1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Examples
1093 Fade in first 15 seconds of audio:
1095 afade=t=in:ss=0:d=15
1099 Fade out last 25 seconds of a 900 seconds audio:
1101 afade=t=out:st=875:d=25
1106 Denoise audio samples with FFT.
1108 A description of the accepted parameters follows.
1112 Set the noise reduction in dB, allowed range is 0.01 to 97.
1113 Default value is 12 dB.
1116 Set the noise floor in dB, allowed range is -80 to -20.
1117 Default value is -50 dB.
1122 It accepts the following values:
1131 Select shellac noise.
1134 Select custom noise, defined in @code{bn} option.
1136 Default value is white noise.
1140 Set custom band noise for every one of 15 bands.
1141 Bands are separated by ' ' or '|'.
1144 Set the residual floor in dB, allowed range is -80 to -20.
1145 Default value is -38 dB.
1148 Enable noise tracking. By default is disabled.
1149 With this enabled, noise floor is automatically adjusted.
1152 Enable residual tracking. By default is disabled.
1155 Set the output mode.
1157 It accepts the following values:
1160 Pass input unchanged.
1163 Pass noise filtered out.
1168 Default value is @var{o}.
1172 @subsection Commands
1174 This filter supports the following commands:
1176 @item sample_noise, sn
1177 Start or stop measuring noise profile.
1178 Syntax for the command is : "start" or "stop" string.
1179 After measuring noise profile is stopped it will be
1180 automatically applied in filtering.
1182 @item noise_reduction, nr
1183 Change noise reduction. Argument is single float number.
1184 Syntax for the command is : "@var{noise_reduction}"
1186 @item noise_floor, nf
1187 Change noise floor. Argument is single float number.
1188 Syntax for the command is : "@var{noise_floor}"
1190 @item output_mode, om
1191 Change output mode operation.
1192 Syntax for the command is : "i", "o" or "n" string.
1196 Apply arbitrary expressions to samples in frequency domain.
1200 Set frequency domain real expression for each separate channel separated
1201 by '|'. Default is "re".
1202 If the number of input channels is greater than the number of
1203 expressions, the last specified expression is used for the remaining
1207 Set frequency domain imaginary expression for each separate channel
1208 separated by '|'. Default is "im".
1210 Each expression in @var{real} and @var{imag} can contain the following
1211 constants and functions:
1218 current frequency bin number
1221 number of available bins
1224 channel number of the current expression
1233 current real part of frequency bin of current channel
1236 current imaginary part of frequency bin of current channel
1239 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1242 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1246 Set window size. Allowed range is from 16 to 131072.
1247 Default is @code{4096}
1250 Set window function. Default is @code{hann}.
1253 Set window overlap. If set to 1, the recommended overlap for selected
1254 window function will be picked. Default is @code{0.75}.
1257 @subsection Examples
1261 Leave almost only low frequencies in audio:
1263 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1267 Apply robotize effect:
1269 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1273 Apply whisper effect:
1275 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1282 Apply an arbitrary Finite Impulse Response filter.
1284 This filter is designed for applying long FIR filters,
1285 up to 60 seconds long.
1287 It can be used as component for digital crossover filters,
1288 room equalization, cross talk cancellation, wavefield synthesis,
1289 auralization, ambiophonics, ambisonics and spatialization.
1291 This filter uses the streams higher than first one as FIR coefficients.
1292 If the non-first stream holds a single channel, it will be used
1293 for all input channels in the first stream, otherwise
1294 the number of channels in the non-first stream must be same as
1295 the number of channels in the first stream.
1297 It accepts the following parameters:
1301 Set dry gain. This sets input gain.
1304 Set wet gain. This sets final output gain.
1307 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1310 Enable applying gain measured from power of IR.
1312 Set which approach to use for auto gain measurement.
1316 Do not apply any gain.
1319 select peak gain, very conservative approach. This is default value.
1322 select DC gain, limited application.
1325 select gain to noise approach, this is most popular one.
1329 Set gain to be applied to IR coefficients before filtering.
1330 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1333 Set format of IR stream. Can be @code{mono} or @code{input}.
1334 Default is @code{input}.
1337 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1338 Allowed range is 0.1 to 60 seconds.
1341 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1342 By default it is disabled.
1345 Set for which IR channel to display frequency response. By default is first channel
1346 displayed. This option is used only when @var{response} is enabled.
