
4635 as the generated audio is always cut at the end of a complete frame.
4636 If not specified, or the expressed duration is negative, the audio is
4637 supposed to be generated forever.
4638 Only used if plugin have zero inputs.
4641 Enable latency compensation, by default is disabled.
4642 Only used if plugin have inputs.
4645 @subsection Examples
4649 List all available plugins within amp (LADSPA example plugin) library:
4655 List all available controls and their valid ranges for @code{vcf_notch}
4656 plugin from @code{VCF} library:
4658 ladspa=f=vcf:p=vcf_notch:c=help
4662 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4665 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4669 Add reverberation to the audio using TAP-plugins
4670 (Tom's Audio Processing plugins):
4672 ladspa=file=tap_reverb:tap_reverb
4676 Generate white noise, with 0.2 amplitude:
4678 ladspa=file=cmt:noise_source_white:c=c0=.2
4682 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4683 @code{C* Audio Plugin Suite} (CAPS) library:
4685 ladspa=file=caps:Click:c=c1=20'
4689 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4691 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4695 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4696 @code{SWH Plugins} collection:
4698 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4702 Attenuate low frequencies using Multiband EQ from Steve Harris
4703 @code{SWH Plugins} collection:
4705 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4709 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4712 ladspa=caps:Narrower
4716 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4718 ladspa=caps:White:.2
4722 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4724 ladspa=caps:Fractal:c=c1=1
4728 Dynamic volume normalization using @code{VLevel} plugin:
4730 ladspa=vlevel-ladspa:vlevel_mono
4734 @subsection Commands
4736 This filter supports the following commands:
4739 Modify the @var{N}-th control value.
4741 If the specified value is not valid, it is ignored and prior one is kept.
4746 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4747 Support for both single pass (livestreams, files) and double pass (files) modes.
4748 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4749 detect true peaks, the audio stream will be upsampled to 192 kHz.
4750 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4752 The filter accepts the following options:
4756 Set integrated loudness target.
4757 Range is -70.0 - -5.0. Default value is -24.0.
4760 Set loudness range target.
4761 Range is 1.0 - 20.0. Default value is 7.0.
4764 Set maximum true peak.
4765 Range is -9.0 - +0.0. Default value is -2.0.
4767 @item measured_I, measured_i
4768 Measured IL of input file.
4769 Range is -99.0 - +0.0.
4771 @item measured_LRA, measured_lra
4772 Measured LRA of input file.
4773 Range is 0.0 - 99.0.
4775 @item measured_TP, measured_tp
4776 Measured true peak of input file.
4777 Range is -99.0 - +99.0.
4779 @item measured_thresh
4780 Measured threshold of input file.
4781 Range is -99.0 - +0.0.
4784 Set offset gain. Gain is applied before the true-peak limiter.
4785 Range is -99.0 - +99.0. Default is +0.0.
4788 Normalize by linearly scaling the source audio.
4789 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4790 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4791 be lower than source LRA and the change in integrated loudness shouldn't
4792 result in a true peak which exceeds the target TP. If any of these
4793 conditions aren't met, normalization mode will revert to @var{dynamic}.
4794 Options are @code{true} or @code{false}. Default is @code{true}.
4797 Treat mono input files as "dual-mono". If a mono file is intended for playback
4798 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4799 If set to @code{true}, this option will compensate for this effect.
4800 Multi-channel input files are not affected by this option.
4801 Options are true or false. Default is false.
4804 Set print format for stats. Options are summary, json, or none.
4805 Default value is none.
4810 Apply a low-pass filter with 3dB point frequency.
4811 The filter can be either single-pole or double-pole (the default).
4812 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4814 The filter accepts the following options:
4818 Set frequency in Hz. Default is 500.
4821 Set number of poles. Default is 2.
4824 Set method to specify band-width of filter.
4839 Specify the band-width of a filter in width_type units.
4840 Applies only to double-pole filter.
4841 The default is 0.707q and gives a Butterworth response.
4844 How much to use filtered signal in output. Default is 1.
4845 Range is between 0 and 1.
4848 Specify which channels to filter, by default all available are filtered.
4851 Normalize biquad coefficients, by default is disabled.
4852 Enabling it will normalize magnitude response at DC to 0dB.
4855 Set transform type of IIR filter.
4864 Set precison of filtering.
4867 Pick automatic sample format depending on surround filters.
4869 Always use signed 16-bit.
4871 Always use signed 32-bit.
4873 Always use float 32-bit.
4875 Always use float 64-bit.
4879 @subsection Examples
4882 Lowpass only LFE channel, it LFE is not present it does nothing:
4888 @subsection Commands
4890 This filter supports the following commands:
4893 Change lowpass frequency.
4894 Syntax for the command is : "@var{frequency}"
4897 Change lowpass width_type.
4898 Syntax for the command is : "@var{width_type}"
4901 Change lowpass width.
4902 Syntax for the command is : "@var{width}"
4906 Syntax for the command is : "@var{mix}"
4911 Load a LV2 (LADSPA Version 2) plugin.
4913 To enable compilation of this filter you need to configure FFmpeg with
4914 @code{--enable-lv2}.
4918 Specifies the plugin URI. You may need to escape ':'.
4921 Set the '|' separated list of controls which are zero or more floating point
4922 values that determine the behavior of the loaded plugin (for example delay,
4924 If @option{controls} is set to @code{help}, all available controls and
4925 their valid ranges are printed.
4927 @item sample_rate, s
4928 Specify the sample rate, default to 44100. Only used if plugin have
4932 Set the number of samples per channel per each output frame, default
4933 is 1024. Only used if plugin have zero inputs.
4936 Set the minimum duration of the sourced audio. See
4937 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4938 for the accepted syntax.
4939 Note that the resulting duration may be greater than the specified duration,
4940 as the generated audio is always cut at the end of a complete frame.
4941 If not specified, or the expressed duration is negative, the audio is
4942 supposed to be generated forever.
4943 Only used if plugin have zero inputs.
4946 @subsection Examples
4950 Apply bass enhancer plugin from Calf:
4952 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4956 Apply vinyl plugin from Calf:
4958 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4962 Apply bit crusher plugin from ArtyFX:
4964 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4969 Multiband Compress or expand the audio's dynamic range.
4971 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4972 This is akin to the crossover of a loudspeaker, and results in flat frequency
4973 response when absent compander action.
4975 It accepts the following parameters:
4979 This option syntax is:
4980 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4981 For explanation of each item refer to compand filter documentation.
4987 Mix channels with specific gain levels. The filter accepts the output
4988 channel layout followed by a set of channels definitions.
4990 This filter is also designed to efficiently remap the channels of an audio
4993 The filter accepts parameters of the form:
4994 "@var{l}|@var{outdef}|@var{outdef}|..."
4998 output channel layout or number of channels
5001 output channel specification, of the form:
5002 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5005 output channel to define, either a channel name (FL, FR, etc.) or a channel
5006 number (c0, c1, etc.)
5009 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5012 input channel to use, see out_name for details; it is not possible to mix
5013 named and numbered input channels
5016 If the `=' in a channel specification is replaced by `<', then the gains for
5017 that specification will be renormalized so that the total is 1, thus
5018 avoiding clipping noise.
5020 @subsection Mixing examples
5022 For example, if you want to down-mix from stereo to mono, but with a bigger
5023 factor for the left channel:
5025 pan=1c|c0=0.9*c0+0.1*c1
5028 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5029 7-channels surround:
5031 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5034 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5035 that should be preferred (see "-ac" option) unless you have very specific
5038 @subsection Remapping examples
5040 The channel remapping will be effective if, and only if:
5043 @item gain coefficients are zeroes or ones,
5044 @item only one input per channel output,
5047 If all these conditions are satisfied, the filter will notify the user ("Pure
5048 channel mapping detected"), and use an optimized and lossless method to do the
5051 For example, if you have a 5.1 source and want a stereo audio stream by
5052 dropping the extra channels:
5054 pan="stereo| c0=FL | c1=FR"
5057 Given the same source, you can also switch front left and front right channels
5058 and keep the input channel layout:
5060 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5063 If the input is a stereo audio stream, you can mute the front left channel (and
5064 still keep the stereo channel layout) with:
5069 Still with a stereo audio stream input, you can copy the right channel in both
5070 front left and right:
5072 pan="stereo| c0=FR | c1=FR"
5077 ReplayGain scanner filter. This filter takes an audio stream as an input and
5078 outputs it unchanged.
5079 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5083 Convert the audio sample format, sample rate and channel layout. It is
5084 not meant to be used directly.
5087 Apply time-stretching and pitch-shifting with librubberband.
5089 To enable compilation of this filter, you need to configure FFmpeg with
5090 @code{--enable-librubberband}.
5092 The filter accepts the following options:
5096 Set tempo scale factor.
5099 Set pitch scale factor.
5102 Set transients detector.
5103 Possible values are:
5112 Possible values are:
5121 Possible values are:
5128 Set processing window size.
5129 Possible values are:
5138 Possible values are:
5145 Enable formant preservation when shift pitching.
5146 Possible values are:
5154 Possible values are:
5163 Possible values are:
5170 @subsection Commands
5172 This filter supports the following commands:
5175 Change filter tempo scale factor.
5176 Syntax for the command is : "@var{tempo}"
5179 Change filter pitch scale factor.
5180 Syntax for the command is : "@var{pitch}"
5183 @section sidechaincompress
5185 This filter acts like normal compressor but has the ability to compress
5186 detected signal using second input signal.
5187 It needs two input streams and returns one output stream.
5188 First input stream will be processed depending on second stream signal.
5189 The filtered signal then can be filtered with other filters in later stages of
5190 processing. See @ref{pan} and @ref{amerge} filter.
5192 The filter accepts the following options:
5196 Set input gain. Default is 1. Range is between 0.015625 and 64.
5199 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5200 Default is @code{downward}.
5203 If a signal of second stream raises above this level it will affect the gain
5204 reduction of first stream.
5205 By default is 0.125. Range is between 0.00097563 and 1.
5208 Set a ratio about which the signal is reduced. 1:2 means that if the level
5209 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5210 Default is 2. Range is between 1 and 20.
5213 Amount of milliseconds the signal has to rise above the threshold before gain
5214 reduction starts. Default is 20. Range is between 0.01 and 2000.
5217 Amount of milliseconds the signal has to fall below the threshold before
5218 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5221 Set the amount by how much signal will be amplified after processing.
5222 Default is 1. Range is from 1 to 64.
5225 Curve the sharp knee around the threshold to enter gain reduction more softly.
5226 Default is 2.82843. Range is between 1 and 8.
5229 Choose if the @code{average} level between all channels of side-chain stream
5230 or the louder(@code{maximum}) channel of side-chain stream affects the
5231 reduction. Default is @code{average}.
5234 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5235 of @code{rms}. Default is @code{rms} which is mainly smoother.
5238 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5241 How much to use compressed signal in output. Default is 1.
5242 Range is between 0 and 1.
5245 @subsection Commands
5247 This filter supports the all above options as @ref{commands}.
5249 @subsection Examples
5253 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5254 depending on the signal of 2nd input and later compressed signal to be
5255 merged with 2nd input:
5257 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5261 @section sidechaingate
5263 A sidechain gate acts like a normal (wideband) gate but has the ability to
5264 filter the detected signal before sending it to the gain reduction stage.
5265 Normally a gate uses the full range signal to detect a level above the
5267 For example: If you cut all lower frequencies from your sidechain signal
5268 the gate will decrease the volume of your track only if not enough highs
5269 appear. With this technique you are able to reduce the resonation of a
5270 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5272 It needs two input streams and returns one output stream.
5273 First input stream will be processed depending on second stream signal.
5275 The filter accepts the following options:
5279 Set input level before filtering.
5280 Default is 1. Allowed range is from 0.015625 to 64.
5283 Set the mode of operation. Can be @code{upward} or @code{downward}.
5284 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5285 will be amplified, expanding dynamic range in upward direction.
5286 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5289 Set the level of gain reduction when the signal is below the threshold.
5290 Default is 0.06125. Allowed range is from 0 to 1.
5291 Setting this to 0 disables reduction and then filter behaves like expander.
5294 If a signal rises above this level the gain reduction is released.
5295 Default is 0.125. Allowed range is from 0 to 1.
5298 Set a ratio about which the signal is reduced.
5299 Default is 2. Allowed range is from 1 to 9000.
5302 Amount of milliseconds the signal has to rise above the threshold before gain
5304 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5307 Amount of milliseconds the signal has to fall below the threshold before the
5308 reduction is increased again. Default is 250 milliseconds.
5309 Allowed range is from 0.01 to 9000.
5312 Set amount of amplification of signal after processing.
5313 Default is 1. Allowed range is from 1 to 64.
5316 Curve the sharp knee around the threshold to enter gain reduction more softly.
5317 Default is 2.828427125. Allowed range is from 1 to 8.
5320 Choose if exact signal should be taken for detection or an RMS like one.
5321 Default is rms. Can be peak or rms.
5324 Choose if the average level between all channels or the louder channel affects
5326 Default is average. Can be average or maximum.
5329 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5332 @subsection Commands
5334 This filter supports the all above options as @ref{commands}.
5336 @section silencedetect
5338 Detect silence in an audio stream.
5340 This filter logs a message when it detects that the input audio volume is less
5341 or equal to a noise tolerance value for a duration greater or equal to the
5342 minimum detected noise duration.
5344 The printed times and duration are expressed in seconds. The
5345 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5346 is set on the first frame whose timestamp equals or exceeds the detection
5347 duration and it contains the timestamp of the first frame of the silence.
5349 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5350 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5351 keys are set on the first frame after the silence. If @option{mono} is
5352 enabled, and each channel is evaluated separately, the @code{.X}
5353 suffixed keys are used, and @code{X} corresponds to the channel number.
5355 The filter accepts the following options:
5359 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5360 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5363 Set silence duration until notification (default is 2 seconds). See
5364 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5365 for the accepted syntax.
5368 Process each channel separately, instead of combined. By default is disabled.
5371 @subsection Examples
5375 Detect 5 seconds of silence with -50dB noise tolerance:
5377 silencedetect=n=-50dB:d=5
5381 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5382 tolerance in @file{silence.mp3}:
5384 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5388 @section silenceremove
5390 Remove silence from the beginning, middle or end of the audio.
5392 The filter accepts the following options:
5396 This value is used to indicate if audio should be trimmed at beginning of
5397 the audio. A value of zero indicates no silence should be trimmed from the
5398 beginning. When specifying a non-zero value, it trims audio up until it
5399 finds non-silence. Normally, when trimming silence from beginning of audio
5400 the @var{start_periods} will be @code{1} but it can be increased to higher
5401 values to trim all audio up to specific count of non-silence periods.
5402 Default value is @code{0}.
5404 @item start_duration
5405 Specify the amount of time that non-silence must be detected before it stops
5406 trimming audio. By increasing the duration, bursts of noises can be treated
5407 as silence and trimmed off. Default value is @code{0}.
5409 @item start_threshold
5410 This indicates what sample value should be treated as silence. For digital
5411 audio, a value of @code{0} may be fine but for audio recorded from analog,
5412 you may wish to increase the value to account for background noise.
5413 Can be specified in dB (in case "dB" is appended to the specified value)
5414 or amplitude ratio. Default value is @code{0}.
5417 Specify max duration of silence at beginning that will be kept after
5418 trimming. Default is 0, which is equal to trimming all samples detected
5422 Specify mode of detection of silence end in start of multi-channel audio.
5423 Can be @var{any} or @var{all}. Default is @var{any}.
5424 With @var{any}, any sample that is detected as non-silence will cause
5425 stopped trimming of silence.
5426 With @var{all}, only if all channels are detected as non-silence will cause
5427 stopped trimming of silence.
5430 Set the count for trimming silence from the end of audio.
5431 To remove silence from the middle of a file, specify a @var{stop_periods}
5432 that is negative. This value is then treated as a positive value and is
5433 used to indicate the effect should restart processing as specified by
5434 @var{start_periods}, making it suitable for removing periods of silence
5435 in the middle of the audio.
5436 Default value is @code{0}.
5439 Specify a duration of silence that must exist before audio is not copied any
5440 more. By specifying a higher duration, silence that is wanted can be left in
5442 Default value is @code{0}.
5444 @item stop_threshold
5445 This is the same as @option{start_threshold} but for trimming silence from
5447 Can be specified in dB (in case "dB" is appended to the specified value)
5448 or amplitude ratio. Default value is @code{0}.
5451 Specify max duration of silence at end that will be kept after
5452 trimming. Default is 0, which is equal to trimming all samples detected
5456 Specify mode of detection of silence start in end of multi-channel audio.
5457 Can be @var{any} or @var{all}. Default is @var{any}.
5458 With @var{any}, any sample that is detected as non-silence will cause
5459 stopped trimming of silence.
5460 With @var{all}, only if all channels are detected as non-silence will cause
5461 stopped trimming of silence.
5464 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5465 and works better with digital silence which is exactly 0.
5466 Default value is @code{rms}.
5469 Set duration in number of seconds used to calculate size of window in number
5470 of samples for detecting silence.
5471 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5474 @subsection Examples
5478 The following example shows how this filter can be used to start a recording
5479 that does not contain the delay at the start which usually occurs between
5480 pressing the record button and the start of the performance:
5482 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5486 Trim all silence encountered from beginning to end where there is more than 1
5487 second of silence in audio:
5489 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5493 Trim all digital silence samples, using peak detection, from beginning to end
5494 where there is more than 0 samples of digital silence in audio and digital
5495 silence is detected in all channels at same positions in stream:
5497 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5503 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5504 loudspeakers around the user for binaural listening via headphones (audio
5505 formats up to 9 channels supported).
5506 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5507 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5508 Austrian Academy of Sciences.
5510 To enable compilation of this filter you need to configure FFmpeg with
5511 @code{--enable-libmysofa}.
5513 The filter accepts the following options:
5517 Set the SOFA file used for rendering.
5520 Set gain applied to audio. Value is in dB. Default is 0.
5523 Set rotation of virtual loudspeakers in deg. Default is 0.
5526 Set elevation of virtual speakers in deg. Default is 0.
5529 Set distance in meters between loudspeakers and the listener with near-field
5530 HRTFs. Default is 1.
5533 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5534 processing audio in time domain which is slow.
5535 @var{freq} is processing audio in frequency domain which is fast.
5536 Default is @var{freq}.
5539 Set custom positions of virtual loudspeakers. Syntax for this option is:
5540 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5541 Each virtual loudspeaker is described with short channel name following with
5542 azimuth and elevation in degrees.
5543 Each virtual loudspeaker description is separated by '|'.
5544 For example to override front left and front right channel positions use:
5545 'speakers=FL 45 15|FR 345 15'.
5546 Descriptions with unrecognised channel names are ignored.
5549 Set custom gain for LFE channels. Value is in dB. Default is 0.
5552 Set custom frame size in number of samples. Default is 1024.
5553 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5554 is set to @var{freq}.
5557 Should all IRs be normalized upon importing SOFA file.
5558 By default is enabled.
5561 Should nearest IRs be interpolated with neighbor IRs if exact position
5562 does not match. By default is disabled.
5565 Minphase all IRs upon loading of SOFA file. By default is disabled.
5568 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5571 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5574 @subsection Examples
5578 Using ClubFritz6 sofa file:
5580 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5584 Using ClubFritz12 sofa file and bigger radius with small rotation:
5586 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5590 Similar as above but with custom speaker positions for front left, front right, back left and back right
5591 and also with custom gain:
5593 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5600 This filter expands or compresses each half-cycle of audio samples
5601 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5602 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5604 The filter accepts the following options:
5608 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5609 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5612 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5613 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5614 would be such that local peak value reaches target peak value but never to surpass it and that
5615 ratio between new and previous peak value does not surpass this option value.
5617 @item compression, c
5618 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5619 This option controls maximum local half-cycle of samples compression. This option is used
5620 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5621 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5622 that peak's half-cycle will be compressed by current compression factor.
5625 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5626 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5627 Any half-cycle samples with their local peak value below or same as this option value will be
5628 compressed by current compression factor, otherwise, if greater than threshold value they will be
5629 expanded with expansion factor so that it could reach peak target value but never surpass it.
5632 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5633 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5634 each new half-cycle until it reaches @option{expansion} value.
5635 Setting this options too high may lead to distortions.
5638 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5639 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5640 each new half-cycle until it reaches @option{compression} value.
5643 Specify which channels to filter, by default all available channels are filtered.
5646 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5647 option. When enabled any half-cycle of samples with their local peak value below or same as
5648 @option{threshold} option will be expanded otherwise it will be compressed.
5651 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5652 When disabled each filtered channel gain calculation is independent, otherwise when this option
5653 is enabled the minimum of all possible gains for each filtered channel is used.
5656 @subsection Commands
5658 This filter supports the all above options as @ref{commands}.
5660 @section stereotools
5662 This filter has some handy utilities to manage stereo signals, for converting
5663 M/S stereo recordings to L/R signal while having control over the parameters
5664 or spreading the stereo image of master track.
5666 The filter accepts the following options:
5670 Set input level before filtering for both channels. Defaults is 1.
5671 Allowed range is from 0.015625 to 64.
5674 Set output level after filtering for both channels. Defaults is 1.
5675 Allowed range is from 0.015625 to 64.
5678 Set input balance between both channels. Default is 0.
5679 Allowed range is from -1 to 1.
5682 Set output balance between both channels. Default is 0.
5683 Allowed range is from -1 to 1.
5686 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5687 clipping. Disabled by default.
5690 Mute the left channel. Disabled by default.
5693 Mute the right channel. Disabled by default.
5696 Change the phase of the left channel. Disabled by default.
5699 Change the phase of the right channel. Disabled by default.
5702 Set stereo mode. Available values are:
5706 Left/Right to Left/Right, this is default.
5709 Left/Right to Mid/Side.
5712 Mid/Side to Left/Right.
5715 Left/Right to Left/Left.
5718 Left/Right to Right/Right.
5721 Left/Right to Left + Right.
5724 Left/Right to Right/Left.
5727 Mid/Side to Left/Left.
5730 Mid/Side to Right/Right.
5733 Mid/Side to Right/Left.
5736 Left/Right to Left - Right.
5740 Set level of side signal. Default is 1.
5741 Allowed range is from 0.015625 to 64.
5744 Set balance of side signal. Default is 0.
5745 Allowed range is from -1 to 1.
5748 Set level of the middle signal. Default is 1.
5749 Allowed range is from 0.015625 to 64.
5752 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5755 Set stereo base between mono and inversed channels. Default is 0.
5756 Allowed range is from -1 to 1.
5759 Set delay in milliseconds how much to delay left from right channel and
5760 vice versa. Default is 0. Allowed range is from -20 to 20.
5763 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5766 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5768 @item bmode_in, bmode_out
5769 Set balance mode for balance_in/balance_out option.
5771 Can be one of the following:
5775 Classic balance mode. Attenuate one channel at time.
5776 Gain is raised up to 1.
5779 Similar as classic mode above but gain is raised up to 2.
5782 Equal power distribution, from -6dB to +6dB range.
5786 @subsection Commands
5788 This filter supports the all above options as @ref{commands}.
5790 @subsection Examples
5794 Apply karaoke like effect:
5796 stereotools=mlev=0.015625
5800 Convert M/S signal to L/R:
5802 "stereotools=mode=ms>lr"
5806 @section stereowiden
5808 This filter enhance the stereo effect by suppressing signal common to both
5809 channels and by delaying the signal of left into right and vice versa,
5810 thereby widening the stereo effect.
5812 The filter accepts the following options:
5816 Time in milliseconds of the delay of left signal into right and vice versa.
5817 Default is 20 milliseconds.
5820 Amount of gain in delayed signal into right and vice versa. Gives a delay
5821 effect of left signal in right output and vice versa which gives widening
5822 effect. Default is 0.3.
5825 Cross feed of left into right with inverted phase. This helps in suppressing
5826 the mono. If the value is 1 it will cancel all the signal common to both
5827 channels. Default is 0.3.
5830 Set level of input signal of original channel. Default is 0.8.
5833 @subsection Commands
5835 This filter supports the all above options except @code{delay} as @ref{commands}.
5837 @section superequalizer
5838 Apply 18 band equalizer.
5840 The filter accepts the following options:
5847 Set 131Hz band gain.
5849 Set 185Hz band gain.
5851 Set 262Hz band gain.
5853 Set 370Hz band gain.
5855 Set 523Hz band gain.
5857 Set 740Hz band gain.
5859 Set 1047Hz band gain.
5861 Set 1480Hz band gain.
5863 Set 2093Hz band gain.
5865 Set 2960Hz band gain.
5867 Set 4186Hz band gain.
5869 Set 5920Hz band gain.
5871 Set 8372Hz band gain.
5873 Set 11840Hz band gain.
5875 Set 16744Hz band gain.
5877 Set 20000Hz band gain.
5881 Apply audio surround upmix filter.
5883 This filter allows to produce multichannel output from audio stream.
5885 The filter accepts the following options:
5889 Set output channel layout. By default, this is @var{5.1}.
5891 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5892 for the required syntax.
5895 Set input channel layout. By default, this is @var{stereo}.
5897 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5898 for the required syntax.
5901 Set input volume level. By default, this is @var{1}.
5904 Set output volume level. By default, this is @var{1}.
5907 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5910 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5913 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5916 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5917 In @var{add} mode, LFE channel is created from input audio and added to output.
5918 In @var{sub} mode, LFE channel is created from input audio and added to output but
5919 also all non-LFE output channels are subtracted with output LFE channel.
5922 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5923 Default is @var{90}.
5926 Set front center input volume. By default, this is @var{1}.
5929 Set front center output volume. By default, this is @var{1}.
5932 Set front left input volume. By default, this is @var{1}.
5935 Set front left output volume. By default, this is @var{1}.
5938 Set front right input volume. By default, this is @var{1}.
5941 Set front right output volume. By default, this is @var{1}.
5944 Set side left input volume. By default, this is @var{1}.
5947 Set side left output volume. By default, this is @var{1}.
5950 Set side right input volume. By default, this is @var{1}.
5953 Set side right output volume. By default, this is @var{1}.
5956 Set back left input volume. By default, this is @var{1}.
5959 Set back left output volume. By default, this is @var{1}.
5962 Set back right input volume. By default, this is @var{1}.
5965 Set back right output volume. By default, this is @var{1}.
5968 Set back center input volume. By default, this is @var{1}.
5971 Set back center output volume. By default, this is @var{1}.
5974 Set LFE input volume. By default, this is @var{1}.
5977 Set LFE output volume. By default, this is @var{1}.
5980 Set spread usage of stereo image across X axis for all channels.
5983 Set spread usage of stereo image across Y axis for all channels.
5985 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5986 Set spread usage of stereo image across X axis for each channel.
5988 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5989 Set spread usage of stereo image across Y axis for each channel.
5992 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5995 Set window function.
5997 It accepts the following values:
6020 Default is @code{hann}.
6023 Set window overlap. If set to 1, the recommended overlap for selected
6024 window function will be picked. Default is @code{0.5}.
6027 @section treble, highshelf
6029 Boost or cut treble (upper) frequencies of the audio using a two-pole
6030 shelving filter with a response similar to that of a standard
6031 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6033 The filter accepts the following options:
6037 Give the gain at whichever is the lower of ~22 kHz and the
6038 Nyquist frequency. Its useful range is about -20 (for a large cut)
6039 to +20 (for a large boost). Beware of clipping when using a positive gain.
6042 Set the filter's central frequency and so can be used
6043 to extend or reduce the frequency range to be boosted or cut.
6044 The default value is @code{3000} Hz.
6047 Set method to specify band-width of filter.
6062 Determine how steep is the filter's shelf transition.
6065 Set number of poles. Default is 2.
6068 How much to use filtered signal in output. Default is 1.
6069 Range is between 0 and 1.
6072 Specify which channels to filter, by default all available are filtered.
6075 Normalize biquad coefficients, by default is disabled.
6076 Enabling it will normalize magnitude response at DC to 0dB.
6079 Set transform type of IIR filter.
6088 Set precison of filtering.
6091 Pick automatic sample format depending on surround filters.
6093 Always use signed 16-bit.
6095 Always use signed 32-bit.
6097 Always use float 32-bit.
6099 Always use float 64-bit.
6103 @subsection Commands
6105 This filter supports the following commands:
6108 Change treble frequency.
6109 Syntax for the command is : "@var{frequency}"
6112 Change treble width_type.
6113 Syntax for the command is : "@var{width_type}"
6116 Change treble width.
6117 Syntax for the command is : "@var{width}"
6121 Syntax for the command is : "@var{gain}"
6125 Syntax for the command is : "@var{mix}"
6130 Sinusoidal amplitude modulation.
6132 The filter accepts the following options:
6136 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6137 (20 Hz or lower) will result in a tremolo effect.
6138 This filter may also be used as a ring modulator by specifying
6139 a modulation frequency higher than 20 Hz.
6140 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6143 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6144 Default value is 0.5.
6149 Sinusoidal phase modulation.
6151 The filter accepts the following options:
6155 Modulation frequency in Hertz.
6156 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6159 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6160 Default value is 0.5.
6165 Adjust the input audio volume.
6167 It accepts the following parameters:
6171 Set audio volume expression.
6173 Output values are clipped to the maximum value.
6175 The output audio volume is given by the relation:
6177 @var{output_volume} = @var{volume} * @var{input_volume}
6180 The default value for @var{volume} is "1.0".
6183 This parameter represents the mathematical precision.
6185 It determines which input sample formats will be allowed, which affects the
6186 precision of the volume scaling.
6190 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6192 32-bit floating-point; this limits input sample format to FLT. (default)
6194 64-bit floating-point; this limits input sample format to DBL.
6198 Choose the behaviour on encountering ReplayGain side data in input frames.
6202 Remove ReplayGain side data, ignoring its contents (the default).
6205 Ignore ReplayGain side data, but leave it in the frame.
6208 Prefer the track gain, if present.
6211 Prefer the album gain, if present.
6214 @item replaygain_preamp
6215 Pre-amplification gain in dB to apply to the selected replaygain gain.
6217 Default value for @var{replaygain_preamp} is 0.0.
6219 @item replaygain_noclip
6220 Prevent clipping by limiting the gain applied.
6222 Default value for @var{replaygain_noclip} is 1.
6225 Set when the volume expression is evaluated.
6227 It accepts the following values:
6230 only evaluate expression once during the filter initialization, or
6231 when the @samp{volume} command is sent
6234 evaluate expression for each incoming frame
6237 Default value is @samp{once}.
6240 The volume expression can contain the following parameters.
6244 frame number (starting at zero)
6247 @item nb_consumed_samples
6248 number of samples consumed by the filter
6250 number of samples in the current frame
6252 original frame position in the file
6258 PTS at start of stream
6260 time at start of stream
6266 last set volume value
6269 Note that when @option{eval} is set to @samp{once} only the
6270 @var{sample_rate} and @var{tb} variables are available, all other
6271 variables will evaluate to NAN.
6273 @subsection Commands
6275 This filter supports the following commands:
6278 Modify the volume expression.
6279 The command accepts the same syntax of the corresponding option.
6281 If the specified expression is not valid, it is kept at its current
6285 @subsection Examples
6289 Halve the input audio volume:
6293 volume=volume=-6.0206dB
6296 In all the above example the named key for @option{volume} can be
6297 omitted, for example like in:
6303 Increase input audio power by 6 decibels using fixed-point precision:
6305 volume=volume=6dB:precision=fixed
6309 Fade volume after time 10 with an annihilation period of 5 seconds:
6311 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6315 @section volumedetect
6317 Detect the volume of the input video.
6319 The filter has no parameters. The input is not modified. Statistics about
6320 the volume will be printed in the log when the input stream end is reached.
6322 In particular it will show the mean volume (root mean square), maximum
6323 volume (on a per-sample basis), and the beginning of a histogram of the
6324 registered volume values (from the maximum value to a cumulated 1/1000 of
6327 All volumes are in decibels relative to the maximum PCM value.
6329 @subsection Examples
6331 Here is an excerpt of the output:
6333 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6334 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6347 The mean square energy is approximately -27 dB, or 10^-2.7.
6349 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6351 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6354 In other words, raising the volume by +4 dB does not cause any clipping,
6355 raising it by +5 dB causes clipping for 6 samples, etc.
6357 @c man end AUDIO FILTERS
6359 @chapter Audio Sources
6360 @c man begin AUDIO SOURCES
6362 Below is a description of the currently available audio sources.
6366 Buffer audio frames, and make them available to the filter chain.
6368 This source is mainly intended for a programmatic use, in particular
6369 through the interface defined in @file{libavfilter/buffersrc.h}.
6371 It accepts the following parameters:
6375 The timebase which will be used for timestamps of submitted frames. It must be
6376 either a floating-point number or in @var{numerator}/@var{denominator} form.
6379 The sample rate of the incoming audio buffers.
6382 The sample format of the incoming audio buffers.
6383 Either a sample format name or its corresponding integer representation from
6384 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6386 @item channel_layout
6387 The channel layout of the incoming audio buffers.
6388 Either a channel layout name from channel_layout_map in
6389 @file{libavutil/channel_layout.c} or its corresponding integer representation
6390 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6393 The number of channels of the incoming audio buffers.
6394 If both @var{channels} and @var{channel_layout} are specified, then they
6399 @subsection Examples
6402 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6405 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6406 Since the sample format with name "s16p" corresponds to the number
6407 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6410 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6415 Generate an audio signal specified by an expression.
6417 This source accepts in input one or more expressions (one for each
6418 channel), which are evaluated and used to generate a corresponding
6421 This source accepts the following options:
6425 Set the '|'-separated expressions list for each separate channel. In case the
6426 @option{channel_layout} option is not specified, the selected channel layout
6427 depends on the number of provided expressions. Otherwise the last
6428 specified expression is applied to the remaining output channels.
6430 @item channel_layout, c
6431 Set the channel layout. The number of channels in the specified layout
6432 must be equal to the number of specified expressions.
6435 Set the minimum duration of the sourced audio. See
6436 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6437 for the accepted syntax.
6438 Note that the resulting duration may be greater than the specified
6439 duration, as the generated audio is always cut at the end of a
6442 If not specified, or the expressed duration is negative, the audio is
6443 supposed to be generated forever.
6446 Set the number of samples per channel per each output frame,
6449 @item sample_rate, s
6450 Specify the sample rate, default to 44100.
6453 Each expression in @var{exprs} can contain the following constants:
6457 number of the evaluated sample, starting from 0
6460 time of the evaluated sample expressed in seconds, starting from 0
6467 @subsection Examples
6477 Generate a sin signal with frequency of 440 Hz, set sample rate to
6480 aevalsrc="sin(440*2*PI*t):s=8000"
6484 Generate a two channels signal, specify the channel layout (Front
6485 Center + Back Center) explicitly:
6487 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6491 Generate white noise:
6493 aevalsrc="-2+random(0)"
6497 Generate an amplitude modulated signal:
6499 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6503 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6505 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6512 Generate a FIR coefficients using frequency sampling method.
6514 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6516 The filter accepts the following options:
6520 Set number of filter coefficents in output audio stream.
6521 Default value is 1025.
6524 Set frequency points from where magnitude and phase are set.
6525 This must be in non decreasing order, and first element must be 0, while last element
6526 must be 1. Elements are separated by white spaces.
6529 Set magnitude value for every frequency point set by @option{frequency}.
6530 Number of values must be same as number of frequency points.
6531 Values are separated by white spaces.
6534 Set phase value for every frequency point set by @option{frequency}.
6535 Number of values must be same as number of frequency points.
6536 Values are separated by white spaces.
6538 @item sample_rate, r
6539 Set sample rate, default is 44100.
6542 Set number of samples per each frame. Default is 1024.
6545 Set window function. Default is blackman.
6550 The null audio source, return unprocessed audio frames. It is mainly useful
6551 as a template and to be employed in analysis / debugging tools, or as
6552 the source for filters which ignore the input data (for example the sox
6555 This source accepts the following options:
6559 @item channel_layout, cl
6561 Specifies the channel layout, and can be either an integer or a string
6562 representing a channel layout. The default value of @var{channel_layout}
6565 Check the channel_layout_map definition in
6566 @file{libavutil/channel_layout.c} for the mapping between strings and
6567 channel layout values.
6569 @item sample_rate, r
6570 Specifies the sample rate, and defaults to 44100.
6573 Set the number of samples per requested frames.
6576 Set the duration of the sourced audio. See
6577 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6578 for the accepted syntax.
6580 If not specified, or the expressed duration is negative, the audio is
6581 supposed to be generated forever.
6584 @subsection Examples
6588 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6590 anullsrc=r=48000:cl=4
6594 Do the same operation with a more obvious syntax:
6596 anullsrc=r=48000:cl=mono
6600 All the parameters need to be explicitly defined.
6604 Synthesize a voice utterance using the libflite library.
6606 To enable compilation of this filter you need to configure FFmpeg with
6607 @code{--enable-libflite}.
6609 Note that versions of the flite library prior to 2.0 are not thread-safe.
6611 The filter accepts the following options:
6616 If set to 1, list the names of the available voices and exit
6617 immediately. Default value is 0.
6620 Set the maximum number of samples per frame. Default value is 512.
6623 Set the filename containing the text to speak.
6626 Set the text to speak.
6629 Set the voice to use for the speech synthesis. Default value is
6630 @code{kal}. See also the @var{list_voices} option.
6633 @subsection Examples
6637 Read from file @file{speech.txt}, and synthesize the text using the
6638 standard flite voice:
6640 flite=textfile=speech.txt
6644 Read the specified text selecting the @code{slt} voice:
6646 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6650 Input text to ffmpeg:
6652 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6656 Make @file{ffplay} speak the specified text, using @code{flite} and
6657 the @code{lavfi} device:
6659 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6663 For more information about libflite, check:
6664 @url{http://www.festvox.org/flite/}
6668 Generate a noise audio signal.
6670 The filter accepts the following options:
6673 @item sample_rate, r
6674 Specify the sample rate. Default value is 48000 Hz.
6677 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6681 Specify the duration of the generated audio stream. Not specifying this option
6682 results in noise with an infinite length.
6684 @item color, colour, c
6685 Specify the color of noise. Available noise colors are white, pink, brown,
6686 blue, violet and velvet. Default color is white.
6689 Specify a value used to seed the PRNG.
6692 Set the number of samples per each output frame, default is 1024.
6695 @subsection Examples
6700 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6702 anoisesrc=d=60:c=pink:r=44100:a=0.5
6708 Generate odd-tap Hilbert transform FIR coefficients.
6710 The resulting stream can be used with @ref{afir} filter for phase-shifting
6711 the signal by 90 degrees.
6713 This is used in many matrix coding schemes and for analytic signal generation.
6714 The process is often written as a multiplication by i (or j), the imaginary unit.
6716 The filter accepts the following options:
6720 @item sample_rate, s
6721 Set sample rate, default is 44100.
6724 Set length of FIR filter, default is 22051.
6727 Set number of samples per each frame.
6730 Set window function to be used when generating FIR coefficients.
6735 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6737 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6739 The filter accepts the following options:
6742 @item sample_rate, r
6743 Set sample rate, default is 44100.
6746 Set number of samples per each frame. Default is 1024.
6749 Set high-pass frequency. Default is 0.
6752 Set low-pass frequency. Default is 0.
6753 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6754 is higher than 0 then filter will create band-pass filter coefficients,
6755 otherwise band-reject filter coefficients.
6758 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6761 Set Kaiser window beta.
6764 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6767 Enable rounding, by default is disabled.
6770 Set number of taps for high-pass filter.
6773 Set number of taps for low-pass filter.
6778 Generate an audio signal made of a sine wave with amplitude 1/8.
6780 The audio signal is bit-exact.
6782 The filter accepts the following options:
6787 Set the carrier frequency. Default is 440 Hz.
6789 @item beep_factor, b
6790 Enable a periodic beep every second with frequency @var{beep_factor} times
6791 the carrier frequency. Default is 0, meaning the beep is disabled.
6793 @item sample_rate, r
6794 Specify the sample rate, default is 44100.
6797 Specify the duration of the generated audio stream.
6799 @item samples_per_frame
6800 Set the number of samples per output frame.
6802 The expression can contain the following constants:
6806 The (sequential) number of the output audio frame, starting from 0.
6809 The PTS (Presentation TimeStamp) of the output audio frame,
6810 expressed in @var{TB} units.
6813 The PTS of the output audio frame, expressed in seconds.
6816 The timebase of the output audio frames.
6819 Default is @code{1024}.
6822 @subsection Examples
6827 Generate a simple 440 Hz sine wave:
6833 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6837 sine=frequency=220:beep_factor=4:duration=5
6841 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6844 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6848 @c man end AUDIO SOURCES
6850 @chapter Audio Sinks
6851 @c man begin AUDIO SINKS
6853 Below is a description of the currently available audio sinks.
6855 @section abuffersink
6857 Buffer audio frames, and make them available to the end of filter chain.
6859 This sink is mainly intended for programmatic use, in particular
6860 through the interface defined in @file{libavfilter/buffersink.h}
6861 or the options system.
6863 It accepts a pointer to an AVABufferSinkContext structure, which
6864 defines the incoming buffers' formats, to be passed as the opaque
6865 parameter to @code{avfilter_init_filter} for initialization.
6868 Null audio sink; do absolutely nothing with the input audio. It is
6869 mainly useful as a template and for use in analysis / debugging
6872 @c man end AUDIO SINKS
6874 @chapter Video Filters
6875 @c man begin VIDEO FILTERS
6877 When you configure your FFmpeg build, you can disable any of the
6878 existing filters using @code{--disable-filters}.
6879 The configure output will show the video filters included in your
6882 Below is a description of the currently available video filters.
6886 Mark a region of interest in a video frame.
6888 The frame data is passed through unchanged, but metadata is attached
6889 to the frame indicating regions of interest which can affect the
6890 behaviour of later encoding. Multiple regions can be marked by
6891 applying the filter multiple times.
6895 Region distance in pixels from the left edge of the frame.
6897 Region distance in pixels from the top edge of the frame.
6899 Region width in pixels.
6901 Region height in pixels.
6903 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6904 and may contain the following variables:
6907 Width of the input frame.
6909 Height of the input frame.
6913 Quantisation offset to apply within the region.
6915 This must be a real value in the range -1 to +1. A value of zero
6916 indicates no quality change. A negative value asks for better quality
6917 (less quantisation), while a positive value asks for worse quality
6918 (greater quantisation).
6920 The range is calibrated so that the extreme values indicate the
6921 largest possible offset - if the rest of the frame is encoded with the
6922 worst possible quality, an offset of -1 indicates that this region
6923 should be encoded with the best possible quality anyway. Intermediate
6924 values are then interpolated in some codec-dependent way.
6926 For example, in 10-bit H.264 the quantisation parameter varies between
6927 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6928 this region should be encoded with a QP around one-tenth of the full
6929 range better than the rest of the frame. So, if most of the frame
6930 were to be encoded with a QP of around 30, this region would get a QP
6931 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6932 An extreme value of -1 would indicate that this region should be
6933 encoded with the best possible quality regardless of the treatment of
6934 the rest of the frame - that is, should be encoded at a QP of -12.
6936 If set to true, remove any existing regions of interest marked on the
6937 frame before adding the new one.
6940 @subsection Examples
6944 Mark the centre quarter of the frame as interesting.
6946 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6949 Mark the 100-pixel-wide region on the left edge of the frame as very
6950 uninteresting (to be encoded at much lower quality than the rest of
6953 addroi=0:0:100:ih:+1/5
6957 @section alphaextract
6959 Extract the alpha component from the input as a grayscale video. This
6960 is especially useful with the @var{alphamerge} filter.
6964 Add or replace the alpha component of the primary input with the
6965 grayscale value of a second input. This is intended for use with
6966 @var{alphaextract} to allow the transmission or storage of frame
6967 sequences that have alpha in a format that doesn't support an alpha
6970 For example, to reconstruct full frames from a normal YUV-encoded video
6971 and a separate video created with @var{alphaextract}, you might use:
6973 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6978 Amplify differences between current pixel and pixels of adjacent frames in
6979 same pixel location.
6981 This filter accepts the following options:
6985 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6986 For example radius of 3 will instruct filter to calculate average of 7 frames.
6989 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6992 Set threshold for difference amplification. Any difference greater or equal to
6993 this value will not alter source pixel. Default is 10.
6994 Allowed range is from 0 to 65535.
6997 Set tolerance for difference amplification. Any difference lower to
6998 this value will not alter source pixel. Default is 0.
6999 Allowed range is from 0 to 65535.
7002 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7003 This option controls maximum possible value that will decrease source pixel value.
7006 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7007 This option controls maximum possible value that will increase source pixel value.
7010 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7013 @subsection Commands
7015 This filter supports the following @ref{commands} that corresponds to option of same name:
7027 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7028 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7029 Substation Alpha) subtitles files.
7031 This filter accepts the following option in addition to the common options from
7032 the @ref{subtitles} filter:
7036 Set the shaping engine
7038 Available values are:
7041 The default libass shaping engine, which is the best available.
7043 Fast, font-agnostic shaper that can do only substitutions
7045 Slower shaper using OpenType for substitutions and positioning
7048 The default is @code{auto}.
7052 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7054 The filter accepts the following options:
7058 Set threshold A for 1st plane. Default is 0.02.
7059 Valid range is 0 to 0.3.
7062 Set threshold B for 1st plane. Default is 0.04.
7063 Valid range is 0 to 5.
7066 Set threshold A for 2nd plane. Default is 0.02.
7067 Valid range is 0 to 0.3.
7070 Set threshold B for 2nd plane. Default is 0.04.
7071 Valid range is 0 to 5.
7074 Set threshold A for 3rd plane. Default is 0.02.
7075 Valid range is 0 to 0.3.
7078 Set threshold B for 3rd plane. Default is 0.04.
7079 Valid range is 0 to 5.
7081 Threshold A is designed to react on abrupt changes in the input signal and
7082 threshold B is designed to react on continuous changes in the input signal.
7085 Set number of frames filter will use for averaging. Default is 9. Must be odd
7086 number in range [5, 129].
7089 Set what planes of frame filter will use for averaging. Default is all.
7092 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7093 Alternatively can be set to @code{s} serial.
7095 Parallel can be faster then serial, while other way around is never true.
7096 Parallel will abort early on first change being greater then thresholds, while serial
7097 will continue processing other side of frames if they are equal or below thresholds.
7100 @subsection Commands
7101 This filter supports same @ref{commands} as options except option @code{s}.
7102 The command accepts the same syntax of the corresponding option.
7106 Apply average blur filter.
7108 The filter accepts the following options:
7112 Set horizontal radius size.
7115 Set which planes to filter. By default all planes are filtered.
7118 Set vertical radius size, if zero it will be same as @code{sizeX}.
7119 Default is @code{0}.
7122 @subsection Commands
7123 This filter supports same commands as options.
7124 The command accepts the same syntax of the corresponding option.
7126 If the specified expression is not valid, it is kept at its current
7131 Compute the bounding box for the non-black pixels in the input frame
7134 This filter computes the bounding box containing all the pixels with a
7135 luminance value greater than the minimum allowed value.
7136 The parameters describing the bounding box are printed on the filter
7139 The filter accepts the following option:
7143 Set the minimal luminance value. Default is @code{16}.
7147 Apply bilateral filter, spatial smoothing while preserving edges.
7149 The filter accepts the following options:
7152 Set sigma of gaussian function to calculate spatial weight.
7153 Allowed range is 0 to 512. Default is 0.1.
7156 Set sigma of gaussian function to calculate range weight.
7157 Allowed range is 0 to 1. Default is 0.1.
7160 Set planes to filter. Default is first only.
7163 @subsection Commands
7165 This filter supports the all above options as @ref{commands}.
7167 @section bitplanenoise
7169 Show and measure bit plane noise.
7171 The filter accepts the following options:
7175 Set which plane to analyze. Default is @code{1}.
7178 Filter out noisy pixels from @code{bitplane} set above.
7179 Default is disabled.
7182 @section blackdetect
7184 Detect video intervals that are (almost) completely black. Can be
7185 useful to detect chapter transitions, commercials, or invalid
7188 The filter outputs its detection analysis to both the log as well as
7189 frame metadata. If a black segment of at least the specified minimum
7190 duration is found, a line with the start and end timestamps as well
7191 as duration is printed to the log with level @code{info}. In addition,
7192 a log line with level @code{debug} is printed per frame showing the
7193 black amount detected for that frame.
7195 The filter also attaches metadata to the first frame of a black
7196 segment with key @code{lavfi.black_start} and to the first frame
7197 after the black segment ends with key @code{lavfi.black_end}. The
7198 value is the frame's timestamp. This metadata is added regardless
7199 of the minimum duration specified.