1349 Set video stream size. This option is used only when @var{response} is enabled.
1352 Set video stream frame rate. This option is used only when @var{response} is enabled.
1355 Set minimal partition size used for convolution. Default is @var{8192}.
1356 Allowed range is from @var{1} to @var{32768}.
1357 Lower values decreases latency at cost of higher CPU usage.
1360 Set maximal partition size used for convolution. Default is @var{8192}.
1361 Allowed range is from @var{8} to @var{32768}.
1362 Lower values may increase CPU usage.
1365 Set number of input impulse responses streams which will be switchable at runtime.
1366 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1369 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1370 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1371 This option can be changed at runtime via @ref{commands}.
1374 @subsection Examples
1378 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1380 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1387 Set output format constraints for the input audio. The framework will
1388 negotiate the most appropriate format to minimize conversions.
1390 It accepts the following parameters:
1393 @item sample_fmts, f
1394 A '|'-separated list of requested sample formats.
1396 @item sample_rates, r
1397 A '|'-separated list of requested sample rates.
1399 @item channel_layouts, cl
1400 A '|'-separated list of requested channel layouts.
1402 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1403 for the required syntax.
1406 If a parameter is omitted, all values are allowed.
1408 Force the output to either unsigned 8-bit or signed 16-bit stereo
1410 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1414 Apply frequency shift to input audio samples.
1416 The filter accepts the following options:
1420 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1421 Default value is 0.0.
1424 @subsection Commands
1426 This filter supports the above option as @ref{commands}.
1430 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1431 processing reduces disturbing noise between useful signals.
1433 Gating is done by detecting the volume below a chosen level @var{threshold}
1434 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1435 floor is set via @var{range}. Because an exact manipulation of the signal
1436 would cause distortion of the waveform the reduction can be levelled over
1437 time. This is done by setting @var{attack} and @var{release}.
1439 @var{attack} determines how long the signal has to fall below the threshold
1440 before any reduction will occur and @var{release} sets the time the signal
1441 has to rise above the threshold to reduce the reduction again.
1442 Shorter signals than the chosen attack time will be left untouched.
1446 Set input level before filtering.
1447 Default is 1. Allowed range is from 0.015625 to 64.
1450 Set the mode of operation. Can be @code{upward} or @code{downward}.
1451 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1452 will be amplified, expanding dynamic range in upward direction.
1453 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1456 Set the level of gain reduction when the signal is below the threshold.
1457 Default is 0.06125. Allowed range is from 0 to 1.
1458 Setting this to 0 disables reduction and then filter behaves like expander.
1461 If a signal rises above this level the gain reduction is released.
1462 Default is 0.125. Allowed range is from 0 to 1.
1465 Set a ratio by which the signal is reduced.
1466 Default is 2. Allowed range is from 1 to 9000.
1469 Amount of milliseconds the signal has to rise above the threshold before gain
1471 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1474 Amount of milliseconds the signal has to fall below the threshold before the
1475 reduction is increased again. Default is 250 milliseconds.
1476 Allowed range is from 0.01 to 9000.
1479 Set amount of amplification of signal after processing.
1480 Default is 1. Allowed range is from 1 to 64.
1483 Curve the sharp knee around the threshold to enter gain reduction more softly.
1484 Default is 2.828427125. Allowed range is from 1 to 8.
1487 Choose if exact signal should be taken for detection or an RMS like one.
1488 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1491 Choose if the average level between all channels or the louder channel affects
1493 Default is @code{average}. Can be @code{average} or @code{maximum}.
1496 @subsection Commands
1498 This filter supports the all above options as @ref{commands}.
1502 Apply an arbitrary Infinite Impulse Response filter.
1504 It accepts the following parameters:
1508 Set B/numerator/zeros/reflection coefficients.
1511 Set A/denominator/poles/ladder coefficients.
1523 Set coefficients format.
1527 lattice-ladder function
1529 analog transfer function
1531 digital transfer function
1533 Z-plane zeros/poles, cartesian (default)
1535 Z-plane zeros/poles, polar radians
1537 Z-plane zeros/poles, polar degrees
1543 Set type of processing.
1555 Set filtering precision.
1559 double-precision floating-point (default)
1561 single-precision floating-point
1569 Normalize filter coefficients, by default is enabled.
1570 Enabling it will normalize magnitude response at DC to 0dB.
1573 How much to use filtered signal in output. Default is 1.
1574 Range is between 0 and 1.
1577 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1578 By default it is disabled.