7201 The filter accepts the following options:
7204 @item black_min_duration, d
7205 Set the minimum detected black duration expressed in seconds. It must
7206 be a non-negative floating point number.
7208 Default value is 2.0.
7210 @item picture_black_ratio_th, pic_th
7211 Set the threshold for considering a picture "black".
7212 Express the minimum value for the ratio:
7214 @var{nb_black_pixels} / @var{nb_pixels}
7217 for which a picture is considered black.
7218 Default value is 0.98.
7220 @item pixel_black_th, pix_th
7221 Set the threshold for considering a pixel "black".
7223 The threshold expresses the maximum pixel luminance value for which a
7224 pixel is considered "black". The provided value is scaled according to
7225 the following equation:
7227 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7230 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7231 the input video format, the range is [0-255] for YUV full-range
7232 formats and [16-235] for YUV non full-range formats.
7234 Default value is 0.10.
7237 The following example sets the maximum pixel threshold to the minimum
7238 value, and detects only black intervals of 2 or more seconds:
7240 blackdetect=d=2:pix_th=0.00
7245 Detect frames that are (almost) completely black. Can be useful to
7246 detect chapter transitions or commercials. Output lines consist of
7247 the frame number of the detected frame, the percentage of blackness,
7248 the position in the file if known or -1 and the timestamp in seconds.
7250 In order to display the output lines, you need to set the loglevel at
7251 least to the AV_LOG_INFO value.
7253 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7254 The value represents the percentage of pixels in the picture that
7255 are below the threshold value.
7257 It accepts the following parameters:
7262 The percentage of the pixels that have to be below the threshold; it defaults to
7265 @item threshold, thresh
7266 The threshold below which a pixel value is considered black; it defaults to
7274 Blend two video frames into each other.
7276 The @code{blend} filter takes two input streams and outputs one
7277 stream, the first input is the "top" layer and second input is
7278 "bottom" layer. By default, the output terminates when the longest input terminates.
7280 The @code{tblend} (time blend) filter takes two consecutive frames
7281 from one single stream, and outputs the result obtained by blending
7282 the new frame on top of the old frame.
7284 A description of the accepted options follows.
7292 Set blend mode for specific pixel component or all pixel components in case
7293 of @var{all_mode}. Default value is @code{normal}.
7295 Available values for component modes are:
7337 Set blend opacity for specific pixel component or all pixel components in case
7338 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7345 Set blend expression for specific pixel component or all pixel components in case
7346 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7348 The expressions can use the following variables:
7352 The sequential number of the filtered frame, starting from @code{0}.
7356 the coordinates of the current sample
7360 the width and height of currently filtered plane
7364 Width and height scale for the plane being filtered. It is the
7365 ratio between the dimensions of the current plane to the luma plane,
7366 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7367 the luma plane and @code{0.5,0.5} for the chroma planes.
7370 Time of the current frame, expressed in seconds.
7373 Value of pixel component at current location for first video frame (top layer).
7376 Value of pixel component at current location for second video frame (bottom layer).
7380 The @code{blend} filter also supports the @ref{framesync} options.
7382 @subsection Examples
7386 Apply transition from bottom layer to top layer in first 10 seconds:
7388 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7392 Apply linear horizontal transition from top layer to bottom layer:
7394 blend=all_expr='A*(X/W)+B*(1-X/W)'
7398 Apply 1x1 checkerboard effect:
7400 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7404 Apply uncover left effect:
7406 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7410 Apply uncover down effect:
7412 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7416 Apply uncover up-left effect:
7418 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7422 Split diagonally video and shows top and bottom layer on each side:
7424 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7428 Display differences between the current and the previous frame:
7430 tblend=all_mode=grainextract
7436 Denoise frames using Block-Matching 3D algorithm.
7438 The filter accepts the following options.
7442 Set denoising strength. Default value is 1.
7443 Allowed range is from 0 to 999.9.
7444 The denoising algorithm is very sensitive to sigma, so adjust it
7445 according to the source.
7448 Set local patch size. This sets dimensions in 2D.
7451 Set sliding step for processing blocks. Default value is 4.
7452 Allowed range is from 1 to 64.
7453 Smaller values allows processing more reference blocks and is slower.
7456 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7457 When set to 1, no block matching is done. Larger values allows more blocks
7459 Allowed range is from 1 to 256.
7462 Set radius for search block matching. Default is 9.
7463 Allowed range is from 1 to INT32_MAX.
7466 Set step between two search locations for block matching. Default is 1.
7467 Allowed range is from 1 to 64. Smaller is slower.
7470 Set threshold of mean square error for block matching. Valid range is 0 to
7474 Set thresholding parameter for hard thresholding in 3D transformed domain.
7475 Larger values results in stronger hard-thresholding filtering in frequency
7479 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7480 Default is @code{basic}.
7483 If enabled, filter will use 2nd stream for block matching.
7484 Default is disabled for @code{basic} value of @var{estim} option,
7485 and always enabled if value of @var{estim} is @code{final}.
7488 Set planes to filter. Default is all available except alpha.
7491 @subsection Examples
7495 Basic filtering with bm3d:
7497 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7501 Same as above, but filtering only luma:
7503 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7507 Same as above, but with both estimation modes:
7509 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
7513 Same as above, but prefilter with @ref{nlmeans} filter instead:
7515 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
7521 Apply a boxblur algorithm to the input video.
7523 It accepts the following parameters:
7527 @item luma_radius, lr
7528 @item luma_power, lp
7529 @item chroma_radius, cr
7530 @item chroma_power, cp
7531 @item alpha_radius, ar
7532 @item alpha_power, ap
7536 A description of the accepted options follows.
7539 @item luma_radius, lr
7540 @item chroma_radius, cr
7541 @item alpha_radius, ar
7542 Set an expression for the box radius in pixels used for blurring the
7543 corresponding input plane.
7545 The radius value must be a non-negative number, and must not be
7546 greater than the value of the expression @code{min(w,h)/2} for the
7547 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7550 Default value for @option{luma_radius} is "2". If not specified,
7551 @option{chroma_radius} and @option{alpha_radius} default to the
7552 corresponding value set for @option{luma_radius}.
7554 The expressions can contain the following constants:
7558 The input width and height in pixels.
7562 The input chroma image width and height in pixels.
7566 The horizontal and vertical chroma subsample values. For example, for the
7567 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7570 @item luma_power, lp
7571 @item chroma_power, cp
7572 @item alpha_power, ap
7573 Specify how many times the boxblur filter is applied to the
7574 corresponding plane.
7576 Default value for @option{luma_power} is 2. If not specified,
7577 @option{chroma_power} and @option{alpha_power} default to the
7578 corresponding value set for @option{luma_power}.
7580 A value of 0 will disable the effect.
7583 @subsection Examples
7587 Apply a boxblur filter with the luma, chroma, and alpha radii
7590 boxblur=luma_radius=2:luma_power=1
7595 Set the luma radius to 2, and alpha and chroma radius to 0:
7597 boxblur=2:1:cr=0:ar=0
7601 Set the luma and chroma radii to a fraction of the video dimension:
7603 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7609 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7610 Deinterlacing Filter").
7612 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7613 interpolation algorithms.
7614 It accepts the following parameters:
7618 The interlacing mode to adopt. It accepts one of the following values:
7622 Output one frame for each frame.
7624 Output one frame for each field.
7627 The default value is @code{send_field}.
7630 The picture field parity assumed for the input interlaced video. It accepts one
7631 of the following values:
7635 Assume the top field is first.
7637 Assume the bottom field is first.
7639 Enable automatic detection of field parity.
7642 The default value is @code{auto}.
7643 If the interlacing is unknown or the decoder does not export this information,
7644 top field first will be assumed.
7647 Specify which frames to deinterlace. Accepts one of the following
7652 Deinterlace all frames.
7654 Only deinterlace frames marked as interlaced.
7657 The default value is @code{all}.
7662 Apply Contrast Adaptive Sharpen filter to video stream.
7664 The filter accepts the following options:
7668 Set the sharpening strength. Default value is 0.
7671 Set planes to filter. Default value is to filter all
7672 planes except alpha plane.
7675 @subsection Commands
7676 This filter supports same @ref{commands} as options.
7679 Remove all color information for all colors except for certain one.
7681 The filter accepts the following options:
7685 The color which will not be replaced with neutral chroma.
7688 Similarity percentage with the above color.
7689 0.01 matches only the exact key color, while 1.0 matches everything.
7693 0.0 makes pixels either fully gray, or not gray at all.
7694 Higher values result in more preserved color.
7697 Signals that the color passed is already in YUV instead of RGB.
7699 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7700 This can be used to pass exact YUV values as hexadecimal numbers.
7703 @subsection Commands
7704 This filter supports same @ref{commands} as options.
7705 The command accepts the same syntax of the corresponding option.
7707 If the specified expression is not valid, it is kept at its current
7711 YUV colorspace color/chroma keying.
7713 The filter accepts the following options:
7717 The color which will be replaced with transparency.
7720 Similarity percentage with the key color.
7722 0.01 matches only the exact key color, while 1.0 matches everything.
7727 0.0 makes pixels either fully transparent, or not transparent at all.
7729 Higher values result in semi-transparent pixels, with a higher transparency
7730 the more similar the pixels color is to the key color.
7733 Signals that the color passed is already in YUV instead of RGB.
7735 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7736 This can be used to pass exact YUV values as hexadecimal numbers.
7739 @subsection Commands
7740 This filter supports same @ref{commands} as options.
7741 The command accepts the same syntax of the corresponding option.
7743 If the specified expression is not valid, it is kept at its current
7746 @subsection Examples
7750 Make every green pixel in the input image transparent:
7752 ffmpeg -i input.png -vf chromakey=green out.png
7756 Overlay a greenscreen-video on top of a static black background.
7758 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
7763 Reduce chrominance noise.
7765 The filter accepts the following options:
7769 Set threshold for averaging chrominance values.
7770 Sum of absolute difference of Y, U and V pixel components of current
7771 pixel and neighbour pixels lower than this threshold will be used in
7772 averaging. Luma component is left unchanged and is copied to output.
7773 Default value is 30. Allowed range is from 1 to 200.
7776 Set horizontal radius of rectangle used for averaging.
7777 Allowed range is from 1 to 100. Default value is 5.
7780 Set vertical radius of rectangle used for averaging.
7781 Allowed range is from 1 to 100. Default value is 5.
7784 Set horizontal step when averaging. Default value is 1.
7785 Allowed range is from 1 to 50.
7786 Mostly useful to speed-up filtering.
7789 Set vertical step when averaging. Default value is 1.
7790 Allowed range is from 1 to 50.
7791 Mostly useful to speed-up filtering.
7794 Set Y threshold for averaging chrominance values.
7795 Set finer control for max allowed difference between Y components
7796 of current pixel and neigbour pixels.
7797 Default value is 200. Allowed range is from 1 to 200.
7800 Set U threshold for averaging chrominance values.
7801 Set finer control for max allowed difference between U components
7802 of current pixel and neigbour pixels.
7803 Default value is 200. Allowed range is from 1 to 200.
7806 Set V threshold for averaging chrominance values.
7807 Set finer control for max allowed difference between V components
7808 of current pixel and neigbour pixels.
7809 Default value is 200. Allowed range is from 1 to 200.
7812 @subsection Commands
7813 This filter supports same @ref{commands} as options.
7814 The command accepts the same syntax of the corresponding option.
7816 @section chromashift
7817 Shift chroma pixels horizontally and/or vertically.
7819 The filter accepts the following options:
7822 Set amount to shift chroma-blue horizontally.
7824 Set amount to shift chroma-blue vertically.
7826 Set amount to shift chroma-red horizontally.
7828 Set amount to shift chroma-red vertically.
7830 Set edge mode, can be @var{smear}, default, or @var{warp}.
7833 @subsection Commands
7835 This filter supports the all above options as @ref{commands}.
7839 Display CIE color diagram with pixels overlaid onto it.
7841 The filter accepts the following options:
7856 @item uhdtv, rec2020
7870 Set what gamuts to draw.
7872 See @code{system} option for available values.
7875 Set ciescope size, by default set to 512.
7878 Set intensity used to map input pixel values to CIE diagram.
7881 Set contrast used to draw tongue colors that are out of active color system gamut.
7884 Correct gamma displayed on scope, by default enabled.
7887 Show white point on CIE diagram, by default disabled.
7890 Set input gamma. Used only with XYZ input color space.
7895 Visualize information exported by some codecs.
7897 Some codecs can export information through frames using side-data or other
7898 means. For example, some MPEG based codecs export motion vectors through the
7899 @var{export_mvs} flag in the codec @option{flags2} option.
7901 The filter accepts the following option:
7905 Set motion vectors to visualize.
7907 Available flags for @var{mv} are:
7911 forward predicted MVs of P-frames
7913 forward predicted MVs of B-frames
7915 backward predicted MVs of B-frames
7919 Display quantization parameters using the chroma planes.
7922 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7924 Available flags for @var{mv_type} are:
7928 forward predicted MVs
7930 backward predicted MVs
7933 @item frame_type, ft
7934 Set frame type to visualize motion vectors of.
7936 Available flags for @var{frame_type} are:
7940 intra-coded frames (I-frames)
7942 predicted frames (P-frames)
7944 bi-directionally predicted frames (B-frames)
7948 @subsection Examples
7952 Visualize forward predicted MVs of all frames using @command{ffplay}:
7954 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7958 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7960 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7964 @section colorbalance
7965 Modify intensity of primary colors (red, green and blue) of input frames.
7967 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7968 regions for the red-cyan, green-magenta or blue-yellow balance.
7970 A positive adjustment value shifts the balance towards the primary color, a negative
7971 value towards the complementary color.
7973 The filter accepts the following options:
7979 Adjust red, green and blue shadows (darkest pixels).
7984 Adjust red, green and blue midtones (medium pixels).
7989 Adjust red, green and blue highlights (brightest pixels).
7991 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7994 Preserve lightness when changing color balance. Default is disabled.
7997 @subsection Examples
8001 Add red color cast to shadows:
8007 @subsection Commands
8009 This filter supports the all above options as @ref{commands}.
8011 @section colorchannelmixer
8013 Adjust video input frames by re-mixing color channels.
8015 This filter modifies a color channel by adding the values associated to
8016 the other channels of the same pixels. For example if the value to
8017 modify is red, the output value will be:
8019 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8022 The filter accepts the following options:
8029 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8030 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8036 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8037 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8043 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8044 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8050 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8051 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8053 Allowed ranges for options are @code{[-2.0, 2.0]}.
8056 @subsection Examples
8060 Convert source to grayscale:
8062 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8065 Simulate sepia tones:
8067 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8071 @subsection Commands
8073 This filter supports the all above options as @ref{commands}.
8076 RGB colorspace color keying.
8078 The filter accepts the following options:
8082 The color which will be replaced with transparency.
8085 Similarity percentage with the key color.
8087 0.01 matches only the exact key color, while 1.0 matches everything.
8092 0.0 makes pixels either fully transparent, or not transparent at all.
8094 Higher values result in semi-transparent pixels, with a higher transparency
8095 the more similar the pixels color is to the key color.
8098 @subsection Examples
8102 Make every green pixel in the input image transparent:
8104 ffmpeg -i input.png -vf colorkey=green out.png
8108 Overlay a greenscreen-video on top of a static background image.
8110 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
8114 @subsection Commands
8115 This filter supports same @ref{commands} as options.
8116 The command accepts the same syntax of the corresponding option.
8118 If the specified expression is not valid, it is kept at its current
8122 Remove all color information for all RGB colors except for certain one.
8124 The filter accepts the following options:
8128 The color which will not be replaced with neutral gray.
8131 Similarity percentage with the above color.
8132 0.01 matches only the exact key color, while 1.0 matches everything.
8135 Blend percentage. 0.0 makes pixels fully gray.
8136 Higher values result in more preserved color.
8139 @subsection Commands
8140 This filter supports same @ref{commands} as options.
8141 The command accepts the same syntax of the corresponding option.
8143 If the specified expression is not valid, it is kept at its current
8146 @section colorlevels
8148 Adjust video input frames using levels.
8150 The filter accepts the following options:
8157 Adjust red, green, blue and alpha input black point.
8158 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8164 Adjust red, green, blue and alpha input white point.
8165 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8167 Input levels are used to lighten highlights (bright tones), darken shadows
8168 (dark tones), change the balance of bright and dark tones.
8174 Adjust red, green, blue and alpha output black point.
8175 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8181 Adjust red, green, blue and alpha output white point.
8182 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8184 Output levels allows manual selection of a constrained output level range.
8187 @subsection Examples
8191 Make video output darker:
8193 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8199 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8203 Make video output lighter:
8205 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8209 Increase brightness:
8211 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8215 @subsection Commands
8217 This filter supports the all above options as @ref{commands}.
8219 @section colormatrix
8221 Convert color matrix.
8223 The filter accepts the following options:
8228 Specify the source and destination color matrix. Both values must be
8231 The accepted values are:
8259 For example to convert from BT.601 to SMPTE-240M, use the command:
8261 colormatrix=bt601:smpte240m
8266 Convert colorspace, transfer characteristics or color primaries.
8267 Input video needs to have an even size.
8269 The filter accepts the following options:
8274 Specify all color properties at once.
8276 The accepted values are:
8306 Specify output colorspace.
8308 The accepted values are:
8317 BT.470BG or BT.601-6 625
8320 SMPTE-170M or BT.601-6 525
8329 BT.2020 with non-constant luminance
8335 Specify output transfer characteristics.
8337 The accepted values are:
8349 Constant gamma of 2.2
8352 Constant gamma of 2.8
8355 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8373 BT.2020 for 10-bits content
8376 BT.2020 for 12-bits content
8382 Specify output color primaries.
8384 The accepted values are:
8393 BT.470BG or BT.601-6 625
8396 SMPTE-170M or BT.601-6 525
8420 Specify output color range.
8422 The accepted values are:
8425 TV (restricted) range
8428 MPEG (restricted) range
8439 Specify output color format.
8441 The accepted values are:
8444 YUV 4:2:0 planar 8-bits
8447 YUV 4:2:0 planar 10-bits
8450 YUV 4:2:0 planar 12-bits
8453 YUV 4:2:2 planar 8-bits
8456 YUV 4:2:2 planar 10-bits
8459 YUV 4:2:2 planar 12-bits
8462 YUV 4:4:4 planar 8-bits
8465 YUV 4:4:4 planar 10-bits
8468 YUV 4:4:4 planar 12-bits
8473 Do a fast conversion, which skips gamma/primary correction. This will take
8474 significantly less CPU, but will be mathematically incorrect. To get output
8475 compatible with that produced by the colormatrix filter, use fast=1.
8478 Specify dithering mode.
8480 The accepted values are:
8486 Floyd-Steinberg dithering
8490 Whitepoint adaptation mode.
8492 The accepted values are:
8495 Bradford whitepoint adaptation
8498 von Kries whitepoint adaptation
8501 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8505 Override all input properties at once. Same accepted values as @ref{all}.
8508 Override input colorspace. Same accepted values as @ref{space}.
8511 Override input color primaries. Same accepted values as @ref{primaries}.
8514 Override input transfer characteristics. Same accepted values as @ref{trc}.
8517 Override input color range. Same accepted values as @ref{range}.
8521 The filter converts the transfer characteristics, color space and color
8522 primaries to the specified user values. The output value, if not specified,
8523 is set to a default value based on the "all" property. If that property is
8524 also not specified, the filter will log an error. The output color range and
8525 format default to the same value as the input color range and format. The
8526 input transfer characteristics, color space, color primaries and color range
8527 should be set on the input data. If any of these are missing, the filter will
8528 log an error and no conversion will take place.
8530 For example to convert the input to SMPTE-240M, use the command:
8532 colorspace=smpte240m
8535 @section convolution
8537 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8539 The filter accepts the following options:
8546 Set matrix for each plane.
8547 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8548 and from 1 to 49 odd number of signed integers in @var{row} mode.
8554 Set multiplier for calculated value for each plane.
8555 If unset or 0, it will be sum of all matrix elements.
8561 Set bias for each plane. This value is added to the result of the multiplication.
8562 Useful for making the overall image brighter or darker. Default is 0.0.
8568 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8569 Default is @var{square}.
8572 @subsection Commands
8574 This filter supports the all above options as @ref{commands}.
8576 @subsection Examples
8582 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"
8588 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"
8594 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"
8600 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"
8604 Apply laplacian edge detector which includes diagonals:
8606 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"
8612 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"
8618 Apply 2D convolution of video stream in frequency domain using second stream
8621 The filter accepts the following options:
8625 Set which planes to process.
8628 Set which impulse video frames will be processed, can be @var{first}
8629 or @var{all}. Default is @var{all}.
8632 The @code{convolve} filter also supports the @ref{framesync} options.
8636 Copy the input video source unchanged to the output. This is mainly useful for
8641 Video filtering on GPU using Apple's CoreImage API on OSX.
8643 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8644 processed by video hardware. However, software-based OpenGL implementations
8645 exist which means there is no guarantee for hardware processing. It depends on
8648 There are many filters and image generators provided by Apple that come with a
8649 large variety of options. The filter has to be referenced by its name along
8652 The coreimage filter accepts the following options:
8655 List all available filters and generators along with all their respective
8656 options as well as possible minimum and maximum values along with the default
8663 Specify all filters by their respective name and options.
8664 Use @var{list_filters} to determine all valid filter names and options.
8665 Numerical options are specified by a float value and are automatically clamped
8666 to their respective value range. Vector and color options have to be specified
8667 by a list of space separated float values. Character escaping has to be done.
8668 A special option name @code{default} is available to use default options for a
8671 It is required to specify either @code{default} or at least one of the filter options.
8672 All omitted options are used with their default values.
8673 The syntax of the filter string is as follows:
8675 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8679 Specify a rectangle where the output of the filter chain is copied into the
8680 input image. It is given by a list of space separated float values:
8682 output_rect=x\ y\ width\ height
8684 If not given, the output rectangle equals the dimensions of the input image.
8685 The output rectangle is automatically cropped at the borders of the input
8686 image. Negative values are valid for each component.
8688 output_rect=25\ 25\ 100\ 100
8692 Several filters can be chained for successive processing without GPU-HOST
8693 transfers allowing for fast processing of complex filter chains.
8694 Currently, only filters with zero (generators) or exactly one (filters) input
8695 image and one output image are supported. Also, transition filters are not yet
8698 Some filters generate output images with additional padding depending on the
8699 respective filter kernel. The padding is automatically removed to ensure the
8700 filter output has the same size as the input image.
8702 For image generators, the size of the output image is determined by the
8703 previous output image of the filter chain or the input image of the whole
8704 filterchain, respectively. The generators do not use the pixel information of
8705 this image to generate their output. However, the generated output is
8706 blended onto this image, resulting in partial or complete coverage of the
8709 The @ref{coreimagesrc} video source can be used for generating input images
8710 which are directly fed into the filter chain. By using it, providing input
8711 images by another video source or an input video is not required.
8713 @subsection Examples
8718 List all filters available:
8720 coreimage=list_filters=true
8724 Use the CIBoxBlur filter with default options to blur an image:
8726 coreimage=filter=CIBoxBlur@@default
8730 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8731 its center at 100x100 and a radius of 50 pixels:
8733 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8737 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8738 given as complete and escaped command-line for Apple's standard bash shell:
8740 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8746 Cover a rectangular object
8748 It accepts the following options:
8752 Filepath of the optional cover image, needs to be in yuv420.
8757 It accepts the following values:
8760 cover it by the supplied image
8762 cover it by interpolating the surrounding pixels
8765 Default value is @var{blur}.
8768 @subsection Examples
8772 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8774 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8780 Crop the input video to given dimensions.
8782 It accepts the following parameters:
8786 The width of the output video. It defaults to @code{iw}.
8787 This expression is evaluated only once during the filter
8788 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8791 The height of the output video. It defaults to @code{ih}.
8792 This expression is evaluated only once during the filter
8793 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8796 The horizontal position, in the input video, of the left edge of the output
8797 video. It defaults to @code{(in_w-out_w)/2}.
8798 This expression is evaluated per-frame.
8801 The vertical position, in the input video, of the top edge of the output video.
8802 It defaults to @code{(in_h-out_h)/2}.
8803 This expression is evaluated per-frame.
8806 If set to 1 will force the output display aspect ratio
8807 to be the same of the input, by changing the output sample aspect
8808 ratio. It defaults to 0.
8811 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8812 width/height/x/y as specified and will not be rounded to nearest smaller value.
8816 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8817 expressions containing the following constants:
8822 The computed values for @var{x} and @var{y}. They are evaluated for
8827 The input width and height.
8831 These are the same as @var{in_w} and @var{in_h}.
8835 The output (cropped) width and height.
8839 These are the same as @var{out_w} and @var{out_h}.
8842 same as @var{iw} / @var{ih}
8845 input sample aspect ratio
8848 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8852 horizontal and vertical chroma subsample values. For example for the
8853 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8856 The number of the input frame, starting from 0.
8859 the position in the file of the input frame, NAN if unknown
8862 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8866 The expression for @var{out_w} may depend on the value of @var{out_h},
8867 and the expression for @var{out_h} may depend on @var{out_w}, but they
8868 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8869 evaluated after @var{out_w} and @var{out_h}.
8871 The @var{x} and @var{y} parameters specify the expressions for the
8872 position of the top-left corner of the output (non-cropped) area. They
8873 are evaluated for each frame. If the evaluated value is not valid, it
8874 is approximated to the nearest valid value.
8876 The expression for @var{x} may depend on @var{y}, and the expression
8877 for @var{y} may depend on @var{x}.
8879 @subsection Examples
8883 Crop area with size 100x100 at position (12,34).
8888 Using named options, the example above becomes:
8890 crop=w=100:h=100:x=12:y=34
8894 Crop the central input area with size 100x100:
8900 Crop the central input area with size 2/3 of the input video:
8902 crop=2/3*in_w:2/3*in_h
8906 Crop the input video central square:
8913 Delimit the rectangle with the top-left corner placed at position
8914 100:100 and the right-bottom corner corresponding to the right-bottom
8915 corner of the input image.
8917 crop=in_w-100:in_h-100:100:100
8921 Crop 10 pixels from the left and right borders, and 20 pixels from
8922 the top and bottom borders
8924 crop=in_w-2*10:in_h-2*20
8928 Keep only the bottom right quarter of the input image:
8930 crop=in_w/2:in_h/2:in_w/2:in_h/2
8934 Crop height for getting Greek harmony:
8936 crop=in_w:1/PHI*in_w
8940 Apply trembling effect:
8942 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)
8946 Apply erratic camera effect depending on timestamp:
8948 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)"
8952 Set x depending on the value of y:
8954 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8958 @subsection Commands
8960 This filter supports the following commands:
8966 Set width/height of the output video and the horizontal/vertical position
8968 The command accepts the same syntax of the corresponding option.
8970 If the specified expression is not valid, it is kept at its current
8976 Auto-detect the crop size.
8978 It calculates the necessary cropping parameters and prints the
8979 recommended parameters via the logging system. The detected dimensions
8980 correspond to the non-black area of the input video.
8982 It accepts the following parameters:
8987 Set higher black value threshold, which can be optionally specified
8988 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8989 value greater to the set value is considered non-black. It defaults to 24.
8990 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8991 on the bitdepth of the pixel format.
8994 The value which the width/height should be divisible by. It defaults to
8995 16. The offset is automatically adjusted to center the video. Use 2 to
8996 get only even dimensions (needed for 4:2:2 video). 16 is best when
8997 encoding to most video codecs.
9000 Set the number of initial frames for which evaluation is skipped.
9001 Default is 2. Range is 0 to INT_MAX.
9003 @item reset_count, reset
9004 Set the counter that determines after how many frames cropdetect will
9005 reset the previously detected largest video area and start over to
9006 detect the current optimal crop area. Default value is 0.
9008 This can be useful when channel logos distort the video area. 0
9009 indicates 'never reset', and returns the largest area encountered during
9016 Delay video filtering until a given wallclock timestamp. The filter first
9017 passes on @option{preroll} amount of frames, then it buffers at most
9018 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9019 it forwards the buffered frames and also any subsequent frames coming in its
9022 The filter can be used synchronize the output of multiple ffmpeg processes for
9023 realtime output devices like decklink. By putting the delay in the filtering
9024 chain and pre-buffering frames the process can pass on data to output almost
9025 immediately after the target wallclock timestamp is reached.
9027 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9033 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9036 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9039 The maximum duration of content to buffer before waiting for the cue expressed
9040 in seconds. Default is 0.
9047 Apply color adjustments using curves.
9049 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9050 component (red, green and blue) has its values defined by @var{N} key points
9051 tied from each other using a smooth curve. The x-axis represents the pixel
9052 values from the input frame, and the y-axis the new pixel values to be set for
9055 By default, a component curve is defined by the two points @var{(0;0)} and
9056 @var{(1;1)}. This creates a straight line where each original pixel value is
9057 "adjusted" to its own value, which means no change to the image.
9059 The filter allows you to redefine these two points and add some more. A new
9060 curve (using a natural cubic spline interpolation) will be define to pass
9061 smoothly through all these new coordinates. The new defined points needs to be
9062 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9063 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9064 the vector spaces, the values will be clipped accordingly.
9066 The filter accepts the following options:
9070 Select one of the available color presets. This option can be used in addition
9071 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9072 options takes priority on the preset values.
9073 Available presets are:
9076 @item color_negative
9079 @item increase_contrast
9081 @item linear_contrast
9082 @item medium_contrast
9084 @item strong_contrast
9087 Default is @code{none}.
9089 Set the master key points. These points will define a second pass mapping. It
9090 is sometimes called a "luminance" or "value" mapping. It can be used with
9091 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9092 post-processing LUT.
9094 Set the key points for the red component.
9096 Set the key points for the green component.
9098 Set the key points for the blue component.
9100 Set the key points for all components (not including master).
9101 Can be used in addition to the other key points component
9102 options. In this case, the unset component(s) will fallback on this
9103 @option{all} setting.
9105 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9107 Save Gnuplot script of the curves in specified file.
9110 To avoid some filtergraph syntax conflicts, each key points list need to be
9111 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9113 @subsection Examples
9117 Increase slightly the middle level of blue:
9119 curves=blue='0/0 0.5/0.58 1/1'
9125 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'
9127 Here we obtain the following coordinates for each components:
9130 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9132 @code{(0;0) (0.50;0.48) (1;1)}
9134 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9138 The previous example can also be achieved with the associated built-in preset:
9140 curves=preset=vintage
9150 Use a Photoshop preset and redefine the points of the green component:
9152 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9156 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9157 and @command{gnuplot}:
9159 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9160 gnuplot -p /tmp/curves.plt
9166 Video data analysis filter.
9168 This filter shows hexadecimal pixel values of part of video.
9170 The filter accepts the following options:
9174 Set output video size.
9177 Set x offset from where to pick pixels.
9180 Set y offset from where to pick pixels.
9183 Set scope mode, can be one of the following:
9186 Draw hexadecimal pixel values with white color on black background.
9189 Draw hexadecimal pixel values with input video pixel color on black
9193 Draw hexadecimal pixel values on color background picked from input video,
9194 the text color is picked in such way so its always visible.
9198 Draw rows and columns numbers on left and top of video.
9201 Set background opacity.
9204 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9208 Apply Directional blur filter.
9210 The filter accepts the following options:
9214 Set angle of directional blur. Default is @code{45}.
9217 Set radius of directional blur. Default is @code{5}.
9220 Set which planes to filter. By default all planes are filtered.
9223 @subsection Commands
9224 This filter supports same @ref{commands} as options.
9225 The command accepts the same syntax of the corresponding option.
9227 If the specified expression is not valid, it is kept at its current
9232 Denoise frames using 2D DCT (frequency domain filtering).
9234 This filter is not designed for real time.
9236 The filter accepts the following options:
9240 Set the noise sigma constant.
9242 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9243 coefficient (absolute value) below this threshold with be dropped.
9245 If you need a more advanced filtering, see @option{expr}.
9247 Default is @code{0}.
9250 Set number overlapping pixels for each block. Since the filter can be slow, you
9251 may want to reduce this value, at the cost of a less effective filter and the
9252 risk of various artefacts.
9254 If the overlapping value doesn't permit processing the whole input width or
9255 height, a warning will be displayed and according borders won't be denoised.
9257 Default value is @var{blocksize}-1, which is the best possible setting.
9260 Set the coefficient factor expression.
9262 For each coefficient of a DCT block, this expression will be evaluated as a
9263 multiplier value for the coefficient.
9265 If this is option is set, the @option{sigma} option will be ignored.
9267 The absolute value of the coefficient can be accessed through the @var{c}
9271 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9272 @var{blocksize}, which is the width and height of the processed blocks.
9274 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9275 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9276 on the speed processing. Also, a larger block size does not necessarily means a
9280 @subsection Examples
9282 Apply a denoise with a @option{sigma} of @code{4.5}:
9287 The same operation can be achieved using the expression system:
9289 dctdnoiz=e='gte(c, 4.5*3)'
9292 Violent denoise using a block size of @code{16x16}:
9299 Remove banding artifacts from input video.
9300 It works by replacing banded pixels with average value of referenced pixels.
9302 The filter accepts the following options:
9309 Set banding detection threshold for each plane. Default is 0.02.
9310 Valid range is 0.00003 to 0.5.
9311 If difference between current pixel and reference pixel is less than threshold,
9312 it will be considered as banded.
9315 Banding detection range in pixels. Default is 16. If positive, random number
9316 in range 0 to set value will be used. If negative, exact absolute value
9318 The range defines square of four pixels around current pixel.
9321 Set direction in radians from which four pixel will be compared. If positive,
9322 random direction from 0 to set direction will be picked. If negative, exact of
9323 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9324 will pick only pixels on same row and -PI/2 will pick only pixels on same
9328 If enabled, current pixel is compared with average value of all four
9329 surrounding pixels. The default is enabled. If disabled current pixel is
9330 compared with all four surrounding pixels. The pixel is considered banded
9331 if only all four differences with surrounding pixels are less than threshold.
9334 If enabled, current pixel is changed if and only if all pixel components are banded,
9335 e.g. banding detection threshold is triggered for all color components.
9336 The default is disabled.
9341 Remove blocking artifacts from input video.
9343 The filter accepts the following options:
9347 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9348 This controls what kind of deblocking is applied.
9351 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9357 Set blocking detection thresholds. Allowed range is 0 to 1.
9358 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9359 Using higher threshold gives more deblocking strength.
9360 Setting @var{alpha} controls threshold detection at exact edge of block.
9361 Remaining options controls threshold detection near the edge. Each one for
9362 below/above or left/right. Setting any of those to @var{0} disables
9366 Set planes to filter. Default is to filter all available planes.
9369 @subsection Examples
9373 Deblock using weak filter and block size of 4 pixels.
9375 deblock=filter=weak:block=4
9379 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9380 deblocking more edges.
9382 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9386 Similar as above, but filter only first plane.
9388 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9392 Similar as above, but filter only second and third plane.
9394 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9401 Drop duplicated frames at regular intervals.
9403 The filter accepts the following options:
9407 Set the number of frames from which one will be dropped. Setting this to
9408 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9409 Default is @code{5}.
9412 Set the threshold for duplicate detection. If the difference metric for a frame
9413 is less than or equal to this value, then it is declared as duplicate. Default
9417 Set scene change threshold. Default is @code{15}.
9421 Set the size of the x and y-axis blocks used during metric calculations.
9422 Larger blocks give better noise suppression, but also give worse detection of
9423 small movements. Must be a power of two. Default is @code{32}.
9426 Mark main input as a pre-processed input and activate clean source input
9427 stream. This allows the input to be pre-processed with various filters to help
9428 the metrics calculation while keeping the frame selection lossless. When set to
9429 @code{1}, the first stream is for the pre-processed input, and the second
9430 stream is the clean source from where the kept frames are chosen. Default is
9434 Set whether or not chroma is considered in the metric calculations. Default is
9440 Apply 2D deconvolution of video stream in frequency domain using second stream
9443 The filter accepts the following options:
9447 Set which planes to process.
9450 Set which impulse video frames will be processed, can be @var{first}
9451 or @var{all}. Default is @var{all}.
9454 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9455 and height are not same and not power of 2 or if stream prior to convolving
9459 The @code{deconvolve} filter also supports the @ref{framesync} options.
9463 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9465 It accepts the following options:
9469 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9470 @var{rainbows} for cross-color reduction.
9473 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9476 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9479 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9482 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9487 Apply deflate effect to the video.
9489 This filter replaces the pixel by the local(3x3) average by taking into account
9490 only values lower than the pixel.
9492 It accepts the following options:
9499 Limit the maximum change for each plane, default is 65535.
9500 If 0, plane will remain unchanged.
9503 @subsection Commands
9505 This filter supports the all above options as @ref{commands}.
9509 Remove temporal frame luminance variations.
9511 It accepts the following options:
9515 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9518 Set averaging mode to smooth temporal luminance variations.
9520 Available values are:
9545 Do not actually modify frame. Useful when one only wants metadata.
9550 Remove judder produced by partially interlaced telecined content.
9552 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9553 source was partially telecined content then the output of @code{pullup,dejudder}
9554 will have a variable frame rate. May change the recorded frame rate of the
9555 container. Aside from that change, this filter will not affect constant frame
9558 The option available in this filter is:
9562 Specify the length of the window over which the judder repeats.
9564 Accepts any integer greater than 1. Useful values are:
9568 If the original was telecined from 24 to 30 fps (Film to NTSC).
9571 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9574 If a mixture of the two.
9577 The default is @samp{4}.
9582 Suppress a TV station logo by a simple interpolation of the surrounding
9583 pixels. Just set a rectangle covering the logo and watch it disappear
9584 (and sometimes something even uglier appear - your mileage may vary).
9586 It accepts the following parameters:
9591 Specify the top left corner coordinates of the logo. They must be
9596 Specify the width and height of the logo to clear. They must be
9600 Specify the thickness of the fuzzy edge of the rectangle (added to
9601 @var{w} and @var{h}). The default value is 1. This option is
9602 deprecated, setting higher values should no longer be necessary and
9606 When set to 1, a green rectangle is drawn on the screen to simplify
9607 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9608 The default value is 0.
9610 The rectangle is drawn on the outermost pixels which will be (partly)
9611 replaced with interpolated values. The values of the next pixels
9612 immediately outside this rectangle in each direction will be used to
9613 compute the interpolated pixel values inside the rectangle.
9617 @subsection Examples
9621 Set a rectangle covering the area with top left corner coordinates 0,0
9622 and size 100x77, and a band of size 10:
9624 delogo=x=0:y=0:w=100:h=77:band=10
9632 Remove the rain in the input image/video by applying the derain methods based on
9633 convolutional neural networks. Supported models:
9637 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9638 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9641 Training as well as model generation scripts are provided in
9642 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9644 Native model files (.model) can be generated from TensorFlow model
9645 files (.pb) by using tools/python/convert.py
9647 The filter accepts the following options:
9651 Specify which filter to use. This option accepts the following values:
9655 Derain filter. To conduct derain filter, you need to use a derain model.
9658 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9660 Default value is @samp{derain}.
9663 Specify which DNN backend to use for model loading and execution. This option accepts
9664 the following values:
9668 Native implementation of DNN loading and execution.
9671 TensorFlow backend. To enable this backend you
9672 need to install the TensorFlow for C library (see
9673 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9674 @code{--enable-libtensorflow}
9676 Default value is @samp{native}.
9679 Set path to model file specifying network architecture and its parameters.
9680 Note that different backends use different file formats. TensorFlow and native
9681 backend can load files for only its format.
9684 It can also be finished with @ref{dnn_processing} filter.
9688 Attempt to fix small changes in horizontal and/or vertical shift. This
9689 filter helps remove camera shake from hand-holding a camera, bumping a
9690 tripod, moving on a vehicle, etc.
9692 The filter accepts the following options:
9700 Specify a rectangular area where to limit the search for motion
9702 If desired the search for motion vectors can be limited to a
9703 rectangular area of the frame defined by its top left corner, width
9704 and height. These parameters have the same meaning as the drawbox
9705 filter which can be used to visualise the position of the bounding
9708 This is useful when simultaneous movement of subjects within the frame
9709 might be confused for camera motion by the motion vector search.
9711 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9712 then the full frame is used. This allows later options to be set
9713 without specifying the bounding box for the motion vector search.
9715 Default - search the whole frame.
9719 Specify the maximum extent of movement in x and y directions in the
9720 range 0-64 pixels. Default 16.
9723 Specify how to generate pixels to fill blanks at the edge of the
9724 frame. Available values are:
9727 Fill zeroes at blank locations
9729 Original image at blank locations
9731 Extruded edge value at blank locations
9733 Mirrored edge at blank locations
9735 Default value is @samp{mirror}.
9738 Specify the blocksize to use for motion search. Range 4-128 pixels,
9742 Specify the contrast threshold for blocks. Only blocks with more than
9743 the specified contrast (difference between darkest and lightest
9744 pixels) will be considered. Range 1-255, default 125.
9747 Specify the search strategy. Available values are:
9750 Set exhaustive search
9752 Set less exhaustive search.
9754 Default value is @samp{exhaustive}.
9757 If set then a detailed log of the motion search is written to the
9764 Remove unwanted contamination of foreground colors, caused by reflected color of
9765 greenscreen or bluescreen.
9767 This filter accepts the following options:
9771 Set what type of despill to use.
9774 Set how spillmap will be generated.
9777 Set how much to get rid of still remaining spill.
9780 Controls amount of red in spill area.
9783 Controls amount of green in spill area.
9784 Should be -1 for greenscreen.
9787 Controls amount of blue in spill area.
9788 Should be -1 for bluescreen.
9791 Controls brightness of spill area, preserving colors.
9794 Modify alpha from generated spillmap.
9797 @subsection Commands
9799 This filter supports the all above options as @ref{commands}.
9803 Apply an exact inverse of the telecine operation. It requires a predefined
9804 pattern specified using the pattern option which must be the same as that passed
9805 to the telecine filter.
9807 This filter accepts the following options:
9816 The default value is @code{top}.
9820 A string of numbers representing the pulldown pattern you wish to apply.
9821 The default value is @code{23}.
9824 A number representing position of the first frame with respect to the telecine
9825 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9830 Apply dilation effect to the video.
9832 This filter replaces the pixel by the local(3x3) maximum.
9834 It accepts the following options:
9841 Limit the maximum change for each plane, default is 65535.
9842 If 0, plane will remain unchanged.
9845 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9848 Flags to local 3x3 coordinates maps like this:
9855 @subsection Commands
9857 This filter supports the all above options as @ref{commands}.
9861 Displace pixels as indicated by second and third input stream.
9863 It takes three input streams and outputs one stream, the first input is the
9864 source, and second and third input are displacement maps.
9866 The second input specifies how much to displace pixels along the
9867 x-axis, while the third input specifies how much to displace pixels
9869 If one of displacement map streams terminates, last frame from that
9870 displacement map will be used.
9872 Note that once generated, displacements maps can be reused over and over again.
9874 A description of the accepted options follows.
9878 Set displace behavior for pixels that are out of range.
9880 Available values are:
9883 Missing pixels are replaced by black pixels.
9886 Adjacent pixels will spread out to replace missing pixels.
9889 Out of range pixels are wrapped so they point to pixels of other side.
9892 Out of range pixels will be replaced with mirrored pixels.
9894 Default is @samp{smear}.
9898 @subsection Examples
9902 Add ripple effect to rgb input of video size hd720:
9904 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
9908 Add wave effect to rgb input of video size hd720:
9910 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
9914 @anchor{dnn_processing}
9915 @section dnn_processing
9917 Do image processing with deep neural networks. It works together with another filter
9918 which converts the pixel format of the Frame to what the dnn network requires.
9920 The filter accepts the following options:
9924 Specify which DNN backend to use for model loading and execution. This option accepts
9925 the following values:
9929 Native implementation of DNN loading and execution.
9932 TensorFlow backend. To enable this backend you
9933 need to install the TensorFlow for C library (see
9934 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9935 @code{--enable-libtensorflow}
9938 OpenVINO backend. To enable this backend you
9939 need to build and install the OpenVINO for C library (see
9940 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9941 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9942 be needed if the header files and libraries are not installed into system path)
9946 Default value is @samp{native}.
9949 Set path to model file specifying network architecture and its parameters.
9950 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9951 backend can load files for only its format.
9953 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9956 Set the input name of the dnn network.
9959 Set the output name of the dnn network.
9963 @subsection Examples
9967 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9969 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9973 Halve the pixel value of the frame with format gray32f:
9975 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
9979 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9981 ./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
9985 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9987 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9994 Draw a colored box on the input image.
9996 It accepts the following parameters:
10001 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10005 The expressions which specify the width and height of the box; if 0 they are interpreted as
10006 the input width and height. It defaults to 0.
10009 Specify the color of the box to write. For the general syntax of this option,
10010 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10011 value @code{invert} is used, the box edge color is the same as the
10012 video with inverted luma.
10015 The expression which sets the thickness of the box edge.
10016 A value of @code{fill} will create a filled box. Default value is @code{3}.
10018 See below for the list of accepted constants.
10021 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10022 will overwrite the video's color and alpha pixels.
10023 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10026 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10027 following constants:
10031 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10035 horizontal and vertical chroma subsample values. For example for the
10036 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10040 The input width and height.
10043 The input sample aspect ratio.
10047 The x and y offset coordinates where the box is drawn.
10051 The width and height of the drawn box.
10054 The thickness of the drawn box.
10056 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10057 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10061 @subsection Examples
10065 Draw a black box around the edge of the input image:
10071 Draw a box with color red and an opacity of 50%:
10073 drawbox=10:20:200:60:red@@0.5
10076 The previous example can be specified as:
10078 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10082 Fill the box with pink color:
10084 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10088 Draw a 2-pixel red 2.40:1 mask:
10090 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
10094 @subsection Commands
10095 This filter supports same commands as options.
10096 The command accepts the same syntax of the corresponding option.
10098 If the specified expression is not valid, it is kept at its current
10103 Draw a graph using input video metadata.
10105 It accepts the following parameters:
10109 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10112 Set 1st foreground color expression.
10115 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10118 Set 2nd foreground color expression.
10121 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10124 Set 3rd foreground color expression.
10127 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10130 Set 4th foreground color expression.
10133 Set minimal value of metadata value.
10136 Set maximal value of metadata value.
10139 Set graph background color. Default is white.
10144 Available values for mode is:
10151 Default is @code{line}.
10156 Available values for slide is:
10159 Draw new frame when right border is reached.
10162 Replace old columns with new ones.
10165 Scroll from right to left.
10168 Scroll from left to right.
10171 Draw single picture.
10174 Default is @code{frame}.
10177 Set size of graph video. For the syntax of this option, check the
10178 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10179 The default value is @code{900x256}.
10182 Set the output frame rate. Default value is @code{25}.
10184 The foreground color expressions can use the following variables:
10187 Minimal value of metadata value.
10190 Maximal value of metadata value.
10193 Current metadata key value.
10196 The color is defined as 0xAABBGGRR.
10199 Example using metadata from @ref{signalstats} filter:
10201 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10204 Example using metadata from @ref{ebur128} filter:
10206 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10211 Draw a grid on the input image.
10213 It accepts the following parameters:
10218 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10222 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10223 input width and height, respectively, minus @code{thickness}, so image gets
10224 framed. Default to 0.
10227 Specify the color of the grid. For the general syntax of this option,
10228 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10229 value @code{invert} is used, the grid color is the same as the
10230 video with inverted luma.
10233 The expression which sets the thickness of the grid line. Default value is @code{1}.
10235 See below for the list of accepted constants.
10238 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10239 will overwrite the video's color and alpha pixels.
10240 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10243 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10244 following constants:
10248 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10252 horizontal and vertical chroma subsample values. For example for the
10253 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10257 The input grid cell width and height.
10260 The input sample aspect ratio.
10264 The x and y coordinates of some point of grid intersection (meant to configure offset).
10268 The width and height of the drawn cell.
10271 The thickness of the drawn cell.
10273 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10274 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10278 @subsection Examples
10282 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10284 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10288 Draw a white 3x3 grid with an opacity of 50%:
10290 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10294 @subsection Commands
10295 This filter supports same commands as options.
10296 The command accepts the same syntax of the corresponding option.
10298 If the specified expression is not valid, it is kept at its current
10304 Draw a text string or text from a specified file on top of a video, using the
10305 libfreetype library.
10307 To enable compilation of this filter, you need to configure FFmpeg with
10308 @code{--enable-libfreetype}.
10309 To enable default font fallback and the @var{font} option you need to
10310 configure FFmpeg with @code{--enable-libfontconfig}.
10311 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10312 @code{--enable-libfribidi}.
10316 It accepts the following parameters:
10321 Used to draw a box around text using the background color.
10322 The value must be either 1 (enable) or 0 (disable).
10323 The default value of @var{box} is 0.
10326 Set the width of the border to be drawn around the box using @var{boxcolor}.
10327 The default value of @var{boxborderw} is 0.
10330 The color to be used for drawing box around text. For the syntax of this
10331 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10333 The default value of @var{boxcolor} is "white".
10336 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10337 The default value of @var{line_spacing} is 0.
10340 Set the width of the border to be drawn around the text using @var{bordercolor}.
10341 The default value of @var{borderw} is 0.
10344 Set the color to be used for drawing border around text. For the syntax of this
10345 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10347 The default value of @var{bordercolor} is "black".
10350 Select how the @var{text} is expanded. Can be either @code{none},
10351 @code{strftime} (deprecated) or
10352 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10356 Set a start time for the count. Value is in microseconds. Only applied
10357 in the deprecated strftime expansion mode. To emulate in normal expansion
10358 mode use the @code{pts} function, supplying the start time (in seconds)
10359 as the second argument.
10362 If true, check and fix text coords to avoid clipping.
10365 The color to be used for drawing fonts. For the syntax of this option, check
10366 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10368 The default value of @var{fontcolor} is "black".
10370 @item fontcolor_expr
10371 String which is expanded the same way as @var{text} to obtain dynamic
10372 @var{fontcolor} value. By default this option has empty value and is not
10373 processed. When this option is set, it overrides @var{fontcolor} option.
10376 The font family to be used for drawing text. By default Sans.
10379 The font file to be used for drawing text. The path must be included.
10380 This parameter is mandatory if the fontconfig support is disabled.
10383 Draw the text applying alpha blending. The value can
10384 be a number between 0.0 and 1.0.
10385 The expression accepts the same variables @var{x, y} as well.
10386 The default value is 1.
10387 Please see @var{fontcolor_expr}.
10390 The font size to be used for drawing text.
10391 The default value of @var{fontsize} is 16.
10394 If set to 1, attempt to shape the text (for example, reverse the order of
10395 right-to-left text and join Arabic characters) before drawing it.
10396 Otherwise, just draw the text exactly as given.
10397 By default 1 (if supported).
10399 @item ft_load_flags
10400 The flags to be used for loading the fonts.
10402 The flags map the corresponding flags supported by libfreetype, and are
10403 a combination of the following values:
10410 @item vertical_layout
10411 @item force_autohint
10414 @item ignore_global_advance_width
10416 @item ignore_transform
10418 @item linear_design
10422 Default value is "default".
10424 For more information consult the documentation for the FT_LOAD_*
10428 The color to be used for drawing a shadow behind the drawn text. For the
10429 syntax of this option, check the @ref{color syntax,,"Color" section in the
10430 ffmpeg-utils manual,ffmpeg-utils}.
10432 The default value of @var{shadowcolor} is "black".
10436 The x and y offsets for the text shadow position with respect to the
10437 position of the text. They can be either positive or negative
10438 values. The default value for both is "0".
10441 The starting frame number for the n/frame_num variable. The default value
10445 The size in number of spaces to use for rendering the tab.
10446 Default value is 4.
10449 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10450 format. It can be used with or without text parameter. @var{timecode_rate}
10451 option must be specified.
10453 @item timecode_rate, rate, r
10454 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10455 integer. Minimum value is "1".
10456 Drop-frame timecode is supported for frame rates 30 & 60.
10459 If set to 1, the output of the timecode option will wrap around at 24 hours.
10460 Default is 0 (disabled).
10463 The text string to be drawn. The text must be a sequence of UTF-8
10464 encoded characters.
10465 This parameter is mandatory if no file is specified with the parameter
10469 A text file containing text to be drawn. The text must be a sequence
10470 of UTF-8 encoded characters.
10472 This parameter is mandatory if no text string is specified with the
10473 parameter @var{text}.
10475 If both @var{text} and @var{textfile} are specified, an error is thrown.
10478 If set to 1, the @var{textfile} will be reloaded before each frame.
10479 Be sure to update it atomically, or it may be read partially, or even fail.
10483 The expressions which specify the offsets where text will be drawn
10484 within the video frame. They are relative to the top/left border of the
10487 The default value of @var{x} and @var{y} is "0".
10489 See below for the list of accepted constants and functions.
10492 The parameters for @var{x} and @var{y} are expressions containing the
10493 following constants and functions:
10497 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10501 horizontal and vertical chroma subsample values. For example for the
10502 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10505 the height of each text line
10513 @item max_glyph_a, ascent
10514 the maximum distance from the baseline to the highest/upper grid
10515 coordinate used to place a glyph outline point, for all the rendered
10517 It is a positive value, due to the grid's orientation with the Y axis
10520 @item max_glyph_d, descent
10521 the maximum distance from the baseline to the lowest grid coordinate
10522 used to place a glyph outline point, for all the rendered glyphs.
10523 This is a negative value, due to the grid's orientation, with the Y axis
10527 maximum glyph height, that is the maximum height for all the glyphs
10528 contained in the rendered text, it is equivalent to @var{ascent} -
10532 maximum glyph width, that is the maximum width for all the glyphs
10533 contained in the rendered text
10536 the number of input frame, starting from 0
10538 @item rand(min, max)
10539 return a random number included between @var{min} and @var{max}
10542 The input sample aspect ratio.
10545 timestamp expressed in seconds, NAN if the input timestamp is unknown
10548 the height of the rendered text
10551 the width of the rendered text
10555 the x and y offset coordinates where the text is drawn.
10557 These parameters allow the @var{x} and @var{y} expressions to refer
10558 to each other, so you can for example specify @code{y=x/dar}.
10561 A one character description of the current frame's picture type.
10564 The current packet's position in the input file or stream
10565 (in bytes, from the start of the input). A value of -1 indicates
10566 this info is not available.
10569 The current packet's duration, in seconds.
10572 The current packet's size (in bytes).
10575 @anchor{drawtext_expansion}
10576 @subsection Text expansion
10578 If @option{expansion} is set to @code{strftime},
10579 the filter recognizes strftime() sequences in the provided text and
10580 expands them accordingly. Check the documentation of strftime(). This
10581 feature is deprecated.
10583 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10585 If @option{expansion} is set to @code{normal} (which is the default),
10586 the following expansion mechanism is used.
10588 The backslash character @samp{\}, followed by any character, always expands to
10589 the second character.
10591 Sequences of the form @code{%@{...@}} are expanded. The text between the
10592 braces is a function name, possibly followed by arguments separated by ':'.
10593 If the arguments contain special characters or delimiters (':' or '@}'),
10594 they should be escaped.
10596 Note that they probably must also be escaped as the value for the
10597 @option{text} option in the filter argument string and as the filter
10598 argument in the filtergraph description, and possibly also for the shell,
10599 that makes up to four levels of escaping; using a text file avoids these
10602 The following functions are available:
10607 The expression evaluation result.
10609 It must take one argument specifying the expression to be evaluated,
10610 which accepts the same constants and functions as the @var{x} and
10611 @var{y} values. Note that not all constants should be used, for
10612 example the text size is not known when evaluating the expression, so
10613 the constants @var{text_w} and @var{text_h} will have an undefined
10616 @item expr_int_format, eif
10617 Evaluate the expression's value and output as formatted integer.
10619 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10620 The second argument specifies the output format. Allowed values are @samp{x},
10621 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10622 @code{printf} function.
10623 The third parameter is optional and sets the number of positions taken by the output.
10624 It can be used to add padding with zeros from the left.
10627 The time at which the filter is running, expressed in UTC.
10628 It can accept an argument: a strftime() format string.
10631 The time at which the filter is running, expressed in the local time zone.
10632 It can accept an argument: a strftime() format string.
10635 Frame metadata. Takes one or two arguments.
10637 The first argument is mandatory and specifies the metadata key.
10639 The second argument is optional and specifies a default value, used when the
10640 metadata key is not found or empty.
10642 Available metadata can be identified by inspecting entries
10643 starting with TAG included within each frame section
10644 printed by running @code{ffprobe -show_frames}.
10646 String metadata generated in filters leading to
10647 the drawtext filter are also available.
10650 The frame number, starting from 0.
10653 A one character description of the current picture type.
10656 The timestamp of the current frame.
10657 It can take up to three arguments.
10659 The first argument is the format of the timestamp; it defaults to @code{flt}
10660 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10661 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10662 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10663 @code{localtime} stands for the timestamp of the frame formatted as
10664 local time zone time.
10666 The second argument is an offset added to the timestamp.
10668 If the format is set to @code{hms}, a third argument @code{24HH} may be
10669 supplied to present the hour part of the formatted timestamp in 24h format
10672 If the format is set to @code{localtime} or @code{gmtime},
10673 a third argument may be supplied: a strftime() format string.
10674 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10677 @subsection Commands
10679 This filter supports altering parameters via commands:
10682 Alter existing filter parameters.
10684 Syntax for the argument is the same as for filter invocation, e.g.
10687 fontsize=56:fontcolor=green:text='Hello World'
10690 Full filter invocation with sendcmd would look like this:
10693 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10697 If the entire argument can't be parsed or applied as valid values then the filter will
10698 continue with its existing parameters.
10700 @subsection Examples
10704 Draw "Test Text" with font FreeSerif, using the default values for the
10705 optional parameters.
10708 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10712 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10713 and y=50 (counting from the top-left corner of the screen), text is
10714 yellow with a red box around it. Both the text and the box have an
10718 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10719 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10722 Note that the double quotes are not necessary if spaces are not used
10723 within the parameter list.
10726 Show the text at the center of the video frame:
10728 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10732 Show the text at a random position, switching to a new position every 30 seconds:
10734 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)"
10738 Show a text line sliding from right to left in the last row of the video
10739 frame. The file @file{LONG_LINE} is assumed to contain a single line
10742 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10746 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10748 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10752 Draw a single green letter "g", at the center of the input video.
10753 The glyph baseline is placed at half screen height.
10755 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10759 Show text for 1 second every 3 seconds:
10761 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10765 Use fontconfig to set the font. Note that the colons need to be escaped.
10767 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10771 Draw "Test Text" with font size dependent on height of the video.
10773 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10777 Print the date of a real-time encoding (see strftime(3)):
10779 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10783 Show text fading in and out (appearing/disappearing):
10786 DS=1.0 # display start
10787 DE=10.0 # display end
10788 FID=1.5 # fade in duration
10789 FOD=5 # fade out duration
10790 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 @}"
10794 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10795 and the @option{fontsize} value are included in the @option{y} offset.
10797 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10798 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10802 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10803 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10804 must have option @option{-export_path_metadata 1} for the special metadata fields
10805 to be available for filters.
10807 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10812 For more information about libfreetype, check:
10813 @url{http://www.freetype.org/}.
10815 For more information about fontconfig, check:
10816 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10818 For more information about libfribidi, check:
10819 @url{http://fribidi.org/}.
10821 @section edgedetect
10823 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10825 The filter accepts the following options:
10830 Set low and high threshold values used by the Canny thresholding
10833 The high threshold selects the "strong" edge pixels, which are then
10834 connected through 8-connectivity with the "weak" edge pixels selected
10835 by the low threshold.
10837 @var{low} and @var{high} threshold values must be chosen in the range
10838 [0,1], and @var{low} should be lesser or equal to @var{high}.
10840 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10844 Define the drawing mode.
10848 Draw white/gray wires on black background.
10851 Mix the colors to create a paint/cartoon effect.
10854 Apply Canny edge detector on all selected planes.
10856 Default value is @var{wires}.
10859 Select planes for filtering. By default all available planes are filtered.
10862 @subsection Examples
10866 Standard edge detection with custom values for the hysteresis thresholding:
10868 edgedetect=low=0.1:high=0.4
10872 Painting effect without thresholding:
10874 edgedetect=mode=colormix:high=0
10880 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10882 For each input image, the filter will compute the optimal mapping from
10883 the input to the output given the codebook length, that is the number
10884 of distinct output colors.
10886 This filter accepts the following options.
10889 @item codebook_length, l
10890 Set codebook length. The value must be a positive integer, and
10891 represents the number of distinct output colors. Default value is 256.
10894 Set the maximum number of iterations to apply for computing the optimal
10895 mapping. The higher the value the better the result and the higher the
10896 computation time. Default value is 1.
10899 Set a random seed, must be an integer included between 0 and
10900 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10901 will try to use a good random seed on a best effort basis.
10904 Set pal8 output pixel format. This option does not work with codebook
10905 length greater than 256.
10910 Measure graylevel entropy in histogram of color channels of video frames.
10912 It accepts the following parameters:
10916 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10918 @var{diff} mode measures entropy of histogram delta values, absolute differences
10919 between neighbour histogram values.
10923 Set brightness, contrast, saturation and approximate gamma adjustment.
10925 The filter accepts the following options:
10929 Set the contrast expression. The value must be a float value in range
10930 @code{-1000.0} to @code{1000.0}. The default value is "1".
10933 Set the brightness expression. The value must be a float value in
10934 range @code{-1.0} to @code{1.0}. The default value is "0".
10937 Set the saturation expression. The value must be a float in
10938 range @code{0.0} to @code{3.0}. The default value is "1".
10941 Set the gamma expression. The value must be a float in range
10942 @code{0.1} to @code{10.0}. The default value is "1".
10945 Set the gamma expression for red. The value must be a float in
10946 range @code{0.1} to @code{10.0}. The default value is "1".
10949 Set the gamma expression for green. The value must be a float in range
10950 @code{0.1} to @code{10.0}. The default value is "1".
10953 Set the gamma expression for blue. The value must be a float in range
10954 @code{0.1} to @code{10.0}. The default value is "1".
10957 Set the gamma weight expression. It can be used to reduce the effect
10958 of a high gamma value on bright image areas, e.g. keep them from
10959 getting overamplified and just plain white. The value must be a float
10960 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10961 gamma correction all the way down while @code{1.0} leaves it at its
10962 full strength. Default is "1".
10965 Set when the expressions for brightness, contrast, saturation and
10966 gamma expressions are evaluated.
10968 It accepts the following values:
10971 only evaluate expressions once during the filter initialization or
10972 when a command is processed
10975 evaluate expressions for each incoming frame
10978 Default value is @samp{init}.
10981 The expressions accept the following parameters:
10984 frame count of the input frame starting from 0
10987 byte position of the corresponding packet in the input file, NAN if
10991 frame rate of the input video, NAN if the input frame rate is unknown
10994 timestamp expressed in seconds, NAN if the input timestamp is unknown
10997 @subsection Commands
10998 The filter supports the following commands:
11002 Set the contrast expression.
11005 Set the brightness expression.
11008 Set the saturation expression.
11011 Set the gamma expression.
11014 Set the gamma_r expression.
11017 Set gamma_g expression.
11020 Set gamma_b expression.
11023 Set gamma_weight expression.
11025 The command accepts the same syntax of the corresponding option.
11027 If the specified expression is not valid, it is kept at its current
11034 Apply erosion effect to the video.
11036 This filter replaces the pixel by the local(3x3) minimum.
11038 It accepts the following options:
11045 Limit the maximum change for each plane, default is 65535.
11046 If 0, plane will remain unchanged.
11049 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11052 Flags to local 3x3 coordinates maps like this:
11059 @subsection Commands
11061 This filter supports the all above options as @ref{commands}.
11063 @section extractplanes
11065 Extract color channel components from input video stream into
11066 separate grayscale video streams.
11068 The filter accepts the following option:
11072 Set plane(s) to extract.
11074 Available values for planes are:
11085 Choosing planes not available in the input will result in an error.
11086 That means you cannot select @code{r}, @code{g}, @code{b} planes
11087 with @code{y}, @code{u}, @code{v} planes at same time.
11090 @subsection Examples
11094 Extract luma, u and v color channel component from input video frame
11095 into 3 grayscale outputs:
11097 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
11103 Apply a fade-in/out effect to the input video.
11105 It accepts the following parameters:
11109 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11111 Default is @code{in}.
11113 @item start_frame, s
11114 Specify the number of the frame to start applying the fade
11115 effect at. Default is 0.
11118 The number of frames that the fade effect lasts. At the end of the
11119 fade-in effect, the output video will have the same intensity as the input video.
11120 At the end of the fade-out transition, the output video will be filled with the
11121 selected @option{color}.
11125 If set to 1, fade only alpha channel, if one exists on the input.
11126 Default value is 0.
11128 @item start_time, st
11129 Specify the timestamp (in seconds) of the frame to start to apply the fade
11130 effect. If both start_frame and start_time are specified, the fade will start at
11131 whichever comes last. Default is 0.
11134 The number of seconds for which the fade effect has to last. At the end of the
11135 fade-in effect the output video will have the same intensity as the input video,
11136 at the end of the fade-out transition the output video will be filled with the
11137 selected @option{color}.
11138 If both duration and nb_frames are specified, duration is used. Default is 0
11139 (nb_frames is used by default).
11142 Specify the color of the fade. Default is "black".
11145 @subsection Examples
11149 Fade in the first 30 frames of video:
11154 The command above is equivalent to:
11160 Fade out the last 45 frames of a 200-frame video:
11163 fade=type=out:start_frame=155:nb_frames=45
11167 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11169 fade=in:0:25, fade=out:975:25
11173 Make the first 5 frames yellow, then fade in from frame 5-24:
11175 fade=in:5:20:color=yellow
11179 Fade in alpha over first 25 frames of video:
11181 fade=in:0:25:alpha=1
11185 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11187 fade=t=in:st=5.5:d=0.5
11193 Denoise frames using 3D FFT (frequency domain filtering).
11195 The filter accepts the following options:
11199 Set the noise sigma constant. This sets denoising strength.
11200 Default value is 1. Allowed range is from 0 to 30.
11201 Using very high sigma with low overlap may give blocking artifacts.
11204 Set amount of denoising. By default all detected noise is reduced.
11205 Default value is 1. Allowed range is from 0 to 1.
11208 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11209 Actual size of block in pixels is 2 to power of @var{block}, so by default
11210 block size in pixels is 2^4 which is 16.
11213 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11216 Set number of previous frames to use for denoising. By default is set to 0.
11219 Set number of next frames to to use for denoising. By default is set to 0.
11222 Set planes which will be filtered, by default are all available filtered
11227 Apply arbitrary expressions to samples in frequency domain
11231 Adjust the dc value (gain) of the luma plane of the image. The filter
11232 accepts an integer value in range @code{0} to @code{1000}. The default
11233 value is set to @code{0}.
11236 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11237 filter accepts an integer value in range @code{0} to @code{1000}. The
11238 default value is set to @code{0}.
11241 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11242 filter accepts an integer value in range @code{0} to @code{1000}. The
11243 default value is set to @code{0}.
11246 Set the frequency domain weight expression for the luma plane.
11249 Set the frequency domain weight expression for the 1st chroma plane.
11252 Set the frequency domain weight expression for the 2nd chroma plane.
11255 Set when the expressions are evaluated.
11257 It accepts the following values:
11260 Only evaluate expressions once during the filter initialization.
11263 Evaluate expressions for each incoming frame.
11266 Default value is @samp{init}.
11268 The filter accepts the following variables:
11271 The coordinates of the current sample.
11275 The width and height of the image.
11278 The number of input frame, starting from 0.
11281 @subsection Examples
11287 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11293 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11299 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11305 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11312 Extract a single field from an interlaced image using stride
11313 arithmetic to avoid wasting CPU time. The output frames are marked as
11316 The filter accepts the following options:
11320 Specify whether to extract the top (if the value is @code{0} or
11321 @code{top}) or the bottom field (if the value is @code{1} or
11327 Create new frames by copying the top and bottom fields from surrounding frames
11328 supplied as numbers by the hint file.
11332 Set file containing hints: absolute/relative frame numbers.
11334 There must be one line for each frame in a clip. Each line must contain two
11335 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11336 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11337 is current frame number for @code{absolute} mode or out of [-1, 1] range
11338 for @code{relative} mode. First number tells from which frame to pick up top
11339 field and second number tells from which frame to pick up bottom field.
11341 If optionally followed by @code{+} output frame will be marked as interlaced,
11342 else if followed by @code{-} output frame will be marked as progressive, else
11343 it will be marked same as input frame.
11344 If optionally followed by @code{t} output frame will use only top field, or in
11345 case of @code{b} it will use only bottom field.
11346 If line starts with @code{#} or @code{;} that line is skipped.
11349 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11352 Example of first several lines of @code{hint} file for @code{relative} mode:
11354 0,0 - # first frame
11355 1,0 - # second frame, use third's frame top field and second's frame bottom field
11356 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11371 @section fieldmatch
11373 Field matching filter for inverse telecine. It is meant to reconstruct the
11374 progressive frames from a telecined stream. The filter does not drop duplicated
11375 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11376 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11378 The separation of the field matching and the decimation is notably motivated by
11379 the possibility of inserting a de-interlacing filter fallback between the two.
11380 If the source has mixed telecined and real interlaced content,
11381 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11382 But these remaining combed frames will be marked as interlaced, and thus can be
11383 de-interlaced by a later filter such as @ref{yadif} before decimation.
11385 In addition to the various configuration options, @code{fieldmatch} can take an
11386 optional second stream, activated through the @option{ppsrc} option. If
11387 enabled, the frames reconstruction will be based on the fields and frames from
11388 this second stream. This allows the first input to be pre-processed in order to
11389 help the various algorithms of the filter, while keeping the output lossless
11390 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11391 or brightness/contrast adjustments can help.
11393 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11394 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11395 which @code{fieldmatch} is based on. While the semantic and usage are very
11396 close, some behaviour and options names can differ.
11398 The @ref{decimate} filter currently only works for constant frame rate input.
11399 If your input has mixed telecined (30fps) and progressive content with a lower
11400 framerate like 24fps use the following filterchain to produce the necessary cfr
11401 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11403 The filter accepts the following options:
11407 Specify the assumed field order of the input stream. Available values are:
11411 Auto detect parity (use FFmpeg's internal parity value).
11413 Assume bottom field first.
11415 Assume top field first.
11418 Note that it is sometimes recommended not to trust the parity announced by the
11421 Default value is @var{auto}.
11424 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11425 sense that it won't risk creating jerkiness due to duplicate frames when
11426 possible, but if there are bad edits or blended fields it will end up
11427 outputting combed frames when a good match might actually exist. On the other
11428 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11429 but will almost always find a good frame if there is one. The other values are
11430 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11431 jerkiness and creating duplicate frames versus finding good matches in sections
11432 with bad edits, orphaned fields, blended fields, etc.
11434 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11436 Available values are:
11440 2-way matching (p/c)
11442 2-way matching, and trying 3rd match if still combed (p/c + n)
11444 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11446 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11447 still combed (p/c + n + u/b)
11449 3-way matching (p/c/n)
11451 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11452 detected as combed (p/c/n + u/b)
11455 The parenthesis at the end indicate the matches that would be used for that
11456 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11459 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11462 Default value is @var{pc_n}.
11465 Mark the main input stream as a pre-processed input, and enable the secondary
11466 input stream as the clean source to pick the fields from. See the filter
11467 introduction for more details. It is similar to the @option{clip2} feature from
11470 Default value is @code{0} (disabled).
11473 Set the field to match from. It is recommended to set this to the same value as
11474 @option{order} unless you experience matching failures with that setting. In
11475 certain circumstances changing the field that is used to match from can have a
11476 large impact on matching performance. Available values are:
11480 Automatic (same value as @option{order}).
11482 Match from the bottom field.
11484 Match from the top field.
11487 Default value is @var{auto}.
11490 Set whether or not chroma is included during the match comparisons. In most
11491 cases it is recommended to leave this enabled. You should set this to @code{0}
11492 only if your clip has bad chroma problems such as heavy rainbowing or other
11493 artifacts. Setting this to @code{0} could also be used to speed things up at
11494 the cost of some accuracy.
11496 Default value is @code{1}.
11500 These define an exclusion band which excludes the lines between @option{y0} and
11501 @option{y1} from being included in the field matching decision. An exclusion
11502 band can be used to ignore subtitles, a logo, or other things that may
11503 interfere with the matching. @option{y0} sets the starting scan line and
11504 @option{y1} sets the ending line; all lines in between @option{y0} and
11505 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11506 @option{y0} and @option{y1} to the same value will disable the feature.
11507 @option{y0} and @option{y1} defaults to @code{0}.
11510 Set the scene change detection threshold as a percentage of maximum change on
11511 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11512 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11513 @option{scthresh} is @code{[0.0, 100.0]}.
11515 Default value is @code{12.0}.
11518 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11519 account the combed scores of matches when deciding what match to use as the
11520 final match. Available values are:
11524 No final matching based on combed scores.
11526 Combed scores are only used when a scene change is detected.
11528 Use combed scores all the time.
11531 Default is @var{sc}.
11534 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11535 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11536 Available values are:
11540 No forced calculation.
11542 Force p/c/n calculations.
11544 Force p/c/n/u/b calculations.
11547 Default value is @var{none}.
11550 This is the area combing threshold used for combed frame detection. This
11551 essentially controls how "strong" or "visible" combing must be to be detected.
11552 Larger values mean combing must be more visible and smaller values mean combing
11553 can be less visible or strong and still be detected. Valid settings are from
11554 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11555 be detected as combed). This is basically a pixel difference value. A good
11556 range is @code{[8, 12]}.
11558 Default value is @code{9}.
11561 Sets whether or not chroma is considered in the combed frame decision. Only
11562 disable this if your source has chroma problems (rainbowing, etc.) that are
11563 causing problems for the combed frame detection with chroma enabled. Actually,
11564 using @option{chroma}=@var{0} is usually more reliable, except for the case
11565 where there is chroma only combing in the source.
11567 Default value is @code{0}.
11571 Respectively set the x-axis and y-axis size of the window used during combed
11572 frame detection. This has to do with the size of the area in which
11573 @option{combpel} pixels are required to be detected as combed for a frame to be
11574 declared combed. See the @option{combpel} parameter description for more info.
11575 Possible values are any number that is a power of 2 starting at 4 and going up
11578 Default value is @code{16}.
11581 The number of combed pixels inside any of the @option{blocky} by
11582 @option{blockx} size blocks on the frame for the frame to be detected as
11583 combed. While @option{cthresh} controls how "visible" the combing must be, this
11584 setting controls "how much" combing there must be in any localized area (a
11585 window defined by the @option{blockx} and @option{blocky} settings) on the
11586 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11587 which point no frames will ever be detected as combed). This setting is known
11588 as @option{MI} in TFM/VFM vocabulary.
11590 Default value is @code{80}.
11593 @anchor{p/c/n/u/b meaning}
11594 @subsection p/c/n/u/b meaning
11596 @subsubsection p/c/n
11598 We assume the following telecined stream:
11601 Top fields: 1 2 2 3 4
11602 Bottom fields: 1 2 3 4 4
11605 The numbers correspond to the progressive frame the fields relate to. Here, the
11606 first two frames are progressive, the 3rd and 4th are combed, and so on.
11608 When @code{fieldmatch} is configured to run a matching from bottom
11609 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11614 B 1 2 3 4 4 <-- matching reference
11623 As a result of the field matching, we can see that some frames get duplicated.
11624 To perform a complete inverse telecine, you need to rely on a decimation filter
11625 after this operation. See for instance the @ref{decimate} filter.
11627 The same operation now matching from top fields (@option{field}=@var{top})
11632 T 1 2 2 3 4 <-- matching reference
11642 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11643 basically, they refer to the frame and field of the opposite parity:
11646 @item @var{p} matches the field of the opposite parity in the previous frame
11647 @item @var{c} matches the field of the opposite parity in the current frame
11648 @item @var{n} matches the field of the opposite parity in the next frame
11653 The @var{u} and @var{b} matching are a bit special in the sense that they match
11654 from the opposite parity flag. In the following examples, we assume that we are
11655 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11656 'x' is placed above and below each matched fields.
11658 With bottom matching (@option{field}=@var{bottom}):
11663 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11664 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11672 With top matching (@option{field}=@var{top}):
11677 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11678 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11686 @subsection Examples
11688 Simple IVTC of a top field first telecined stream:
11690 fieldmatch=order=tff:combmatch=none, decimate
11693 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11695 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11698 @section fieldorder
11700 Transform the field order of the input video.
11702 It accepts the following parameters:
11707 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11708 for bottom field first.
11711 The default value is @samp{tff}.
11713 The transformation is done by shifting the picture content up or down
11714 by one line, and filling the remaining line with appropriate picture content.
11715 This method is consistent with most broadcast field order converters.
11717 If the input video is not flagged as being interlaced, or it is already
11718 flagged as being of the required output field order, then this filter does
11719 not alter the incoming video.
11721 It is very useful when converting to or from PAL DV material,
11722 which is bottom field first.
11726 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11729 @section fifo, afifo
11731 Buffer input images and send them when they are requested.
11733 It is mainly useful when auto-inserted by the libavfilter
11736 It does not take parameters.
11738 @section fillborders
11740 Fill borders of the input video, without changing video stream dimensions.
11741 Sometimes video can have garbage at the four edges and you may not want to
11742 crop video input to keep size multiple of some number.
11744 This filter accepts the following options:
11748 Number of pixels to fill from left border.
11751 Number of pixels to fill from right border.
11754 Number of pixels to fill from top border.
11757 Number of pixels to fill from bottom border.
11762 It accepts the following values:
11765 fill pixels using outermost pixels
11768 fill pixels using mirroring (half sample symmetric)
11771 fill pixels with constant value
11774 fill pixels using reflecting (whole sample symmetric)
11777 fill pixels using wrapping
11780 fade pixels to constant value
11783 Default is @var{smear}.
11786 Set color for pixels in fixed or fade mode. Default is @var{black}.
11789 @subsection Commands
11790 This filter supports same @ref{commands} as options.
11791 The command accepts the same syntax of the corresponding option.
11793 If the specified expression is not valid, it is kept at its current
11798 Find a rectangular object
11800 It accepts the following options:
11804 Filepath of the object image, needs to be in gray8.
11807 Detection threshold, default is 0.5.
11810 Number of mipmaps, default is 3.
11812 @item xmin, ymin, xmax, ymax
11813 Specifies the rectangle in which to search.
11816 @subsection Examples
11820 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11822 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11828 Flood area with values of same pixel components with another values.
11830 It accepts the following options:
11833 Set pixel x coordinate.
11836 Set pixel y coordinate.
11839 Set source #0 component value.
11842 Set source #1 component value.
11845 Set source #2 component value.
11848 Set source #3 component value.
11851 Set destination #0 component value.
11854 Set destination #1 component value.
11857 Set destination #2 component value.
11860 Set destination #3 component value.
11866 Convert the input video to one of the specified pixel formats.
11867 Libavfilter will try to pick one that is suitable as input to
11870 It accepts the following parameters:
11874 A '|'-separated list of pixel format names, such as
11875 "pix_fmts=yuv420p|monow|rgb24".
11879 @subsection Examples
11883 Convert the input video to the @var{yuv420p} format
11885 format=pix_fmts=yuv420p
11888 Convert the input video to any of the formats in the list
11890 format=pix_fmts=yuv420p|yuv444p|yuv410p
11897 Convert the video to specified constant frame rate by duplicating or dropping
11898 frames as necessary.
11900 It accepts the following parameters:
11904 The desired output frame rate. The default is @code{25}.
11907 Assume the first PTS should be the given value, in seconds. This allows for
11908 padding/trimming at the start of stream. By default, no assumption is made
11909 about the first frame's expected PTS, so no padding or trimming is done.
11910 For example, this could be set to 0 to pad the beginning with duplicates of
11911 the first frame if a video stream starts after the audio stream or to trim any
11912 frames with a negative PTS.
11915 Timestamp (PTS) rounding method.
11917 Possible values are:
11924 round towards -infinity
11926 round towards +infinity
11930 The default is @code{near}.
11933 Action performed when reading the last frame.
11935 Possible values are:
11938 Use same timestamp rounding method as used for other frames.
11940 Pass through last frame if input duration has not been reached yet.
11942 The default is @code{round}.
11946 Alternatively, the options can be specified as a flat string:
11947 @var{fps}[:@var{start_time}[:@var{round}]].
11949 See also the @ref{setpts} filter.
11951 @subsection Examples
11955 A typical usage in order to set the fps to 25:
11961 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11963 fps=fps=film:round=near
11969 Pack two different video streams into a stereoscopic video, setting proper
11970 metadata on supported codecs. The two views should have the same size and
11971 framerate and processing will stop when the shorter video ends. Please note
11972 that you may conveniently adjust view properties with the @ref{scale} and
11975 It accepts the following parameters:
11979 The desired packing format. Supported values are:
11984 The views are next to each other (default).
11987 The views are on top of each other.
11990 The views are packed by line.
11993 The views are packed by column.
11996 The views are temporally interleaved.
12005 # Convert left and right views into a frame-sequential video
12006 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12008 # Convert views into a side-by-side video with the same output resolution as the input
12009 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
12014 Change the frame rate by interpolating new video output frames from the source
12017 This filter is not designed to function correctly with interlaced media. If
12018 you wish to change the frame rate of interlaced media then you are required
12019 to deinterlace before this filter and re-interlace after this filter.
12021 A description of the accepted options follows.
12025 Specify the output frames per second. This option can also be specified
12026 as a value alone. The default is @code{50}.
12029 Specify the start of a range where the output frame will be created as a
12030 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12031 the default is @code{15}.
12034 Specify the end of a range where the output frame will be created as a
12035 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12036 the default is @code{240}.
12039 Specify the level at which a scene change is detected as a value between
12040 0 and 100 to indicate a new scene; a low value reflects a low
12041 probability for the current frame to introduce a new scene, while a higher
12042 value means the current frame is more likely to be one.
12043 The default is @code{8.2}.
12046 Specify flags influencing the filter process.
12048 Available value for @var{flags} is:
12051 @item scene_change_detect, scd
12052 Enable scene change detection using the value of the option @var{scene}.
12053 This flag is enabled by default.
12059 Select one frame every N-th frame.
12061 This filter accepts the following option:
12064 Select frame after every @code{step} frames.
12065 Allowed values are positive integers higher than 0. Default value is @code{1}.
12068 @section freezedetect
12070 Detect frozen video.
12072 This filter logs a message and sets frame metadata when it detects that the
12073 input video has no significant change in content during a specified duration.
12074 Video freeze detection calculates the mean average absolute difference of all
12075 the components of video frames and compares it to a noise floor.
12077 The printed times and duration are expressed in seconds. The
12078 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12079 whose timestamp equals or exceeds the detection duration and it contains the
12080 timestamp of the first frame of the freeze. The
12081 @code{lavfi.freezedetect.freeze_duration} and
12082 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12085 The filter accepts the following options:
12089 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12090 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12094 Set freeze duration until notification (default is 2 seconds).
12097 @section freezeframes
12099 Freeze video frames.
12101 This filter freezes video frames using frame from 2nd input.
12103 The filter accepts the following options:
12107 Set number of first frame from which to start freeze.
12110 Set number of last frame from which to end freeze.
12113 Set number of frame from 2nd input which will be used instead of replaced frames.
12119 Apply a frei0r effect to the input video.
12121 To enable the compilation of this filter, you need to install the frei0r
12122 header and configure FFmpeg with @code{--enable-frei0r}.
12124 It accepts the following parameters:
12129 The name of the frei0r effect to load. If the environment variable
12130 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12131 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12132 Otherwise, the standard frei0r paths are searched, in this order:
12133 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12134 @file{/usr/lib/frei0r-1/}.
12136 @item filter_params
12137 A '|'-separated list of parameters to pass to the frei0r effect.
12141 A frei0r effect parameter can be a boolean (its value is either
12142 "y" or "n"), a double, a color (specified as
12143 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12144 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12145 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12146 a position (specified as @var{X}/@var{Y}, where
12147 @var{X} and @var{Y} are floating point numbers) and/or a string.
12149 The number and types of parameters depend on the loaded effect. If an
12150 effect parameter is not specified, the default value is set.
12152 @subsection Examples
12156 Apply the distort0r effect, setting the first two double parameters:
12158 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12162 Apply the colordistance effect, taking a color as the first parameter:
12164 frei0r=colordistance:0.2/0.3/0.4
12165 frei0r=colordistance:violet
12166 frei0r=colordistance:0x112233
12170 Apply the perspective effect, specifying the top left and top right image
12173 frei0r=perspective:0.2/0.2|0.8/0.2
12177 For more information, see
12178 @url{http://frei0r.dyne.org}
12180 @subsection Commands
12182 This filter supports the @option{filter_params} option as @ref{commands}.
12186 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12188 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12189 processing filter, one of them is performed once per block, not per pixel.
12190 This allows for much higher speed.
12192 The filter accepts the following options:
12196 Set quality. This option defines the number of levels for averaging. It accepts
12197 an integer in the range 4-5. Default value is @code{4}.
12200 Force a constant quantization parameter. It accepts an integer in range 0-63.
12201 If not set, the filter will use the QP from the video stream (if available).
12204 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12205 more details but also more artifacts, while higher values make the image smoother
12206 but also blurrier. Default value is @code{0} − PSNR optimal.
12208 @item use_bframe_qp
12209 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12210 option may cause flicker since the B-Frames have often larger QP. Default is
12211 @code{0} (not enabled).
12217 Apply Gaussian blur filter.
12219 The filter accepts the following options:
12223 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12226 Set number of steps for Gaussian approximation. Default is @code{1}.
12229 Set which planes to filter. By default all planes are filtered.
12232 Set vertical sigma, if negative it will be same as @code{sigma}.
12233 Default is @code{-1}.
12236 @subsection Commands
12237 This filter supports same commands as options.
12238 The command accepts the same syntax of the corresponding option.
12240 If the specified expression is not valid, it is kept at its current
12245 Apply generic equation to each pixel.
12247 The filter accepts the following options:
12250 @item lum_expr, lum
12251 Set the luminance expression.
12253 Set the chrominance blue expression.
12255 Set the chrominance red expression.
12256 @item alpha_expr, a
12257 Set the alpha expression.
12259 Set the red expression.
12260 @item green_expr, g
12261 Set the green expression.
12263 Set the blue expression.
12266 The colorspace is selected according to the specified options. If one
12267 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12268 options is specified, the filter will automatically select a YCbCr
12269 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12270 @option{blue_expr} options is specified, it will select an RGB
12273 If one of the chrominance expression is not defined, it falls back on the other
12274 one. If no alpha expression is specified it will evaluate to opaque value.
12275 If none of chrominance expressions are specified, they will evaluate
12276 to the luminance expression.
12278 The expressions can use the following variables and functions:
12282 The sequential number of the filtered frame, starting from @code{0}.
12286 The coordinates of the current sample.
12290 The width and height of the image.
12294 Width and height scale depending on the currently filtered plane. It is the
12295 ratio between the corresponding luma plane number of pixels and the current
12296 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12297 @code{0.5,0.5} for chroma planes.
12300 Time of the current frame, expressed in seconds.
12303 Return the value of the pixel at location (@var{x},@var{y}) of the current
12307 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12311 Return the value of the pixel at location (@var{x},@var{y}) of the
12312 blue-difference chroma plane. Return 0 if there is no such plane.
12315 Return the value of the pixel at location (@var{x},@var{y}) of the
12316 red-difference chroma plane. Return 0 if there is no such plane.
12321 Return the value of the pixel at location (@var{x},@var{y}) of the
12322 red/green/blue component. Return 0 if there is no such component.
12325 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12326 plane. Return 0 if there is no such plane.
12328 @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)
12329 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12330 sums of samples within a rectangle. See the functions without the sum postfix.
12332 @item interpolation
12333 Set one of interpolation methods:
12338 Default is bilinear.
12341 For functions, if @var{x} and @var{y} are outside the area, the value will be
12342 automatically clipped to the closer edge.
12344 Please note that this filter can use multiple threads in which case each slice
12345 will have its own expression state. If you want to use only a single expression
12346 state because your expressions depend on previous state then you should limit
12347 the number of filter threads to 1.
12349 @subsection Examples
12353 Flip the image horizontally:
12359 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12360 wavelength of 100 pixels:
12362 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12366 Generate a fancy enigmatic moving light:
12368 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
12372 Generate a quick emboss effect:
12374 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12378 Modify RGB components depending on pixel position:
12380 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12384 Create a radial gradient that is the same size as the input (also see
12385 the @ref{vignette} filter):
12387 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12393 Fix the banding artifacts that are sometimes introduced into nearly flat
12394 regions by truncation to 8-bit color depth.
12395 Interpolate the gradients that should go where the bands are, and
12398 It is designed for playback only. Do not use it prior to
12399 lossy compression, because compression tends to lose the dither and
12400 bring back the bands.
12402 It accepts the following parameters:
12407 The maximum amount by which the filter will change any one pixel. This is also
12408 the threshold for detecting nearly flat regions. Acceptable values range from
12409 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12413 The neighborhood to fit the gradient to. A larger radius makes for smoother
12414 gradients, but also prevents the filter from modifying the pixels near detailed
12415 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12416 values will be clipped to the valid range.
12420 Alternatively, the options can be specified as a flat string:
12421 @var{strength}[:@var{radius}]
12423 @subsection Examples
12427 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12433 Specify radius, omitting the strength (which will fall-back to the default
12441 @anchor{graphmonitor}
12442 @section graphmonitor
12443 Show various filtergraph stats.
12445 With this filter one can debug complete filtergraph.
12446 Especially issues with links filling with queued frames.
12448 The filter accepts the following options:
12452 Set video output size. Default is @var{hd720}.
12455 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12458 Set output mode, can be @var{fulll} or @var{compact}.
12459 In @var{compact} mode only filters with some queued frames have displayed stats.
12462 Set flags which enable which stats are shown in video.
12464 Available values for flags are:
12467 Display number of queued frames in each link.
12469 @item frame_count_in
12470 Display number of frames taken from filter.
12472 @item frame_count_out
12473 Display number of frames given out from filter.
12476 Display current filtered frame pts.
12479 Display current filtered frame time.
12482 Display time base for filter link.
12485 Display used format for filter link.
12488 Display video size or number of audio channels in case of audio used by filter link.
12491 Display video frame rate or sample rate in case of audio used by filter link.
12494 Display link output status.
12498 Set upper limit for video rate of output stream, Default value is @var{25}.
12499 This guarantee that output video frame rate will not be higher than this value.
12503 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12504 and corrects the scene colors accordingly.
12506 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12508 The filter accepts the following options:
12512 The order of differentiation to be applied on the scene. Must be chosen in the range
12513 [0,2] and default value is 1.
12516 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12517 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12518 max value instead of calculating Minkowski distance.
12521 The standard deviation of Gaussian blur to be applied on the scene. Must be
12522 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12523 can't be equal to 0 if @var{difford} is greater than 0.
12526 @subsection Examples
12532 greyedge=difford=1:minknorm=5:sigma=2
12538 greyedge=difford=1:minknorm=0:sigma=2
12546 Apply a Hald CLUT to a video stream.
12548 First input is the video stream to process, and second one is the Hald CLUT.
12549 The Hald CLUT input can be a simple picture or a complete video stream.
12551 The filter accepts the following options:
12555 Force termination when the shortest input terminates. Default is @code{0}.
12557 Continue applying the last CLUT after the end of the stream. A value of
12558 @code{0} disable the filter after the last frame of the CLUT is reached.
12559 Default is @code{1}.
12562 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12563 filters share the same internals).
12565 This filter also supports the @ref{framesync} options.
12567 More information about the Hald CLUT can be found on Eskil Steenberg's website
12568 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12570 @subsection Workflow examples
12572 @subsubsection Hald CLUT video stream
12574 Generate an identity Hald CLUT stream altered with various effects:
12576 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
12579 Note: make sure you use a lossless codec.
12581 Then use it with @code{haldclut} to apply it on some random stream:
12583 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12586 The Hald CLUT will be applied to the 10 first seconds (duration of
12587 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12588 to the remaining frames of the @code{mandelbrot} stream.
12590 @subsubsection Hald CLUT with preview
12592 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12593 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12594 biggest possible square starting at the top left of the picture. The remaining
12595 padding pixels (bottom or right) will be ignored. This area can be used to add
12596 a preview of the Hald CLUT.
12598 Typically, the following generated Hald CLUT will be supported by the
12599 @code{haldclut} filter:
12602 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12603 pad=iw+320 [padded_clut];
12604 smptebars=s=320x256, split [a][b];
12605 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12606 [main][b] overlay=W-320" -frames:v 1 clut.png
12609 It contains the original and a preview of the effect of the CLUT: SMPTE color
12610 bars are displayed on the right-top, and below the same color bars processed by
12613 Then, the effect of this Hald CLUT can be visualized with:
12615 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12620 Flip the input video horizontally.
12622 For example, to horizontally flip the input video with @command{ffmpeg}:
12624 ffmpeg -i in.avi -vf "hflip" out.avi
12628 This filter applies a global color histogram equalization on a
12631 It can be used to correct video that has a compressed range of pixel
12632 intensities. The filter redistributes the pixel intensities to
12633 equalize their distribution across the intensity range. It may be
12634 viewed as an "automatically adjusting contrast filter". This filter is
12635 useful only for correcting degraded or poorly captured source
12638 The filter accepts the following options:
12642 Determine the amount of equalization to be applied. As the strength
12643 is reduced, the distribution of pixel intensities more-and-more
12644 approaches that of the input frame. The value must be a float number
12645 in the range [0,1] and defaults to 0.200.
12648 Set the maximum intensity that can generated and scale the output
12649 values appropriately. The strength should be set as desired and then
12650 the intensity can be limited if needed to avoid washing-out. The value
12651 must be a float number in the range [0,1] and defaults to 0.210.
12654 Set the antibanding level. If enabled the filter will randomly vary
12655 the luminance of output pixels by a small amount to avoid banding of
12656 the histogram. Possible values are @code{none}, @code{weak} or
12657 @code{strong}. It defaults to @code{none}.
12663 Compute and draw a color distribution histogram for the input video.
12665 The computed histogram is a representation of the color component
12666 distribution in an image.
12668 Standard histogram displays the color components distribution in an image.
12669 Displays color graph for each color component. Shows distribution of
12670 the Y, U, V, A or R, G, B components, depending on input format, in the
12671 current frame. Below each graph a color component scale meter is shown.
12673 The filter accepts the following options:
12677 Set height of level. Default value is @code{200}.
12678 Allowed range is [50, 2048].
12681 Set height of color scale. Default value is @code{12}.
12682 Allowed range is [0, 40].
12686 It accepts the following values:
12689 Per color component graphs are placed below each other.
12692 Per color component graphs are placed side by side.
12695 Presents information identical to that in the @code{parade}, except
12696 that the graphs representing color components are superimposed directly
12699 Default is @code{stack}.
12702 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12703 Default is @code{linear}.
12706 Set what color components to display.
12707 Default is @code{7}.
12710 Set foreground opacity. Default is @code{0.7}.
12713 Set background opacity. Default is @code{0.5}.
12716 @subsection Examples
12721 Calculate and draw histogram:
12723 ffplay -i input -vf histogram
12731 This is a high precision/quality 3d denoise filter. It aims to reduce
12732 image noise, producing smooth images and making still images really
12733 still. It should enhance compressibility.
12735 It accepts the following optional parameters:
12739 A non-negative floating point number which specifies spatial luma strength.
12740 It defaults to 4.0.
12742 @item chroma_spatial
12743 A non-negative floating point number which specifies spatial chroma strength.
12744 It defaults to 3.0*@var{luma_spatial}/4.0.
12747 A floating point number which specifies luma temporal strength. It defaults to
12748 6.0*@var{luma_spatial}/4.0.
12751 A floating point number which specifies chroma temporal strength. It defaults to
12752 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12755 @subsection Commands
12756 This filter supports same @ref{commands} as options.
12757 The command accepts the same syntax of the corresponding option.
12759 If the specified expression is not valid, it is kept at its current
12762 @anchor{hwdownload}
12763 @section hwdownload
12765 Download hardware frames to system memory.
12767 The input must be in hardware frames, and the output a non-hardware format.
12768 Not all formats will be supported on the output - it may be necessary to insert
12769 an additional @option{format} filter immediately following in the graph to get
12770 the output in a supported format.
12774 Map hardware frames to system memory or to another device.
12776 This filter has several different modes of operation; which one is used depends
12777 on the input and output formats:
12780 Hardware frame input, normal frame output
12782 Map the input frames to system memory and pass them to the output. If the
12783 original hardware frame is later required (for example, after overlaying
12784 something else on part of it), the @option{hwmap} filter can be used again
12785 in the next mode to retrieve it.
12787 Normal frame input, hardware frame output
12789 If the input is actually a software-mapped hardware frame, then unmap it -
12790 that is, return the original hardware frame.
12792 Otherwise, a device must be provided. Create new hardware surfaces on that
12793 device for the output, then map them back to the software format at the input
12794 and give those frames to the preceding filter. This will then act like the
12795 @option{hwupload} filter, but may be able to avoid an additional copy when
12796 the input is already in a compatible format.