1581 Set for which IR channel to display frequency response. By default is first channel
1582 displayed. This option is used only when @var{response} is enabled.
1585 Set video stream size. This option is used only when @var{response} is enabled.
1588 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1591 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1592 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1595 Different coefficients and gains can be provided for every channel, in such case
1596 use '|' to separate coefficients or gains. Last provided coefficients will be
1597 used for all remaining channels.
1599 @subsection Examples
1603 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1605 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
1609 Same as above but in @code{zp} format:
1611 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
1615 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1617 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1623 The limiter prevents an input signal from rising over a desired threshold.
1624 This limiter uses lookahead technology to prevent your signal from distorting.
1625 It means that there is a small delay after the signal is processed. Keep in mind
1626 that the delay it produces is the attack time you set.
1628 The filter accepts the following options:
1632 Set input gain. Default is 1.
1635 Set output gain. Default is 1.
1638 Don't let signals above this level pass the limiter. Default is 1.
1641 The limiter will reach its attenuation level in this amount of time in
1642 milliseconds. Default is 5 milliseconds.
1645 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1646 Default is 50 milliseconds.
1649 When gain reduction is always needed ASC takes care of releasing to an
1650 average reduction level rather than reaching a reduction of 0 in the release
1654 Select how much the release time is affected by ASC, 0 means nearly no changes
1655 in release time while 1 produces higher release times.
1658 Auto level output signal. Default is enabled.
1659 This normalizes audio back to 0dB if enabled.
1662 Depending on picked setting it is recommended to upsample input 2x or 4x times
1663 with @ref{aresample} before applying this filter.
1667 Apply a two-pole all-pass filter with central frequency (in Hz)
1668 @var{frequency}, and filter-width @var{width}.
1669 An all-pass filter changes the audio's frequency to phase relationship
1670 without changing its frequency to amplitude relationship.
1672 The filter accepts the following options:
1676 Set frequency in Hz.
1679 Set method to specify band-width of filter.
1694 Specify the band-width of a filter in width_type units.
1697 How much to use filtered signal in output. Default is 1.
1698 Range is between 0 and 1.
1701 Specify which channels to filter, by default all available are filtered.
1704 Normalize biquad coefficients, by default is disabled.
1705 Enabling it will normalize magnitude response at DC to 0dB.
1708 Set the filter order, can be 1 or 2. Default is 2.
1711 Set transform type of IIR filter.
1720 @subsection Commands
1722 This filter supports the following commands:
1725 Change allpass frequency.
1726 Syntax for the command is : "@var{frequency}"
1729 Change allpass width_type.
1730 Syntax for the command is : "@var{width_type}"
1733 Change allpass width.
1734 Syntax for the command is : "@var{width}"
1738 Syntax for the command is : "@var{mix}"
1745 The filter accepts the following options:
1749 Set the number of loops. Setting this value to -1 will result in infinite loops.
1753 Set maximal number of samples. Default is 0.
1756 Set first sample of loop. Default is 0.
1762 Merge two or more audio streams into a single multi-channel stream.
1764 The filter accepts the following options:
1769 Set the number of inputs. Default is 2.
1773 If the channel layouts of the inputs are disjoint, and therefore compatible,
1774 the channel layout of the output will be set accordingly and the channels
1775 will be reordered as necessary. If the channel layouts of the inputs are not
1776 disjoint, the output will have all the channels of the first input then all
1777 the channels of the second input, in that order, and the channel layout of
1778 the output will be the default value corresponding to the total number of
1781 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1782 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1783 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1784 first input, b1 is the first channel of the second input).
1786 On the other hand, if both input are in stereo, the output channels will be
1787 in the default order: a1, a2, b1, b2, and the channel layout will be
1788 arbitrarily set to 4.0, which may or may not be the expected value.
1790 All inputs must have the same sample rate, and format.
1792 If inputs do not have the same duration, the output will stop with the
1795 @subsection Examples
1799 Merge two mono files into a stereo stream:
1801 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1805 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1807 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
1813 Mixes multiple audio inputs into a single output.
1815 Note that this filter only supports float samples (the @var{amerge}
1816 and @var{pan} audio filters support many formats). If the @var{amix}
1817 input has integer samples then @ref{aresample} will be automatically
1818 inserted to perform the conversion to float samples.
1822 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1824 will mix 3 input audio streams to a single output with the same duration as the
1825 first input and a dropout transition time of 3 seconds.