12798 Hardware frame input and output
12800 A device must be supplied for the output, either directly or with the
12801 @option{derive_device} option. The input and output devices must be of
12802 different types and compatible - the exact meaning of this is
12803 system-dependent, but typically it means that they must refer to the same
12804 underlying hardware context (for example, refer to the same graphics card).
12806 If the input frames were originally created on the output device, then unmap
12807 to retrieve the original frames.
12809 Otherwise, map the frames to the output device - create new hardware frames
12810 on the output corresponding to the frames on the input.
12813 The following additional parameters are accepted:
12817 Set the frame mapping mode. Some combination of:
12820 The mapped frame should be readable.
12822 The mapped frame should be writeable.
12824 The mapping will always overwrite the entire frame.
12826 This may improve performance in some cases, as the original contents of the
12827 frame need not be loaded.
12829 The mapping must not involve any copying.
12831 Indirect mappings to copies of frames are created in some cases where either
12832 direct mapping is not possible or it would have unexpected properties.
12833 Setting this flag ensures that the mapping is direct and will fail if that is
12836 Defaults to @var{read+write} if not specified.
12838 @item derive_device @var{type}
12839 Rather than using the device supplied at initialisation, instead derive a new
12840 device of type @var{type} from the device the input frames exist on.
12843 In a hardware to hardware mapping, map in reverse - create frames in the sink
12844 and map them back to the source. This may be necessary in some cases where
12845 a mapping in one direction is required but only the opposite direction is
12846 supported by the devices being used.
12848 This option is dangerous - it may break the preceding filter in undefined
12849 ways if there are any additional constraints on that filter's output.
12850 Do not use it without fully understanding the implications of its use.
12856 Upload system memory frames to hardware surfaces.
12858 The device to upload to must be supplied when the filter is initialised. If
12859 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12860 option or with the @option{derive_device} option. The input and output devices
12861 must be of different types and compatible - the exact meaning of this is
12862 system-dependent, but typically it means that they must refer to the same
12863 underlying hardware context (for example, refer to the same graphics card).
12865 The following additional parameters are accepted:
12868 @item derive_device @var{type}
12869 Rather than using the device supplied at initialisation, instead derive a new
12870 device of type @var{type} from the device the input frames exist on.
12873 @anchor{hwupload_cuda}
12874 @section hwupload_cuda
12876 Upload system memory frames to a CUDA device.
12878 It accepts the following optional parameters:
12882 The number of the CUDA device to use
12887 Apply a high-quality magnification filter designed for pixel art. This filter
12888 was originally created by Maxim Stepin.
12890 It accepts the following option:
12894 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12895 @code{hq3x} and @code{4} for @code{hq4x}.
12896 Default is @code{3}.
12900 Stack input videos horizontally.
12902 All streams must be of same pixel format and of same height.
12904 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12905 to create same output.
12907 The filter accepts the following option:
12911 Set number of input streams. Default is 2.
12914 If set to 1, force the output to terminate when the shortest input
12915 terminates. Default value is 0.
12920 Modify the hue and/or the saturation of the input.
12922 It accepts the following parameters:
12926 Specify the hue angle as a number of degrees. It accepts an expression,
12927 and defaults to "0".
12930 Specify the saturation in the [-10,10] range. It accepts an expression and
12934 Specify the hue angle as a number of radians. It accepts an
12935 expression, and defaults to "0".
12938 Specify the brightness in the [-10,10] range. It accepts an expression and
12942 @option{h} and @option{H} are mutually exclusive, and can't be
12943 specified at the same time.
12945 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12946 expressions containing the following constants:
12950 frame count of the input frame starting from 0
12953 presentation timestamp of the input frame expressed in time base units
12956 frame rate of the input video, NAN if the input frame rate is unknown
12959 timestamp expressed in seconds, NAN if the input timestamp is unknown
12962 time base of the input video
12965 @subsection Examples
12969 Set the hue to 90 degrees and the saturation to 1.0:
12975 Same command but expressing the hue in radians:
12981 Rotate hue and make the saturation swing between 0
12982 and 2 over a period of 1 second:
12984 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12988 Apply a 3 seconds saturation fade-in effect starting at 0:
12990 hue="s=min(t/3\,1)"
12993 The general fade-in expression can be written as:
12995 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12999 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13001 hue="s=max(0\, min(1\, (8-t)/3))"
13004 The general fade-out expression can be written as:
13006 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13011 @subsection Commands
13013 This filter supports the following commands:
13019 Modify the hue and/or the saturation and/or brightness of the input video.
13020 The command accepts the same syntax of the corresponding option.
13022 If the specified expression is not valid, it is kept at its current
13026 @section hysteresis
13028 Grow first stream into second stream by connecting components.
13029 This makes it possible to build more robust edge masks.
13031 This filter accepts the following options:
13035 Set which planes will be processed as bitmap, unprocessed planes will be
13036 copied from first stream.
13037 By default value 0xf, all planes will be processed.
13040 Set threshold which is used in filtering. If pixel component value is higher than
13041 this value filter algorithm for connecting components is activated.
13042 By default value is 0.
13045 The @code{hysteresis} filter also supports the @ref{framesync} options.
13049 Detect video interlacing type.
13051 This filter tries to detect if the input frames are interlaced, progressive,
13052 top or bottom field first. It will also try to detect fields that are
13053 repeated between adjacent frames (a sign of telecine).
13055 Single frame detection considers only immediately adjacent frames when classifying each frame.
13056 Multiple frame detection incorporates the classification history of previous frames.
13058 The filter will log these metadata values:
13061 @item single.current_frame
13062 Detected type of current frame using single-frame detection. One of:
13063 ``tff'' (top field first), ``bff'' (bottom field first),
13064 ``progressive'', or ``undetermined''
13067 Cumulative number of frames detected as top field first using single-frame detection.
13070 Cumulative number of frames detected as top field first using multiple-frame detection.
13073 Cumulative number of frames detected as bottom field first using single-frame detection.
13075 @item multiple.current_frame
13076 Detected type of current frame using multiple-frame detection. One of:
13077 ``tff'' (top field first), ``bff'' (bottom field first),
13078 ``progressive'', or ``undetermined''
13081 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13083 @item single.progressive
13084 Cumulative number of frames detected as progressive using single-frame detection.
13086 @item multiple.progressive
13087 Cumulative number of frames detected as progressive using multiple-frame detection.
13089 @item single.undetermined
13090 Cumulative number of frames that could not be classified using single-frame detection.
13092 @item multiple.undetermined
13093 Cumulative number of frames that could not be classified using multiple-frame detection.
13095 @item repeated.current_frame
13096 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13098 @item repeated.neither
13099 Cumulative number of frames with no repeated field.
13102 Cumulative number of frames with the top field repeated from the previous frame's top field.
13104 @item repeated.bottom
13105 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13108 The filter accepts the following options:
13112 Set interlacing threshold.
13114 Set progressive threshold.
13116 Threshold for repeated field detection.
13118 Number of frames after which a given frame's contribution to the
13119 statistics is halved (i.e., it contributes only 0.5 to its
13120 classification). The default of 0 means that all frames seen are given
13121 full weight of 1.0 forever.
13122 @item analyze_interlaced_flag
13123 When this is not 0 then idet will use the specified number of frames to determine
13124 if the interlaced flag is accurate, it will not count undetermined frames.
13125 If the flag is found to be accurate it will be used without any further
13126 computations, if it is found to be inaccurate it will be cleared without any
13127 further computations. This allows inserting the idet filter as a low computational
13128 method to clean up the interlaced flag
13133 Deinterleave or interleave fields.
13135 This filter allows one to process interlaced images fields without
13136 deinterlacing them. Deinterleaving splits the input frame into 2
13137 fields (so called half pictures). Odd lines are moved to the top
13138 half of the output image, even lines to the bottom half.
13139 You can process (filter) them independently and then re-interleave them.
13141 The filter accepts the following options:
13145 @item chroma_mode, c
13146 @item alpha_mode, a
13147 Available values for @var{luma_mode}, @var{chroma_mode} and
13148 @var{alpha_mode} are:
13154 @item deinterleave, d
13155 Deinterleave fields, placing one above the other.
13157 @item interleave, i
13158 Interleave fields. Reverse the effect of deinterleaving.
13160 Default value is @code{none}.
13162 @item luma_swap, ls
13163 @item chroma_swap, cs
13164 @item alpha_swap, as
13165 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13168 @subsection Commands
13170 This filter supports the all above options as @ref{commands}.
13174 Apply inflate effect to the video.
13176 This filter replaces the pixel by the local(3x3) average by taking into account
13177 only values higher than the pixel.
13179 It accepts the following options:
13186 Limit the maximum change for each plane, default is 65535.
13187 If 0, plane will remain unchanged.
13190 @subsection Commands
13192 This filter supports the all above options as @ref{commands}.
13196 Simple interlacing filter from progressive contents. This interleaves upper (or
13197 lower) lines from odd frames with lower (or upper) lines from even frames,
13198 halving the frame rate and preserving image height.
13201 Original Original New Frame
13202 Frame 'j' Frame 'j+1' (tff)
13203 ========== =========== ==================
13204 Line 0 --------------------> Frame 'j' Line 0
13205 Line 1 Line 1 ----> Frame 'j+1' Line 1
13206 Line 2 ---------------------> Frame 'j' Line 2
13207 Line 3 Line 3 ----> Frame 'j+1' Line 3
13209 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13212 It accepts the following optional parameters:
13216 This determines whether the interlaced frame is taken from the even
13217 (tff - default) or odd (bff) lines of the progressive frame.
13220 Vertical lowpass filter to avoid twitter interlacing and
13221 reduce moire patterns.
13225 Disable vertical lowpass filter
13228 Enable linear filter (default)
13231 Enable complex filter. This will slightly less reduce twitter and moire
13232 but better retain detail and subjective sharpness impression.
13239 Deinterlace input video by applying Donald Graft's adaptive kernel
13240 deinterling. Work on interlaced parts of a video to produce
13241 progressive frames.
13243 The description of the accepted parameters follows.
13247 Set the threshold which affects the filter's tolerance when
13248 determining if a pixel line must be processed. It must be an integer
13249 in the range [0,255] and defaults to 10. A value of 0 will result in
13250 applying the process on every pixels.
13253 Paint pixels exceeding the threshold value to white if set to 1.
13257 Set the fields order. Swap fields if set to 1, leave fields alone if
13261 Enable additional sharpening if set to 1. Default is 0.
13264 Enable twoway sharpening if set to 1. Default is 0.
13267 @subsection Examples
13271 Apply default values:
13273 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13277 Enable additional sharpening:
13283 Paint processed pixels in white:
13291 Slowly update darker pixels.
13293 This filter makes short flashes of light appear longer.
13294 This filter accepts the following options:
13298 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13301 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13304 @section lenscorrection
13306 Correct radial lens distortion
13308 This filter can be used to correct for radial distortion as can result from the use
13309 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13310 one can use tools available for example as part of opencv or simply trial-and-error.
13311 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13312 and extract the k1 and k2 coefficients from the resulting matrix.
13314 Note that effectively the same filter is available in the open-source tools Krita and
13315 Digikam from the KDE project.
13317 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13318 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13319 brightness distribution, so you may want to use both filters together in certain
13320 cases, though you will have to take care of ordering, i.e. whether vignetting should
13321 be applied before or after lens correction.
13323 @subsection Options
13325 The filter accepts the following options:
13329 Relative x-coordinate of the focal point of the image, and thereby the center of the
13330 distortion. This value has a range [0,1] and is expressed as fractions of the image
13331 width. Default is 0.5.
13333 Relative y-coordinate of the focal point of the image, and thereby the center of the
13334 distortion. This value has a range [0,1] and is expressed as fractions of the image
13335 height. Default is 0.5.
13337 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13338 no correction. Default is 0.
13340 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13341 0 means no correction. Default is 0.
13344 The formula that generates the correction is:
13346 @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)
13348 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13349 distances from the focal point in the source and target images, respectively.
13353 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13355 The @code{lensfun} filter requires the camera make, camera model, and lens model
13356 to apply the lens correction. The filter will load the lensfun database and
13357 query it to find the corresponding camera and lens entries in the database. As
13358 long as these entries can be found with the given options, the filter can
13359 perform corrections on frames. Note that incomplete strings will result in the
13360 filter choosing the best match with the given options, and the filter will
13361 output the chosen camera and lens models (logged with level "info"). You must
13362 provide the make, camera model, and lens model as they are required.
13364 The filter accepts the following options:
13368 The make of the camera (for example, "Canon"). This option is required.
13371 The model of the camera (for example, "Canon EOS 100D"). This option is
13375 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13376 option is required.
13379 The type of correction to apply. The following values are valid options:
13383 Enables fixing lens vignetting.
13386 Enables fixing lens geometry. This is the default.
13389 Enables fixing chromatic aberrations.
13392 Enables fixing lens vignetting and lens geometry.
13395 Enables fixing lens vignetting and chromatic aberrations.
13398 Enables fixing both lens geometry and chromatic aberrations.
13401 Enables all possible corrections.
13405 The focal length of the image/video (zoom; expected constant for video). For
13406 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13407 range should be chosen when using that lens. Default 18.
13410 The aperture of the image/video (expected constant for video). Note that
13411 aperture is only used for vignetting correction. Default 3.5.
13413 @item focus_distance
13414 The focus distance of the image/video (expected constant for video). Note that
13415 focus distance is only used for vignetting and only slightly affects the
13416 vignetting correction process. If unknown, leave it at the default value (which
13420 The scale factor which is applied after transformation. After correction the
13421 video is no longer necessarily rectangular. This parameter controls how much of
13422 the resulting image is visible. The value 0 means that a value will be chosen
13423 automatically such that there is little or no unmapped area in the output
13424 image. 1.0 means that no additional scaling is done. Lower values may result
13425 in more of the corrected image being visible, while higher values may avoid
13426 unmapped areas in the output.
13428 @item target_geometry
13429 The target geometry of the output image/video. The following values are valid
13433 @item rectilinear (default)
13436 @item equirectangular
13437 @item fisheye_orthographic
13438 @item fisheye_stereographic
13439 @item fisheye_equisolid
13440 @item fisheye_thoby
13443 Apply the reverse of image correction (instead of correcting distortion, apply
13446 @item interpolation
13447 The type of interpolation used when correcting distortion. The following values
13452 @item linear (default)
13457 @subsection Examples
13461 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13462 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13466 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
13470 Apply the same as before, but only for the first 5 seconds of video.
13473 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
13480 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13481 score between two input videos.
13483 The obtained VMAF score is printed through the logging system.
13485 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13486 After installing the library it can be enabled using:
13487 @code{./configure --enable-libvmaf}.
13488 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13490 The filter has following options:
13494 Set the model path which is to be used for SVM.
13495 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13498 Set the file path to be used to store logs.
13501 Set the format of the log file (csv, json or xml).
13503 @item enable_transform
13504 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13505 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13506 Default value: @code{false}
13509 Invokes the phone model which will generate VMAF scores higher than in the
13510 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13511 Default value: @code{false}
13514 Enables computing psnr along with vmaf.
13515 Default value: @code{false}
13518 Enables computing ssim along with vmaf.
13519 Default value: @code{false}
13522 Enables computing ms_ssim along with vmaf.
13523 Default value: @code{false}
13526 Set the pool method to be used for computing vmaf.
13527 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13530 Set number of threads to be used when computing vmaf.
13531 Default value: @code{0}, which makes use of all available logical processors.
13534 Set interval for frame subsampling used when computing vmaf.
13535 Default value: @code{1}
13537 @item enable_conf_interval
13538 Enables confidence interval.
13539 Default value: @code{false}
13542 This filter also supports the @ref{framesync} options.
13544 @subsection Examples
13547 On the below examples the input file @file{main.mpg} being processed is
13548 compared with the reference file @file{ref.mpg}.
13551 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13555 Example with options:
13557 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13561 Example with options and different containers:
13563 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 -
13569 Limits the pixel components values to the specified range [min, max].
13571 The filter accepts the following options:
13575 Lower bound. Defaults to the lowest allowed value for the input.
13578 Upper bound. Defaults to the highest allowed value for the input.
13581 Specify which planes will be processed. Defaults to all available.
13584 @subsection Commands
13586 This filter supports the all above options as @ref{commands}.
13592 The filter accepts the following options:
13596 Set the number of loops. Setting this value to -1 will result in infinite loops.
13600 Set maximal size in number of frames. Default is 0.
13603 Set first frame of loop. Default is 0.
13606 @subsection Examples
13610 Loop single first frame infinitely:
13612 loop=loop=-1:size=1:start=0
13616 Loop single first frame 10 times:
13618 loop=loop=10:size=1:start=0
13622 Loop 10 first frames 5 times:
13624 loop=loop=5:size=10:start=0
13630 Apply a 1D LUT to an input video.
13632 The filter accepts the following options:
13636 Set the 1D LUT file name.
13638 Currently supported formats:
13647 Select interpolation mode.
13649 Available values are:
13653 Use values from the nearest defined point.
13655 Interpolate values using the linear interpolation.
13657 Interpolate values using the cosine interpolation.
13659 Interpolate values using the cubic interpolation.
13661 Interpolate values using the spline interpolation.
13668 Apply a 3D LUT to an input video.
13670 The filter accepts the following options:
13674 Set the 3D LUT file name.
13676 Currently supported formats:
13690 Select interpolation mode.
13692 Available values are:
13696 Use values from the nearest defined point.
13698 Interpolate values using the 8 points defining a cube.
13700 Interpolate values using a tetrahedron.
13706 Turn certain luma values into transparency.
13708 The filter accepts the following options:
13712 Set the luma which will be used as base for transparency.
13713 Default value is @code{0}.
13716 Set the range of luma values to be keyed out.
13717 Default value is @code{0.01}.
13720 Set the range of softness. Default value is @code{0}.
13721 Use this to control gradual transition from zero to full transparency.
13724 @subsection Commands
13725 This filter supports same @ref{commands} as options.
13726 The command accepts the same syntax of the corresponding option.
13728 If the specified expression is not valid, it is kept at its current
13731 @section lut, lutrgb, lutyuv
13733 Compute a look-up table for binding each pixel component input value
13734 to an output value, and apply it to the input video.
13736 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13737 to an RGB input video.
13739 These filters accept the following parameters:
13742 set first pixel component expression
13744 set second pixel component expression
13746 set third pixel component expression
13748 set fourth pixel component expression, corresponds to the alpha component
13751 set red component expression
13753 set green component expression
13755 set blue component expression
13757 alpha component expression
13760 set Y/luminance component expression
13762 set U/Cb component expression
13764 set V/Cr component expression
13767 Each of them specifies the expression to use for computing the lookup table for
13768 the corresponding pixel component values.
13770 The exact component associated to each of the @var{c*} options depends on the
13773 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13774 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13776 The expressions can contain the following constants and functions:
13781 The input width and height.
13784 The input value for the pixel component.
13787 The input value, clipped to the @var{minval}-@var{maxval} range.
13790 The maximum value for the pixel component.
13793 The minimum value for the pixel component.
13796 The negated value for the pixel component value, clipped to the
13797 @var{minval}-@var{maxval} range; it corresponds to the expression
13798 "maxval-clipval+minval".
13801 The computed value in @var{val}, clipped to the
13802 @var{minval}-@var{maxval} range.
13804 @item gammaval(gamma)
13805 The computed gamma correction value of the pixel component value,
13806 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13808 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13812 All expressions default to "val".
13814 @subsection Examples
13818 Negate input video:
13820 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13821 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13824 The above is the same as:
13826 lutrgb="r=negval:g=negval:b=negval"
13827 lutyuv="y=negval:u=negval:v=negval"
13837 Remove chroma components, turning the video into a graytone image:
13839 lutyuv="u=128:v=128"
13843 Apply a luma burning effect:
13849 Remove green and blue components:
13855 Set a constant alpha channel value on input:
13857 format=rgba,lutrgb=a="maxval-minval/2"
13861 Correct luminance gamma by a factor of 0.5:
13863 lutyuv=y=gammaval(0.5)
13867 Discard least significant bits of luma:
13869 lutyuv=y='bitand(val, 128+64+32)'
13873 Technicolor like effect:
13875 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13879 @section lut2, tlut2
13881 The @code{lut2} filter takes two input streams and outputs one
13884 The @code{tlut2} (time lut2) filter takes two consecutive frames
13885 from one single stream.
13887 This filter accepts the following parameters:
13890 set first pixel component expression
13892 set second pixel component expression
13894 set third pixel component expression
13896 set fourth pixel component expression, corresponds to the alpha component
13899 set output bit depth, only available for @code{lut2} filter. By default is 0,
13900 which means bit depth is automatically picked from first input format.
13903 The @code{lut2} filter also supports the @ref{framesync} options.
13905 Each of them specifies the expression to use for computing the lookup table for
13906 the corresponding pixel component values.
13908 The exact component associated to each of the @var{c*} options depends on the
13911 The expressions can contain the following constants:
13916 The input width and height.
13919 The first input value for the pixel component.
13922 The second input value for the pixel component.
13925 The first input video bit depth.
13928 The second input video bit depth.
13931 All expressions default to "x".
13933 @subsection Examples
13937 Highlight differences between two RGB video streams:
13939 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)'
13943 Highlight differences between two YUV video streams:
13945 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)'
13949 Show max difference between two video streams:
13951 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)))'
13955 @section maskedclamp
13957 Clamp the first input stream with the second input and third input stream.
13959 Returns the value of first stream to be between second input
13960 stream - @code{undershoot} and third input stream + @code{overshoot}.
13962 This filter accepts the following options:
13965 Default value is @code{0}.
13968 Default value is @code{0}.
13971 Set which planes will be processed as bitmap, unprocessed planes will be
13972 copied from first stream.
13973 By default value 0xf, all planes will be processed.
13976 @subsection Commands
13978 This filter supports the all above options as @ref{commands}.
13982 Merge the second and third input stream into output stream using absolute differences
13983 between second input stream and first input stream and absolute difference between
13984 third input stream and first input stream. The picked value will be from second input
13985 stream if second absolute difference is greater than first one or from third input stream
13988 This filter accepts the following options:
13991 Set which planes will be processed as bitmap, unprocessed planes will be
13992 copied from first stream.
13993 By default value 0xf, all planes will be processed.
13996 @subsection Commands
13998 This filter supports the all above options as @ref{commands}.
14000 @section maskedmerge
14002 Merge the first input stream with the second input stream using per pixel
14003 weights in the third input stream.
14005 A value of 0 in the third stream pixel component means that pixel component
14006 from first stream is returned unchanged, while maximum value (eg. 255 for
14007 8-bit videos) means that pixel component from second stream is returned
14008 unchanged. Intermediate values define the amount of merging between both
14009 input stream's pixel components.
14011 This filter accepts the following options:
14014 Set which planes will be processed as bitmap, unprocessed planes will be
14015 copied from first stream.
14016 By default value 0xf, all planes will be processed.
14021 Merge the second and third input stream into output stream using absolute differences
14022 between second input stream and first input stream and absolute difference between
14023 third input stream and first input stream. The picked value will be from second input
14024 stream if second absolute difference is less than first one or from third input stream
14027 This filter accepts the following options:
14030 Set which planes will be processed as bitmap, unprocessed planes will be
14031 copied from first stream.
14032 By default value 0xf, all planes will be processed.
14035 @subsection Commands
14037 This filter supports the all above options as @ref{commands}.
14039 @section maskedthreshold
14040 Pick pixels comparing absolute difference of two video streams with fixed
14043 If absolute difference between pixel component of first and second video
14044 stream is equal or lower than user supplied threshold than pixel component
14045 from first video stream is picked, otherwise pixel component from second
14046 video stream is picked.
14048 This filter accepts the following options:
14051 Set threshold used when picking pixels from absolute difference from two input
14055 Set which planes will be processed as bitmap, unprocessed planes will be
14056 copied from second stream.
14057 By default value 0xf, all planes will be processed.
14060 @subsection Commands
14062 This filter supports the all above options as @ref{commands}.
14065 Create mask from input video.
14067 For example it is useful to create motion masks after @code{tblend} filter.
14069 This filter accepts the following options:
14073 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14076 Set high threshold. Any pixel component higher than this value will be set to max value
14077 allowed for current pixel format.
14080 Set planes to filter, by default all available planes are filtered.
14083 Fill all frame pixels with this value.
14086 Set max average pixel value for frame. If sum of all pixel components is higher that this
14087 average, output frame will be completely filled with value set by @var{fill} option.
14088 Typically useful for scene changes when used in combination with @code{tblend} filter.
14093 Apply motion-compensation deinterlacing.
14095 It needs one field per frame as input and must thus be used together
14096 with yadif=1/3 or equivalent.
14098 This filter accepts the following options:
14101 Set the deinterlacing mode.
14103 It accepts one of the following values:
14108 use iterative motion estimation
14110 like @samp{slow}, but use multiple reference frames.
14112 Default value is @samp{fast}.
14115 Set the picture field parity assumed for the input video. It must be
14116 one of the following values:
14120 assume top field first
14122 assume bottom field first
14125 Default value is @samp{bff}.
14128 Set per-block quantization parameter (QP) used by the internal
14131 Higher values should result in a smoother motion vector field but less
14132 optimal individual vectors. Default value is 1.
14137 Pick median pixel from certain rectangle defined by radius.
14139 This filter accepts the following options:
14143 Set horizontal radius size. Default value is @code{1}.
14144 Allowed range is integer from 1 to 127.
14147 Set which planes to process. Default is @code{15}, which is all available planes.
14150 Set vertical radius size. Default value is @code{0}.
14151 Allowed range is integer from 0 to 127.
14152 If it is 0, value will be picked from horizontal @code{radius} option.
14155 Set median percentile. Default value is @code{0.5}.
14156 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14157 minimum values, and @code{1} maximum values.
14160 @subsection Commands
14161 This filter supports same @ref{commands} as options.
14162 The command accepts the same syntax of the corresponding option.
14164 If the specified expression is not valid, it is kept at its current
14167 @section mergeplanes
14169 Merge color channel components from several video streams.
14171 The filter accepts up to 4 input streams, and merge selected input
14172 planes to the output video.
14174 This filter accepts the following options:
14177 Set input to output plane mapping. Default is @code{0}.
14179 The mappings is specified as a bitmap. It should be specified as a
14180 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14181 mapping for the first plane of the output stream. 'A' sets the number of
14182 the input stream to use (from 0 to 3), and 'a' the plane number of the
14183 corresponding input to use (from 0 to 3). The rest of the mappings is
14184 similar, 'Bb' describes the mapping for the output stream second
14185 plane, 'Cc' describes the mapping for the output stream third plane and
14186 'Dd' describes the mapping for the output stream fourth plane.
14189 Set output pixel format. Default is @code{yuva444p}.
14192 @subsection Examples
14196 Merge three gray video streams of same width and height into single video stream:
14198 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14202 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14204 [a0][a1]mergeplanes=0x00010210:yuva444p
14208 Swap Y and A plane in yuva444p stream:
14210 format=yuva444p,mergeplanes=0x03010200:yuva444p
14214 Swap U and V plane in yuv420p stream:
14216 format=yuv420p,mergeplanes=0x000201:yuv420p
14220 Cast a rgb24 clip to yuv444p:
14222 format=rgb24,mergeplanes=0x000102:yuv444p
14228 Estimate and export motion vectors using block matching algorithms.
14229 Motion vectors are stored in frame side data to be used by other filters.
14231 This filter accepts the following options:
14234 Specify the motion estimation method. Accepts one of the following values:
14238 Exhaustive search algorithm.
14240 Three step search algorithm.
14242 Two dimensional logarithmic search algorithm.
14244 New three step search algorithm.
14246 Four step search algorithm.
14248 Diamond search algorithm.
14250 Hexagon-based search algorithm.
14252 Enhanced predictive zonal search algorithm.
14254 Uneven multi-hexagon search algorithm.
14256 Default value is @samp{esa}.
14259 Macroblock size. Default @code{16}.
14262 Search parameter. Default @code{7}.
14265 @section midequalizer
14267 Apply Midway Image Equalization effect using two video streams.
14269 Midway Image Equalization adjusts a pair of images to have the same
14270 histogram, while maintaining their dynamics as much as possible. It's
14271 useful for e.g. matching exposures from a pair of stereo cameras.
14273 This filter has two inputs and one output, which must be of same pixel format, but
14274 may be of different sizes. The output of filter is first input adjusted with
14275 midway histogram of both inputs.
14277 This filter accepts the following option:
14281 Set which planes to process. Default is @code{15}, which is all available planes.
14284 @section minterpolate
14286 Convert the video to specified frame rate using motion interpolation.
14288 This filter accepts the following options:
14291 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}.
14294 Motion interpolation mode. Following values are accepted:
14297 Duplicate previous or next frame for interpolating new ones.
14299 Blend source frames. Interpolated frame is mean of previous and next frames.
14301 Motion compensated interpolation. Following options are effective when this mode is selected:
14305 Motion compensation mode. Following values are accepted:
14308 Overlapped block motion compensation.
14310 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14312 Default mode is @samp{obmc}.
14315 Motion estimation mode. Following values are accepted:
14318 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14320 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14322 Default mode is @samp{bilat}.
14325 The algorithm to be used for motion estimation. Following values are accepted:
14328 Exhaustive search algorithm.
14330 Three step search algorithm.
14332 Two dimensional logarithmic search algorithm.
14334 New three step search algorithm.
14336 Four step search algorithm.
14338 Diamond search algorithm.
14340 Hexagon-based search algorithm.
14342 Enhanced predictive zonal search algorithm.
14344 Uneven multi-hexagon search algorithm.
14346 Default algorithm is @samp{epzs}.
14349 Macroblock size. Default @code{16}.
14352 Motion estimation search parameter. Default @code{32}.
14355 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).
14360 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:
14363 Disable scene change detection.
14365 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14367 Default method is @samp{fdiff}.
14369 @item scd_threshold
14370 Scene change detection threshold. Default is @code{10.}.
14375 Mix several video input streams into one video stream.
14377 A description of the accepted options follows.
14381 The number of inputs. If unspecified, it defaults to 2.
14384 Specify weight of each input video stream as sequence.
14385 Each weight is separated by space. If number of weights
14386 is smaller than number of @var{frames} last specified
14387 weight will be used for all remaining unset weights.
14390 Specify scale, if it is set it will be multiplied with sum
14391 of each weight multiplied with pixel values to give final destination
14392 pixel value. By default @var{scale} is auto scaled to sum of weights.
14395 Specify how end of stream is determined.
14398 The duration of the longest input. (default)
14401 The duration of the shortest input.
14404 The duration of the first input.
14408 @section mpdecimate
14410 Drop frames that do not differ greatly from the previous frame in
14411 order to reduce frame rate.
14413 The main use of this filter is for very-low-bitrate encoding
14414 (e.g. streaming over dialup modem), but it could in theory be used for
14415 fixing movies that were inverse-telecined incorrectly.
14417 A description of the accepted options follows.
14421 Set the maximum number of consecutive frames which can be dropped (if
14422 positive), or the minimum interval between dropped frames (if
14423 negative). If the value is 0, the frame is dropped disregarding the
14424 number of previous sequentially dropped frames.
14426 Default value is 0.
14431 Set the dropping threshold values.
14433 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14434 represent actual pixel value differences, so a threshold of 64
14435 corresponds to 1 unit of difference for each pixel, or the same spread
14436 out differently over the block.
14438 A frame is a candidate for dropping if no 8x8 blocks differ by more
14439 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14440 meaning the whole image) differ by more than a threshold of @option{lo}.
14442 Default value for @option{hi} is 64*12, default value for @option{lo} is
14443 64*5, and default value for @option{frac} is 0.33.
14449 Negate (invert) the input video.
14451 It accepts the following option:
14456 With value 1, it negates the alpha component, if present. Default value is 0.
14462 Denoise frames using Non-Local Means algorithm.
14464 Each pixel is adjusted by looking for other pixels with similar contexts. This
14465 context similarity is defined by comparing their surrounding patches of size
14466 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14469 Note that the research area defines centers for patches, which means some
14470 patches will be made of pixels outside that research area.
14472 The filter accepts the following options.
14476 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14479 Set patch size. Default is 7. Must be odd number in range [0, 99].
14482 Same as @option{p} but for chroma planes.
14484 The default value is @var{0} and means automatic.
14487 Set research size. Default is 15. Must be odd number in range [0, 99].
14490 Same as @option{r} but for chroma planes.
14492 The default value is @var{0} and means automatic.
14497 Deinterlace video using neural network edge directed interpolation.
14499 This filter accepts the following options:
14503 Mandatory option, without binary file filter can not work.
14504 Currently file can be found here:
14505 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14508 Set which frames to deinterlace, by default it is @code{all}.
14509 Can be @code{all} or @code{interlaced}.
14512 Set mode of operation.
14514 Can be one of the following:
14518 Use frame flags, both fields.
14520 Use frame flags, single field.
14522 Use top field only.
14524 Use bottom field only.
14526 Use both fields, top first.
14528 Use both fields, bottom first.
14532 Set which planes to process, by default filter process all frames.
14535 Set size of local neighborhood around each pixel, used by the predictor neural
14538 Can be one of the following:
14551 Set the number of neurons in predictor neural network.
14552 Can be one of the following:
14563 Controls the number of different neural network predictions that are blended
14564 together to compute the final output value. Can be @code{fast}, default or
14568 Set which set of weights to use in the predictor.
14569 Can be one of the following:
14573 weights trained to minimize absolute error
14575 weights trained to minimize squared error
14579 Controls whether or not the prescreener neural network is used to decide
14580 which pixels should be processed by the predictor neural network and which
14581 can be handled by simple cubic interpolation.
14582 The prescreener is trained to know whether cubic interpolation will be
14583 sufficient for a pixel or whether it should be predicted by the predictor nn.
14584 The computational complexity of the prescreener nn is much less than that of
14585 the predictor nn. Since most pixels can be handled by cubic interpolation,
14586 using the prescreener generally results in much faster processing.
14587 The prescreener is pretty accurate, so the difference between using it and not
14588 using it is almost always unnoticeable.
14590 Can be one of the following:
14598 Default is @code{new}.
14601 Set various debugging flags.
14606 Force libavfilter not to use any of the specified pixel formats for the
14607 input to the next filter.
14609 It accepts the following parameters:
14613 A '|'-separated list of pixel format names, such as
14614 pix_fmts=yuv420p|monow|rgb24".
14618 @subsection Examples
14622 Force libavfilter to use a format different from @var{yuv420p} for the
14623 input to the vflip filter:
14625 noformat=pix_fmts=yuv420p,vflip
14629 Convert the input video to any of the formats not contained in the list:
14631 noformat=yuv420p|yuv444p|yuv410p
14637 Add noise on video input frame.
14639 The filter accepts the following options:
14647 Set noise seed for specific pixel component or all pixel components in case
14648 of @var{all_seed}. Default value is @code{123457}.
14650 @item all_strength, alls
14651 @item c0_strength, c0s
14652 @item c1_strength, c1s
14653 @item c2_strength, c2s
14654 @item c3_strength, c3s
14655 Set noise strength for specific pixel component or all pixel components in case
14656 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14658 @item all_flags, allf
14659 @item c0_flags, c0f
14660 @item c1_flags, c1f
14661 @item c2_flags, c2f
14662 @item c3_flags, c3f
14663 Set pixel component flags or set flags for all components if @var{all_flags}.
14664 Available values for component flags are:
14667 averaged temporal noise (smoother)
14669 mix random noise with a (semi)regular pattern
14671 temporal noise (noise pattern changes between frames)
14673 uniform noise (gaussian otherwise)
14677 @subsection Examples
14679 Add temporal and uniform noise to input video:
14681 noise=alls=20:allf=t+u
14686 Normalize RGB video (aka histogram stretching, contrast stretching).
14687 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14689 For each channel of each frame, the filter computes the input range and maps
14690 it linearly to the user-specified output range. The output range defaults
14691 to the full dynamic range from pure black to pure white.
14693 Temporal smoothing can be used on the input range to reduce flickering (rapid
14694 changes in brightness) caused when small dark or bright objects enter or leave
14695 the scene. This is similar to the auto-exposure (automatic gain control) on a
14696 video camera, and, like a video camera, it may cause a period of over- or
14697 under-exposure of the video.
14699 The R,G,B channels can be normalized independently, which may cause some
14700 color shifting, or linked together as a single channel, which prevents
14701 color shifting. Linked normalization preserves hue. Independent normalization
14702 does not, so it can be used to remove some color casts. Independent and linked
14703 normalization can be combined in any ratio.
14705 The normalize filter accepts the following options:
14710 Colors which define the output range. The minimum input value is mapped to
14711 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14712 The defaults are black and white respectively. Specifying white for
14713 @var{blackpt} and black for @var{whitept} will give color-inverted,
14714 normalized video. Shades of grey can be used to reduce the dynamic range
14715 (contrast). Specifying saturated colors here can create some interesting
14719 The number of previous frames to use for temporal smoothing. The input range
14720 of each channel is smoothed using a rolling average over the current frame
14721 and the @var{smoothing} previous frames. The default is 0 (no temporal
14725 Controls the ratio of independent (color shifting) channel normalization to
14726 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14727 independent. Defaults to 1.0 (fully independent).
14730 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14731 expensive no-op. Defaults to 1.0 (full strength).
14735 @subsection Commands
14736 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14737 The command accepts the same syntax of the corresponding option.
14739 If the specified expression is not valid, it is kept at its current
14742 @subsection Examples
14744 Stretch video contrast to use the full dynamic range, with no temporal
14745 smoothing; may flicker depending on the source content:
14747 normalize=blackpt=black:whitept=white:smoothing=0
14750 As above, but with 50 frames of temporal smoothing; flicker should be
14751 reduced, depending on the source content:
14753 normalize=blackpt=black:whitept=white:smoothing=50
14756 As above, but with hue-preserving linked channel normalization:
14758 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14761 As above, but with half strength:
14763 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14766 Map the darkest input color to red, the brightest input color to cyan:
14768 normalize=blackpt=red:whitept=cyan
14773 Pass the video source unchanged to the output.
14776 Optical Character Recognition
14778 This filter uses Tesseract for optical character recognition. To enable
14779 compilation of this filter, you need to configure FFmpeg with
14780 @code{--enable-libtesseract}.
14782 It accepts the following options:
14786 Set datapath to tesseract data. Default is to use whatever was
14787 set at installation.
14790 Set language, default is "eng".
14793 Set character whitelist.
14796 Set character blacklist.
14799 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14800 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14804 Apply a video transform using libopencv.
14806 To enable this filter, install the libopencv library and headers and
14807 configure FFmpeg with @code{--enable-libopencv}.
14809 It accepts the following parameters:
14814 The name of the libopencv filter to apply.
14816 @item filter_params
14817 The parameters to pass to the libopencv filter. If not specified, the default
14818 values are assumed.
14822 Refer to the official libopencv documentation for more precise
14824 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14826 Several libopencv filters are supported; see the following subsections.
14831 Dilate an image by using a specific structuring element.
14832 It corresponds to the libopencv function @code{cvDilate}.
14834 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14836 @var{struct_el} represents a structuring element, and has the syntax:
14837 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14839 @var{cols} and @var{rows} represent the number of columns and rows of
14840 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14841 point, and @var{shape} the shape for the structuring element. @var{shape}
14842 must be "rect", "cross", "ellipse", or "custom".
14844 If the value for @var{shape} is "custom", it must be followed by a
14845 string of the form "=@var{filename}". The file with name
14846 @var{filename} is assumed to represent a binary image, with each
14847 printable character corresponding to a bright pixel. When a custom
14848 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14849 or columns and rows of the read file are assumed instead.
14851 The default value for @var{struct_el} is "3x3+0x0/rect".
14853 @var{nb_iterations} specifies the number of times the transform is
14854 applied to the image, and defaults to 1.
14858 # Use the default values
14861 # Dilate using a structuring element with a 5x5 cross, iterating two times
14862 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14864 # Read the shape from the file diamond.shape, iterating two times.
14865 # The file diamond.shape may contain a pattern of characters like this
14871 # The specified columns and rows are ignored
14872 # but the anchor point coordinates are not
14873 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14878 Erode an image by using a specific structuring element.
14879 It corresponds to the libopencv function @code{cvErode}.
14881 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14882 with the same syntax and semantics as the @ref{dilate} filter.
14886 Smooth the input video.
14888 The filter takes the following parameters:
14889 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14891 @var{type} is the type of smooth filter to apply, and must be one of
14892 the following values: "blur", "blur_no_scale", "median", "gaussian",
14893 or "bilateral". The default value is "gaussian".
14895 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14896 depends on the smooth type. @var{param1} and
14897 @var{param2} accept integer positive values or 0. @var{param3} and
14898 @var{param4} accept floating point values.
14900 The default value for @var{param1} is 3. The default value for the
14901 other parameters is 0.
14903 These parameters correspond to the parameters assigned to the
14904 libopencv function @code{cvSmooth}.
14906 @section oscilloscope
14908 2D Video Oscilloscope.
14910 Useful to measure spatial impulse, step responses, chroma delays, etc.
14912 It accepts the following parameters:
14916 Set scope center x position.
14919 Set scope center y position.
14922 Set scope size, relative to frame diagonal.
14925 Set scope tilt/rotation.
14931 Set trace center x position.
14934 Set trace center y position.
14937 Set trace width, relative to width of frame.
14940 Set trace height, relative to height of frame.
14943 Set which components to trace. By default it traces first three components.
14946 Draw trace grid. By default is enabled.
14949 Draw some statistics. By default is enabled.
14952 Draw scope. By default is enabled.
14955 @subsection Commands
14956 This filter supports same @ref{commands} as options.
14957 The command accepts the same syntax of the corresponding option.
14959 If the specified expression is not valid, it is kept at its current
14962 @subsection Examples
14966 Inspect full first row of video frame.
14968 oscilloscope=x=0.5:y=0:s=1
14972 Inspect full last row of video frame.
14974 oscilloscope=x=0.5:y=1:s=1
14978 Inspect full 5th line of video frame of height 1080.
14980 oscilloscope=x=0.5:y=5/1080:s=1
14984 Inspect full last column of video frame.
14986 oscilloscope=x=1:y=0.5:s=1:t=1
14994 Overlay one video on top of another.
14996 It takes two inputs and has one output. The first input is the "main"
14997 video on which the second input is overlaid.
14999 It accepts the following parameters:
15001 A description of the accepted options follows.
15006 Set the expression for the x and y coordinates of the overlaid video
15007 on the main video. Default value is "0" for both expressions. In case
15008 the expression is invalid, it is set to a huge value (meaning that the
15009 overlay will not be displayed within the output visible area).
15012 See @ref{framesync}.
15015 Set when the expressions for @option{x}, and @option{y} are evaluated.
15017 It accepts the following values:
15020 only evaluate expressions once during the filter initialization or
15021 when a command is processed
15024 evaluate expressions for each incoming frame
15027 Default value is @samp{frame}.
15030 See @ref{framesync}.
15033 Set the format for the output video.
15035 It accepts the following values:
15038 force YUV420 output
15041 force YUV420p10 output
15044 force YUV422 output
15047 force YUV422p10 output
15050 force YUV444 output
15053 force packed RGB output
15056 force planar RGB output
15059 automatically pick format
15062 Default value is @samp{yuv420}.
15065 See @ref{framesync}.
15068 Set format of alpha of the overlaid video, it can be @var{straight} or
15069 @var{premultiplied}. Default is @var{straight}.
15072 The @option{x}, and @option{y} expressions can contain the following
15078 The main input width and height.
15082 The overlay input width and height.
15086 The computed values for @var{x} and @var{y}. They are evaluated for
15091 horizontal and vertical chroma subsample values of the output
15092 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15096 the number of input frame, starting from 0
15099 the position in the file of the input frame, NAN if unknown
15102 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15106 This filter also supports the @ref{framesync} options.
15108 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15109 when evaluation is done @emph{per frame}, and will evaluate to NAN
15110 when @option{eval} is set to @samp{init}.
15112 Be aware that frames are taken from each input video in timestamp
15113 order, hence, if their initial timestamps differ, it is a good idea
15114 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15115 have them begin in the same zero timestamp, as the example for
15116 the @var{movie} filter does.
15118 You can chain together more overlays but you should test the
15119 efficiency of such approach.
15121 @subsection Commands
15123 This filter supports the following commands:
15127 Modify the x and y of the overlay input.
15128 The command accepts the same syntax of the corresponding option.
15130 If the specified expression is not valid, it is kept at its current
15134 @subsection Examples
15138 Draw the overlay at 10 pixels from the bottom right corner of the main
15141 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15144 Using named options the example above becomes:
15146 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15150 Insert a transparent PNG logo in the bottom left corner of the input,
15151 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15153 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15157 Insert 2 different transparent PNG logos (second logo on bottom
15158 right corner) using the @command{ffmpeg} tool:
15160 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
15164 Add a transparent color layer on top of the main video; @code{WxH}
15165 must specify the size of the main input to the overlay filter:
15167 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15171 Play an original video and a filtered version (here with the deshake
15172 filter) side by side using the @command{ffplay} tool:
15174 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15177 The above command is the same as:
15179 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15183 Make a sliding overlay appearing from the left to the right top part of the
15184 screen starting since time 2:
15186 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15190 Compose output by putting two input videos side to side:
15192 ffmpeg -i left.avi -i right.avi -filter_complex "
15193 nullsrc=size=200x100 [background];
15194 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15195 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15196 [background][left] overlay=shortest=1 [background+left];
15197 [background+left][right] overlay=shortest=1:x=100 [left+right]
15202 Mask 10-20 seconds of a video by applying the delogo filter to a section
15204 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15205 -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]'
15210 Chain several overlays in cascade:
15212 nullsrc=s=200x200 [bg];
15213 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15214 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15215 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15216 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15217 [in3] null, [mid2] overlay=100:100 [out0]
15222 @anchor{overlay_cuda}
15223 @section overlay_cuda
15225 Overlay one video on top of another.
15227 This is the CUDA variant of the @ref{overlay} filter.
15228 It only accepts CUDA frames. The underlying input pixel formats have to match.
15230 It takes two inputs and has one output. The first input is the "main"
15231 video on which the second input is overlaid.
15233 It accepts the following parameters:
15238 Set the x and y coordinates of the overlaid video on the main video.
15239 Default value is "0" for both expressions.
15242 See @ref{framesync}.
15245 See @ref{framesync}.
15248 See @ref{framesync}.
15252 This filter also supports the @ref{framesync} options.
15256 Apply Overcomplete Wavelet denoiser.
15258 The filter accepts the following options:
15264 Larger depth values will denoise lower frequency components more, but
15265 slow down filtering.
15267 Must be an int in the range 8-16, default is @code{8}.
15269 @item luma_strength, ls
15272 Must be a double value in the range 0-1000, default is @code{1.0}.
15274 @item chroma_strength, cs
15275 Set chroma strength.
15277 Must be a double value in the range 0-1000, default is @code{1.0}.
15283 Add paddings to the input image, and place the original input at the
15284 provided @var{x}, @var{y} coordinates.
15286 It accepts the following parameters:
15291 Specify an expression for the size of the output image with the
15292 paddings added. If the value for @var{width} or @var{height} is 0, the
15293 corresponding input size is used for the output.
15295 The @var{width} expression can reference the value set by the
15296 @var{height} expression, and vice versa.
15298 The default value of @var{width} and @var{height} is 0.
15302 Specify the offsets to place the input image at within the padded area,
15303 with respect to the top/left border of the output image.
15305 The @var{x} expression can reference the value set by the @var{y}
15306 expression, and vice versa.
15308 The default value of @var{x} and @var{y} is 0.
15310 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15311 so the input image is centered on the padded area.
15314 Specify the color of the padded area. For the syntax of this option,
15315 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15316 manual,ffmpeg-utils}.
15318 The default value of @var{color} is "black".
15321 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15323 It accepts the following values:
15327 Only evaluate expressions once during the filter initialization or when
15328 a command is processed.
15331 Evaluate expressions for each incoming frame.
15335 Default value is @samp{init}.
15338 Pad to aspect instead to a resolution.
15342 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15343 options are expressions containing the following constants:
15348 The input video width and height.
15352 These are the same as @var{in_w} and @var{in_h}.
15356 The output width and height (the size of the padded area), as
15357 specified by the @var{width} and @var{height} expressions.
15361 These are the same as @var{out_w} and @var{out_h}.
15365 The x and y offsets as specified by the @var{x} and @var{y}
15366 expressions, or NAN if not yet specified.
15369 same as @var{iw} / @var{ih}
15372 input sample aspect ratio
15375 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15379 The horizontal and vertical chroma subsample values. For example for the
15380 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15383 @subsection Examples
15387 Add paddings with the color "violet" to the input video. The output video
15388 size is 640x480, and the top-left corner of the input video is placed at
15391 pad=640:480:0:40:violet
15394 The example above is equivalent to the following command:
15396 pad=width=640:height=480:x=0:y=40:color=violet
15400 Pad the input to get an output with dimensions increased by 3/2,
15401 and put the input video at the center of the padded area:
15403 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15407 Pad the input to get a squared output with size equal to the maximum
15408 value between the input width and height, and put the input video at
15409 the center of the padded area:
15411 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15415 Pad the input to get a final w/h ratio of 16:9:
15417 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15421 In case of anamorphic video, in order to set the output display aspect
15422 correctly, it is necessary to use @var{sar} in the expression,
15423 according to the relation:
15425 (ih * X / ih) * sar = output_dar
15426 X = output_dar / sar
15429 Thus the previous example needs to be modified to:
15431 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15435 Double the output size and put the input video in the bottom-right
15436 corner of the output padded area:
15438 pad="2*iw:2*ih:ow-iw:oh-ih"
15442 @anchor{palettegen}
15443 @section palettegen
15445 Generate one palette for a whole video stream.
15447 It accepts the following options:
15451 Set the maximum number of colors to quantize in the palette.
15452 Note: the palette will still contain 256 colors; the unused palette entries
15455 @item reserve_transparent
15456 Create a palette of 255 colors maximum and reserve the last one for
15457 transparency. Reserving the transparency color is useful for GIF optimization.
15458 If not set, the maximum of colors in the palette will be 256. You probably want
15459 to disable this option for a standalone image.
15462 @item transparency_color
15463 Set the color that will be used as background for transparency.
15466 Set statistics mode.
15468 It accepts the following values:
15471 Compute full frame histograms.
15473 Compute histograms only for the part that differs from previous frame. This
15474 might be relevant to give more importance to the moving part of your input if
15475 the background is static.
15477 Compute new histogram for each frame.
15480 Default value is @var{full}.
15483 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15484 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15485 color quantization of the palette. This information is also visible at
15486 @var{info} logging level.
15488 @subsection Examples
15492 Generate a representative palette of a given video using @command{ffmpeg}:
15494 ffmpeg -i input.mkv -vf palettegen palette.png
15498 @section paletteuse
15500 Use a palette to downsample an input video stream.
15502 The filter takes two inputs: one video stream and a palette. The palette must
15503 be a 256 pixels image.
15505 It accepts the following options:
15509 Select dithering mode. Available algorithms are:
15512 Ordered 8x8 bayer dithering (deterministic)
15514 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15515 Note: this dithering is sometimes considered "wrong" and is included as a
15517 @item floyd_steinberg
15518 Floyd and Steingberg dithering (error diffusion)
15520 Frankie Sierra dithering v2 (error diffusion)
15522 Frankie Sierra dithering v2 "Lite" (error diffusion)
15525 Default is @var{sierra2_4a}.
15528 When @var{bayer} dithering is selected, this option defines the scale of the
15529 pattern (how much the crosshatch pattern is visible). A low value means more
15530 visible pattern for less banding, and higher value means less visible pattern
15531 at the cost of more banding.
15533 The option must be an integer value in the range [0,5]. Default is @var{2}.
15536 If set, define the zone to process
15540 Only the changing rectangle will be reprocessed. This is similar to GIF
15541 cropping/offsetting compression mechanism. This option can be useful for speed
15542 if only a part of the image is changing, and has use cases such as limiting the
15543 scope of the error diffusal @option{dither} to the rectangle that bounds the
15544 moving scene (it leads to more deterministic output if the scene doesn't change
15545 much, and as a result less moving noise and better GIF compression).
15548 Default is @var{none}.
15551 Take new palette for each output frame.
15553 @item alpha_threshold
15554 Sets the alpha threshold for transparency. Alpha values above this threshold
15555 will be treated as completely opaque, and values below this threshold will be
15556 treated as completely transparent.
15558 The option must be an integer value in the range [0,255]. Default is @var{128}.
15561 @subsection Examples
15565 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15566 using @command{ffmpeg}:
15568 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15572 @section perspective
15574 Correct perspective of video not recorded perpendicular to the screen.
15576 A description of the accepted parameters follows.
15587 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15588 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15589 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15590 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15591 then the corners of the source will be sent to the specified coordinates.
15593 The expressions can use the following variables:
15598 the width and height of video frame.
15602 Output frame count.
15605 @item interpolation
15606 Set interpolation for perspective correction.
15608 It accepts the following values:
15614 Default value is @samp{linear}.
15617 Set interpretation of coordinate options.
15619 It accepts the following values:
15623 Send point in the source specified by the given coordinates to
15624 the corners of the destination.
15626 @item 1, destination
15628 Send the corners of the source to the point in the destination specified
15629 by the given coordinates.
15631 Default value is @samp{source}.
15635 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15637 It accepts the following values:
15640 only evaluate expressions once during the filter initialization or
15641 when a command is processed
15644 evaluate expressions for each incoming frame
15647 Default value is @samp{init}.
15652 Delay interlaced video by one field time so that the field order changes.
15654 The intended use is to fix PAL movies that have been captured with the
15655 opposite field order to the film-to-video transfer.
15657 A description of the accepted parameters follows.
15663 It accepts the following values:
15666 Capture field order top-first, transfer bottom-first.
15667 Filter will delay the bottom field.
15670 Capture field order bottom-first, transfer top-first.
15671 Filter will delay the top field.
15674 Capture and transfer with the same field order. This mode only exists
15675 for the documentation of the other options to refer to, but if you
15676 actually select it, the filter will faithfully do nothing.
15679 Capture field order determined automatically by field flags, transfer
15681 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15682 basis using field flags. If no field information is available,
15683 then this works just like @samp{u}.
15686 Capture unknown or varying, transfer opposite.
15687 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15688 analyzing the images and selecting the alternative that produces best
15689 match between the fields.
15692 Capture top-first, transfer unknown or varying.
15693 Filter selects among @samp{t} and @samp{p} using image analysis.
15696 Capture bottom-first, transfer unknown or varying.
15697 Filter selects among @samp{b} and @samp{p} using image analysis.
15700 Capture determined by field flags, transfer unknown or varying.
15701 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15702 image analysis. If no field information is available, then this works just
15703 like @samp{U}. This is the default mode.
15706 Both capture and transfer unknown or varying.
15707 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15711 @section photosensitivity
15712 Reduce various flashes in video, so to help users with epilepsy.
15714 It accepts the following options:
15717 Set how many frames to use when filtering. Default is 30.
15720 Set detection threshold factor. Default is 1.
15724 Set how many pixels to skip when sampling frames. Default is 1.
15725 Allowed range is from 1 to 1024.
15728 Leave frames unchanged. Default is disabled.
15731 @section pixdesctest
15733 Pixel format descriptor test filter, mainly useful for internal
15734 testing. The output video should be equal to the input video.
15738 format=monow, pixdesctest
15741 can be used to test the monowhite pixel format descriptor definition.
15745 Display sample values of color channels. Mainly useful for checking color
15746 and levels. Minimum supported resolution is 640x480.
15748 The filters accept the following options:
15752 Set scope X position, relative offset on X axis.
15755 Set scope Y position, relative offset on Y axis.
15764 Set window opacity. This window also holds statistics about pixel area.
15767 Set window X position, relative offset on X axis.
15770 Set window Y position, relative offset on Y axis.
15775 Enable the specified chain of postprocessing subfilters using libpostproc. This
15776 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15777 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15778 Each subfilter and some options have a short and a long name that can be used
15779 interchangeably, i.e. dr/dering are the same.
15781 The filters accept the following options:
15785 Set postprocessing subfilters string.
15788 All subfilters share common options to determine their scope:
15792 Honor the quality commands for this subfilter.
15795 Do chrominance filtering, too (default).
15798 Do luminance filtering only (no chrominance).
15801 Do chrominance filtering only (no luminance).
15804 These options can be appended after the subfilter name, separated by a '|'.
15806 Available subfilters are:
15809 @item hb/hdeblock[|difference[|flatness]]
15810 Horizontal deblocking filter
15813 Difference factor where higher values mean more deblocking (default: @code{32}).
15815 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15818 @item vb/vdeblock[|difference[|flatness]]
15819 Vertical deblocking filter
15822 Difference factor where higher values mean more deblocking (default: @code{32}).
15824 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15827 @item ha/hadeblock[|difference[|flatness]]
15828 Accurate horizontal deblocking filter
15831 Difference factor where higher values mean more deblocking (default: @code{32}).
15833 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15836 @item va/vadeblock[|difference[|flatness]]
15837 Accurate vertical deblocking filter
15840 Difference factor where higher values mean more deblocking (default: @code{32}).
15842 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15846 The horizontal and vertical deblocking filters share the difference and
15847 flatness values so you cannot set different horizontal and vertical
15851 @item h1/x1hdeblock
15852 Experimental horizontal deblocking filter
15854 @item v1/x1vdeblock
15855 Experimental vertical deblocking filter
15860 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15863 larger -> stronger filtering
15865 larger -> stronger filtering
15867 larger -> stronger filtering
15870 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15873 Stretch luminance to @code{0-255}.
15876 @item lb/linblenddeint
15877 Linear blend deinterlacing filter that deinterlaces the given block by
15878 filtering all lines with a @code{(1 2 1)} filter.
15880 @item li/linipoldeint
15881 Linear interpolating deinterlacing filter that deinterlaces the given block by
15882 linearly interpolating every second line.
15884 @item ci/cubicipoldeint
15885 Cubic interpolating deinterlacing filter deinterlaces the given block by
15886 cubically interpolating every second line.
15888 @item md/mediandeint
15889 Median deinterlacing filter that deinterlaces the given block by applying a
15890 median filter to every second line.
15892 @item fd/ffmpegdeint
15893 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15894 second line with a @code{(-1 4 2 4 -1)} filter.
15897 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15898 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15900 @item fq/forceQuant[|quantizer]
15901 Overrides the quantizer table from the input with the constant quantizer you
15909 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15912 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15915 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15918 @subsection Examples
15922 Apply horizontal and vertical deblocking, deringing and automatic
15923 brightness/contrast:
15929 Apply default filters without brightness/contrast correction:
15935 Apply default filters and temporal denoiser:
15937 pp=default/tmpnoise|1|2|3
15941 Apply deblocking on luminance only, and switch vertical deblocking on or off
15942 automatically depending on available CPU time:
15949 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15950 similar to spp = 6 with 7 point DCT, where only the center sample is
15953 The filter accepts the following options:
15957 Force a constant quantization parameter. It accepts an integer in range
15958 0 to 63. If not set, the filter will use the QP from the video stream
15962 Set thresholding mode. Available modes are:
15966 Set hard thresholding.
15968 Set soft thresholding (better de-ringing effect, but likely blurrier).
15970 Set medium thresholding (good results, default).
15974 @section premultiply
15975 Apply alpha premultiply effect to input video stream using first plane
15976 of second stream as alpha.
15978 Both streams must have same dimensions and same pixel format.
15980 The filter accepts the following option:
15984 Set which planes will be processed, unprocessed planes will be copied.
15985 By default value 0xf, all planes will be processed.
15988 Do not require 2nd input for processing, instead use alpha plane from input stream.
15992 Apply prewitt operator to input video stream.
15994 The filter accepts the following option:
15998 Set which planes will be processed, unprocessed planes will be copied.
15999 By default value 0xf, all planes will be processed.
16002 Set value which will be multiplied with filtered result.
16005 Set value which will be added to filtered result.
16008 @subsection Commands
16010 This filter supports the all above options as @ref{commands}.
16012 @section pseudocolor
16014 Alter frame colors in video with pseudocolors.
16016 This filter accepts the following options:
16020 set pixel first component expression
16023 set pixel second component expression
16026 set pixel third component expression
16029 set pixel fourth component expression, corresponds to the alpha component
16032 set component to use as base for altering colors
16035 Each of them specifies the expression to use for computing the lookup table for
16036 the corresponding pixel component values.
16038 The expressions can contain the following constants and functions:
16043 The input width and height.
16046 The input value for the pixel component.
16048 @item ymin, umin, vmin, amin
16049 The minimum allowed component value.
16051 @item ymax, umax, vmax, amax
16052 The maximum allowed component value.
16055 All expressions default to "val".
16057 @subsection Examples
16061 Change too high luma values to gradient:
16063 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'"
16069 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16070 Ratio) between two input videos.
16072 This filter takes in input two input videos, the first input is
16073 considered the "main" source and is passed unchanged to the
16074 output. The second input is used as a "reference" video for computing
16077 Both video inputs must have the same resolution and pixel format for
16078 this filter to work correctly. Also it assumes that both inputs
16079 have the same number of frames, which are compared one by one.
16081 The obtained average PSNR is printed through the logging system.
16083 The filter stores the accumulated MSE (mean squared error) of each
16084 frame, and at the end of the processing it is averaged across all frames
16085 equally, and the following formula is applied to obtain the PSNR:
16088 PSNR = 10*log10(MAX^2/MSE)
16091 Where MAX is the average of the maximum values of each component of the
16094 The description of the accepted parameters follows.
16097 @item stats_file, f
16098 If specified the filter will use the named file to save the PSNR of
16099 each individual frame. When filename equals "-" the data is sent to
16102 @item stats_version
16103 Specifies which version of the stats file format to use. Details of
16104 each format are written below.
16105 Default value is 1.
16107 @item stats_add_max
16108 Determines whether the max value is output to the stats log.
16109 Default value is 0.
16110 Requires stats_version >= 2. If this is set and stats_version < 2,
16111 the filter will return an error.
16114 This filter also supports the @ref{framesync} options.
16116 The file printed if @var{stats_file} is selected, contains a sequence of
16117 key/value pairs of the form @var{key}:@var{value} for each compared
16120 If a @var{stats_version} greater than 1 is specified, a header line precedes
16121 the list of per-frame-pair stats, with key value pairs following the frame
16122 format with the following parameters:
16125 @item psnr_log_version
16126 The version of the log file format. Will match @var{stats_version}.
16129 A comma separated list of the per-frame-pair parameters included in
16133 A description of each shown per-frame-pair parameter follows:
16137 sequential number of the input frame, starting from 1
16140 Mean Square Error pixel-by-pixel average difference of the compared
16141 frames, averaged over all the image components.
16143 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16144 Mean Square Error pixel-by-pixel average difference of the compared
16145 frames for the component specified by the suffix.
16147 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16148 Peak Signal to Noise ratio of the compared frames for the component
16149 specified by the suffix.
16151 @item max_avg, max_y, max_u, max_v
16152 Maximum allowed value for each channel, and average over all
16156 @subsection Examples
16161 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16162 [main][ref] psnr="stats_file=stats.log" [out]
16165 On this example the input file being processed is compared with the
16166 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16167 is stored in @file{stats.log}.
16170 Another example with different containers:
16172 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 -
16179 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16180 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16183 The pullup filter is designed to take advantage of future context in making
16184 its decisions. This filter is stateless in the sense that it does not lock
16185 onto a pattern to follow, but it instead looks forward to the following
16186 fields in order to identify matches and rebuild progressive frames.
16188 To produce content with an even framerate, insert the fps filter after
16189 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16190 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16192 The filter accepts the following options:
16199 These options set the amount of "junk" to ignore at the left, right, top, and
16200 bottom of the image, respectively. Left and right are in units of 8 pixels,
16201 while top and bottom are in units of 2 lines.
16202 The default is 8 pixels on each side.
16205 Set the strict breaks. Setting this option to 1 will reduce the chances of
16206 filter generating an occasional mismatched frame, but it may also cause an
16207 excessive number of frames to be dropped during high motion sequences.
16208 Conversely, setting it to -1 will make filter match fields more easily.
16209 This may help processing of video where there is slight blurring between
16210 the fields, but may also cause there to be interlaced frames in the output.
16211 Default value is @code{0}.
16214 Set the metric plane to use. It accepts the following values:
16220 Use chroma blue plane.
16223 Use chroma red plane.
16226 This option may be set to use chroma plane instead of the default luma plane
16227 for doing filter's computations. This may improve accuracy on very clean
16228 source material, but more likely will decrease accuracy, especially if there
16229 is chroma noise (rainbow effect) or any grayscale video.
16230 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16231 load and make pullup usable in realtime on slow machines.
16234 For best results (without duplicated frames in the output file) it is
16235 necessary to change the output frame rate. For example, to inverse
16236 telecine NTSC input:
16238 ffmpeg -i input -vf pullup -r 24000/1001 ...
16243 Change video quantization parameters (QP).
16245 The filter accepts the following option:
16249 Set expression for quantization parameter.
16252 The expression is evaluated through the eval API and can contain, among others,
16253 the following constants:
16257 1 if index is not 129, 0 otherwise.
16260 Sequential index starting from -129 to 128.
16263 @subsection Examples
16267 Some equation like:
16275 Flush video frames from internal cache of frames into a random order.
16276 No frame is discarded.
16277 Inspired by @ref{frei0r} nervous filter.
16281 Set size in number of frames of internal cache, in range from @code{2} to
16282 @code{512}. Default is @code{30}.
16285 Set seed for random number generator, must be an integer included between
16286 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16287 less than @code{0}, the filter will try to use a good random seed on a
16291 @section readeia608
16293 Read closed captioning (EIA-608) information from the top lines of a video frame.
16295 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16296 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16297 with EIA-608 data (starting from 0). A description of each metadata value follows:
16300 @item lavfi.readeia608.X.cc
16301 The two bytes stored as EIA-608 data (printed in hexadecimal).
16303 @item lavfi.readeia608.X.line
16304 The number of the line on which the EIA-608 data was identified and read.
16307 This filter accepts the following options:
16311 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16314 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16317 Set the ratio of width reserved for sync code detection.
16318 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16321 Enable checking the parity bit. In the event of a parity error, the filter will output
16322 @code{0x00} for that character. Default is false.
16325 Lowpass lines prior to further processing. Default is enabled.
16328 @subsection Commands
16330 This filter supports the all above options as @ref{commands}.
16332 @subsection Examples
16336 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16338 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
16344 Read vertical interval timecode (VITC) information from the top lines of a
16347 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16348 timecode value, if a valid timecode has been detected. Further metadata key
16349 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16350 timecode data has been found or not.
16352 This filter accepts the following options:
16356 Set the maximum number of lines to scan for VITC data. If the value is set to
16357 @code{-1} the full video frame is scanned. Default is @code{45}.
16360 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16361 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16364 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16365 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16368 @subsection Examples
16372 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16373 draw @code{--:--:--:--} as a placeholder:
16375 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16381 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16383 Destination pixel at position (X, Y) will be picked from source (x, y) position
16384 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16385 value for pixel will be used for destination pixel.
16387 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16388 will have Xmap/Ymap video stream dimensions.
16389 Xmap and Ymap input video streams are 16bit depth, single channel.
16393 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16394 Default is @code{color}.
16397 Specify the color of the unmapped pixels. For the syntax of this option,
16398 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16399 manual,ffmpeg-utils}. Default color is @code{black}.
16402 @section removegrain
16404 The removegrain filter is a spatial denoiser for progressive video.
16408 Set mode for the first plane.
16411 Set mode for the second plane.
16414 Set mode for the third plane.
16417 Set mode for the fourth plane.
16420 Range of mode is from 0 to 24. Description of each mode follows:
16424 Leave input plane unchanged. Default.
16427 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16430 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16433 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16436 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16437 This is equivalent to a median filter.
16440 Line-sensitive clipping giving the minimal change.
16443 Line-sensitive clipping, intermediate.
16446 Line-sensitive clipping, intermediate.
16449 Line-sensitive clipping, intermediate.
16452 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16455 Replaces the target pixel with the closest neighbour.
16458 [1 2 1] horizontal and vertical kernel blur.
16464 Bob mode, interpolates top field from the line where the neighbours
16465 pixels are the closest.
16468 Bob mode, interpolates bottom field from the line where the neighbours
16469 pixels are the closest.
16472 Bob mode, interpolates top field. Same as 13 but with a more complicated
16473 interpolation formula.
16476 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16477 interpolation formula.
16480 Clips the pixel with the minimum and maximum of respectively the maximum and
16481 minimum of each pair of opposite neighbour pixels.
16484 Line-sensitive clipping using opposite neighbours whose greatest distance from
16485 the current pixel is minimal.
16488 Replaces the pixel with the average of its 8 neighbours.
16491 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16494 Clips pixels using the averages of opposite neighbour.
16497 Same as mode 21 but simpler and faster.
16500 Small edge and halo removal, but reputed useless.
16506 @section removelogo
16508 Suppress a TV station logo, using an image file to determine which
16509 pixels comprise the logo. It works by filling in the pixels that
16510 comprise the logo with neighboring pixels.
16512 The filter accepts the following options:
16516 Set the filter bitmap file, which can be any image format supported by
16517 libavformat. The width and height of the image file must match those of the
16518 video stream being processed.
16521 Pixels in the provided bitmap image with a value of zero are not
16522 considered part of the logo, non-zero pixels are considered part of
16523 the logo. If you use white (255) for the logo and black (0) for the
16524 rest, you will be safe. For making the filter bitmap, it is
16525 recommended to take a screen capture of a black frame with the logo
16526 visible, and then using a threshold filter followed by the erode
16527 filter once or twice.
16529 If needed, little splotches can be fixed manually. Remember that if
16530 logo pixels are not covered, the filter quality will be much
16531 reduced. Marking too many pixels as part of the logo does not hurt as
16532 much, but it will increase the amount of blurring needed to cover over
16533 the image and will destroy more information than necessary, and extra
16534 pixels will slow things down on a large logo.
16536 @section repeatfields
16538 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16539 fields based on its value.
16543 Reverse a video clip.
16545 Warning: This filter requires memory to buffer the entire clip, so trimming
16548 @subsection Examples
16552 Take the first 5 seconds of a clip, and reverse it.
16559 Shift R/G/B/A pixels horizontally and/or vertically.
16561 The filter accepts the following options:
16564 Set amount to shift red horizontally.
16566 Set amount to shift red vertically.
16568 Set amount to shift green horizontally.
16570 Set amount to shift green vertically.
16572 Set amount to shift blue horizontally.
16574 Set amount to shift blue vertically.
16576 Set amount to shift alpha horizontally.
16578 Set amount to shift alpha vertically.
16580 Set edge mode, can be @var{smear}, default, or @var{warp}.
16583 @subsection Commands
16585 This filter supports the all above options as @ref{commands}.
16588 Apply roberts cross operator to input video stream.
16590 The filter accepts the following option:
16594 Set which planes will be processed, unprocessed planes will be copied.
16595 By default value 0xf, all planes will be processed.
16598 Set value which will be multiplied with filtered result.
16601 Set value which will be added to filtered result.
16604 @subsection Commands
16606 This filter supports the all above options as @ref{commands}.
16610 Rotate video by an arbitrary angle expressed in radians.
16612 The filter accepts the following options:
16614 A description of the optional parameters follows.
16617 Set an expression for the angle by which to rotate the input video
16618 clockwise, expressed as a number of radians. A negative value will
16619 result in a counter-clockwise rotation. By default it is set to "0".
16621 This expression is evaluated for each frame.
16624 Set the output width expression, default value is "iw".
16625 This expression is evaluated just once during configuration.
16628 Set the output height expression, default value is "ih".
16629 This expression is evaluated just once during configuration.
16632 Enable bilinear interpolation if set to 1, a value of 0 disables
16633 it. Default value is 1.
16636 Set the color used to fill the output area not covered by the rotated
16637 image. For the general syntax of this option, check the
16638 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16639 If the special value "none" is selected then no
16640 background is printed (useful for example if the background is never shown).
16642 Default value is "black".
16645 The expressions for the angle and the output size can contain the
16646 following constants and functions:
16650 sequential number of the input frame, starting from 0. It is always NAN
16651 before the first frame is filtered.
16654 time in seconds of the input frame, it is set to 0 when the filter is
16655 configured. It is always NAN before the first frame is filtered.
16659 horizontal and vertical chroma subsample values. For example for the
16660 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16664 the input video width and height
16668 the output width and height, that is the size of the padded area as
16669 specified by the @var{width} and @var{height} expressions
16673 the minimal width/height required for completely containing the input
16674 video rotated by @var{a} radians.
16676 These are only available when computing the @option{out_w} and
16677 @option{out_h} expressions.
16680 @subsection Examples
16684 Rotate the input by PI/6 radians clockwise:
16690 Rotate the input by PI/6 radians counter-clockwise:
16696 Rotate the input by 45 degrees clockwise:
16702 Apply a constant rotation with period T, starting from an angle of PI/3:
16704 rotate=PI/3+2*PI*t/T
16708 Make the input video rotation oscillating with a period of T
16709 seconds and an amplitude of A radians:
16711 rotate=A*sin(2*PI/T*t)
16715 Rotate the video, output size is chosen so that the whole rotating
16716 input video is always completely contained in the output:
16718 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16722 Rotate the video, reduce the output size so that no background is ever
16725 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16729 @subsection Commands
16731 The filter supports the following commands:
16735 Set the angle expression.
16736 The command accepts the same syntax of the corresponding option.
16738 If the specified expression is not valid, it is kept at its current
16744 Apply Shape Adaptive Blur.
16746 The filter accepts the following options:
16749 @item luma_radius, lr
16750 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16751 value is 1.0. A greater value will result in a more blurred image, and
16752 in slower processing.
16754 @item luma_pre_filter_radius, lpfr
16755 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16758 @item luma_strength, ls
16759 Set luma maximum difference between pixels to still be considered, must
16760 be a value in the 0.1-100.0 range, default value is 1.0.
16762 @item chroma_radius, cr
16763 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16764 greater value will result in a more blurred image, and in slower
16767 @item chroma_pre_filter_radius, cpfr
16768 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16770 @item chroma_strength, cs
16771 Set chroma maximum difference between pixels to still be considered,
16772 must be a value in the -0.9-100.0 range.
16775 Each chroma option value, if not explicitly specified, is set to the
16776 corresponding luma option value.
16781 Scale (resize) the input video, using the libswscale library.
16783 The scale filter forces the output display aspect ratio to be the same
16784 of the input, by changing the output sample aspect ratio.
16786 If the input image format is different from the format requested by
16787 the next filter, the scale filter will convert the input to the
16790 @subsection Options
16791 The filter accepts the following options, or any of the options
16792 supported by the libswscale scaler.
16794 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16795 the complete list of scaler options.
16800 Set the output video dimension expression. Default value is the input
16803 If the @var{width} or @var{w} value is 0, the input width is used for
16804 the output. If the @var{height} or @var{h} value is 0, the input height
16805 is used for the output.
16807 If one and only one of the values is -n with n >= 1, the scale filter
16808 will use a value that maintains the aspect ratio of the input image,
16809 calculated from the other specified dimension. After that it will,
16810 however, make sure that the calculated dimension is divisible by n and
16811 adjust the value if necessary.
16813 If both values are -n with n >= 1, the behavior will be identical to
16814 both values being set to 0 as previously detailed.
16816 See below for the list of accepted constants for use in the dimension
16820 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16824 Only evaluate expressions once during the filter initialization or when a command is processed.
16827 Evaluate expressions for each incoming frame.
16831 Default value is @samp{init}.
16835 Set the interlacing mode. It accepts the following values:
16839 Force interlaced aware scaling.
16842 Do not apply interlaced scaling.
16845 Select interlaced aware scaling depending on whether the source frames
16846 are flagged as interlaced or not.
16849 Default value is @samp{0}.
16852 Set libswscale scaling flags. See
16853 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16854 complete list of values. If not explicitly specified the filter applies
16858 @item param0, param1
16859 Set libswscale input parameters for scaling algorithms that need them. See
16860 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16861 complete documentation. If not explicitly specified the filter applies
16867 Set the video size. For the syntax of this option, check the
16868 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16870 @item in_color_matrix
16871 @item out_color_matrix
16872 Set in/output YCbCr color space type.
16874 This allows the autodetected value to be overridden as well as allows forcing
16875 a specific value used for the output and encoder.
16877 If not specified, the color space type depends on the pixel format.
16883 Choose automatically.
16886 Format conforming to International Telecommunication Union (ITU)
16887 Recommendation BT.709.
16890 Set color space conforming to the United States Federal Communications
16891 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16896 Set color space conforming to:
16900 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16903 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16906 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16911 Set color space conforming to SMPTE ST 240:1999.
16914 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16919 Set in/output YCbCr sample range.
16921 This allows the autodetected value to be overridden as well as allows forcing
16922 a specific value used for the output and encoder. If not specified, the
16923 range depends on the pixel format. Possible values:
16927 Choose automatically.
16930 Set full range (0-255 in case of 8-bit luma).
16932 @item mpeg/limited/tv
16933 Set "MPEG" range (16-235 in case of 8-bit luma).
16936 @item force_original_aspect_ratio
16937 Enable decreasing or increasing output video width or height if necessary to
16938 keep the original aspect ratio. Possible values:
16942 Scale the video as specified and disable this feature.
16945 The output video dimensions will automatically be decreased if needed.
16948 The output video dimensions will automatically be increased if needed.
16952 One useful instance of this option is that when you know a specific device's
16953 maximum allowed resolution, you can use this to limit the output video to
16954 that, while retaining the aspect ratio. For example, device A allows
16955 1280x720 playback, and your video is 1920x800. Using this option (set it to
16956 decrease) and specifying 1280x720 to the command line makes the output
16959 Please note that this is a different thing than specifying -1 for @option{w}
16960 or @option{h}, you still need to specify the output resolution for this option
16963 @item force_divisible_by
16964 Ensures that both the output dimensions, width and height, are divisible by the
16965 given integer when used together with @option{force_original_aspect_ratio}. This
16966 works similar to using @code{-n} in the @option{w} and @option{h} options.
16968 This option respects the value set for @option{force_original_aspect_ratio},
16969 increasing or decreasing the resolution accordingly. The video's aspect ratio
16970 may be slightly modified.
16972 This option can be handy if you need to have a video fit within or exceed
16973 a defined resolution using @option{force_original_aspect_ratio} but also have
16974 encoder restrictions on width or height divisibility.
16978 The values of the @option{w} and @option{h} options are expressions
16979 containing the following constants:
16984 The input width and height
16988 These are the same as @var{in_w} and @var{in_h}.
16992 The output (scaled) width and height
16996 These are the same as @var{out_w} and @var{out_h}
16999 The same as @var{iw} / @var{ih}
17002 input sample aspect ratio
17005 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17009 horizontal and vertical input chroma subsample values. For example for the
17010 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17014 horizontal and vertical output chroma subsample values. For example for the
17015 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17018 The (sequential) number of the input frame, starting from 0.
17019 Only available with @code{eval=frame}.
17022 The presentation timestamp of the input frame, expressed as a number of
17023 seconds. Only available with @code{eval=frame}.
17026 The position (byte offset) of the frame in the input stream, or NaN if
17027 this information is unavailable and/or meaningless (for example in case of synthetic video).
17028 Only available with @code{eval=frame}.
17031 @subsection Examples
17035 Scale the input video to a size of 200x100
17040 This is equivalent to:
17051 Specify a size abbreviation for the output size:
17056 which can also be written as:
17062 Scale the input to 2x:
17064 scale=w=2*iw:h=2*ih
17068 The above is the same as:
17070 scale=2*in_w:2*in_h
17074 Scale the input to 2x with forced interlaced scaling:
17076 scale=2*iw:2*ih:interl=1
17080 Scale the input to half size:
17082 scale=w=iw/2:h=ih/2
17086 Increase the width, and set the height to the same size:
17092 Seek Greek harmony:
17099 Increase the height, and set the width to 3/2 of the height:
17101 scale=w=3/2*oh:h=3/5*ih
17105 Increase the size, making the size a multiple of the chroma
17108 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17112 Increase the width to a maximum of 500 pixels,
17113 keeping the same aspect ratio as the input:
17115 scale=w='min(500\, iw*3/2):h=-1'
17119 Make pixels square by combining scale and setsar:
17121 scale='trunc(ih*dar):ih',setsar=1/1
17125 Make pixels square by combining scale and setsar,
17126 making sure the resulting resolution is even (required by some codecs):
17128 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17132 @subsection Commands
17134 This filter supports the following commands:
17138 Set the output video dimension expression.
17139 The command accepts the same syntax of the corresponding option.
17141 If the specified expression is not valid, it is kept at its current
17147 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17148 format conversion on CUDA video frames. Setting the output width and height
17149 works in the same way as for the @var{scale} filter.
17151 The following additional options are accepted:
17154 The pixel format of the output CUDA frames. If set to the string "same" (the
17155 default), the input format will be kept. Note that automatic format negotiation
17156 and conversion is not yet supported for hardware frames
17159 The interpolation algorithm used for resizing. One of the following:
17166 @item cubic2p_bspline
17167 2-parameter cubic (B=1, C=0)
17169 @item cubic2p_catmullrom
17170 2-parameter cubic (B=0, C=1/2)
17172 @item cubic2p_b05c03
17173 2-parameter cubic (B=1/2, C=3/10)
17181 @item force_original_aspect_ratio
17182 Enable decreasing or increasing output video width or height if necessary to
17183 keep the original aspect ratio. Possible values:
17187 Scale the video as specified and disable this feature.
17190 The output video dimensions will automatically be decreased if needed.
17193 The output video dimensions will automatically be increased if needed.
17197 One useful instance of this option is that when you know a specific device's
17198 maximum allowed resolution, you can use this to limit the output video to
17199 that, while retaining the aspect ratio. For example, device A allows
17200 1280x720 playback, and your video is 1920x800. Using this option (set it to
17201 decrease) and specifying 1280x720 to the command line makes the output
17204 Please note that this is a different thing than specifying -1 for @option{w}
17205 or @option{h}, you still need to specify the output resolution for this option
17208 @item force_divisible_by
17209 Ensures that both the output dimensions, width and height, are divisible by the
17210 given integer when used together with @option{force_original_aspect_ratio}. This
17211 works similar to using @code{-n} in the @option{w} and @option{h} options.
17213 This option respects the value set for @option{force_original_aspect_ratio},
17214 increasing or decreasing the resolution accordingly. The video's aspect ratio
17215 may be slightly modified.
17217 This option can be handy if you need to have a video fit within or exceed
17218 a defined resolution using @option{force_original_aspect_ratio} but also have
17219 encoder restrictions on width or height divisibility.
17225 Scale (resize) the input video, based on a reference video.
17227 See the scale filter for available options, scale2ref supports the same but
17228 uses the reference video instead of the main input as basis. scale2ref also
17229 supports the following additional constants for the @option{w} and
17230 @option{h} options:
17235 The main input video's width and height
17238 The same as @var{main_w} / @var{main_h}
17241 The main input video's sample aspect ratio
17243 @item main_dar, mdar
17244 The main input video's display aspect ratio. Calculated from
17245 @code{(main_w / main_h) * main_sar}.
17249 The main input video's horizontal and vertical chroma subsample values.
17250 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17254 The (sequential) number of the main input frame, starting from 0.
17255 Only available with @code{eval=frame}.
17258 The presentation timestamp of the main input frame, expressed as a number of
17259 seconds. Only available with @code{eval=frame}.
17262 The position (byte offset) of the frame in the main input stream, or NaN if
17263 this information is unavailable and/or meaningless (for example in case of synthetic video).
17264 Only available with @code{eval=frame}.
17267 @subsection Examples
17271 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17273 'scale2ref[b][a];[a][b]overlay'
17277 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17279 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17283 @subsection Commands
17285 This filter supports the following commands:
17289 Set the output video dimension expression.
17290 The command accepts the same syntax of the corresponding option.
17292 If the specified expression is not valid, it is kept at its current
17297 Scroll input video horizontally and/or vertically by constant speed.
17299 The filter accepts the following options:
17301 @item horizontal, h
17302 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17303 Negative values changes scrolling direction.
17306 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17307 Negative values changes scrolling direction.
17310 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17313 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17316 @subsection Commands
17318 This filter supports the following @ref{commands}:
17320 @item horizontal, h
17321 Set the horizontal scrolling speed.
17323 Set the vertical scrolling speed.
17329 Detect video scene change.
17331 This filter sets frame metadata with mafd between frame, the scene score, and
17332 forward the frame to the next filter, so they can use these metadata to detect
17333 scene change or others.
17335 In addition, this filter logs a message and sets frame metadata when it detects
17336 a scene change by @option{threshold}.
17338 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17340 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17341 to detect scene change.
17343 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17344 detect scene change with @option{threshold}.
17346 The filter accepts the following options:
17350 Set the scene change detection threshold as a percentage of maximum change. Good
17351 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17354 Default value is @code{10.}.
17357 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17358 You can enable it if you want to get snapshot of scene change frames only.
17361 @anchor{selectivecolor}
17362 @section selectivecolor
17364 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17365 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17366 by the "purity" of the color (that is, how saturated it already is).
17368 This filter is similar to the Adobe Photoshop Selective Color tool.
17370 The filter accepts the following options:
17373 @item correction_method
17374 Select color correction method.
17376 Available values are:
17379 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17382 Specified adjustments are relative to the original component value.
17384 Default is @code{absolute}.
17386 Adjustments for red pixels (pixels where the red component is the maximum)
17388 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17390 Adjustments for green pixels (pixels where the green component is the maximum)
17392 Adjustments for cyan pixels (pixels where the red component is the minimum)
17394 Adjustments for blue pixels (pixels where the blue component is the maximum)
17396 Adjustments for magenta pixels (pixels where the green component is the minimum)
17398 Adjustments for white pixels (pixels where all components are greater than 128)
17400 Adjustments for all pixels except pure black and pure white
17402 Adjustments for black pixels (pixels where all components are lesser than 128)
17404 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17407 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17408 4 space separated floating point adjustment values in the [-1,1] range,
17409 respectively to adjust the amount of cyan, magenta, yellow and black for the
17410 pixels of its range.
17412 @subsection Examples
17416 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17417 increase magenta by 27% in blue areas:
17419 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17423 Use a Photoshop selective color preset:
17425 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17429 @anchor{separatefields}
17430 @section separatefields
17432 The @code{separatefields} takes a frame-based video input and splits
17433 each frame into its components fields, producing a new half height clip
17434 with twice the frame rate and twice the frame count.
17436 This filter use field-dominance information in frame to decide which
17437 of each pair of fields to place first in the output.
17438 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17440 @section setdar, setsar
17442 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17445 This is done by changing the specified Sample (aka Pixel) Aspect
17446 Ratio, according to the following equation:
17448 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17451 Keep in mind that the @code{setdar} filter does not modify the pixel
17452 dimensions of the video frame. Also, the display aspect ratio set by
17453 this filter may be changed by later filters in the filterchain,
17454 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17457 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17458 the filter output video.
17460 Note that as a consequence of the application of this filter, the
17461 output display aspect ratio will change according to the equation
17464 Keep in mind that the sample aspect ratio set by the @code{setsar}
17465 filter may be changed by later filters in the filterchain, e.g. if
17466 another "setsar" or a "setdar" filter is applied.
17468 It accepts the following parameters:
17471 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17472 Set the aspect ratio used by the filter.
17474 The parameter can be a floating point number string, an expression, or
17475 a string of the form @var{num}:@var{den}, where @var{num} and
17476 @var{den} are the numerator and denominator of the aspect ratio. If
17477 the parameter is not specified, it is assumed the value "0".
17478 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17482 Set the maximum integer value to use for expressing numerator and
17483 denominator when reducing the expressed aspect ratio to a rational.
17484 Default value is @code{100}.
17488 The parameter @var{sar} is an expression containing
17489 the following constants:
17493 These are approximated values for the mathematical constants e
17494 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17497 The input width and height.
17500 These are the same as @var{w} / @var{h}.
17503 The input sample aspect ratio.
17506 The input display aspect ratio. It is the same as
17507 (@var{w} / @var{h}) * @var{sar}.
17510 Horizontal and vertical chroma subsample values. For example, for the
17511 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17514 @subsection Examples
17519 To change the display aspect ratio to 16:9, specify one of the following:
17526 To change the sample aspect ratio to 10:11, specify:
17532 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17533 1000 in the aspect ratio reduction, use the command:
17535 setdar=ratio=16/9:max=1000
17543 Force field for the output video frame.
17545 The @code{setfield} filter marks the interlace type field for the
17546 output frames. It does not change the input frame, but only sets the
17547 corresponding property, which affects how the frame is treated by
17548 following filters (e.g. @code{fieldorder} or @code{yadif}).
17550 The filter accepts the following options:
17555 Available values are:
17559 Keep the same field property.
17562 Mark the frame as bottom-field-first.
17565 Mark the frame as top-field-first.
17568 Mark the frame as progressive.
17575 Force frame parameter for the output video frame.
17577 The @code{setparams} filter marks interlace and color range for the
17578 output frames. It does not change the input frame, but only sets the
17579 corresponding property, which affects how the frame is treated by
17584 Available values are:
17588 Keep the same field property (default).
17591 Mark the frame as bottom-field-first.
17594 Mark the frame as top-field-first.
17597 Mark the frame as progressive.
17601 Available values are:
17605 Keep the same color range property (default).
17607 @item unspecified, unknown
17608 Mark the frame as unspecified color range.
17610 @item limited, tv, mpeg
17611 Mark the frame as limited range.
17613 @item full, pc, jpeg
17614 Mark the frame as full range.
17617 @item color_primaries
17618 Set the color primaries.
17619 Available values are:
17623 Keep the same color primaries property (default).
17640 Set the color transfer.
17641 Available values are:
17645 Keep the same color trc property (default).
17667 Set the colorspace.
17668 Available values are:
17672 Keep the same colorspace property (default).
17685 @item chroma-derived-nc
17686 @item chroma-derived-c
17693 Show a line containing various information for each input video frame.
17694 The input video is not modified.
17696 This filter supports the following options:
17700 Calculate checksums of each plane. By default enabled.
17703 The shown line contains a sequence of key/value pairs of the form
17704 @var{key}:@var{value}.
17706 The following values are shown in the output:
17710 The (sequential) number of the input frame, starting from 0.
17713 The Presentation TimeStamp of the input frame, expressed as a number of
17714 time base units. The time base unit depends on the filter input pad.
17717 The Presentation TimeStamp of the input frame, expressed as a number of
17721 The position of the frame in the input stream, or -1 if this information is
17722 unavailable and/or meaningless (for example in case of synthetic video).
17725 The pixel format name.
17728 The sample aspect ratio of the input frame, expressed in the form
17729 @var{num}/@var{den}.
17732 The size of the input frame. For the syntax of this option, check the
17733 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17736 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17737 for bottom field first).
17740 This is 1 if the frame is a key frame, 0 otherwise.
17743 The picture type of the input frame ("I" for an I-frame, "P" for a
17744 P-frame, "B" for a B-frame, or "?" for an unknown type).
17745 Also refer to the documentation of the @code{AVPictureType} enum and of
17746 the @code{av_get_picture_type_char} function defined in
17747 @file{libavutil/avutil.h}.
17750 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17752 @item plane_checksum
17753 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17754 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17757 The mean value of pixels in each plane of the input frame, expressed in the form
17758 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17761 The standard deviation of pixel values in each plane of the input frame, expressed
17762 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17766 @section showpalette
17768 Displays the 256 colors palette of each frame. This filter is only relevant for
17769 @var{pal8} pixel format frames.
17771 It accepts the following option:
17775 Set the size of the box used to represent one palette color entry. Default is
17776 @code{30} (for a @code{30x30} pixel box).
17779 @section shuffleframes
17781 Reorder and/or duplicate and/or drop video frames.
17783 It accepts the following parameters:
17787 Set the destination indexes of input frames.
17788 This is space or '|' separated list of indexes that maps input frames to output
17789 frames. Number of indexes also sets maximal value that each index may have.
17790 '-1' index have special meaning and that is to drop frame.
17793 The first frame has the index 0. The default is to keep the input unchanged.
17795 @subsection Examples
17799 Swap second and third frame of every three frames of the input:
17801 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17805 Swap 10th and 1st frame of every ten frames of the input:
17807 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17811 @section shufflepixels
17813 Reorder pixels in video frames.
17815 This filter accepts the following options:
17819 Set shuffle direction. Can be forward or inverse direction.
17820 Default direction is forward.
17823 Set shuffle mode. Can be horizontal, vertical or block mode.
17827 Set shuffle block_size. In case of horizontal shuffle mode only width
17828 part of size is used, and in case of vertical shuffle mode only height
17829 part of size is used.
17832 Set random seed used with shuffling pixels. Mainly useful to set to be able
17833 to reverse filtering process to get original input.
17834 For example, to reverse forward shuffle you need to use same parameters
17835 and exact same seed and to set direction to inverse.
17838 @section shuffleplanes
17840 Reorder and/or duplicate video planes.
17842 It accepts the following parameters:
17847 The index of the input plane to be used as the first output plane.
17850 The index of the input plane to be used as the second output plane.
17853 The index of the input plane to be used as the third output plane.
17856 The index of the input plane to be used as the fourth output plane.
17860 The first plane has the index 0. The default is to keep the input unchanged.
17862 @subsection Examples
17866 Swap the second and third planes of the input:
17868 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17872 @anchor{signalstats}
17873 @section signalstats
17874 Evaluate various visual metrics that assist in determining issues associated
17875 with the digitization of analog video media.
17877 By default the filter will log these metadata values:
17881 Display the minimal Y value contained within the input frame. Expressed in
17885 Display the Y value at the 10% percentile within the input frame. Expressed in
17889 Display the average Y value within the input frame. Expressed in range of
17893 Display the Y value at the 90% percentile within the input frame. Expressed in
17897 Display the maximum Y value contained within the input frame. Expressed in
17901 Display the minimal U value contained within the input frame. Expressed in
17905 Display the U value at the 10% percentile within the input frame. Expressed in
17909 Display the average U value within the input frame. Expressed in range of
17913 Display the U value at the 90% percentile within the input frame. Expressed in
17917 Display the maximum U value contained within the input frame. Expressed in
17921 Display the minimal V value contained within the input frame. Expressed in
17925 Display the V value at the 10% percentile within the input frame. Expressed in
17929 Display the average V value within the input frame. Expressed in range of
17933 Display the V value at the 90% percentile within the input frame. Expressed in
17937 Display the maximum V value contained within the input frame. Expressed in
17941 Display the minimal saturation value contained within the input frame.
17942 Expressed in range of [0-~181.02].
17945 Display the saturation value at the 10% percentile within the input frame.
17946 Expressed in range of [0-~181.02].
17949 Display the average saturation value within the input frame. Expressed in range
17953 Display the saturation value at the 90% percentile within the input frame.
17954 Expressed in range of [0-~181.02].
17957 Display the maximum saturation value contained within the input frame.
17958 Expressed in range of [0-~181.02].
17961 Display the median value for hue within the input frame. Expressed in range of
17965 Display the average value for hue within the input frame. Expressed in range of
17969 Display the average of sample value difference between all values of the Y
17970 plane in the current frame and corresponding values of the previous input frame.
17971 Expressed in range of [0-255].
17974 Display the average of sample value difference between all values of the U
17975 plane in the current frame and corresponding values of the previous input frame.
17976 Expressed in range of [0-255].
17979 Display the average of sample value difference between all values of the V
17980 plane in the current frame and corresponding values of the previous input frame.
17981 Expressed in range of [0-255].
17984 Display bit depth of Y plane in current frame.
17985 Expressed in range of [0-16].
17988 Display bit depth of U plane in current frame.
17989 Expressed in range of [0-16].
17992 Display bit depth of V plane in current frame.
17993 Expressed in range of [0-16].
17996 The filter accepts the following options:
18002 @option{stat} specify an additional form of image analysis.
18003 @option{out} output video with the specified type of pixel highlighted.
18005 Both options accept the following values:
18009 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18010 unlike the neighboring pixels of the same field. Examples of temporal outliers
18011 include the results of video dropouts, head clogs, or tape tracking issues.
18014 Identify @var{vertical line repetition}. Vertical line repetition includes
18015 similar rows of pixels within a frame. In born-digital video vertical line
18016 repetition is common, but this pattern is uncommon in video digitized from an
18017 analog source. When it occurs in video that results from the digitization of an
18018 analog source it can indicate concealment from a dropout compensator.
18021 Identify pixels that fall outside of legal broadcast range.
18025 Set the highlight color for the @option{out} option. The default color is
18029 @subsection Examples
18033 Output data of various video metrics:
18035 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18039 Output specific data about the minimum and maximum values of the Y plane per frame:
18041 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18045 Playback video while highlighting pixels that are outside of broadcast range in red.
18047 ffplay example.mov -vf signalstats="out=brng:color=red"
18051 Playback video with signalstats metadata drawn over the frame.
18053 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18056 The contents of signalstat_drawtext.txt used in the command are:
18059 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18060 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18061 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18062 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18070 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18071 input. In this case the matching between the inputs can be calculated additionally.
18072 The filter always passes through the first input. The signature of each stream can
18073 be written into a file.
18075 It accepts the following options:
18079 Enable or disable the matching process.
18081 Available values are:
18085 Disable the calculation of a matching (default).
18087 Calculate the matching for the whole video and output whether the whole video
18088 matches or only parts.
18090 Calculate only until a matching is found or the video ends. Should be faster in
18095 Set the number of inputs. The option value must be a non negative integer.
18096 Default value is 1.
18099 Set the path to which the output is written. If there is more than one input,
18100 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18101 integer), that will be replaced with the input number. If no filename is
18102 specified, no output will be written. This is the default.
18105 Choose the output format.
18107 Available values are:
18111 Use the specified binary representation (default).
18113 Use the specified xml representation.
18117 Set threshold to detect one word as similar. The option value must be an integer
18118 greater than zero. The default value is 9000.
18121 Set threshold to detect all words as similar. The option value must be an integer
18122 greater than zero. The default value is 60000.
18125 Set threshold to detect frames as similar. The option value must be an integer
18126 greater than zero. The default value is 116.
18129 Set the minimum length of a sequence in frames to recognize it as matching
18130 sequence. The option value must be a non negative integer value.
18131 The default value is 0.
18134 Set the minimum relation, that matching frames to all frames must have.
18135 The option value must be a double value between 0 and 1. The default value is 0.5.
18138 @subsection Examples
18142 To calculate the signature of an input video and store it in signature.bin:
18144 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18148 To detect whether two videos match and store the signatures in XML format in
18149 signature0.xml and signature1.xml:
18151 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 -
18159 Blur the input video without impacting the outlines.
18161 It accepts the following options:
18164 @item luma_radius, lr
18165 Set the luma radius. The option value must be a float number in
18166 the range [0.1,5.0] that specifies the variance of the gaussian filter
18167 used to blur the image (slower if larger). Default value is 1.0.
18169 @item luma_strength, ls
18170 Set the luma strength. The option value must be a float number
18171 in the range [-1.0,1.0] that configures the blurring. A value included
18172 in [0.0,1.0] will blur the image whereas a value included in
18173 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18175 @item luma_threshold, lt
18176 Set the luma threshold used as a coefficient to determine
18177 whether a pixel should be blurred or not. The option value must be an
18178 integer in the range [-30,30]. A value of 0 will filter all the image,
18179 a value included in [0,30] will filter flat areas and a value included
18180 in [-30,0] will filter edges. Default value is 0.
18182 @item chroma_radius, cr
18183 Set the chroma radius. The option value must be a float number in
18184 the range [0.1,5.0] that specifies the variance of the gaussian filter
18185 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18187 @item chroma_strength, cs
18188 Set the chroma strength. The option value must be a float number
18189 in the range [-1.0,1.0] that configures the blurring. A value included
18190 in [0.0,1.0] will blur the image whereas a value included in
18191 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18193 @item chroma_threshold, ct
18194 Set the chroma threshold used as a coefficient to determine
18195 whether a pixel should be blurred or not. The option value must be an
18196 integer in the range [-30,30]. A value of 0 will filter all the image,
18197 a value included in [0,30] will filter flat areas and a value included
18198 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18201 If a chroma option is not explicitly set, the corresponding luma value
18205 Apply sobel operator to input video stream.
18207 The filter accepts the following option:
18211 Set which planes will be processed, unprocessed planes will be copied.
18212 By default value 0xf, all planes will be processed.
18215 Set value which will be multiplied with filtered result.
18218 Set value which will be added to filtered result.
18221 @subsection Commands
18223 This filter supports the all above options as @ref{commands}.
18228 Apply a simple postprocessing filter that compresses and decompresses the image
18229 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18230 and average the results.
18232 The filter accepts the following options:
18236 Set quality. This option defines the number of levels for averaging. It accepts
18237 an integer in the range 0-6. If set to @code{0}, the filter will have no
18238 effect. A value of @code{6} means the higher quality. For each increment of
18239 that value the speed drops by a factor of approximately 2. Default value is
18243 Force a constant quantization parameter. If not set, the filter will use the QP
18244 from the video stream (if available).
18247 Set thresholding mode. Available modes are:
18251 Set hard thresholding (default).
18253 Set soft thresholding (better de-ringing effect, but likely blurrier).
18256 @item use_bframe_qp
18257 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18258 option may cause flicker since the B-Frames have often larger QP. Default is
18259 @code{0} (not enabled).
18262 @subsection Commands
18264 This filter supports the following commands:
18266 @item quality, level
18267 Set quality level. The value @code{max} can be used to set the maximum level,
18268 currently @code{6}.
18274 Scale the input by applying one of the super-resolution methods based on
18275 convolutional neural networks. Supported models:
18279 Super-Resolution Convolutional Neural Network model (SRCNN).
18280 See @url{https://arxiv.org/abs/1501.00092}.
18283 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18284 See @url{https://arxiv.org/abs/1609.05158}.
18287 Training scripts as well as scripts for model file (.pb) saving can be found at
18288 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18289 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18291 Native model files (.model) can be generated from TensorFlow model
18292 files (.pb) by using tools/python/convert.py
18294 The filter accepts the following options:
18298 Specify which DNN backend to use for model loading and execution. This option accepts
18299 the following values:
18303 Native implementation of DNN loading and execution.
18306 TensorFlow backend. To enable this backend you
18307 need to install the TensorFlow for C library (see
18308 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18309 @code{--enable-libtensorflow}
18312 Default value is @samp{native}.
18315 Set path to model file specifying network architecture and its parameters.
18316 Note that different backends use different file formats. TensorFlow backend
18317 can load files for both formats, while native backend can load files for only
18321 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18322 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18323 input upscaled using bicubic upscaling with proper scale factor.
18326 This feature can also be finished with @ref{dnn_processing} filter.
18330 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18332 This filter takes in input two input videos, the first input is
18333 considered the "main" source and is passed unchanged to the
18334 output. The second input is used as a "reference" video for computing
18337 Both video inputs must have the same resolution and pixel format for
18338 this filter to work correctly. Also it assumes that both inputs
18339 have the same number of frames, which are compared one by one.
18341 The filter stores the calculated SSIM of each frame.
18343 The description of the accepted parameters follows.
18346 @item stats_file, f
18347 If specified the filter will use the named file to save the SSIM of
18348 each individual frame. When filename equals "-" the data is sent to
18352 The file printed if @var{stats_file} is selected, contains a sequence of
18353 key/value pairs of the form @var{key}:@var{value} for each compared
18356 A description of each shown parameter follows:
18360 sequential number of the input frame, starting from 1
18362 @item Y, U, V, R, G, B
18363 SSIM of the compared frames for the component specified by the suffix.
18366 SSIM of the compared frames for the whole frame.
18369 Same as above but in dB representation.
18372 This filter also supports the @ref{framesync} options.
18374 @subsection Examples
18379 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18380 [main][ref] ssim="stats_file=stats.log" [out]
18383 On this example the input file being processed is compared with the
18384 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18385 is stored in @file{stats.log}.
18388 Another example with both psnr and ssim at same time:
18390 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18394 Another example with different containers:
18396 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 -
18402 Convert between different stereoscopic image formats.
18404 The filters accept the following options:
18408 Set stereoscopic image format of input.
18410 Available values for input image formats are:
18413 side by side parallel (left eye left, right eye right)
18416 side by side crosseye (right eye left, left eye right)
18419 side by side parallel with half width resolution
18420 (left eye left, right eye right)
18423 side by side crosseye with half width resolution
18424 (right eye left, left eye right)
18428 above-below (left eye above, right eye below)
18432 above-below (right eye above, left eye below)
18436 above-below with half height resolution
18437 (left eye above, right eye below)
18441 above-below with half height resolution
18442 (right eye above, left eye below)
18445 alternating frames (left eye first, right eye second)
18448 alternating frames (right eye first, left eye second)
18451 interleaved rows (left eye has top row, right eye starts on next row)
18454 interleaved rows (right eye has top row, left eye starts on next row)
18457 interleaved columns, left eye first
18460 interleaved columns, right eye first
18462 Default value is @samp{sbsl}.
18466 Set stereoscopic image format of output.
18470 side by side parallel (left eye left, right eye right)
18473 side by side crosseye (right eye left, left eye right)
18476 side by side parallel with half width resolution
18477 (left eye left, right eye right)
18480 side by side crosseye with half width resolution
18481 (right eye left, left eye right)
18485 above-below (left eye above, right eye below)
18489 above-below (right eye above, left eye below)
18493 above-below with half height resolution
18494 (left eye above, right eye below)
18498 above-below with half height resolution
18499 (right eye above, left eye below)
18502 alternating frames (left eye first, right eye second)
18505 alternating frames (right eye first, left eye second)
18508 interleaved rows (left eye has top row, right eye starts on next row)
18511 interleaved rows (right eye has top row, left eye starts on next row)
18514 anaglyph red/blue gray
18515 (red filter on left eye, blue filter on right eye)
18518 anaglyph red/green gray
18519 (red filter on left eye, green filter on right eye)
18522 anaglyph red/cyan gray
18523 (red filter on left eye, cyan filter on right eye)
18526 anaglyph red/cyan half colored
18527 (red filter on left eye, cyan filter on right eye)
18530 anaglyph red/cyan color
18531 (red filter on left eye, cyan filter on right eye)
18534 anaglyph red/cyan color optimized with the least squares projection of dubois
18535 (red filter on left eye, cyan filter on right eye)
18538 anaglyph green/magenta gray
18539 (green filter on left eye, magenta filter on right eye)
18542 anaglyph green/magenta half colored
18543 (green filter on left eye, magenta filter on right eye)
18546 anaglyph green/magenta colored
18547 (green filter on left eye, magenta filter on right eye)
18550 anaglyph green/magenta color optimized with the least squares projection of dubois
18551 (green filter on left eye, magenta filter on right eye)
18554 anaglyph yellow/blue gray
18555 (yellow filter on left eye, blue filter on right eye)
18558 anaglyph yellow/blue half colored
18559 (yellow filter on left eye, blue filter on right eye)
18562 anaglyph yellow/blue colored
18563 (yellow filter on left eye, blue filter on right eye)
18566 anaglyph yellow/blue color optimized with the least squares projection of dubois
18567 (yellow filter on left eye, blue filter on right eye)
18570 mono output (left eye only)
18573 mono output (right eye only)
18576 checkerboard, left eye first
18579 checkerboard, right eye first
18582 interleaved columns, left eye first
18585 interleaved columns, right eye first
18591 Default value is @samp{arcd}.
18594 @subsection Examples
18598 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18604 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18610 @section streamselect, astreamselect
18611 Select video or audio streams.
18613 The filter accepts the following options:
18617 Set number of inputs. Default is 2.
18620 Set input indexes to remap to outputs.
18623 @subsection Commands
18625 The @code{streamselect} and @code{astreamselect} filter supports the following
18630 Set input indexes to remap to outputs.
18633 @subsection Examples
18637 Select first 5 seconds 1st stream and rest of time 2nd stream:
18639 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18643 Same as above, but for audio:
18645 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18652 Draw subtitles on top of input video using the libass library.
18654 To enable compilation of this filter you need to configure FFmpeg with
18655 @code{--enable-libass}. This filter also requires a build with libavcodec and
18656 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18657 Alpha) subtitles format.
18659 The filter accepts the following options:
18663 Set the filename of the subtitle file to read. It must be specified.
18665 @item original_size
18666 Specify the size of the original video, the video for which the ASS file
18667 was composed. For the syntax of this option, check the
18668 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18669 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18670 correctly scale the fonts if the aspect ratio has been changed.
18673 Set a directory path containing fonts that can be used by the filter.
18674 These fonts will be used in addition to whatever the font provider uses.
18677 Process alpha channel, by default alpha channel is untouched.
18680 Set subtitles input character encoding. @code{subtitles} filter only. Only
18681 useful if not UTF-8.
18683 @item stream_index, si
18684 Set subtitles stream index. @code{subtitles} filter only.
18687 Override default style or script info parameters of the subtitles. It accepts a
18688 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18691 If the first key is not specified, it is assumed that the first value
18692 specifies the @option{filename}.
18694 For example, to render the file @file{sub.srt} on top of the input
18695 video, use the command:
18700 which is equivalent to:
18702 subtitles=filename=sub.srt
18705 To render the default subtitles stream from file @file{video.mkv}, use:
18707 subtitles=video.mkv
18710 To render the second subtitles stream from that file, use:
18712 subtitles=video.mkv:si=1
18715 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18716 @code{DejaVu Serif}, use:
18718 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18721 @section super2xsai
18723 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18724 Interpolate) pixel art scaling algorithm.
18726 Useful for enlarging pixel art images without reducing sharpness.
18730 Swap two rectangular objects in video.
18732 This filter accepts the following options:
18742 Set 1st rect x coordinate.
18745 Set 1st rect y coordinate.
18748 Set 2nd rect x coordinate.
18751 Set 2nd rect y coordinate.
18753 All expressions are evaluated once for each frame.
18756 The all options are expressions containing the following constants:
18761 The input width and height.
18764 same as @var{w} / @var{h}
18767 input sample aspect ratio
18770 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18773 The number of the input frame, starting from 0.
18776 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18779 the position in the file of the input frame, NAN if unknown
18786 Blend successive video frames.
18792 Apply telecine process to the video.
18794 This filter accepts the following options:
18803 The default value is @code{top}.
18807 A string of numbers representing the pulldown pattern you wish to apply.
18808 The default value is @code{23}.
18812 Some typical patterns:
18817 24p: 2332 (preferred)
18824 24p: 222222222223 ("Euro pulldown")
18829 @section thistogram
18831 Compute and draw a color distribution histogram for the input video across time.
18833 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18834 at certain time, this filter shows also past histograms of number of frames defined
18835 by @code{width} option.
18837 The computed histogram is a representation of the color component
18838 distribution in an image.
18840 The filter accepts the following options:
18844 Set width of single color component output. Default value is @code{0}.
18845 Value of @code{0} means width will be picked from input video.
18846 This also set number of passed histograms to keep.
18847 Allowed range is [0, 8192].
18849 @item display_mode, d
18851 It accepts the following values:
18854 Per color component graphs are placed below each other.
18857 Per color component graphs are placed side by side.
18860 Presents information identical to that in the @code{parade}, except
18861 that the graphs representing color components are superimposed directly
18864 Default is @code{stack}.
18866 @item levels_mode, m
18867 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18868 Default is @code{linear}.
18870 @item components, c
18871 Set what color components to display.
18872 Default is @code{7}.
18875 Set background opacity. Default is @code{0.9}.
18878 Show envelope. Default is disabled.
18881 Set envelope color. Default is @code{gold}.
18886 Available values for slide is:
18889 Draw new frame when right border is reached.
18892 Replace old columns with new ones.
18895 Scroll from right to left.
18898 Scroll from left to right.
18901 Draw single picture.
18904 Default is @code{replace}.
18909 Apply threshold effect to video stream.
18911 This filter needs four video streams to perform thresholding.
18912 First stream is stream we are filtering.
18913 Second stream is holding threshold values, third stream is holding min values,
18914 and last, fourth stream is holding max values.
18916 The filter accepts the following option:
18920 Set which planes will be processed, unprocessed planes will be copied.
18921 By default value 0xf, all planes will be processed.
18924 For example if first stream pixel's component value is less then threshold value
18925 of pixel component from 2nd threshold stream, third stream value will picked,
18926 otherwise fourth stream pixel component value will be picked.
18928 Using color source filter one can perform various types of thresholding:
18930 @subsection Examples
18934 Binary threshold, using gray color as threshold:
18936 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18940 Inverted binary threshold, using gray color as threshold:
18942 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18946 Truncate binary threshold, using gray color as threshold:
18948 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18952 Threshold to zero, using gray color as threshold:
18954 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18958 Inverted threshold to zero, using gray color as threshold:
18960 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18965 Select the most representative frame in a given sequence of consecutive frames.
18967 The filter accepts the following options:
18971 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18972 will pick one of them, and then handle the next batch of @var{n} frames until
18973 the end. Default is @code{100}.
18976 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18977 value will result in a higher memory usage, so a high value is not recommended.
18979 @subsection Examples
18983 Extract one picture each 50 frames:
18989 Complete example of a thumbnail creation with @command{ffmpeg}:
18991 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18998 Tile several successive frames together.
19000 The @ref{untile} filter can do the reverse.
19002 The filter accepts the following options:
19007 Set the grid size (i.e. the number of lines and columns). For the syntax of
19008 this option, check the
19009 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19012 Set the maximum number of frames to render in the given area. It must be less
19013 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19014 the area will be used.
19017 Set the outer border margin in pixels.
19020 Set the inner border thickness (i.e. the number of pixels between frames). For
19021 more advanced padding options (such as having different values for the edges),
19022 refer to the pad video filter.
19025 Specify the color of the unused area. For the syntax of this option, check the
19026 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19027 The default value of @var{color} is "black".
19030 Set the number of frames to overlap when tiling several successive frames together.
19031 The value must be between @code{0} and @var{nb_frames - 1}.
19034 Set the number of frames to initially be empty before displaying first output frame.
19035 This controls how soon will one get first output frame.
19036 The value must be between @code{0} and @var{nb_frames - 1}.
19039 @subsection Examples
19043 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19045 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19047 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19048 duplicating each output frame to accommodate the originally detected frame
19052 Display @code{5} pictures in an area of @code{3x2} frames,
19053 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19054 mixed flat and named options:
19056 tile=3x2:nb_frames=5:padding=7:margin=2
19060 @section tinterlace
19062 Perform various types of temporal field interlacing.
19064 Frames are counted starting from 1, so the first input frame is
19067 The filter accepts the following options:
19072 Specify the mode of the interlacing. This option can also be specified
19073 as a value alone. See below for a list of values for this option.
19075 Available values are:
19079 Move odd frames into the upper field, even into the lower field,
19080 generating a double height frame at half frame rate.
19084 Frame 1 Frame 2 Frame 3 Frame 4
19086 11111 22222 33333 44444
19087 11111 22222 33333 44444
19088 11111 22222 33333 44444
19089 11111 22222 33333 44444
19103 Only output odd frames, even frames are dropped, generating a frame with
19104 unchanged height at half frame rate.
19109 Frame 1 Frame 2 Frame 3 Frame 4
19111 11111 22222 33333 44444
19112 11111 22222 33333 44444
19113 11111 22222 33333 44444
19114 11111 22222 33333 44444
19124 Only output even frames, odd frames are dropped, generating a frame with
19125 unchanged height at half frame rate.
19130 Frame 1 Frame 2 Frame 3 Frame 4
19132 11111 22222 33333 44444
19133 11111 22222 33333 44444
19134 11111 22222 33333 44444
19135 11111 22222 33333 44444
19145 Expand each frame to full height, but pad alternate lines with black,
19146 generating a frame with double height at the same input frame rate.
19151 Frame 1 Frame 2 Frame 3 Frame 4
19153 11111 22222 33333 44444
19154 11111 22222 33333 44444
19155 11111 22222 33333 44444
19156 11111 22222 33333 44444
19159 11111 ..... 33333 .....
19160 ..... 22222 ..... 44444
19161 11111 ..... 33333 .....
19162 ..... 22222 ..... 44444
19163 11111 ..... 33333 .....
19164 ..... 22222 ..... 44444
19165 11111 ..... 33333 .....
19166 ..... 22222 ..... 44444
19170 @item interleave_top, 4
19171 Interleave the upper field from odd frames with the lower field from
19172 even frames, generating a frame with unchanged height at half frame rate.
19177 Frame 1 Frame 2 Frame 3 Frame 4
19179 11111<- 22222 33333<- 44444
19180 11111 22222<- 33333 44444<-
19181 11111<- 22222 33333<- 44444
19182 11111 22222<- 33333 44444<-
19192 @item interleave_bottom, 5
19193 Interleave the lower field from odd frames with the upper field from
19194 even frames, generating a frame with unchanged height at half frame rate.
19199 Frame 1 Frame 2 Frame 3 Frame 4
19201 11111 22222<- 33333 44444<-
19202 11111<- 22222 33333<- 44444
19203 11111 22222<- 33333 44444<-
19204 11111<- 22222 33333<- 44444
19214 @item interlacex2, 6
19215 Double frame rate with unchanged height. Frames are inserted each
19216 containing the second temporal field from the previous input frame and
19217 the first temporal field from the next input frame. This mode relies on
19218 the top_field_first flag. Useful for interlaced video displays with no
19219 field synchronisation.
19224 Frame 1 Frame 2 Frame 3 Frame 4
19226 11111 22222 33333 44444
19227 11111 22222 33333 44444
19228 11111 22222 33333 44444
19229 11111 22222 33333 44444
19232 11111 22222 22222 33333 33333 44444 44444
19233 11111 11111 22222 22222 33333 33333 44444
19234 11111 22222 22222 33333 33333 44444 44444
19235 11111 11111 22222 22222 33333 33333 44444
19240 Move odd frames into the upper field, even into the lower field,
19241 generating a double height frame at same frame rate.
19246 Frame 1 Frame 2 Frame 3 Frame 4
19248 11111 22222 33333 44444
19249 11111 22222 33333 44444
19250 11111 22222 33333 44444
19251 11111 22222 33333 44444
19254 11111 33333 33333 55555
19255 22222 22222 44444 44444
19256 11111 33333 33333 55555
19257 22222 22222 44444 44444
19258 11111 33333 33333 55555
19259 22222 22222 44444 44444
19260 11111 33333 33333 55555
19261 22222 22222 44444 44444
19266 Numeric values are deprecated but are accepted for backward
19267 compatibility reasons.
19269 Default mode is @code{merge}.
19272 Specify flags influencing the filter process.
19274 Available value for @var{flags} is:
19277 @item low_pass_filter, vlpf
19278 Enable linear vertical low-pass filtering in the filter.
19279 Vertical low-pass filtering is required when creating an interlaced
19280 destination from a progressive source which contains high-frequency
19281 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19284 @item complex_filter, cvlpf
19285 Enable complex vertical low-pass filtering.
19286 This will slightly less reduce interlace 'twitter' and Moire
19287 patterning but better retain detail and subjective sharpness impression.
19290 Bypass already interlaced frames, only adjust the frame rate.
19293 Vertical low-pass filtering and bypassing already interlaced frames can only be
19294 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19299 Pick median pixels from several successive input video frames.
19301 The filter accepts the following options:
19305 Set radius of median filter.
19306 Default is 1. Allowed range is from 1 to 127.
19309 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19312 Set median percentile. Default value is @code{0.5}.
19313 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19314 minimum values, and @code{1} maximum values.
19319 Mix successive video frames.
19321 A description of the accepted options follows.
19325 The number of successive frames to mix. If unspecified, it defaults to 3.
19328 Specify weight of each input video frame.
19329 Each weight is separated by space. If number of weights is smaller than
19330 number of @var{frames} last specified weight will be used for all remaining
19334 Specify scale, if it is set it will be multiplied with sum
19335 of each weight multiplied with pixel values to give final destination
19336 pixel value. By default @var{scale} is auto scaled to sum of weights.
19339 @subsection Examples
19343 Average 7 successive frames:
19345 tmix=frames=7:weights="1 1 1 1 1 1 1"
19349 Apply simple temporal convolution:
19351 tmix=frames=3:weights="-1 3 -1"
19355 Similar as above but only showing temporal differences:
19357 tmix=frames=3:weights="-1 2 -1":scale=1
19363 Tone map colors from different dynamic ranges.
19365 This filter expects data in single precision floating point, as it needs to
19366 operate on (and can output) out-of-range values. Another filter, such as
19367 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19369 The tonemapping algorithms implemented only work on linear light, so input
19370 data should be linearized beforehand (and possibly correctly tagged).
19373 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19376 @subsection Options
19377 The filter accepts the following options.
19381 Set the tone map algorithm to use.
19383 Possible values are:
19386 Do not apply any tone map, only desaturate overbright pixels.
19389 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19390 in-range values, while distorting out-of-range values.
19393 Stretch the entire reference gamut to a linear multiple of the display.
19396 Fit a logarithmic transfer between the tone curves.
19399 Preserve overall image brightness with a simple curve, using nonlinear
19400 contrast, which results in flattening details and degrading color accuracy.
19403 Preserve both dark and bright details better than @var{reinhard}, at the cost
19404 of slightly darkening everything. Use it when detail preservation is more
19405 important than color and brightness accuracy.
19408 Smoothly map out-of-range values, while retaining contrast and colors for
19409 in-range material as much as possible. Use it when color accuracy is more
19410 important than detail preservation.
19416 Tune the tone mapping algorithm.
19418 This affects the following algorithms:
19424 Specifies the scale factor to use while stretching.
19428 Specifies the exponent of the function.
19432 Specify an extra linear coefficient to multiply into the signal before clipping.
19436 Specify the local contrast coefficient at the display peak.
19437 Default to 0.5, which means that in-gamut values will be about half as bright
19444 Specify the transition point from linear to mobius transform. Every value
19445 below this point is guaranteed to be mapped 1:1. The higher the value, the
19446 more accurate the result will be, at the cost of losing bright details.
19447 Default to 0.3, which due to the steep initial slope still preserves in-range
19448 colors fairly accurately.
19452 Apply desaturation for highlights that exceed this level of brightness. The
19453 higher the parameter, the more color information will be preserved. This
19454 setting helps prevent unnaturally blown-out colors for super-highlights, by
19455 (smoothly) turning into white instead. This makes images feel more natural,
19456 at the cost of reducing information about out-of-range colors.
19458 The default of 2.0 is somewhat conservative and will mostly just apply to
19459 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19461 This option works only if the input frame has a supported color tag.
19464 Override signal/nominal/reference peak with this value. Useful when the
19465 embedded peak information in display metadata is not reliable or when tone
19466 mapping from a lower range to a higher range.
19471 Temporarily pad video frames.
19473 The filter accepts the following options:
19477 Specify number of delay frames before input video stream. Default is 0.
19480 Specify number of padding frames after input video stream.
19481 Set to -1 to pad indefinitely. Default is 0.
19484 Set kind of frames added to beginning of stream.
19485 Can be either @var{add} or @var{clone}.
19486 With @var{add} frames of solid-color are added.
19487 With @var{clone} frames are clones of first frame.
19488 Default is @var{add}.
19491 Set kind of frames added to end of stream.
19492 Can be either @var{add} or @var{clone}.
19493 With @var{add} frames of solid-color are added.
19494 With @var{clone} frames are clones of last frame.
19495 Default is @var{add}.
19497 @item start_duration, stop_duration
19498 Specify the duration of the start/stop delay. See
19499 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19500 for the accepted syntax.
19501 These options override @var{start} and @var{stop}. Default is 0.
19504 Specify the color of the padded area. For the syntax of this option,
19505 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19506 manual,ffmpeg-utils}.
19508 The default value of @var{color} is "black".
19514 Transpose rows with columns in the input video and optionally flip it.
19516 It accepts the following parameters:
19521 Specify the transposition direction.
19523 Can assume the following values:
19525 @item 0, 4, cclock_flip
19526 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19534 Rotate by 90 degrees clockwise, that is:
19542 Rotate by 90 degrees counterclockwise, that is:
19549 @item 3, 7, clock_flip
19550 Rotate by 90 degrees clockwise and vertically flip, that is:
19558 For values between 4-7, the transposition is only done if the input
19559 video geometry is portrait and not landscape. These values are
19560 deprecated, the @code{passthrough} option should be used instead.
19562 Numerical values are deprecated, and should be dropped in favor of
19563 symbolic constants.
19566 Do not apply the transposition if the input geometry matches the one
19567 specified by the specified value. It accepts the following values:
19570 Always apply transposition.
19572 Preserve portrait geometry (when @var{height} >= @var{width}).
19574 Preserve landscape geometry (when @var{width} >= @var{height}).
19577 Default value is @code{none}.
19580 For example to rotate by 90 degrees clockwise and preserve portrait
19583 transpose=dir=1:passthrough=portrait
19586 The command above can also be specified as:
19588 transpose=1:portrait
19591 @section transpose_npp
19593 Transpose rows with columns in the input video and optionally flip it.
19594 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19596 It accepts the following parameters:
19601 Specify the transposition direction.
19603 Can assume the following values:
19606 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19609 Rotate by 90 degrees clockwise.
19612 Rotate by 90 degrees counterclockwise.
19615 Rotate by 90 degrees clockwise and vertically flip.
19619 Do not apply the transposition if the input geometry matches the one
19620 specified by the specified value. It accepts the following values:
19623 Always apply transposition. (default)
19625 Preserve portrait geometry (when @var{height} >= @var{width}).
19627 Preserve landscape geometry (when @var{width} >= @var{height}).
19633 Trim the input so that the output contains one continuous subpart of the input.
19635 It accepts the following parameters:
19638 Specify the time of the start of the kept section, i.e. the frame with the
19639 timestamp @var{start} will be the first frame in the output.
19642 Specify the time of the first frame that will be dropped, i.e. the frame
19643 immediately preceding the one with the timestamp @var{end} will be the last
19644 frame in the output.
19647 This is the same as @var{start}, except this option sets the start timestamp
19648 in timebase units instead of seconds.
19651 This is the same as @var{end}, except this option sets the end timestamp
19652 in timebase units instead of seconds.
19655 The maximum duration of the output in seconds.
19658 The number of the first frame that should be passed to the output.
19661 The number of the first frame that should be dropped.
19664 @option{start}, @option{end}, and @option{duration} are expressed as time
19665 duration specifications; see
19666 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19667 for the accepted syntax.
19669 Note that the first two sets of the start/end options and the @option{duration}
19670 option look at the frame timestamp, while the _frame variants simply count the
19671 frames that pass through the filter. Also note that this filter does not modify
19672 the timestamps. If you wish for the output timestamps to start at zero, insert a
19673 setpts filter after the trim filter.
19675 If multiple start or end options are set, this filter tries to be greedy and
19676 keep all the frames that match at least one of the specified constraints. To keep
19677 only the part that matches all the constraints at once, chain multiple trim
19680 The defaults are such that all the input is kept. So it is possible to set e.g.
19681 just the end values to keep everything before the specified time.
19686 Drop everything except the second minute of input:
19688 ffmpeg -i INPUT -vf trim=60:120
19692 Keep only the first second:
19694 ffmpeg -i INPUT -vf trim=duration=1
19699 @section unpremultiply
19700 Apply alpha unpremultiply effect to input video stream using first plane
19701 of second stream as alpha.
19703 Both streams must have same dimensions and same pixel format.
19705 The filter accepts the following option:
19709 Set which planes will be processed, unprocessed planes will be copied.
19710 By default value 0xf, all planes will be processed.
19712 If the format has 1 or 2 components, then luma is bit 0.
19713 If the format has 3 or 4 components:
19714 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19715 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19716 If present, the alpha channel is always the last bit.
19719 Do not require 2nd input for processing, instead use alpha plane from input stream.
19725 Sharpen or blur the input video.
19727 It accepts the following parameters:
19730 @item luma_msize_x, lx
19731 Set the luma matrix horizontal size. It must be an odd integer between
19732 3 and 23. The default value is 5.
19734 @item luma_msize_y, ly
19735 Set the luma matrix vertical size. It must be an odd integer between 3
19736 and 23. The default value is 5.
19738 @item luma_amount, la
19739 Set the luma effect strength. It must be a floating point number, reasonable
19740 values lay between -1.5 and 1.5.
19742 Negative values will blur the input video, while positive values will
19743 sharpen it, a value of zero will disable the effect.
19745 Default value is 1.0.
19747 @item chroma_msize_x, cx
19748 Set the chroma matrix horizontal size. It must be an odd integer
19749 between 3 and 23. The default value is 5.
19751 @item chroma_msize_y, cy
19752 Set the chroma matrix vertical size. It must be an odd integer
19753 between 3 and 23. The default value is 5.
19755 @item chroma_amount, ca
19756 Set the chroma effect strength. It must be a floating point number, reasonable
19757 values lay between -1.5 and 1.5.
19759 Negative values will blur the input video, while positive values will
19760 sharpen it, a value of zero will disable the effect.
19762 Default value is 0.0.
19766 All parameters are optional and default to the equivalent of the
19767 string '5:5:1.0:5:5:0.0'.
19769 @subsection Examples
19773 Apply strong luma sharpen effect:
19775 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19779 Apply a strong blur of both luma and chroma parameters:
19781 unsharp=7:7:-2:7:7:-2
19788 Decompose a video made of tiled images into the individual images.
19790 The frame rate of the output video is the frame rate of the input video
19791 multiplied by the number of tiles.
19793 This filter does the reverse of @ref{tile}.
19795 The filter accepts the following options:
19800 Set the grid size (i.e. the number of lines and columns). For the syntax of
19801 this option, check the
19802 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19805 @subsection Examples
19809 Produce a 1-second video from a still image file made of 25 frames stacked
19810 vertically, like an analogic film reel:
19812 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19818 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19819 the image at several (or - in the case of @option{quality} level @code{8} - all)
19820 shifts and average the results.
19822 The way this differs from the behavior of spp is that uspp actually encodes &
19823 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19824 DCT similar to MJPEG.
19826 The filter accepts the following options:
19830 Set quality. This option defines the number of levels for averaging. It accepts
19831 an integer in the range 0-8. If set to @code{0}, the filter will have no
19832 effect. A value of @code{8} means the higher quality. For each increment of
19833 that value the speed drops by a factor of approximately 2. Default value is
19837 Force a constant quantization parameter. If not set, the filter will use the QP
19838 from the video stream (if available).
19843 Convert 360 videos between various formats.
19845 The filter accepts the following options:
19851 Set format of the input/output video.
19859 Equirectangular projection.
19864 Cubemap with 3x2/6x1/1x6 layout.
19866 Format specific options:
19871 Set padding proportion for the input/output cubemap. Values in decimals.
19878 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)
19881 Default value is @b{@samp{0}}.
19882 Maximum value is @b{@samp{0.1}}.
19886 Set fixed padding for the input/output cubemap. Values in pixels.
19888 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19892 Set order of faces for the input/output cubemap. Choose one direction for each position.
19894 Designation of directions:
19910 Default value is @b{@samp{rludfb}}.
19914 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19916 Designation of angles:
19919 0 degrees clockwise
19921 90 degrees clockwise
19923 180 degrees clockwise
19925 270 degrees clockwise
19928 Default value is @b{@samp{000000}}.
19932 Equi-Angular Cubemap.
19939 Format specific options:
19944 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19946 If diagonal field of view is set it overrides horizontal and vertical field of view.
19951 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19953 If diagonal field of view is set it overrides horizontal and vertical field of view.
19959 Format specific options:
19964 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19966 If diagonal field of view is set it overrides horizontal and vertical field of view.
19971 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19973 If diagonal field of view is set it overrides horizontal and vertical field of view.
19979 Facebook's 360 formats.
19982 Stereographic format.
19984 Format specific options:
19989 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19991 If diagonal field of view is set it overrides horizontal and vertical field of view.
19996 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19998 If diagonal field of view is set it overrides horizontal and vertical field of view.
20005 Ball format, gives significant distortion toward the back.
20008 Hammer-Aitoff map projection format.
20011 Sinusoidal map projection format.
20014 Fisheye projection.
20016 Format specific options:
20021 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20023 If diagonal field of view is set it overrides horizontal and vertical field of view.
20028 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20030 If diagonal field of view is set it overrides horizontal and vertical field of view.
20034 Pannini projection.
20036 Format specific options:
20039 Set output pannini parameter.
20042 Set input pannini parameter.
20046 Cylindrical projection.
20048 Format specific options:
20053 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20055 If diagonal field of view is set it overrides horizontal and vertical field of view.
20060 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20062 If diagonal field of view is set it overrides horizontal and vertical field of view.
20066 Perspective projection. @i{(output only)}
20068 Format specific options:
20071 Set perspective parameter.
20075 Tetrahedron projection.
20078 Truncated square pyramid projection.
20082 Half equirectangular projection.
20087 Format specific options:
20092 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20094 If diagonal field of view is set it overrides horizontal and vertical field of view.
20099 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20101 If diagonal field of view is set it overrides horizontal and vertical field of view.
20105 Orthographic format.
20107 Format specific options:
20112 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20114 If diagonal field of view is set it overrides horizontal and vertical field of view.
20119 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20121 If diagonal field of view is set it overrides horizontal and vertical field of view.
20125 Octahedron projection.
20129 Set interpolation method.@*
20130 @i{Note: more complex interpolation methods require much more memory to run.}
20140 Bilinear interpolation.
20142 Lagrange9 interpolation.
20145 Bicubic interpolation.
20148 Lanczos interpolation.
20151 Spline16 interpolation.
20154 Gaussian interpolation.
20156 Mitchell interpolation.
20159 Default value is @b{@samp{line}}.
20163 Set the output video resolution.
20165 Default resolution depends on formats.
20169 Set the input/output stereo format.
20180 Default value is @b{@samp{2d}} for input and output format.
20185 Set rotation for the output video. Values in degrees.
20188 Set rotation order for the output video. Choose one item for each position.
20199 Default value is @b{@samp{ypr}}.
20204 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20208 Set if input video is flipped horizontally/vertically. Boolean values.
20211 Set if input video is transposed. Boolean value, by default disabled.
20214 Set if output video needs to be transposed. Boolean value, by default disabled.
20217 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20220 @subsection Examples
20224 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20226 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20229 Extract back view of Equi-Angular Cubemap:
20231 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20234 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20236 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20240 @subsection Commands
20242 This filter supports subset of above options as @ref{commands}.
20244 @section vaguedenoiser
20246 Apply a wavelet based denoiser.
20248 It transforms each frame from the video input into the wavelet domain,
20249 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20250 the obtained coefficients. It does an inverse wavelet transform after.
20251 Due to wavelet properties, it should give a nice smoothed result, and
20252 reduced noise, without blurring picture features.
20254 This filter accepts the following options:
20258 The filtering strength. The higher, the more filtered the video will be.
20259 Hard thresholding can use a higher threshold than soft thresholding
20260 before the video looks overfiltered. Default value is 2.
20263 The filtering method the filter will use.
20265 It accepts the following values:
20268 All values under the threshold will be zeroed.
20271 All values under the threshold will be zeroed. All values above will be
20272 reduced by the threshold.
20275 Scales or nullifies coefficients - intermediary between (more) soft and
20276 (less) hard thresholding.
20279 Default is garrote.
20282 Number of times, the wavelet will decompose the picture. Picture can't
20283 be decomposed beyond a particular point (typically, 8 for a 640x480
20284 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20287 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20290 A list of the planes to process. By default all planes are processed.
20293 The threshold type the filter will use.
20295 It accepts the following values:
20298 Threshold used is same for all decompositions.
20301 Threshold used depends also on each decomposition coefficients.
20304 Default is universal.
20307 @section vectorscope
20309 Display 2 color component values in the two dimensional graph (which is called
20312 This filter accepts the following options:
20316 Set vectorscope mode.
20318 It accepts the following values:
20322 Gray values are displayed on graph, higher brightness means more pixels have
20323 same component color value on location in graph. This is the default mode.
20326 Gray values are displayed on graph. Surrounding pixels values which are not
20327 present in video frame are drawn in gradient of 2 color components which are
20328 set by option @code{x} and @code{y}. The 3rd color component is static.
20331 Actual color components values present in video frame are displayed on graph.
20334 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20335 on graph increases value of another color component, which is luminance by
20336 default values of @code{x} and @code{y}.
20339 Actual colors present in video frame are displayed on graph. If two different
20340 colors map to same position on graph then color with higher value of component
20341 not present in graph is picked.
20344 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20345 component picked from radial gradient.
20349 Set which color component will be represented on X-axis. Default is @code{1}.
20352 Set which color component will be represented on Y-axis. Default is @code{2}.
20355 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20356 of color component which represents frequency of (X, Y) location in graph.
20361 No envelope, this is default.
20364 Instant envelope, even darkest single pixel will be clearly highlighted.
20367 Hold maximum and minimum values presented in graph over time. This way you
20368 can still spot out of range values without constantly looking at vectorscope.
20371 Peak and instant envelope combined together.
20375 Set what kind of graticule to draw.
20384 Set graticule opacity.
20387 Set graticule flags.
20391 Draw graticule for white point.
20394 Draw graticule for black point.
20397 Draw color points short names.
20401 Set background opacity.
20403 @item lthreshold, l
20404 Set low threshold for color component not represented on X or Y axis.
20405 Values lower than this value will be ignored. Default is 0.
20406 Note this value is multiplied with actual max possible value one pixel component
20407 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20410 @item hthreshold, h
20411 Set high threshold for color component not represented on X or Y axis.
20412 Values higher than this value will be ignored. Default is 1.
20413 Note this value is multiplied with actual max possible value one pixel component
20414 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20415 is 0.9 * 255 = 230.
20417 @item colorspace, c
20418 Set what kind of colorspace to use when drawing graticule.
20428 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20429 This means no tint, and output will remain gray.
20432 @anchor{vidstabdetect}
20433 @section vidstabdetect
20435 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20436 @ref{vidstabtransform} for pass 2.
20438 This filter generates a file with relative translation and rotation
20439 transform information about subsequent frames, which is then used by
20440 the @ref{vidstabtransform} filter.
20442 To enable compilation of this filter you need to configure FFmpeg with
20443 @code{--enable-libvidstab}.
20445 This filter accepts the following options:
20449 Set the path to the file used to write the transforms information.
20450 Default value is @file{transforms.trf}.
20453 Set how shaky the video is and how quick the camera is. It accepts an
20454 integer in the range 1-10, a value of 1 means little shakiness, a
20455 value of 10 means strong shakiness. Default value is 5.
20458 Set the accuracy of the detection process. It must be a value in the
20459 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20460 accuracy. Default value is 15.
20463 Set stepsize of the search process. The region around minimum is
20464 scanned with 1 pixel resolution. Default value is 6.
20467 Set minimum contrast. Below this value a local measurement field is
20468 discarded. Must be a floating point value in the range 0-1. Default
20472 Set reference frame number for tripod mode.
20474 If enabled, the motion of the frames is compared to a reference frame
20475 in the filtered stream, identified by the specified number. The idea
20476 is to compensate all movements in a more-or-less static scene and keep
20477 the camera view absolutely still.
20479 If set to 0, it is disabled. The frames are counted starting from 1.
20482 Show fields and transforms in the resulting frames. It accepts an
20483 integer in the range 0-2. Default value is 0, which disables any
20487 @subsection Examples
20491 Use default values:
20497 Analyze strongly shaky movie and put the results in file
20498 @file{mytransforms.trf}:
20500 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20504 Visualize the result of internal transformations in the resulting
20507 vidstabdetect=show=1
20511 Analyze a video with medium shakiness using @command{ffmpeg}:
20513 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20517 @anchor{vidstabtransform}
20518 @section vidstabtransform
20520 Video stabilization/deshaking: pass 2 of 2,
20521 see @ref{vidstabdetect} for pass 1.
20523 Read a file with transform information for each frame and
20524 apply/compensate them. Together with the @ref{vidstabdetect}
20525 filter this can be used to deshake videos. See also
20526 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20527 the @ref{unsharp} filter, see below.
20529 To enable compilation of this filter you need to configure FFmpeg with
20530 @code{--enable-libvidstab}.
20532 @subsection Options
20536 Set path to the file used to read the transforms. Default value is
20537 @file{transforms.trf}.
20540 Set the number of frames (value*2 + 1) used for lowpass filtering the
20541 camera movements. Default value is 10.
20543 For example a number of 10 means that 21 frames are used (10 in the
20544 past and 10 in the future) to smoothen the motion in the video. A
20545 larger value leads to a smoother video, but limits the acceleration of
20546 the camera (pan/tilt movements). 0 is a special case where a static
20547 camera is simulated.
20550 Set the camera path optimization algorithm.
20552 Accepted values are:
20555 gaussian kernel low-pass filter on camera motion (default)
20557 averaging on transformations
20561 Set maximal number of pixels to translate frames. Default value is -1,
20565 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20566 value is -1, meaning no limit.
20569 Specify how to deal with borders that may be visible due to movement
20572 Available values are:
20575 keep image information from previous frame (default)
20577 fill the border black
20581 Invert transforms if set to 1. Default value is 0.
20584 Consider transforms as relative to previous frame if set to 1,
20585 absolute if set to 0. Default value is 0.
20588 Set percentage to zoom. A positive value will result in a zoom-in
20589 effect, a negative value in a zoom-out effect. Default value is 0 (no
20593 Set optimal zooming to avoid borders.
20595 Accepted values are:
20600 optimal static zoom value is determined (only very strong movements
20601 will lead to visible borders) (default)
20603 optimal adaptive zoom value is determined (no borders will be
20604 visible), see @option{zoomspeed}
20607 Note that the value given at zoom is added to the one calculated here.
20610 Set percent to zoom maximally each frame (enabled when
20611 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20615 Specify type of interpolation.
20617 Available values are:
20622 linear only horizontal
20624 linear in both directions (default)
20626 cubic in both directions (slow)
20630 Enable virtual tripod mode if set to 1, which is equivalent to
20631 @code{relative=0:smoothing=0}. Default value is 0.
20633 Use also @code{tripod} option of @ref{vidstabdetect}.
20636 Increase log verbosity if set to 1. Also the detected global motions
20637 are written to the temporary file @file{global_motions.trf}. Default
20641 @subsection Examples
20645 Use @command{ffmpeg} for a typical stabilization with default values:
20647 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20650 Note the use of the @ref{unsharp} filter which is always recommended.
20653 Zoom in a bit more and load transform data from a given file:
20655 vidstabtransform=zoom=5:input="mytransforms.trf"
20659 Smoothen the video even more:
20661 vidstabtransform=smoothing=30
20667 Flip the input video vertically.
20669 For example, to vertically flip a video with @command{ffmpeg}:
20671 ffmpeg -i in.avi -vf "vflip" out.avi
20676 Detect variable frame rate video.
20678 This filter tries to detect if the input is variable or constant frame rate.
20680 At end it will output number of frames detected as having variable delta pts,
20681 and ones with constant delta pts.
20682 If there was frames with variable delta, than it will also show min, max and
20683 average delta encountered.
20687 Boost or alter saturation.
20689 The filter accepts the following options:
20692 Set strength of boost if positive value or strength of alter if negative value.
20693 Default is 0. Allowed range is from -2 to 2.
20696 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20699 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20702 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20705 Set the red luma coefficient.
20708 Set the green luma coefficient.
20711 Set the blue luma coefficient.
20714 If @code{intensity} is negative and this is set to 1, colors will change,
20715 otherwise colors will be less saturated, more towards gray.
20718 @subsection Commands
20720 This filter supports the all above options as @ref{commands}.
20725 Make or reverse a natural vignetting effect.
20727 The filter accepts the following options:
20731 Set lens angle expression as a number of radians.
20733 The value is clipped in the @code{[0,PI/2]} range.
20735 Default value: @code{"PI/5"}
20739 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20743 Set forward/backward mode.
20745 Available modes are:
20748 The larger the distance from the central point, the darker the image becomes.
20751 The larger the distance from the central point, the brighter the image becomes.
20752 This can be used to reverse a vignette effect, though there is no automatic
20753 detection to extract the lens @option{angle} and other settings (yet). It can
20754 also be used to create a burning effect.
20757 Default value is @samp{forward}.
20760 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20762 It accepts the following values:
20765 Evaluate expressions only once during the filter initialization.
20768 Evaluate expressions for each incoming frame. This is way slower than the
20769 @samp{init} mode since it requires all the scalers to be re-computed, but it
20770 allows advanced dynamic expressions.
20773 Default value is @samp{init}.
20776 Set dithering to reduce the circular banding effects. Default is @code{1}
20780 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20781 Setting this value to the SAR of the input will make a rectangular vignetting
20782 following the dimensions of the video.
20784 Default is @code{1/1}.
20787 @subsection Expressions
20789 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20790 following parameters.
20795 input width and height
20798 the number of input frame, starting from 0
20801 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20802 @var{TB} units, NAN if undefined
20805 frame rate of the input video, NAN if the input frame rate is unknown
20808 the PTS (Presentation TimeStamp) of the filtered video frame,
20809 expressed in seconds, NAN if undefined
20812 time base of the input video
20816 @subsection Examples
20820 Apply simple strong vignetting effect:
20826 Make a flickering vignetting:
20828 vignette='PI/4+random(1)*PI/50':eval=frame
20833 @section vmafmotion
20835 Obtain the average VMAF motion score of a video.
20836 It is one of the component metrics of VMAF.
20838 The obtained average motion score is printed through the logging system.
20840 The filter accepts the following options:
20844 If specified, the filter will use the named file to save the motion score of
20845 each frame with respect to the previous frame.
20846 When filename equals "-" the data is sent to standard output.
20851 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20855 Stack input videos vertically.
20857 All streams must be of same pixel format and of same width.
20859 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20860 to create same output.
20862 The filter accepts the following options:
20866 Set number of input streams. Default is 2.
20869 If set to 1, force the output to terminate when the shortest input
20870 terminates. Default value is 0.
20875 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20876 Deinterlacing Filter").
20878 Based on the process described by Martin Weston for BBC R&D, and
20879 implemented based on the de-interlace algorithm written by Jim
20880 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20881 uses filter coefficients calculated by BBC R&D.
20883 This filter uses field-dominance information in frame to decide which
20884 of each pair of fields to place first in the output.
20885 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20887 There are two sets of filter coefficients, so called "simple"
20888 and "complex". Which set of filter coefficients is used can
20889 be set by passing an optional parameter:
20893 Set the interlacing filter coefficients. Accepts one of the following values:
20897 Simple filter coefficient set.
20899 More-complex filter coefficient set.
20901 Default value is @samp{complex}.
20904 Specify which frames to deinterlace. Accepts one of the following values:
20908 Deinterlace all frames,
20910 Only deinterlace frames marked as interlaced.
20913 Default value is @samp{all}.
20917 Video waveform monitor.
20919 The waveform monitor plots color component intensity. By default luminance
20920 only. Each column of the waveform corresponds to a column of pixels in the
20923 It accepts the following options:
20927 Can be either @code{row}, or @code{column}. Default is @code{column}.
20928 In row mode, the graph on the left side represents color component value 0 and
20929 the right side represents value = 255. In column mode, the top side represents
20930 color component value = 0 and bottom side represents value = 255.
20933 Set intensity. Smaller values are useful to find out how many values of the same
20934 luminance are distributed across input rows/columns.
20935 Default value is @code{0.04}. Allowed range is [0, 1].
20938 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20939 In mirrored mode, higher values will be represented on the left
20940 side for @code{row} mode and at the top for @code{column} mode. Default is
20941 @code{1} (mirrored).
20945 It accepts the following values:
20948 Presents information identical to that in the @code{parade}, except
20949 that the graphs representing color components are superimposed directly
20952 This display mode makes it easier to spot relative differences or similarities
20953 in overlapping areas of the color components that are supposed to be identical,
20954 such as neutral whites, grays, or blacks.
20957 Display separate graph for the color components side by side in
20958 @code{row} mode or one below the other in @code{column} mode.
20961 Display separate graph for the color components side by side in
20962 @code{column} mode or one below the other in @code{row} mode.
20964 Using this display mode makes it easy to spot color casts in the highlights
20965 and shadows of an image, by comparing the contours of the top and the bottom
20966 graphs of each waveform. Since whites, grays, and blacks are characterized
20967 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20968 should display three waveforms of roughly equal width/height. If not, the
20969 correction is easy to perform by making level adjustments the three waveforms.
20971 Default is @code{stack}.
20973 @item components, c
20974 Set which color components to display. Default is 1, which means only luminance
20975 or red color component if input is in RGB colorspace. If is set for example to
20976 7 it will display all 3 (if) available color components.
20981 No envelope, this is default.
20984 Instant envelope, minimum and maximum values presented in graph will be easily
20985 visible even with small @code{step} value.
20988 Hold minimum and maximum values presented in graph across time. This way you
20989 can still spot out of range values without constantly looking at waveforms.
20992 Peak and instant envelope combined together.
20998 No filtering, this is default.
21001 Luma and chroma combined together.
21004 Similar as above, but shows difference between blue and red chroma.
21007 Similar as above, but use different colors.
21010 Similar as above, but again with different colors.
21013 Displays only chroma.
21016 Displays actual color value on waveform.
21019 Similar as above, but with luma showing frequency of chroma values.
21023 Set which graticule to display.
21027 Do not display graticule.
21030 Display green graticule showing legal broadcast ranges.
21033 Display orange graticule showing legal broadcast ranges.
21036 Display invert graticule showing legal broadcast ranges.
21040 Set graticule opacity.
21043 Set graticule flags.
21047 Draw numbers above lines. By default enabled.
21050 Draw dots instead of lines.
21054 Set scale used for displaying graticule.
21061 Default is digital.
21064 Set background opacity.
21068 Set tint for output.
21069 Only used with lowpass filter and when display is not overlay and input
21070 pixel formats are not RGB.
21073 @section weave, doubleweave
21075 The @code{weave} takes a field-based video input and join
21076 each two sequential fields into single frame, producing a new double
21077 height clip with half the frame rate and half the frame count.
21079 The @code{doubleweave} works same as @code{weave} but without
21080 halving frame rate and frame count.
21082 It accepts the following option:
21086 Set first field. Available values are:
21090 Set the frame as top-field-first.
21093 Set the frame as bottom-field-first.
21097 @subsection Examples
21101 Interlace video using @ref{select} and @ref{separatefields} filter:
21103 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21108 Apply the xBR high-quality magnification filter which is designed for pixel
21109 art. It follows a set of edge-detection rules, see
21110 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21112 It accepts the following option:
21116 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21117 @code{3xBR} and @code{4} for @code{4xBR}.
21118 Default is @code{3}.
21123 Apply cross fade from one input video stream to another input video stream.
21124 The cross fade is applied for specified duration.
21126 The filter accepts the following options:
21130 Set one of available transition effects:
21178 Default transition effect is fade.
21181 Set cross fade duration in seconds.
21182 Default duration is 1 second.
21185 Set cross fade start relative to first input stream in seconds.
21186 Default offset is 0.
21189 Set expression for custom transition effect.
21191 The expressions can use the following variables and functions:
21196 The coordinates of the current sample.
21200 The width and height of the image.
21203 Progress of transition effect.
21206 Currently processed plane.
21209 Return value of first input at current location and plane.
21212 Return value of second input at current location and plane.
21218 Return the value of the pixel at location (@var{x},@var{y}) of the
21219 first/second/third/fourth component of first input.
21225 Return the value of the pixel at location (@var{x},@var{y}) of the
21226 first/second/third/fourth component of second input.
21230 @subsection Examples
21234 Cross fade from one input video to another input video, with fade transition and duration of transition
21235 of 2 seconds starting at offset of 5 seconds:
21237 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21242 Pick median pixels from several input videos.
21244 The filter accepts the following options:
21248 Set number of inputs.
21249 Default is 3. Allowed range is from 3 to 255.
21250 If number of inputs is even number, than result will be mean value between two median values.
21253 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21256 Set median percentile. Default value is @code{0.5}.
21257 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21258 minimum values, and @code{1} maximum values.
21262 Stack video inputs into custom layout.
21264 All streams must be of same pixel format.
21266 The filter accepts the following options:
21270 Set number of input streams. Default is 2.
21273 Specify layout of inputs.
21274 This option requires the desired layout configuration to be explicitly set by the user.
21275 This sets position of each video input in output. Each input
21276 is separated by '|'.
21277 The first number represents the column, and the second number represents the row.
21278 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21279 where X is video input from which to take width or height.
21280 Multiple values can be used when separated by '+'. In such
21281 case values are summed together.
21283 Note that if inputs are of different sizes gaps may appear, as not all of
21284 the output video frame will be filled. Similarly, videos can overlap each
21285 other if their position doesn't leave enough space for the full frame of
21288 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21289 a layout must be set by the user.
21292 If set to 1, force the output to terminate when the shortest input
21293 terminates. Default value is 0.
21296 If set to valid color, all unused pixels will be filled with that color.
21297 By default fill is set to none, so it is disabled.
21300 @subsection Examples
21304 Display 4 inputs into 2x2 grid.
21308 input1(0, 0) | input3(w0, 0)
21309 input2(0, h0) | input4(w0, h0)
21313 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21316 Note that if inputs are of different sizes, gaps or overlaps may occur.
21319 Display 4 inputs into 1x4 grid.
21326 input4(0, h0+h1+h2)
21330 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21333 Note that if inputs are of different widths, unused space will appear.
21336 Display 9 inputs into 3x3 grid.
21340 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21341 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21342 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21346 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
21349 Note that if inputs are of different sizes, gaps or overlaps may occur.
21352 Display 16 inputs into 4x4 grid.
21356 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21357 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21358 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21359 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21363 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|
21364 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
21367 Note that if inputs are of different sizes, gaps or overlaps may occur.
21374 Deinterlace the input video ("yadif" means "yet another deinterlacing
21377 It accepts the following parameters:
21383 The interlacing mode to adopt. It accepts one of the following values:
21386 @item 0, send_frame
21387 Output one frame for each frame.
21388 @item 1, send_field
21389 Output one frame for each field.
21390 @item 2, send_frame_nospatial
21391 Like @code{send_frame}, but it skips the spatial interlacing check.
21392 @item 3, send_field_nospatial
21393 Like @code{send_field}, but it skips the spatial interlacing check.
21396 The default value is @code{send_frame}.
21399 The picture field parity assumed for the input interlaced video. It accepts one
21400 of the following values:
21404 Assume the top field is first.
21406 Assume the bottom field is first.
21408 Enable automatic detection of field parity.
21411 The default value is @code{auto}.
21412 If the interlacing is unknown or the decoder does not export this information,
21413 top field first will be assumed.
21416 Specify which frames to deinterlace. Accepts one of the following
21421 Deinterlace all frames.
21422 @item 1, interlaced
21423 Only deinterlace frames marked as interlaced.
21426 The default value is @code{all}.
21429 @section yadif_cuda
21431 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21432 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21435 It accepts the following parameters:
21441 The interlacing mode to adopt. It accepts one of the following values:
21444 @item 0, send_frame
21445 Output one frame for each frame.
21446 @item 1, send_field
21447 Output one frame for each field.
21448 @item 2, send_frame_nospatial
21449 Like @code{send_frame}, but it skips the spatial interlacing check.
21450 @item 3, send_field_nospatial
21451 Like @code{send_field}, but it skips the spatial interlacing check.
21454 The default value is @code{send_frame}.
21457 The picture field parity assumed for the input interlaced video. It accepts one
21458 of the following values:
21462 Assume the top field is first.
21464 Assume the bottom field is first.
21466 Enable automatic detection of field parity.
21469 The default value is @code{auto}.
21470 If the interlacing is unknown or the decoder does not export this information,
21471 top field first will be assumed.
21474 Specify which frames to deinterlace. Accepts one of the following
21479 Deinterlace all frames.
21480 @item 1, interlaced
21481 Only deinterlace frames marked as interlaced.
21484 The default value is @code{all}.
21489 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21490 The algorithm is described in
21491 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21493 It accepts the following parameters:
21497 Set the window radius. Default value is 3.
21500 Set which planes to filter. Default is only the first plane.
21503 Set blur strength. Default value is 128.
21506 @subsection Commands
21507 This filter supports same @ref{commands} as options.
21511 Apply Zoom & Pan effect.
21513 This filter accepts the following options:
21517 Set the zoom expression. Range is 1-10. Default is 1.
21521 Set the x and y expression. Default is 0.
21524 Set the duration expression in number of frames.
21525 This sets for how many number of frames effect will last for
21526 single input image.
21529 Set the output image size, default is 'hd720'.
21532 Set the output frame rate, default is '25'.
21535 Each expression can contain the following constants:
21554 Output frame count.
21557 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21559 @item out_time, time, ot
21560 The output timestamp expressed in seconds.
21564 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21565 for current input frame.
21569 'x' and 'y' of last output frame of previous input frame or 0 when there was
21570 not yet such frame (first input frame).
21573 Last calculated zoom from 'z' expression for current input frame.
21576 Last calculated zoom of last output frame of previous input frame.
21579 Number of output frames for current input frame. Calculated from 'd' expression
21580 for each input frame.
21583 number of output frames created for previous input frame
21586 Rational number: input width / input height
21589 sample aspect ratio
21592 display aspect ratio
21596 @subsection Examples
21600 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21602 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
21606 Zoom in up to 1.5x and pan always at center of picture:
21608 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21612 Same as above but without pausing:
21614 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21618 Zoom in 2x into center of picture only for the first second of the input video:
21620 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21627 Scale (resize) the input video, using the z.lib library:
21628 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21629 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21631 The zscale filter forces the output display aspect ratio to be the same
21632 as the input, by changing the output sample aspect ratio.
21634 If the input image format is different from the format requested by
21635 the next filter, the zscale filter will convert the input to the
21638 @subsection Options
21639 The filter accepts the following options.
21644 Set the output video dimension expression. Default value is the input
21647 If the @var{width} or @var{w} value is 0, the input width is used for
21648 the output. If the @var{height} or @var{h} value is 0, the input height
21649 is used for the output.
21651 If one and only one of the values is -n with n >= 1, the zscale filter
21652 will use a value that maintains the aspect ratio of the input image,
21653 calculated from the other specified dimension. After that it will,
21654 however, make sure that the calculated dimension is divisible by n and
21655 adjust the value if necessary.
21657 If both values are -n with n >= 1, the behavior will be identical to
21658 both values being set to 0 as previously detailed.
21660 See below for the list of accepted constants for use in the dimension
21664 Set the video size. For the syntax of this option, check the
21665 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21668 Set the dither type.
21670 Possible values are:
21675 @item error_diffusion
21681 Set the resize filter type.
21683 Possible values are:
21693 Default is bilinear.
21696 Set the color range.
21698 Possible values are:
21705 Default is same as input.
21708 Set the color primaries.
21710 Possible values are:
21720 Default is same as input.
21723 Set the transfer characteristics.
21725 Possible values are:
21739 Default is same as input.
21742 Set the colorspace matrix.
21744 Possible value are:
21755 Default is same as input.
21758 Set the input color range.
21760 Possible values are:
21767 Default is same as input.
21769 @item primariesin, pin
21770 Set the input color primaries.
21772 Possible values are:
21782 Default is same as input.
21784 @item transferin, tin
21785 Set the input transfer characteristics.
21787 Possible values are:
21798 Default is same as input.
21800 @item matrixin, min
21801 Set the input colorspace matrix.
21803 Possible value are:
21815 Set the output chroma location.
21817 Possible values are:
21828 @item chromalin, cin
21829 Set the input chroma location.
21831 Possible values are:
21843 Set the nominal peak luminance.
21846 The values of the @option{w} and @option{h} options are expressions
21847 containing the following constants:
21852 The input width and height
21856 These are the same as @var{in_w} and @var{in_h}.
21860 The output (scaled) width and height
21864 These are the same as @var{out_w} and @var{out_h}
21867 The same as @var{iw} / @var{ih}
21870 input sample aspect ratio
21873 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21877 horizontal and vertical input chroma subsample values. For example for the
21878 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21882 horizontal and vertical output chroma subsample values. For example for the
21883 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21886 @subsection Commands
21888 This filter supports the following commands:
21892 Set the output video dimension expression.
21893 The command accepts the same syntax of the corresponding option.
21895 If the specified expression is not valid, it is kept at its current
21899 @c man end VIDEO FILTERS
21901 @chapter OpenCL Video Filters
21902 @c man begin OPENCL VIDEO FILTERS
21904 Below is a description of the currently available OpenCL video filters.
21906 To enable compilation of these filters you need to configure FFmpeg with
21907 @code{--enable-opencl}.
21909 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21912 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21913 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21914 given device parameters.
21916 @item -filter_hw_device @var{name}
21917 Pass the hardware device called @var{name} to all filters in any filter graph.
21921 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21925 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21927 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21931 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.
21933 @section avgblur_opencl
21935 Apply average blur filter.
21937 The filter accepts the following options:
21941 Set horizontal radius size.
21942 Range is @code{[1, 1024]} and default value is @code{1}.
21945 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21948 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21951 @subsection Example
21955 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.
21957 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21961 @section boxblur_opencl
21963 Apply a boxblur algorithm to the input video.
21965 It accepts the following parameters:
21969 @item luma_radius, lr
21970 @item luma_power, lp
21971 @item chroma_radius, cr
21972 @item chroma_power, cp
21973 @item alpha_radius, ar
21974 @item alpha_power, ap
21978 A description of the accepted options follows.
21981 @item luma_radius, lr
21982 @item chroma_radius, cr
21983 @item alpha_radius, ar
21984 Set an expression for the box radius in pixels used for blurring the
21985 corresponding input plane.
21987 The radius value must be a non-negative number, and must not be
21988 greater than the value of the expression @code{min(w,h)/2} for the
21989 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21992 Default value for @option{luma_radius} is "2". If not specified,
21993 @option{chroma_radius} and @option{alpha_radius} default to the
21994 corresponding value set for @option{luma_radius}.
21996 The expressions can contain the following constants:
22000 The input width and height in pixels.
22004 The input chroma image width and height in pixels.
22008 The horizontal and vertical chroma subsample values. For example, for the
22009 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22012 @item luma_power, lp
22013 @item chroma_power, cp
22014 @item alpha_power, ap
22015 Specify how many times the boxblur filter is applied to the
22016 corresponding plane.
22018 Default value for @option{luma_power} is 2. If not specified,
22019 @option{chroma_power} and @option{alpha_power} default to the
22020 corresponding value set for @option{luma_power}.
22022 A value of 0 will disable the effect.
22025 @subsection Examples
22027 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.
22031 Apply a boxblur filter with the luma, chroma, and alpha radius
22032 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.
22034 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22035 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22039 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.
22041 For the luma plane, a 2x2 box radius will be run once.
22043 For the chroma plane, a 4x4 box radius will be run 5 times.
22045 For the alpha plane, a 3x3 box radius will be run 7 times.
22047 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22051 @section colorkey_opencl
22052 RGB colorspace color keying.
22054 The filter accepts the following options:
22058 The color which will be replaced with transparency.
22061 Similarity percentage with the key color.
22063 0.01 matches only the exact key color, while 1.0 matches everything.
22068 0.0 makes pixels either fully transparent, or not transparent at all.
22070 Higher values result in semi-transparent pixels, with a higher transparency
22071 the more similar the pixels color is to the key color.
22074 @subsection Examples
22078 Make every semi-green pixel in the input transparent with some slight blending:
22080 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22084 @section convolution_opencl
22086 Apply convolution of 3x3, 5x5, 7x7 matrix.
22088 The filter accepts the following options:
22095 Set matrix for each plane.
22096 Matrix is sequence of 9, 25 or 49 signed numbers.
22097 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22103 Set multiplier for calculated value for each plane.
22104 If unset or 0, it will be sum of all matrix elements.
22105 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22111 Set bias for each plane. This value is added to the result of the multiplication.
22112 Useful for making the overall image brighter or darker.
22113 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22117 @subsection Examples
22123 -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
22129 -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
22133 Apply edge enhance:
22135 -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
22141 -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
22145 Apply laplacian edge detector which includes diagonals:
22147 -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
22153 -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
22157 @section erosion_opencl
22159 Apply erosion effect to the video.
22161 This filter replaces the pixel by the local(3x3) minimum.
22163 It accepts the following options:
22170 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22171 If @code{0}, plane will remain unchanged.
22174 Flag which specifies the pixel to refer to.
22175 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22177 Flags to local 3x3 coordinates region centered on @code{x}:
22186 @subsection Example
22190 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.
22192 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22196 @section deshake_opencl
22197 Feature-point based video stabilization filter.
22199 The filter accepts the following options:
22203 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22206 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22208 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22210 Viewing point matches in the output video is only supported for RGB input.
22212 Defaults to @code{0}.
22214 @item adaptive_crop
22215 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22217 Defaults to @code{1}.
22219 @item refine_features
22220 Whether or not feature points should be refined at a sub-pixel level.
22222 This can be turned off for a slight performance gain at the cost of precision.
22224 Defaults to @code{1}.
22226 @item smooth_strength
22227 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22229 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22231 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22233 Defaults to @code{0.0}.
22235 @item smooth_window_multiplier
22236 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22238 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22240 Acceptable values range from @code{0.1} to @code{10.0}.
22242 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22243 potentially improving smoothness, but also increase latency and memory usage.
22245 Defaults to @code{2.0}.
22249 @subsection Examples
22253 Stabilize a video with a fixed, medium smoothing strength:
22255 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22259 Stabilize a video with debugging (both in console and in rendered video):
22261 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22265 @section dilation_opencl
22267 Apply dilation effect to the video.
22269 This filter replaces the pixel by the local(3x3) maximum.
22271 It accepts the following options:
22278 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22279 If @code{0}, plane will remain unchanged.
22282 Flag which specifies the pixel to refer to.
22283 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22285 Flags to local 3x3 coordinates region centered on @code{x}:
22294 @subsection Example
22298 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.
22300 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22304 @section nlmeans_opencl
22306 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22308 @section overlay_opencl
22310 Overlay one video on top of another.
22312 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22313 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22315 The filter accepts the following options:
22320 Set the x coordinate of the overlaid video on the main video.
22321 Default value is @code{0}.
22324 Set the y coordinate of the overlaid video on the main video.
22325 Default value is @code{0}.
22329 @subsection Examples
22333 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22335 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22338 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22340 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22345 @section pad_opencl
22347 Add paddings to the input image, and place the original input at the
22348 provided @var{x}, @var{y} coordinates.
22350 It accepts the following options:
22355 Specify an expression for the size of the output image with the
22356 paddings added. If the value for @var{width} or @var{height} is 0, the
22357 corresponding input size is used for the output.
22359 The @var{width} expression can reference the value set by the
22360 @var{height} expression, and vice versa.
22362 The default value of @var{width} and @var{height} is 0.
22366 Specify the offsets to place the input image at within the padded area,
22367 with respect to the top/left border of the output image.
22369 The @var{x} expression can reference the value set by the @var{y}
22370 expression, and vice versa.
22372 The default value of @var{x} and @var{y} is 0.
22374 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22375 so the input image is centered on the padded area.
22378 Specify the color of the padded area. For the syntax of this option,
22379 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22380 manual,ffmpeg-utils}.
22383 Pad to an aspect instead to a resolution.
22386 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22387 options are expressions containing the following constants:
22392 The input video width and height.
22396 These are the same as @var{in_w} and @var{in_h}.
22400 The output width and height (the size of the padded area), as
22401 specified by the @var{width} and @var{height} expressions.
22405 These are the same as @var{out_w} and @var{out_h}.
22409 The x and y offsets as specified by the @var{x} and @var{y}
22410 expressions, or NAN if not yet specified.
22413 same as @var{iw} / @var{ih}
22416 input sample aspect ratio
22419 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22422 @section prewitt_opencl
22424 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22426 The filter accepts the following option:
22430 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22433 Set value which will be multiplied with filtered result.
22434 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22437 Set value which will be added to filtered result.
22438 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22441 @subsection Example
22445 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22447 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22451 @anchor{program_opencl}
22452 @section program_opencl
22454 Filter video using an OpenCL program.
22459 OpenCL program source file.
22462 Kernel name in program.
22465 Number of inputs to the filter. Defaults to 1.
22468 Size of output frames. Defaults to the same as the first input.
22472 The @code{program_opencl} filter also supports the @ref{framesync} options.
22474 The program source file must contain a kernel function with the given name,
22475 which will be run once for each plane of the output. Each run on a plane
22476 gets enqueued as a separate 2D global NDRange with one work-item for each
22477 pixel to be generated. The global ID offset for each work-item is therefore
22478 the coordinates of a pixel in the destination image.
22480 The kernel function needs to take the following arguments:
22483 Destination image, @var{__write_only image2d_t}.
22485 This image will become the output; the kernel should write all of it.
22487 Frame index, @var{unsigned int}.
22489 This is a counter starting from zero and increasing by one for each frame.
22491 Source images, @var{__read_only image2d_t}.
22493 These are the most recent images on each input. The kernel may read from
22494 them to generate the output, but they can't be written to.
22501 Copy the input to the output (output must be the same size as the input).
22503 __kernel void copy(__write_only image2d_t destination,
22504 unsigned int index,
22505 __read_only image2d_t source)
22507 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22509 int2 location = (int2)(get_global_id(0), get_global_id(1));
22511 float4 value = read_imagef(source, sampler, location);
22513 write_imagef(destination, location, value);
22518 Apply a simple transformation, rotating the input by an amount increasing
22519 with the index counter. Pixel values are linearly interpolated by the
22520 sampler, and the output need not have the same dimensions as the input.
22522 __kernel void rotate_image(__write_only image2d_t dst,
22523 unsigned int index,
22524 __read_only image2d_t src)
22526 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22527 CLK_FILTER_LINEAR);
22529 float angle = (float)index / 100.0f;
22531 float2 dst_dim = convert_float2(get_image_dim(dst));
22532 float2 src_dim = convert_float2(get_image_dim(src));
22534 float2 dst_cen = dst_dim / 2.0f;
22535 float2 src_cen = src_dim / 2.0f;
22537 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22539 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22541 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22542 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22544 src_pos = src_pos * src_dim / dst_dim;
22546 float2 src_loc = src_pos + src_cen;
22548 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22549 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22550 write_imagef(dst, dst_loc, 0.5f);
22552 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22557 Blend two inputs together, with the amount of each input used varying
22558 with the index counter.
22560 __kernel void blend_images(__write_only image2d_t dst,
22561 unsigned int index,
22562 __read_only image2d_t src1,
22563 __read_only image2d_t src2)
22565 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22566 CLK_FILTER_LINEAR);
22568 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22570 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22571 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22572 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22574 float4 val1 = read_imagef(src1, sampler, src1_loc);
22575 float4 val2 = read_imagef(src2, sampler, src2_loc);
22577 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22583 @section roberts_opencl
22584 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22586 The filter accepts the following option:
22590 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22593 Set value which will be multiplied with filtered result.
22594 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22597 Set value which will be added to filtered result.
22598 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22601 @subsection Example
22605 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22607 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22611 @section sobel_opencl
22613 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22615 The filter accepts the following option:
22619 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22622 Set value which will be multiplied with filtered result.
22623 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22626 Set value which will be added to filtered result.
22627 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22630 @subsection Example
22634 Apply sobel operator with scale set to 2 and delta set to 10
22636 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22640 @section tonemap_opencl
22642 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22644 It accepts the following parameters:
22648 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22651 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22654 Apply desaturation for highlights that exceed this level of brightness. The
22655 higher the parameter, the more color information will be preserved. This
22656 setting helps prevent unnaturally blown-out colors for super-highlights, by
22657 (smoothly) turning into white instead. This makes images feel more natural,
22658 at the cost of reducing information about out-of-range colors.
22660 The default value is 0.5, and the algorithm here is a little different from
22661 the cpu version tonemap currently. A setting of 0.0 disables this option.
22664 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22665 is used to detect whether the scene has changed or not. If the distance between
22666 the current frame average brightness and the current running average exceeds
22667 a threshold value, we would re-calculate scene average and peak brightness.
22668 The default value is 0.2.
22671 Specify the output pixel format.
22673 Currently supported formats are:
22680 Set the output color range.
22682 Possible values are:
22688 Default is same as input.
22691 Set the output color primaries.
22693 Possible values are:
22699 Default is same as input.
22702 Set the output transfer characteristics.
22704 Possible values are:
22713 Set the output colorspace matrix.
22715 Possible value are:
22721 Default is same as input.
22725 @subsection Example
22729 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22731 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22735 @section unsharp_opencl
22737 Sharpen or blur the input video.
22739 It accepts the following parameters:
22742 @item luma_msize_x, lx
22743 Set the luma matrix horizontal size.
22744 Range is @code{[1, 23]} and default value is @code{5}.
22746 @item luma_msize_y, ly
22747 Set the luma matrix vertical size.
22748 Range is @code{[1, 23]} and default value is @code{5}.
22750 @item luma_amount, la
22751 Set the luma effect strength.
22752 Range is @code{[-10, 10]} and default value is @code{1.0}.
22754 Negative values will blur the input video, while positive values will
22755 sharpen it, a value of zero will disable the effect.
22757 @item chroma_msize_x, cx
22758 Set the chroma matrix horizontal size.
22759 Range is @code{[1, 23]} and default value is @code{5}.
22761 @item chroma_msize_y, cy
22762 Set the chroma matrix vertical size.
22763 Range is @code{[1, 23]} and default value is @code{5}.
22765 @item chroma_amount, ca
22766 Set the chroma effect strength.
22767 Range is @code{[-10, 10]} and default value is @code{0.0}.
22769 Negative values will blur the input video, while positive values will
22770 sharpen it, a value of zero will disable the effect.
22774 All parameters are optional and default to the equivalent of the
22775 string '5:5:1.0:5:5:0.0'.
22777 @subsection Examples
22781 Apply strong luma sharpen effect:
22783 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22787 Apply a strong blur of both luma and chroma parameters:
22789 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22793 @section xfade_opencl
22795 Cross fade two videos with custom transition effect by using OpenCL.
22797 It accepts the following options:
22801 Set one of possible transition effects.
22805 Select custom transition effect, the actual transition description
22806 will be picked from source and kernel options.
22818 Default transition is fade.
22822 OpenCL program source file for custom transition.
22825 Set name of kernel to use for custom transition from program source file.
22828 Set duration of video transition.
22831 Set time of start of transition relative to first video.
22834 The program source file must contain a kernel function with the given name,
22835 which will be run once for each plane of the output. Each run on a plane
22836 gets enqueued as a separate 2D global NDRange with one work-item for each
22837 pixel to be generated. The global ID offset for each work-item is therefore
22838 the coordinates of a pixel in the destination image.
22840 The kernel function needs to take the following arguments:
22843 Destination image, @var{__write_only image2d_t}.
22845 This image will become the output; the kernel should write all of it.
22848 First Source image, @var{__read_only image2d_t}.
22849 Second Source image, @var{__read_only image2d_t}.
22851 These are the most recent images on each input. The kernel may read from
22852 them to generate the output, but they can't be written to.
22855 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22862 Apply dots curtain transition effect:
22864 __kernel void blend_images(__write_only image2d_t dst,
22865 __read_only image2d_t src1,
22866 __read_only image2d_t src2,
22869 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22870 CLK_FILTER_LINEAR);
22871 int2 p = (int2)(get_global_id(0), get_global_id(1));
22872 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22873 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22876 float2 dots = (float2)(20.0, 20.0);
22877 float2 center = (float2)(0,0);
22880 float4 val1 = read_imagef(src1, sampler, p);
22881 float4 val2 = read_imagef(src2, sampler, p);
22882 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22884 write_imagef(dst, p, next ? val1 : val2);
22890 @c man end OPENCL VIDEO FILTERS
22892 @chapter VAAPI Video Filters
22893 @c man begin VAAPI VIDEO FILTERS
22895 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22897 To enable compilation of these filters you need to configure FFmpeg with
22898 @code{--enable-vaapi}.
22900 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}
22902 @section tonemap_vaapi
22904 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22905 It maps the dynamic range of HDR10 content to the SDR content.
22906 It currently only accepts HDR10 as input.
22908 It accepts the following parameters:
22912 Specify the output pixel format.
22914 Currently supported formats are:
22923 Set the output color primaries.
22925 Default is same as input.
22928 Set the output transfer characteristics.
22933 Set the output colorspace matrix.
22935 Default is same as input.
22939 @subsection Example
22943 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22945 tonemap_vaapi=format=p010:t=bt2020-10
22949 @c man end VAAPI VIDEO FILTERS
22951 @chapter Video Sources
22952 @c man begin VIDEO SOURCES
22954 Below is a description of the currently available video sources.
22958 Buffer video frames, and make them available to the filter chain.
22960 This source is mainly intended for a programmatic use, in particular
22961 through the interface defined in @file{libavfilter/buffersrc.h}.
22963 It accepts the following parameters:
22968 Specify the size (width and height) of the buffered video frames. For the
22969 syntax of this option, check the
22970 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22973 The input video width.
22976 The input video height.
22979 A string representing the pixel format of the buffered video frames.
22980 It may be a number corresponding to a pixel format, or a pixel format
22984 Specify the timebase assumed by the timestamps of the buffered frames.
22987 Specify the frame rate expected for the video stream.
22989 @item pixel_aspect, sar
22990 The sample (pixel) aspect ratio of the input video.
22993 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22994 to the filtergraph description to specify swscale flags for automatically
22995 inserted scalers. See @ref{Filtergraph syntax}.
22997 @item hw_frames_ctx
22998 When using a hardware pixel format, this should be a reference to an
22999 AVHWFramesContext describing input frames.
23004 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23007 will instruct the source to accept video frames with size 320x240 and
23008 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23009 square pixels (1:1 sample aspect ratio).
23010 Since the pixel format with name "yuv410p" corresponds to the number 6
23011 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23012 this example corresponds to:
23014 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23017 Alternatively, the options can be specified as a flat string, but this
23018 syntax is deprecated:
23020 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23024 Create a pattern generated by an elementary cellular automaton.
23026 The initial state of the cellular automaton can be defined through the
23027 @option{filename} and @option{pattern} options. If such options are
23028 not specified an initial state is created randomly.
23030 At each new frame a new row in the video is filled with the result of
23031 the cellular automaton next generation. The behavior when the whole
23032 frame is filled is defined by the @option{scroll} option.
23034 This source accepts the following options:
23038 Read the initial cellular automaton state, i.e. the starting row, from
23039 the specified file.
23040 In the file, each non-whitespace character is considered an alive
23041 cell, a newline will terminate the row, and further characters in the
23042 file will be ignored.
23045 Read the initial cellular automaton state, i.e. the starting row, from
23046 the specified string.
23048 Each non-whitespace character in the string is considered an alive
23049 cell, a newline will terminate the row, and further characters in the
23050 string will be ignored.
23053 Set the video rate, that is the number of frames generated per second.
23056 @item random_fill_ratio, ratio
23057 Set the random fill ratio for the initial cellular automaton row. It
23058 is a floating point number value ranging from 0 to 1, defaults to
23061 This option is ignored when a file or a pattern is specified.
23063 @item random_seed, seed
23064 Set the seed for filling randomly the initial row, must be an integer
23065 included between 0 and UINT32_MAX. If not specified, or if explicitly
23066 set to -1, the filter will try to use a good random seed on a best
23070 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23071 Default value is 110.
23074 Set the size of the output video. For the syntax of this option, check the
23075 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23077 If @option{filename} or @option{pattern} is specified, the size is set
23078 by default to the width of the specified initial state row, and the
23079 height is set to @var{width} * PHI.
23081 If @option{size} is set, it must contain the width of the specified
23082 pattern string, and the specified pattern will be centered in the
23085 If a filename or a pattern string is not specified, the size value
23086 defaults to "320x518" (used for a randomly generated initial state).
23089 If set to 1, scroll the output upward when all the rows in the output
23090 have been already filled. If set to 0, the new generated row will be
23091 written over the top row just after the bottom row is filled.
23094 @item start_full, full
23095 If set to 1, completely fill the output with generated rows before
23096 outputting the first frame.
23097 This is the default behavior, for disabling set the value to 0.
23100 If set to 1, stitch the left and right row edges together.
23101 This is the default behavior, for disabling set the value to 0.
23104 @subsection Examples
23108 Read the initial state from @file{pattern}, and specify an output of
23111 cellauto=f=pattern:s=200x400
23115 Generate a random initial row with a width of 200 cells, with a fill
23118 cellauto=ratio=2/3:s=200x200
23122 Create a pattern generated by rule 18 starting by a single alive cell
23123 centered on an initial row with width 100:
23125 cellauto=p=@@:s=100x400:full=0:rule=18
23129 Specify a more elaborated initial pattern:
23131 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23136 @anchor{coreimagesrc}
23137 @section coreimagesrc
23138 Video source generated on GPU using Apple's CoreImage API on OSX.
23140 This video source is a specialized version of the @ref{coreimage} video filter.
23141 Use a core image generator at the beginning of the applied filterchain to
23142 generate the content.
23144 The coreimagesrc video source accepts the following options:
23146 @item list_generators
23147 List all available generators along with all their respective options as well as
23148 possible minimum and maximum values along with the default values.
23150 list_generators=true
23154 Specify the size of the sourced video. For the syntax of this option, check the
23155 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23156 The default value is @code{320x240}.
23159 Specify the frame rate of the sourced video, as the number of frames
23160 generated per second. It has to be a string in the format
23161 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23162 number or a valid video frame rate abbreviation. The default value is
23166 Set the sample aspect ratio of the sourced video.
23169 Set the duration of the sourced video. See
23170 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23171 for the accepted syntax.
23173 If not specified, or the expressed duration is negative, the video is
23174 supposed to be generated forever.
23177 Additionally, all options of the @ref{coreimage} video filter are accepted.
23178 A complete filterchain can be used for further processing of the
23179 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23180 and examples for details.
23182 @subsection Examples
23187 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23188 given as complete and escaped command-line for Apple's standard bash shell:
23190 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23192 This example is equivalent to the QRCode example of @ref{coreimage} without the
23193 need for a nullsrc video source.
23198 Generate several gradients.
23202 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23203 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23206 Set frame rate, expressed as number of frames per second. Default
23209 @item c0, c1, c2, c3, c4, c5, c6, c7
23210 Set 8 colors. Default values for colors is to pick random one.
23212 @item x0, y0, y0, y1
23213 Set gradient line source and destination points. If negative or out of range, random ones
23217 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23220 Set seed for picking gradient line points.
23223 Set the duration of the sourced video. See
23224 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23225 for the accepted syntax.
23227 If not specified, or the expressed duration is negative, the video is
23228 supposed to be generated forever.
23231 Set speed of gradients rotation.
23235 @section mandelbrot
23237 Generate a Mandelbrot set fractal, and progressively zoom towards the
23238 point specified with @var{start_x} and @var{start_y}.
23240 This source accepts the following options:
23245 Set the terminal pts value. Default value is 400.
23248 Set the terminal scale value.
23249 Must be a floating point value. Default value is 0.3.
23252 Set the inner coloring mode, that is the algorithm used to draw the
23253 Mandelbrot fractal internal region.
23255 It shall assume one of the following values:
23260 Show time until convergence.
23262 Set color based on point closest to the origin of the iterations.
23267 Default value is @var{mincol}.
23270 Set the bailout value. Default value is 10.0.
23273 Set the maximum of iterations performed by the rendering
23274 algorithm. Default value is 7189.
23277 Set outer coloring mode.
23278 It shall assume one of following values:
23280 @item iteration_count
23281 Set iteration count mode.
23282 @item normalized_iteration_count
23283 set normalized iteration count mode.
23285 Default value is @var{normalized_iteration_count}.
23288 Set frame rate, expressed as number of frames per second. Default
23292 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23293 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23296 Set the initial scale value. Default value is 3.0.
23299 Set the initial x position. Must be a floating point value between
23300 -100 and 100. Default value is -0.743643887037158704752191506114774.
23303 Set the initial y position. Must be a floating point value between
23304 -100 and 100. Default value is -0.131825904205311970493132056385139.
23309 Generate various test patterns, as generated by the MPlayer test filter.
23311 The size of the generated video is fixed, and is 256x256.
23312 This source is useful in particular for testing encoding features.
23314 This source accepts the following options:
23319 Specify the frame rate of the sourced video, as the number of frames
23320 generated per second. It has to be a string in the format
23321 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23322 number or a valid video frame rate abbreviation. The default value is
23326 Set the duration of the sourced video. See
23327 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23328 for the accepted syntax.
23330 If not specified, or the expressed duration is negative, the video is
23331 supposed to be generated forever.
23335 Set the number or the name of the test to perform. Supported tests are:
23349 @item max_frames, m
23350 Set the maximum number of frames generated for each test, default value is 30.
23354 Default value is "all", which will cycle through the list of all tests.
23359 mptestsrc=t=dc_luma
23362 will generate a "dc_luma" test pattern.
23364 @section frei0r_src
23366 Provide a frei0r source.
23368 To enable compilation of this filter you need to install the frei0r
23369 header and configure FFmpeg with @code{--enable-frei0r}.
23371 This source accepts the following parameters:
23376 The size of the video to generate. For the syntax of this option, check the
23377 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23380 The framerate of the generated video. It may be a string of the form
23381 @var{num}/@var{den} or a frame rate abbreviation.
23384 The name to the frei0r source to load. For more information regarding frei0r and
23385 how to set the parameters, read the @ref{frei0r} section in the video filters
23388 @item filter_params
23389 A '|'-separated list of parameters to pass to the frei0r source.
23393 For example, to generate a frei0r partik0l source with size 200x200
23394 and frame rate 10 which is overlaid on the overlay filter main input:
23396 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23401 Generate a life pattern.
23403 This source is based on a generalization of John Conway's life game.
23405 The sourced input represents a life grid, each pixel represents a cell
23406 which can be in one of two possible states, alive or dead. Every cell
23407 interacts with its eight neighbours, which are the cells that are
23408 horizontally, vertically, or diagonally adjacent.
23410 At each interaction the grid evolves according to the adopted rule,
23411 which specifies the number of neighbor alive cells which will make a
23412 cell stay alive or born. The @option{rule} option allows one to specify
23415 This source accepts the following options:
23419 Set the file from which to read the initial grid state. In the file,
23420 each non-whitespace character is considered an alive cell, and newline
23421 is used to delimit the end of each row.
23423 If this option is not specified, the initial grid is generated
23427 Set the video rate, that is the number of frames generated per second.
23430 @item random_fill_ratio, ratio
23431 Set the random fill ratio for the initial random grid. It is a
23432 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23433 It is ignored when a file is specified.
23435 @item random_seed, seed
23436 Set the seed for filling the initial random grid, must be an integer
23437 included between 0 and UINT32_MAX. If not specified, or if explicitly
23438 set to -1, the filter will try to use a good random seed on a best
23444 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23445 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23446 @var{NS} specifies the number of alive neighbor cells which make a
23447 live cell stay alive, and @var{NB} the number of alive neighbor cells
23448 which make a dead cell to become alive (i.e. to "born").
23449 "s" and "b" can be used in place of "S" and "B", respectively.
23451 Alternatively a rule can be specified by an 18-bits integer. The 9
23452 high order bits are used to encode the next cell state if it is alive
23453 for each number of neighbor alive cells, the low order bits specify
23454 the rule for "borning" new cells. Higher order bits encode for an
23455 higher number of neighbor cells.
23456 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23457 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23459 Default value is "S23/B3", which is the original Conway's game of life
23460 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23461 cells, and will born a new cell if there are three alive cells around
23465 Set the size of the output video. For the syntax of this option, check the
23466 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23468 If @option{filename} is specified, the size is set by default to the
23469 same size of the input file. If @option{size} is set, it must contain
23470 the size specified in the input file, and the initial grid defined in
23471 that file is centered in the larger resulting area.
23473 If a filename is not specified, the size value defaults to "320x240"
23474 (used for a randomly generated initial grid).
23477 If set to 1, stitch the left and right grid edges together, and the
23478 top and bottom edges also. Defaults to 1.
23481 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23482 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23483 value from 0 to 255.
23486 Set the color of living (or new born) cells.
23489 Set the color of dead cells. If @option{mold} is set, this is the first color
23490 used to represent a dead cell.
23493 Set mold color, for definitely dead and moldy cells.
23495 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23496 ffmpeg-utils manual,ffmpeg-utils}.
23499 @subsection Examples
23503 Read a grid from @file{pattern}, and center it on a grid of size
23506 life=f=pattern:s=300x300
23510 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23512 life=ratio=2/3:s=200x200
23516 Specify a custom rule for evolving a randomly generated grid:
23522 Full example with slow death effect (mold) using @command{ffplay}:
23524 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23531 @anchor{haldclutsrc}
23534 @anchor{pal100bars}
23535 @anchor{rgbtestsrc}
23537 @anchor{smptehdbars}
23540 @anchor{yuvtestsrc}
23541 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23543 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23545 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23547 The @code{color} source provides an uniformly colored input.
23549 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23550 @ref{haldclut} filter.
23552 The @code{nullsrc} source returns unprocessed video frames. It is
23553 mainly useful to be employed in analysis / debugging tools, or as the
23554 source for filters which ignore the input data.
23556 The @code{pal75bars} source generates a color bars pattern, based on
23557 EBU PAL recommendations with 75% color levels.
23559 The @code{pal100bars} source generates a color bars pattern, based on
23560 EBU PAL recommendations with 100% color levels.
23562 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23563 detecting RGB vs BGR issues. You should see a red, green and blue
23564 stripe from top to bottom.
23566 The @code{smptebars} source generates a color bars pattern, based on
23567 the SMPTE Engineering Guideline EG 1-1990.
23569 The @code{smptehdbars} source generates a color bars pattern, based on
23570 the SMPTE RP 219-2002.
23572 The @code{testsrc} source generates a test video pattern, showing a
23573 color pattern, a scrolling gradient and a timestamp. This is mainly
23574 intended for testing purposes.
23576 The @code{testsrc2} source is similar to testsrc, but supports more
23577 pixel formats instead of just @code{rgb24}. This allows using it as an
23578 input for other tests without requiring a format conversion.
23580 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23581 see a y, cb and cr stripe from top to bottom.
23583 The sources accept the following parameters:
23588 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23589 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23590 pixels to be used as identity matrix for 3D lookup tables. Each component is
23591 coded on a @code{1/(N*N)} scale.
23594 Specify the color of the source, only available in the @code{color}
23595 source. For the syntax of this option, check the
23596 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23599 Specify the size of the sourced video. For the syntax of this option, check the
23600 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23601 The default value is @code{320x240}.
23603 This option is not available with the @code{allrgb}, @code{allyuv}, and
23604 @code{haldclutsrc} filters.
23607 Specify the frame rate of the sourced video, as the number of frames
23608 generated per second. It has to be a string in the format
23609 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23610 number or a valid video frame rate abbreviation. The default value is
23614 Set the duration of the sourced video. See
23615 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23616 for the accepted syntax.
23618 If not specified, or the expressed duration is negative, the video is
23619 supposed to be generated forever.
23621 Since the frame rate is used as time base, all frames including the last one
23622 will have their full duration. If the specified duration is not a multiple
23623 of the frame duration, it will be rounded up.
23626 Set the sample aspect ratio of the sourced video.
23629 Specify the alpha (opacity) of the background, only available in the
23630 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23631 255 (fully opaque, the default).
23634 Set the number of decimals to show in the timestamp, only available in the
23635 @code{testsrc} source.
23637 The displayed timestamp value will correspond to the original
23638 timestamp value multiplied by the power of 10 of the specified
23639 value. Default value is 0.
23642 @subsection Examples
23646 Generate a video with a duration of 5.3 seconds, with size
23647 176x144 and a frame rate of 10 frames per second:
23649 testsrc=duration=5.3:size=qcif:rate=10
23653 The following graph description will generate a red source
23654 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23657 color=c=red@@0.2:s=qcif:r=10
23661 If the input content is to be ignored, @code{nullsrc} can be used. The
23662 following command generates noise in the luminance plane by employing
23663 the @code{geq} filter:
23665 nullsrc=s=256x256, geq=random(1)*255:128:128
23669 @subsection Commands
23671 The @code{color} source supports the following commands:
23675 Set the color of the created image. Accepts the same syntax of the
23676 corresponding @option{color} option.
23681 Generate video using an OpenCL program.
23686 OpenCL program source file.
23689 Kernel name in program.
23692 Size of frames to generate. This must be set.
23695 Pixel format to use for the generated frames. This must be set.
23698 Number of frames generated every second. Default value is '25'.
23702 For details of how the program loading works, see the @ref{program_opencl}
23709 Generate a colour ramp by setting pixel values from the position of the pixel
23710 in the output image. (Note that this will work with all pixel formats, but
23711 the generated output will not be the same.)
23713 __kernel void ramp(__write_only image2d_t dst,
23714 unsigned int index)
23716 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23719 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23721 write_imagef(dst, loc, val);
23726 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23728 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23729 unsigned int index)
23731 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23733 float4 value = 0.0f;
23734 int x = loc.x + index;
23735 int y = loc.y + index;
23736 while (x > 0 || y > 0) {
23737 if (x % 3 == 1 && y % 3 == 1) {
23745 write_imagef(dst, loc, value);
23751 @section sierpinski
23753 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23755 This source accepts the following options:
23759 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23760 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23763 Set frame rate, expressed as number of frames per second. Default
23767 Set seed which is used for random panning.
23770 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23773 Set fractal type, can be default @code{carpet} or @code{triangle}.
23776 @c man end VIDEO SOURCES
23778 @chapter Video Sinks
23779 @c man begin VIDEO SINKS
23781 Below is a description of the currently available video sinks.
23783 @section buffersink
23785 Buffer video frames, and make them available to the end of the filter
23788 This sink is mainly intended for programmatic use, in particular
23789 through the interface defined in @file{libavfilter/buffersink.h}
23790 or the options system.
23792 It accepts a pointer to an AVBufferSinkContext structure, which
23793 defines the incoming buffers' formats, to be passed as the opaque
23794 parameter to @code{avfilter_init_filter} for initialization.
23798 Null video sink: do absolutely nothing with the input video. It is
23799 mainly useful as a template and for use in analysis / debugging
23802 @c man end VIDEO SINKS
23804 @chapter Multimedia Filters
23805 @c man begin MULTIMEDIA FILTERS
23807 Below is a description of the currently available multimedia filters.
23811 Convert input audio to a video output, displaying the audio bit scope.
23813 The filter accepts the following options:
23817 Set frame rate, expressed as number of frames per second. Default
23821 Specify the video size for the output. For the syntax of this option, check the
23822 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23823 Default value is @code{1024x256}.
23826 Specify list of colors separated by space or by '|' which will be used to
23827 draw channels. Unrecognized or missing colors will be replaced
23831 @section adrawgraph
23832 Draw a graph using input audio metadata.
23834 See @ref{drawgraph}
23836 @section agraphmonitor
23838 See @ref{graphmonitor}.
23840 @section ahistogram
23842 Convert input audio to a video output, displaying the volume histogram.
23844 The filter accepts the following options:
23848 Specify how histogram is calculated.
23850 It accepts the following values:
23853 Use single histogram for all channels.
23855 Use separate histogram for each channel.
23857 Default is @code{single}.
23860 Set frame rate, expressed as number of frames per second. Default
23864 Specify the video size for the output. For the syntax of this option, check the
23865 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23866 Default value is @code{hd720}.
23871 It accepts the following values:
23882 reverse logarithmic
23884 Default is @code{log}.
23887 Set amplitude scale.
23889 It accepts the following values:
23896 Default is @code{log}.
23899 Set how much frames to accumulate in histogram.
23900 Default is 1. Setting this to -1 accumulates all frames.
23903 Set histogram ratio of window height.
23906 Set sonogram sliding.
23908 It accepts the following values:
23911 replace old rows with new ones.
23913 scroll from top to bottom.
23915 Default is @code{replace}.
23918 @section aphasemeter
23920 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23921 representing mean phase of current audio frame. A video output can also be produced and is
23922 enabled by default. The audio is passed through as first output.
23924 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23925 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23926 and @code{1} means channels are in phase.
23928 The filter accepts the following options, all related to its video output:
23932 Set the output frame rate. Default value is @code{25}.
23935 Set the video size for the output. For the syntax of this option, check the
23936 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23937 Default value is @code{800x400}.
23942 Specify the red, green, blue contrast. Default values are @code{2},
23943 @code{7} and @code{1}.
23944 Allowed range is @code{[0, 255]}.
23947 Set color which will be used for drawing median phase. If color is
23948 @code{none} which is default, no median phase value will be drawn.
23951 Enable video output. Default is enabled.
23954 @subsection phasing detection
23956 The filter also detects out of phase and mono sequences in stereo streams.
23957 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23959 The filter accepts the following options for this detection:
23963 Enable mono and out of phase detection. Default is disabled.
23966 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23967 Allowed range is @code{[0, 1]}.
23970 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23971 Allowed range is @code{[90, 180]}.
23974 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23977 @subsection Examples
23981 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23983 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23987 @section avectorscope
23989 Convert input audio to a video output, representing the audio vector
23992 The filter is used to measure the difference between channels of stereo
23993 audio stream. A monaural signal, consisting of identical left and right
23994 signal, results in straight vertical line. Any stereo separation is visible
23995 as a deviation from this line, creating a Lissajous figure.
23996 If the straight (or deviation from it) but horizontal line appears this
23997 indicates that the left and right channels are out of phase.
23999 The filter accepts the following options:
24003 Set the vectorscope mode.
24005 Available values are:
24008 Lissajous rotated by 45 degrees.
24011 Same as above but not rotated.
24014 Shape resembling half of circle.
24017 Default value is @samp{lissajous}.
24020 Set the video size for the output. For the syntax of this option, check the
24021 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24022 Default value is @code{400x400}.
24025 Set the output frame rate. Default value is @code{25}.
24031 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24032 @code{160}, @code{80} and @code{255}.
24033 Allowed range is @code{[0, 255]}.
24039 Specify the red, green, blue and alpha fade. Default values are @code{15},
24040 @code{10}, @code{5} and @code{5}.
24041 Allowed range is @code{[0, 255]}.
24044 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24045 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24048 Set the vectorscope drawing mode.
24050 Available values are:
24053 Draw dot for each sample.
24056 Draw line between previous and current sample.
24059 Default value is @samp{dot}.
24062 Specify amplitude scale of audio samples.
24064 Available values are:
24080 Swap left channel axis with right channel axis.
24090 Mirror only x axis.
24093 Mirror only y axis.
24101 @subsection Examples
24105 Complete example using @command{ffplay}:
24107 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24108 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24112 @section bench, abench
24114 Benchmark part of a filtergraph.
24116 The filter accepts the following options:
24120 Start or stop a timer.
24122 Available values are:
24125 Get the current time, set it as frame metadata (using the key
24126 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24129 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24130 the input frame metadata to get the time difference. Time difference, average,
24131 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24132 @code{min}) are then printed. The timestamps are expressed in seconds.
24136 @subsection Examples
24140 Benchmark @ref{selectivecolor} filter:
24142 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24148 Concatenate audio and video streams, joining them together one after the
24151 The filter works on segments of synchronized video and audio streams. All
24152 segments must have the same number of streams of each type, and that will
24153 also be the number of streams at output.
24155 The filter accepts the following options:
24160 Set the number of segments. Default is 2.
24163 Set the number of output video streams, that is also the number of video
24164 streams in each segment. Default is 1.
24167 Set the number of output audio streams, that is also the number of audio
24168 streams in each segment. Default is 0.
24171 Activate unsafe mode: do not fail if segments have a different format.
24175 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24176 @var{a} audio outputs.
24178 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24179 segment, in the same order as the outputs, then the inputs for the second
24182 Related streams do not always have exactly the same duration, for various
24183 reasons including codec frame size or sloppy authoring. For that reason,
24184 related synchronized streams (e.g. a video and its audio track) should be
24185 concatenated at once. The concat filter will use the duration of the longest
24186 stream in each segment (except the last one), and if necessary pad shorter
24187 audio streams with silence.
24189 For this filter to work correctly, all segments must start at timestamp 0.
24191 All corresponding streams must have the same parameters in all segments; the
24192 filtering system will automatically select a common pixel format for video
24193 streams, and a common sample format, sample rate and channel layout for
24194 audio streams, but other settings, such as resolution, must be converted
24195 explicitly by the user.
24197 Different frame rates are acceptable but will result in variable frame rate
24198 at output; be sure to configure the output file to handle it.
24200 @subsection Examples
24204 Concatenate an opening, an episode and an ending, all in bilingual version
24205 (video in stream 0, audio in streams 1 and 2):
24207 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24208 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24209 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24210 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24214 Concatenate two parts, handling audio and video separately, using the
24215 (a)movie sources, and adjusting the resolution:
24217 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24218 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24219 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24221 Note that a desync will happen at the stitch if the audio and video streams
24222 do not have exactly the same duration in the first file.
24226 @subsection Commands
24228 This filter supports the following commands:
24231 Close the current segment and step to the next one
24237 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24238 level. By default, it logs a message at a frequency of 10Hz with the
24239 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24240 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24242 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24243 sample format is double-precision floating point. The input stream will be converted to
24244 this specification, if needed. Users may need to insert aformat and/or aresample filters
24245 after this filter to obtain the original parameters.
24247 The filter also has a video output (see the @var{video} option) with a real
24248 time graph to observe the loudness evolution. The graphic contains the logged
24249 message mentioned above, so it is not printed anymore when this option is set,
24250 unless the verbose logging is set. The main graphing area contains the
24251 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24252 the momentary loudness (400 milliseconds), but can optionally be configured
24253 to instead display short-term loudness (see @var{gauge}).
24255 The green area marks a +/- 1LU target range around the target loudness
24256 (-23LUFS by default, unless modified through @var{target}).
24258 More information about the Loudness Recommendation EBU R128 on
24259 @url{http://tech.ebu.ch/loudness}.
24261 The filter accepts the following options:
24266 Activate the video output. The audio stream is passed unchanged whether this
24267 option is set or no. The video stream will be the first output stream if
24268 activated. Default is @code{0}.
24271 Set the video size. This option is for video only. For the syntax of this
24273 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24274 Default and minimum resolution is @code{640x480}.
24277 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24278 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24279 other integer value between this range is allowed.
24282 Set metadata injection. If set to @code{1}, the audio input will be segmented
24283 into 100ms output frames, each of them containing various loudness information
24284 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24286 Default is @code{0}.
24289 Force the frame logging level.
24291 Available values are:
24294 information logging level
24296 verbose logging level
24299 By default, the logging level is set to @var{info}. If the @option{video} or
24300 the @option{metadata} options are set, it switches to @var{verbose}.
24305 Available modes can be cumulated (the option is a @code{flag} type). Possible
24309 Disable any peak mode (default).
24311 Enable sample-peak mode.
24313 Simple peak mode looking for the higher sample value. It logs a message
24314 for sample-peak (identified by @code{SPK}).
24316 Enable true-peak mode.
24318 If enabled, the peak lookup is done on an over-sampled version of the input
24319 stream for better peak accuracy. It logs a message for true-peak.
24320 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24321 This mode requires a build with @code{libswresample}.
24325 Treat mono input files as "dual mono". If a mono file is intended for playback
24326 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24327 If set to @code{true}, this option will compensate for this effect.
24328 Multi-channel input files are not affected by this option.
24331 Set a specific pan law to be used for the measurement of dual mono files.
24332 This parameter is optional, and has a default value of -3.01dB.
24335 Set a specific target level (in LUFS) used as relative zero in the visualization.
24336 This parameter is optional and has a default value of -23LUFS as specified
24337 by EBU R128. However, material published online may prefer a level of -16LUFS
24338 (e.g. for use with podcasts or video platforms).
24341 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24342 @code{shortterm}. By default the momentary value will be used, but in certain
24343 scenarios it may be more useful to observe the short term value instead (e.g.
24347 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24348 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24349 video output, not the summary or continuous log output.
24352 @subsection Examples
24356 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24358 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24362 Run an analysis with @command{ffmpeg}:
24364 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24368 @section interleave, ainterleave
24370 Temporally interleave frames from several inputs.
24372 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24374 These filters read frames from several inputs and send the oldest
24375 queued frame to the output.
24377 Input streams must have well defined, monotonically increasing frame
24380 In order to submit one frame to output, these filters need to enqueue
24381 at least one frame for each input, so they cannot work in case one
24382 input is not yet terminated and will not receive incoming frames.
24384 For example consider the case when one input is a @code{select} filter
24385 which always drops input frames. The @code{interleave} filter will keep
24386 reading from that input, but it will never be able to send new frames
24387 to output until the input sends an end-of-stream signal.
24389 Also, depending on inputs synchronization, the filters will drop
24390 frames in case one input receives more frames than the other ones, and
24391 the queue is already filled.
24393 These filters accept the following options:
24397 Set the number of different inputs, it is 2 by default.
24400 How to determine the end-of-stream.
24404 The duration of the longest input. (default)
24407 The duration of the shortest input.
24410 The duration of the first input.
24415 @subsection Examples
24419 Interleave frames belonging to different streams using @command{ffmpeg}:
24421 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24425 Add flickering blur effect:
24427 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24431 @section metadata, ametadata
24433 Manipulate frame metadata.
24435 This filter accepts the following options:
24439 Set mode of operation of the filter.
24441 Can be one of the following:
24445 If both @code{value} and @code{key} is set, select frames
24446 which have such metadata. If only @code{key} is set, select
24447 every frame that has such key in metadata.
24450 Add new metadata @code{key} and @code{value}. If key is already available
24454 Modify value of already present key.
24457 If @code{value} is set, delete only keys that have such value.
24458 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24462 Print key and its value if metadata was found. If @code{key} is not set print all
24463 metadata values available in frame.
24467 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24470 Set metadata value which will be used. This option is mandatory for
24471 @code{modify} and @code{add} mode.
24474 Which function to use when comparing metadata value and @code{value}.
24476 Can be one of following:
24480 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24483 Values are interpreted as strings, returns true if metadata value starts with
24484 the @code{value} option string.
24487 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24490 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24493 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24496 Values are interpreted as floats, returns true if expression from option @code{expr}
24500 Values are interpreted as strings, returns true if metadata value ends with
24501 the @code{value} option string.
24505 Set expression which is used when @code{function} is set to @code{expr}.
24506 The expression is evaluated through the eval API and can contain the following
24511 Float representation of @code{value} from metadata key.
24514 Float representation of @code{value} as supplied by user in @code{value} option.
24518 If specified in @code{print} mode, output is written to the named file. Instead of
24519 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24520 for standard output. If @code{file} option is not set, output is written to the log
24521 with AV_LOG_INFO loglevel.
24524 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24528 @subsection Examples
24532 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24535 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24538 Print silencedetect output to file @file{metadata.txt}.
24540 silencedetect,ametadata=mode=print:file=metadata.txt
24543 Direct all metadata to a pipe with file descriptor 4.
24545 metadata=mode=print:file='pipe\:4'
24549 @section perms, aperms
24551 Set read/write permissions for the output frames.
24553 These filters are mainly aimed at developers to test direct path in the
24554 following filter in the filtergraph.
24556 The filters accept the following options:
24560 Select the permissions mode.
24562 It accepts the following values:
24565 Do nothing. This is the default.
24567 Set all the output frames read-only.
24569 Set all the output frames directly writable.
24571 Make the frame read-only if writable, and writable if read-only.
24573 Set each output frame read-only or writable randomly.
24577 Set the seed for the @var{random} mode, must be an integer included between
24578 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24579 @code{-1}, the filter will try to use a good random seed on a best effort
24583 Note: in case of auto-inserted filter between the permission filter and the
24584 following one, the permission might not be received as expected in that
24585 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24586 perms/aperms filter can avoid this problem.
24588 @section realtime, arealtime
24590 Slow down filtering to match real time approximately.
24592 These filters will pause the filtering for a variable amount of time to
24593 match the output rate with the input timestamps.
24594 They are similar to the @option{re} option to @code{ffmpeg}.
24596 They accept the following options:
24600 Time limit for the pauses. Any pause longer than that will be considered
24601 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24603 Speed factor for processing. The value must be a float larger than zero.
24604 Values larger than 1.0 will result in faster than realtime processing,
24605 smaller will slow processing down. The @var{limit} is automatically adapted
24606 accordingly. Default is 1.0.
24608 A processing speed faster than what is possible without these filters cannot
24613 @section select, aselect
24615 Select frames to pass in output.
24617 This filter accepts the following options:
24622 Set expression, which is evaluated for each input frame.
24624 If the expression is evaluated to zero, the frame is discarded.
24626 If the evaluation result is negative or NaN, the frame is sent to the
24627 first output; otherwise it is sent to the output with index
24628 @code{ceil(val)-1}, assuming that the input index starts from 0.
24630 For example a value of @code{1.2} corresponds to the output with index
24631 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24634 Set the number of outputs. The output to which to send the selected
24635 frame is based on the result of the evaluation. Default value is 1.
24638 The expression can contain the following constants:
24642 The (sequential) number of the filtered frame, starting from 0.
24645 The (sequential) number of the selected frame, starting from 0.
24647 @item prev_selected_n
24648 The sequential number of the last selected frame. It's NAN if undefined.
24651 The timebase of the input timestamps.
24654 The PTS (Presentation TimeStamp) of the filtered video frame,
24655 expressed in @var{TB} units. It's NAN if undefined.
24658 The PTS of the filtered video frame,
24659 expressed in seconds. It's NAN if undefined.
24662 The PTS of the previously filtered video frame. It's NAN if undefined.
24664 @item prev_selected_pts
24665 The PTS of the last previously filtered video frame. It's NAN if undefined.
24667 @item prev_selected_t
24668 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24671 The PTS of the first video frame in the video. It's NAN if undefined.
24674 The time of the first video frame in the video. It's NAN if undefined.
24676 @item pict_type @emph{(video only)}
24677 The type of the filtered frame. It can assume one of the following
24689 @item interlace_type @emph{(video only)}
24690 The frame interlace type. It can assume one of the following values:
24693 The frame is progressive (not interlaced).
24695 The frame is top-field-first.
24697 The frame is bottom-field-first.
24700 @item consumed_sample_n @emph{(audio only)}
24701 the number of selected samples before the current frame
24703 @item samples_n @emph{(audio only)}
24704 the number of samples in the current frame
24706 @item sample_rate @emph{(audio only)}
24707 the input sample rate
24710 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24713 the position in the file of the filtered frame, -1 if the information
24714 is not available (e.g. for synthetic video)
24716 @item scene @emph{(video only)}
24717 value between 0 and 1 to indicate a new scene; a low value reflects a low
24718 probability for the current frame to introduce a new scene, while a higher
24719 value means the current frame is more likely to be one (see the example below)
24721 @item concatdec_select
24722 The concat demuxer can select only part of a concat input file by setting an
24723 inpoint and an outpoint, but the output packets may not be entirely contained
24724 in the selected interval. By using this variable, it is possible to skip frames
24725 generated by the concat demuxer which are not exactly contained in the selected
24728 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24729 and the @var{lavf.concat.duration} packet metadata values which are also
24730 present in the decoded frames.
24732 The @var{concatdec_select} variable is -1 if the frame pts is at least
24733 start_time and either the duration metadata is missing or the frame pts is less
24734 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24737 That basically means that an input frame is selected if its pts is within the
24738 interval set by the concat demuxer.
24742 The default value of the select expression is "1".
24744 @subsection Examples
24748 Select all frames in input:
24753 The example above is the same as:
24765 Select only I-frames:
24767 select='eq(pict_type\,I)'
24771 Select one frame every 100:
24773 select='not(mod(n\,100))'
24777 Select only frames contained in the 10-20 time interval:
24779 select=between(t\,10\,20)
24783 Select only I-frames contained in the 10-20 time interval:
24785 select=between(t\,10\,20)*eq(pict_type\,I)
24789 Select frames with a minimum distance of 10 seconds:
24791 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24795 Use aselect to select only audio frames with samples number > 100:
24797 aselect='gt(samples_n\,100)'
24801 Create a mosaic of the first scenes:
24803 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24806 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24810 Send even and odd frames to separate outputs, and compose them:
24812 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24816 Select useful frames from an ffconcat file which is using inpoints and
24817 outpoints but where the source files are not intra frame only.
24819 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24823 @section sendcmd, asendcmd
24825 Send commands to filters in the filtergraph.
24827 These filters read commands to be sent to other filters in the
24830 @code{sendcmd} must be inserted between two video filters,
24831 @code{asendcmd} must be inserted between two audio filters, but apart
24832 from that they act the same way.
24834 The specification of commands can be provided in the filter arguments
24835 with the @var{commands} option, or in a file specified by the
24836 @var{filename} option.
24838 These filters accept the following options:
24841 Set the commands to be read and sent to the other filters.
24843 Set the filename of the commands to be read and sent to the other
24847 @subsection Commands syntax
24849 A commands description consists of a sequence of interval
24850 specifications, comprising a list of commands to be executed when a
24851 particular event related to that interval occurs. The occurring event
24852 is typically the current frame time entering or leaving a given time
24855 An interval is specified by the following syntax:
24857 @var{START}[-@var{END}] @var{COMMANDS};
24860 The time interval is specified by the @var{START} and @var{END} times.
24861 @var{END} is optional and defaults to the maximum time.
24863 The current frame time is considered within the specified interval if
24864 it is included in the interval [@var{START}, @var{END}), that is when
24865 the time is greater or equal to @var{START} and is lesser than
24868 @var{COMMANDS} consists of a sequence of one or more command
24869 specifications, separated by ",", relating to that interval. The
24870 syntax of a command specification is given by:
24872 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24875 @var{FLAGS} is optional and specifies the type of events relating to
24876 the time interval which enable sending the specified command, and must
24877 be a non-null sequence of identifier flags separated by "+" or "|" and
24878 enclosed between "[" and "]".
24880 The following flags are recognized:
24883 The command is sent when the current frame timestamp enters the
24884 specified interval. In other words, the command is sent when the
24885 previous frame timestamp was not in the given interval, and the
24889 The command is sent when the current frame timestamp leaves the
24890 specified interval. In other words, the command is sent when the
24891 previous frame timestamp was in the given interval, and the
24895 The command @var{ARG} is interpreted as expression and result of
24896 expression is passed as @var{ARG}.
24898 The expression is evaluated through the eval API and can contain the following
24903 Original position in the file of the frame, or undefined if undefined
24904 for the current frame.
24907 The presentation timestamp in input.
24910 The count of the input frame for video or audio, starting from 0.
24913 The time in seconds of the current frame.
24916 The start time in seconds of the current command interval.
24919 The end time in seconds of the current command interval.
24922 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24927 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24930 @var{TARGET} specifies the target of the command, usually the name of
24931 the filter class or a specific filter instance name.
24933 @var{COMMAND} specifies the name of the command for the target filter.
24935 @var{ARG} is optional and specifies the optional list of argument for
24936 the given @var{COMMAND}.
24938 Between one interval specification and another, whitespaces, or
24939 sequences of characters starting with @code{#} until the end of line,
24940 are ignored and can be used to annotate comments.
24942 A simplified BNF description of the commands specification syntax
24945 @var{COMMAND_FLAG} ::= "enter" | "leave"
24946 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24947 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24948 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24949 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24950 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24953 @subsection Examples
24957 Specify audio tempo change at second 4:
24959 asendcmd=c='4.0 atempo tempo 1.5',atempo
24963 Target a specific filter instance:
24965 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24969 Specify a list of drawtext and hue commands in a file.
24971 # show text in the interval 5-10
24972 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24973 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24975 # desaturate the image in the interval 15-20
24976 15.0-20.0 [enter] hue s 0,
24977 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24979 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24981 # apply an exponential saturation fade-out effect, starting from time 25
24982 25 [enter] hue s exp(25-t)
24985 A filtergraph allowing to read and process the above command list
24986 stored in a file @file{test.cmd}, can be specified with:
24988 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24993 @section setpts, asetpts
24995 Change the PTS (presentation timestamp) of the input frames.
24997 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24999 This filter accepts the following options:
25004 The expression which is evaluated for each frame to construct its timestamp.
25008 The expression is evaluated through the eval API and can contain the following
25012 @item FRAME_RATE, FR
25013 frame rate, only defined for constant frame-rate video
25016 The presentation timestamp in input
25019 The count of the input frame for video or the number of consumed samples,
25020 not including the current frame for audio, starting from 0.
25022 @item NB_CONSUMED_SAMPLES
25023 The number of consumed samples, not including the current frame (only
25026 @item NB_SAMPLES, S
25027 The number of samples in the current frame (only audio)
25029 @item SAMPLE_RATE, SR
25030 The audio sample rate.
25033 The PTS of the first frame.
25036 the time in seconds of the first frame
25039 State whether the current frame is interlaced.
25042 the time in seconds of the current frame
25045 original position in the file of the frame, or undefined if undefined
25046 for the current frame
25049 The previous input PTS.
25052 previous input time in seconds
25055 The previous output PTS.
25058 previous output time in seconds
25061 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25065 The wallclock (RTC) time at the start of the movie in microseconds.
25068 The timebase of the input timestamps.
25072 @subsection Examples
25076 Start counting PTS from zero
25078 setpts=PTS-STARTPTS
25082 Apply fast motion effect:
25088 Apply slow motion effect:
25094 Set fixed rate of 25 frames per second:
25100 Set fixed rate 25 fps with some jitter:
25102 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25106 Apply an offset of 10 seconds to the input PTS:
25112 Generate timestamps from a "live source" and rebase onto the current timebase:
25114 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25118 Generate timestamps by counting samples:
25127 Force color range for the output video frame.
25129 The @code{setrange} filter marks the color range property for the
25130 output frames. It does not change the input frame, but only sets the
25131 corresponding property, which affects how the frame is treated by
25134 The filter accepts the following options:
25139 Available values are:
25143 Keep the same color range property.
25145 @item unspecified, unknown
25146 Set the color range as unspecified.
25148 @item limited, tv, mpeg
25149 Set the color range as limited.
25151 @item full, pc, jpeg
25152 Set the color range as full.
25156 @section settb, asettb
25158 Set the timebase to use for the output frames timestamps.
25159 It is mainly useful for testing timebase configuration.
25161 It accepts the following parameters:
25166 The expression which is evaluated into the output timebase.
25170 The value for @option{tb} is an arithmetic expression representing a
25171 rational. The expression can contain the constants "AVTB" (the default
25172 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25173 audio only). Default value is "intb".
25175 @subsection Examples
25179 Set the timebase to 1/25:
25185 Set the timebase to 1/10:
25191 Set the timebase to 1001/1000:
25197 Set the timebase to 2*intb:
25203 Set the default timebase value:
25210 Convert input audio to a video output representing frequency spectrum
25211 logarithmically using Brown-Puckette constant Q transform algorithm with
25212 direct frequency domain coefficient calculation (but the transform itself
25213 is not really constant Q, instead the Q factor is actually variable/clamped),
25214 with musical tone scale, from E0 to D#10.
25216 The filter accepts the following options:
25220 Specify the video size for the output. It must be even. For the syntax of this option,
25221 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25222 Default value is @code{1920x1080}.
25225 Set the output frame rate. Default value is @code{25}.
25228 Set the bargraph height. It must be even. Default value is @code{-1} which
25229 computes the bargraph height automatically.
25232 Set the axis height. It must be even. Default value is @code{-1} which computes
25233 the axis height automatically.
25236 Set the sonogram height. It must be even. Default value is @code{-1} which
25237 computes the sonogram height automatically.
25240 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25241 instead. Default value is @code{1}.
25243 @item sono_v, volume
25244 Specify the sonogram volume expression. It can contain variables:
25247 the @var{bar_v} evaluated expression
25248 @item frequency, freq, f
25249 the frequency where it is evaluated
25250 @item timeclamp, tc
25251 the value of @var{timeclamp} option
25255 @item a_weighting(f)
25256 A-weighting of equal loudness
25257 @item b_weighting(f)
25258 B-weighting of equal loudness
25259 @item c_weighting(f)
25260 C-weighting of equal loudness.
25262 Default value is @code{16}.
25264 @item bar_v, volume2
25265 Specify the bargraph volume expression. It can contain variables:
25268 the @var{sono_v} evaluated expression
25269 @item frequency, freq, f
25270 the frequency where it is evaluated
25271 @item timeclamp, tc
25272 the value of @var{timeclamp} option
25276 @item a_weighting(f)
25277 A-weighting of equal loudness
25278 @item b_weighting(f)
25279 B-weighting of equal loudness
25280 @item c_weighting(f)
25281 C-weighting of equal loudness.
25283 Default value is @code{sono_v}.
25285 @item sono_g, gamma
25286 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25287 higher gamma makes the spectrum having more range. Default value is @code{3}.
25288 Acceptable range is @code{[1, 7]}.
25290 @item bar_g, gamma2
25291 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25295 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25296 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25298 @item timeclamp, tc
25299 Specify the transform timeclamp. At low frequency, there is trade-off between
25300 accuracy in time domain and frequency domain. If timeclamp is lower,
25301 event in time domain is represented more accurately (such as fast bass drum),
25302 otherwise event in frequency domain is represented more accurately
25303 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25306 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25307 limits future samples by applying asymmetric windowing in time domain, useful
25308 when low latency is required. Accepted range is @code{[0, 1]}.
25311 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25312 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25315 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25316 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25319 This option is deprecated and ignored.
25322 Specify the transform length in time domain. Use this option to control accuracy
25323 trade-off between time domain and frequency domain at every frequency sample.
25324 It can contain variables:
25326 @item frequency, freq, f
25327 the frequency where it is evaluated
25328 @item timeclamp, tc
25329 the value of @var{timeclamp} option.
25331 Default value is @code{384*tc/(384+tc*f)}.
25334 Specify the transform count for every video frame. Default value is @code{6}.
25335 Acceptable range is @code{[1, 30]}.
25338 Specify the transform count for every single pixel. Default value is @code{0},
25339 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25342 Specify font file for use with freetype to draw the axis. If not specified,
25343 use embedded font. Note that drawing with font file or embedded font is not
25344 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25348 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25349 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25353 Specify font color expression. This is arithmetic expression that should return
25354 integer value 0xRRGGBB. It can contain variables:
25356 @item frequency, freq, f
25357 the frequency where it is evaluated
25358 @item timeclamp, tc
25359 the value of @var{timeclamp} option
25364 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25365 @item r(x), g(x), b(x)
25366 red, green, and blue value of intensity x.
25368 Default value is @code{st(0, (midi(f)-59.5)/12);
25369 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25370 r(1-ld(1)) + b(ld(1))}.
25373 Specify image file to draw the axis. This option override @var{fontfile} and
25374 @var{fontcolor} option.
25377 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25378 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25379 Default value is @code{1}.
25382 Set colorspace. The accepted values are:
25385 Unspecified (default)
25394 BT.470BG or BT.601-6 625
25397 SMPTE-170M or BT.601-6 525
25403 BT.2020 with non-constant luminance
25408 Set spectrogram color scheme. This is list of floating point values with format
25409 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25410 The default is @code{1|0.5|0|0|0.5|1}.
25414 @subsection Examples
25418 Playing audio while showing the spectrum:
25420 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25424 Same as above, but with frame rate 30 fps:
25426 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25430 Playing at 1280x720:
25432 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25436 Disable sonogram display:
25442 A1 and its harmonics: A1, A2, (near)E3, A3:
25444 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),
25445 asplit[a][out1]; [a] showcqt [out0]'
25449 Same as above, but with more accuracy in frequency domain:
25451 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),
25452 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25458 bar_v=10:sono_v=bar_v*a_weighting(f)
25462 Custom gamma, now spectrum is linear to the amplitude.
25468 Custom tlength equation:
25470 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)))'
25474 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25476 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25480 Custom font using fontconfig:
25482 font='Courier New,Monospace,mono|bold'
25486 Custom frequency range with custom axis using image file:
25488 axisfile=myaxis.png:basefreq=40:endfreq=10000
25494 Convert input audio to video output representing the audio power spectrum.
25495 Audio amplitude is on Y-axis while frequency is on X-axis.
25497 The filter accepts the following options:
25501 Specify size of video. For the syntax of this option, check the
25502 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25503 Default is @code{1024x512}.
25507 This set how each frequency bin will be represented.
25509 It accepts the following values:
25515 Default is @code{bar}.
25518 Set amplitude scale.
25520 It accepts the following values:
25534 Default is @code{log}.
25537 Set frequency scale.
25539 It accepts the following values:
25548 Reverse logarithmic scale.
25550 Default is @code{lin}.
25553 Set window size. Allowed range is from 16 to 65536.
25555 Default is @code{2048}
25558 Set windowing function.
25560 It accepts the following values:
25583 Default is @code{hanning}.
25586 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25587 which means optimal overlap for selected window function will be picked.
25590 Set time averaging. Setting this to 0 will display current maximal peaks.
25591 Default is @code{1}, which means time averaging is disabled.
25594 Specify list of colors separated by space or by '|' which will be used to
25595 draw channel frequencies. Unrecognized or missing colors will be replaced
25599 Set channel display mode.
25601 It accepts the following values:
25606 Default is @code{combined}.
25609 Set minimum amplitude used in @code{log} amplitude scaler.
25612 Set data display mode.
25614 It accepts the following values:
25620 Default is @code{magnitude}.
25623 @section showspatial
25625 Convert stereo input audio to a video output, representing the spatial relationship
25626 between two channels.
25628 The filter accepts the following options:
25632 Specify the video size for the output. For the syntax of this option, check the
25633 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25634 Default value is @code{512x512}.
25637 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25640 Set window function.
25642 It accepts the following values:
25667 Default value is @code{hann}.
25670 Set ratio of overlap window. Default value is @code{0.5}.
25671 When value is @code{1} overlap is set to recommended size for specific
25672 window function currently used.
25675 @anchor{showspectrum}
25676 @section showspectrum
25678 Convert input audio to a video output, representing the audio frequency
25681 The filter accepts the following options:
25685 Specify the video size for the output. For the syntax of this option, check the
25686 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25687 Default value is @code{640x512}.
25690 Specify how the spectrum should slide along the window.
25692 It accepts the following values:
25695 the samples start again on the left when they reach the right
25697 the samples scroll from right to left
25699 frames are only produced when the samples reach the right
25701 the samples scroll from left to right
25704 Default value is @code{replace}.
25707 Specify display mode.
25709 It accepts the following values:
25712 all channels are displayed in the same row
25714 all channels are displayed in separate rows
25717 Default value is @samp{combined}.
25720 Specify display color mode.
25722 It accepts the following values:
25725 each channel is displayed in a separate color
25727 each channel is displayed using the same color scheme
25729 each channel is displayed using the rainbow color scheme
25731 each channel is displayed using the moreland color scheme
25733 each channel is displayed using the nebulae color scheme
25735 each channel is displayed using the fire color scheme
25737 each channel is displayed using the fiery color scheme
25739 each channel is displayed using the fruit color scheme
25741 each channel is displayed using the cool color scheme
25743 each channel is displayed using the magma color scheme
25745 each channel is displayed using the green color scheme
25747 each channel is displayed using the viridis color scheme
25749 each channel is displayed using the plasma color scheme
25751 each channel is displayed using the cividis color scheme
25753 each channel is displayed using the terrain color scheme
25756 Default value is @samp{channel}.
25759 Specify scale used for calculating intensity color values.
25761 It accepts the following values:
25766 square root, default
25777 Default value is @samp{sqrt}.
25780 Specify frequency scale.
25782 It accepts the following values:
25790 Default value is @samp{lin}.
25793 Set saturation modifier for displayed colors. Negative values provide
25794 alternative color scheme. @code{0} is no saturation at all.
25795 Saturation must be in [-10.0, 10.0] range.
25796 Default value is @code{1}.
25799 Set window function.
25801 It accepts the following values:
25826 Default value is @code{hann}.
25829 Set orientation of time vs frequency axis. Can be @code{vertical} or
25830 @code{horizontal}. Default is @code{vertical}.
25833 Set ratio of overlap window. Default value is @code{0}.
25834 When value is @code{1} overlap is set to recommended size for specific
25835 window function currently used.
25838 Set scale gain for calculating intensity color values.
25839 Default value is @code{1}.
25842 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25845 Set color rotation, must be in [-1.0, 1.0] range.
25846 Default value is @code{0}.
25849 Set start frequency from which to display spectrogram. Default is @code{0}.
25852 Set stop frequency to which to display spectrogram. Default is @code{0}.
25855 Set upper frame rate limit. Default is @code{auto}, unlimited.
25858 Draw time and frequency axes and legends. Default is disabled.
25861 The usage is very similar to the showwaves filter; see the examples in that
25864 @subsection Examples
25868 Large window with logarithmic color scaling:
25870 showspectrum=s=1280x480:scale=log
25874 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25876 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25877 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25881 @section showspectrumpic
25883 Convert input audio to a single video frame, representing the audio frequency
25886 The filter accepts the following options:
25890 Specify the video size for the output. For the syntax of this option, check the
25891 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25892 Default value is @code{4096x2048}.
25895 Specify display mode.
25897 It accepts the following values:
25900 all channels are displayed in the same row
25902 all channels are displayed in separate rows
25904 Default value is @samp{combined}.
25907 Specify display color mode.
25909 It accepts the following values:
25912 each channel is displayed in a separate color
25914 each channel is displayed using the same color scheme
25916 each channel is displayed using the rainbow color scheme
25918 each channel is displayed using the moreland color scheme
25920 each channel is displayed using the nebulae color scheme
25922 each channel is displayed using the fire color scheme
25924 each channel is displayed using the fiery color scheme
25926 each channel is displayed using the fruit color scheme
25928 each channel is displayed using the cool color scheme
25930 each channel is displayed using the magma color scheme
25932 each channel is displayed using the green color scheme
25934 each channel is displayed using the viridis color scheme
25936 each channel is displayed using the plasma color scheme
25938 each channel is displayed using the cividis color scheme
25940 each channel is displayed using the terrain color scheme
25942 Default value is @samp{intensity}.
25945 Specify scale used for calculating intensity color values.
25947 It accepts the following values:
25952 square root, default
25962 Default value is @samp{log}.
25965 Specify frequency scale.
25967 It accepts the following values:
25975 Default value is @samp{lin}.
25978 Set saturation modifier for displayed colors. Negative values provide
25979 alternative color scheme. @code{0} is no saturation at all.
25980 Saturation must be in [-10.0, 10.0] range.
25981 Default value is @code{1}.
25984 Set window function.
25986 It accepts the following values:
26010 Default value is @code{hann}.
26013 Set orientation of time vs frequency axis. Can be @code{vertical} or
26014 @code{horizontal}. Default is @code{vertical}.
26017 Set scale gain for calculating intensity color values.
26018 Default value is @code{1}.
26021 Draw time and frequency axes and legends. Default is enabled.
26024 Set color rotation, must be in [-1.0, 1.0] range.
26025 Default value is @code{0}.
26028 Set start frequency from which to display spectrogram. Default is @code{0}.
26031 Set stop frequency to which to display spectrogram. Default is @code{0}.
26034 @subsection Examples
26038 Extract an audio spectrogram of a whole audio track
26039 in a 1024x1024 picture using @command{ffmpeg}:
26041 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26045 @section showvolume
26047 Convert input audio volume to a video output.
26049 The filter accepts the following options:
26056 Set border width, allowed range is [0, 5]. Default is 1.
26059 Set channel width, allowed range is [80, 8192]. Default is 400.
26062 Set channel height, allowed range is [1, 900]. Default is 20.
26065 Set fade, allowed range is [0, 1]. Default is 0.95.
26068 Set volume color expression.
26070 The expression can use the following variables:
26074 Current max volume of channel in dB.
26080 Current channel number, starting from 0.
26084 If set, displays channel names. Default is enabled.
26087 If set, displays volume values. Default is enabled.
26090 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26091 default is @code{h}.
26094 Set step size, allowed range is [0, 5]. Default is 0, which means
26098 Set background opacity, allowed range is [0, 1]. Default is 0.
26101 Set metering mode, can be peak: @code{p} or rms: @code{r},
26102 default is @code{p}.
26105 Set display scale, can be linear: @code{lin} or log: @code{log},
26106 default is @code{lin}.
26110 If set to > 0., display a line for the max level
26111 in the previous seconds.
26112 default is disabled: @code{0.}
26115 The color of the max line. Use when @code{dm} option is set to > 0.
26116 default is: @code{orange}
26121 Convert input audio to a video output, representing the samples waves.
26123 The filter accepts the following options:
26127 Specify the video size for the output. For the syntax of this option, check the
26128 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26129 Default value is @code{600x240}.
26134 Available values are:
26137 Draw a point for each sample.
26140 Draw a vertical line for each sample.
26143 Draw a point for each sample and a line between them.
26146 Draw a centered vertical line for each sample.
26149 Default value is @code{point}.
26152 Set the number of samples which are printed on the same column. A
26153 larger value will decrease the frame rate. Must be a positive
26154 integer. This option can be set only if the value for @var{rate}
26155 is not explicitly specified.
26158 Set the (approximate) output frame rate. This is done by setting the
26159 option @var{n}. Default value is "25".
26161 @item split_channels
26162 Set if channels should be drawn separately or overlap. Default value is 0.
26165 Set colors separated by '|' which are going to be used for drawing of each channel.
26168 Set amplitude scale.
26170 Available values are:
26188 Set the draw mode. This is mostly useful to set for high @var{n}.
26190 Available values are:
26193 Scale pixel values for each drawn sample.
26196 Draw every sample directly.
26199 Default value is @code{scale}.
26202 @subsection Examples
26206 Output the input file audio and the corresponding video representation
26209 amovie=a.mp3,asplit[out0],showwaves[out1]
26213 Create a synthetic signal and show it with showwaves, forcing a
26214 frame rate of 30 frames per second:
26216 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26220 @section showwavespic
26222 Convert input audio to a single video frame, representing the samples waves.
26224 The filter accepts the following options:
26228 Specify the video size for the output. For the syntax of this option, check the
26229 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26230 Default value is @code{600x240}.
26232 @item split_channels
26233 Set if channels should be drawn separately or overlap. Default value is 0.
26236 Set colors separated by '|' which are going to be used for drawing of each channel.
26239 Set amplitude scale.
26241 Available values are:
26261 Available values are:
26264 Scale pixel values for each drawn sample.
26267 Draw every sample directly.
26270 Default value is @code{scale}.
26273 Set the filter mode.
26275 Available values are:
26278 Use average samples values for each drawn sample.
26281 Use peak samples values for each drawn sample.
26284 Default value is @code{average}.
26287 @subsection Examples
26291 Extract a channel split representation of the wave form of a whole audio track
26292 in a 1024x800 picture using @command{ffmpeg}:
26294 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26298 @section sidedata, asidedata
26300 Delete frame side data, or select frames based on it.
26302 This filter accepts the following options:
26306 Set mode of operation of the filter.
26308 Can be one of the following:
26312 Select every frame with side data of @code{type}.
26315 Delete side data of @code{type}. If @code{type} is not set, delete all side
26321 Set side data type used with all modes. Must be set for @code{select} mode. For
26322 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26323 in @file{libavutil/frame.h}. For example, to choose
26324 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26328 @section spectrumsynth
26330 Synthesize audio from 2 input video spectrums, first input stream represents
26331 magnitude across time and second represents phase across time.
26332 The filter will transform from frequency domain as displayed in videos back
26333 to time domain as presented in audio output.
26335 This filter is primarily created for reversing processed @ref{showspectrum}
26336 filter outputs, but can synthesize sound from other spectrograms too.
26337 But in such case results are going to be poor if the phase data is not
26338 available, because in such cases phase data need to be recreated, usually
26339 it's just recreated from random noise.
26340 For best results use gray only output (@code{channel} color mode in
26341 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26342 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26343 @code{data} option. Inputs videos should generally use @code{fullframe}
26344 slide mode as that saves resources needed for decoding video.
26346 The filter accepts the following options:
26350 Specify sample rate of output audio, the sample rate of audio from which
26351 spectrum was generated may differ.
26354 Set number of channels represented in input video spectrums.
26357 Set scale which was used when generating magnitude input spectrum.
26358 Can be @code{lin} or @code{log}. Default is @code{log}.
26361 Set slide which was used when generating inputs spectrums.
26362 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26363 Default is @code{fullframe}.
26366 Set window function used for resynthesis.
26369 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26370 which means optimal overlap for selected window function will be picked.
26373 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26374 Default is @code{vertical}.
26377 @subsection Examples
26381 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26382 then resynthesize videos back to audio with spectrumsynth:
26384 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
26385 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
26386 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26390 @section split, asplit
26392 Split input into several identical outputs.
26394 @code{asplit} works with audio input, @code{split} with video.
26396 The filter accepts a single parameter which specifies the number of outputs. If
26397 unspecified, it defaults to 2.
26399 @subsection Examples
26403 Create two separate outputs from the same input:
26405 [in] split [out0][out1]
26409 To create 3 or more outputs, you need to specify the number of
26412 [in] asplit=3 [out0][out1][out2]
26416 Create two separate outputs from the same input, one cropped and
26419 [in] split [splitout1][splitout2];
26420 [splitout1] crop=100:100:0:0 [cropout];
26421 [splitout2] pad=200:200:100:100 [padout];
26425 Create 5 copies of the input audio with @command{ffmpeg}:
26427 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26433 Receive commands sent through a libzmq client, and forward them to
26434 filters in the filtergraph.
26436 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26437 must be inserted between two video filters, @code{azmq} between two
26438 audio filters. Both are capable to send messages to any filter type.
26440 To enable these filters you need to install the libzmq library and
26441 headers and configure FFmpeg with @code{--enable-libzmq}.
26443 For more information about libzmq see:
26444 @url{http://www.zeromq.org/}
26446 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26447 receives messages sent through a network interface defined by the
26448 @option{bind_address} (or the abbreviation "@option{b}") option.
26449 Default value of this option is @file{tcp://localhost:5555}. You may
26450 want to alter this value to your needs, but do not forget to escape any
26451 ':' signs (see @ref{filtergraph escaping}).
26453 The received message must be in the form:
26455 @var{TARGET} @var{COMMAND} [@var{ARG}]
26458 @var{TARGET} specifies the target of the command, usually the name of
26459 the filter class or a specific filter instance name. The default
26460 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26461 but you can override this by using the @samp{filter_name@@id} syntax
26462 (see @ref{Filtergraph syntax}).
26464 @var{COMMAND} specifies the name of the command for the target filter.
26466 @var{ARG} is optional and specifies the optional argument list for the
26467 given @var{COMMAND}.
26469 Upon reception, the message is processed and the corresponding command
26470 is injected into the filtergraph. Depending on the result, the filter
26471 will send a reply to the client, adopting the format:
26473 @var{ERROR_CODE} @var{ERROR_REASON}
26477 @var{MESSAGE} is optional.
26479 @subsection Examples
26481 Look at @file{tools/zmqsend} for an example of a zmq client which can
26482 be used to send commands processed by these filters.
26484 Consider the following filtergraph generated by @command{ffplay}.
26485 In this example the last overlay filter has an instance name. All other
26486 filters will have default instance names.
26489 ffplay -dumpgraph 1 -f lavfi "
26490 color=s=100x100:c=red [l];
26491 color=s=100x100:c=blue [r];
26492 nullsrc=s=200x100, zmq [bg];
26493 [bg][l] overlay [bg+l];
26494 [bg+l][r] overlay@@my=x=100 "
26497 To change the color of the left side of the video, the following
26498 command can be used:
26500 echo Parsed_color_0 c yellow | tools/zmqsend
26503 To change the right side:
26505 echo Parsed_color_1 c pink | tools/zmqsend
26508 To change the position of the right side:
26510 echo overlay@@my x 150 | tools/zmqsend
26514 @c man end MULTIMEDIA FILTERS
26516 @chapter Multimedia Sources
26517 @c man begin MULTIMEDIA SOURCES
26519 Below is a description of the currently available multimedia sources.
26523 This is the same as @ref{movie} source, except it selects an audio
26529 Read audio and/or video stream(s) from a movie container.
26531 It accepts the following parameters:
26535 The name of the resource to read (not necessarily a file; it can also be a
26536 device or a stream accessed through some protocol).
26538 @item format_name, f
26539 Specifies the format assumed for the movie to read, and can be either
26540 the name of a container or an input device. If not specified, the
26541 format is guessed from @var{movie_name} or by probing.
26543 @item seek_point, sp
26544 Specifies the seek point in seconds. The frames will be output
26545 starting from this seek point. The parameter is evaluated with
26546 @code{av_strtod}, so the numerical value may be suffixed by an IS
26547 postfix. The default value is "0".
26550 Specifies the streams to read. Several streams can be specified,
26551 separated by "+". The source will then have as many outputs, in the
26552 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26553 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26554 respectively the default (best suited) video and audio stream. Default
26555 is "dv", or "da" if the filter is called as "amovie".
26557 @item stream_index, si
26558 Specifies the index of the video stream to read. If the value is -1,
26559 the most suitable video stream will be automatically selected. The default
26560 value is "-1". Deprecated. If the filter is called "amovie", it will select
26561 audio instead of video.
26564 Specifies how many times to read the stream in sequence.
26565 If the value is 0, the stream will be looped infinitely.
26566 Default value is "1".
26568 Note that when the movie is looped the source timestamps are not
26569 changed, so it will generate non monotonically increasing timestamps.
26571 @item discontinuity
26572 Specifies the time difference between frames above which the point is
26573 considered a timestamp discontinuity which is removed by adjusting the later
26577 It allows overlaying a second video on top of the main input of
26578 a filtergraph, as shown in this graph:
26580 input -----------> deltapts0 --> overlay --> output
26583 movie --> scale--> deltapts1 -------+
26585 @subsection Examples
26589 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26590 on top of the input labelled "in":
26592 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26593 [in] setpts=PTS-STARTPTS [main];
26594 [main][over] overlay=16:16 [out]
26598 Read from a video4linux2 device, and overlay it on top of the input
26601 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26602 [in] setpts=PTS-STARTPTS [main];
26603 [main][over] overlay=16:16 [out]
26607 Read the first video stream and the audio stream with id 0x81 from
26608 dvd.vob; the video is connected to the pad named "video" and the audio is
26609 connected to the pad named "audio":
26611 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26615 @subsection Commands
26617 Both movie and amovie support the following commands:
26620 Perform seek using "av_seek_frame".
26621 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26624 @var{stream_index}: If stream_index is -1, a default
26625 stream is selected, and @var{timestamp} is automatically converted
26626 from AV_TIME_BASE units to the stream specific time_base.
26628 @var{timestamp}: Timestamp in AVStream.time_base units
26629 or, if no stream is specified, in AV_TIME_BASE units.
26631 @var{flags}: Flags which select direction and seeking mode.
26635 Get movie duration in AV_TIME_BASE units.
26639 @c man end MULTIMEDIA SOURCES