1827 It accepts the following parameters:
1831 The number of inputs. If unspecified, it defaults to 2.
1834 How to determine the end-of-stream.
1838 The duration of the longest input. (default)
1841 The duration of the shortest input.
1844 The duration of the first input.
1848 @item dropout_transition
1849 The transition time, in seconds, for volume renormalization when an input
1850 stream ends. The default value is 2 seconds.
1853 Specify weight of each input audio stream as sequence.
1854 Each weight is separated by space. By default all inputs have same weight.
1857 @subsection Commands
1859 This filter supports the following commands:
1862 Syntax is same as option with same name.
1867 Multiply first audio stream with second audio stream and store result
1868 in output audio stream. Multiplication is done by multiplying each
1869 sample from first stream with sample at same position from second stream.
1871 With this element-wise multiplication one can create amplitude fades and
1872 amplitude modulations.
1874 @section anequalizer
1876 High-order parametric multiband equalizer for each channel.
1878 It accepts the following parameters:
1882 This option string is in format:
1883 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1884 Each equalizer band is separated by '|'.
1888 Set channel number to which equalization will be applied.
1889 If input doesn't have that channel the entry is ignored.
1892 Set central frequency for band.
1893 If input doesn't have that frequency the entry is ignored.
1896 Set band width in Hertz.
1899 Set band gain in dB.
1902 Set filter type for band, optional, can be:
1906 Butterworth, this is default.
1917 With this option activated frequency response of anequalizer is displayed
1921 Set video stream size. Only useful if curves option is activated.
1924 Set max gain that will be displayed. Only useful if curves option is activated.
1925 Setting this to a reasonable value makes it possible to display gain which is derived from
1926 neighbour bands which are too close to each other and thus produce higher gain
1927 when both are activated.
1930 Set frequency scale used to draw frequency response in video output.
1931 Can be linear or logarithmic. Default is logarithmic.
1934 Set color for each channel curve which is going to be displayed in video stream.
1935 This is list of color names separated by space or by '|'.
1936 Unrecognised or missing colors will be replaced by white color.
1939 @subsection Examples
1943 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1944 for first 2 channels using Chebyshev type 1 filter:
1946 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1950 @subsection Commands
1952 This filter supports the following commands:
1955 Alter existing filter parameters.
1956 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1958 @var{fN} is existing filter number, starting from 0, if no such filter is available
1960 @var{freq} set new frequency parameter.
1961 @var{width} set new width parameter in Hertz.
1962 @var{gain} set new gain parameter in dB.
1964 Full filter invocation with asendcmd may look like this:
1965 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1970 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1972 Each sample is adjusted by looking for other samples with similar contexts. This
1973 context similarity is defined by comparing their surrounding patches of size
1974 @option{p}. Patches are searched in an area of @option{r} around the sample.
1976 The filter accepts the following options:
1980 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
1983 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
1984 Default value is 2 milliseconds.
1987 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
1988 Default value is 6 milliseconds.
1991 Set the output mode.
1993 It accepts the following values:
1996 Pass input unchanged.
1999 Pass noise filtered out.
2004 Default value is @var{o}.
2008 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2011 @subsection Commands
2013 This filter supports the all above options as @ref{commands}.
2016 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2018 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2019 relate to producing the least mean square of the error signal (difference between the desired,
2020 2nd input audio stream and the actual signal, the 1st input audio stream).
2022 A description of the accepted options follows.
2035 Set the filter leakage.
2038 It accepts the following values:
2047 Pass filtered samples.
2050 Pass difference between desired and filtered samples.
2052 Default value is @var{o}.
2056 @subsection Examples
2060 One of many usages of this filter is noise reduction, input audio is filtered
2061 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2063 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2067 @subsection Commands
2069 This filter supports the same commands as options, excluding option @code{order}.
2073 Pass the audio source unchanged to the output.
2077 Pad the end of an audio stream with silence.
2079 This can be used together with @command{ffmpeg} @option{-shortest} to
2080 extend audio streams to the same length as the video stream.
2082 A description of the accepted options follows.
2086 Set silence packet size. Default value is 4096.
2089 Set the number of samples of silence to add to the end. After the
2090 value is reached, the stream is terminated. This option is mutually
2091 exclusive with @option{whole_len}.
2094 Set the minimum total number of samples in the output audio stream. If
2095 the value is longer than the input audio length, silence is added to
2096 the end, until the value is reached. This option is mutually exclusive
2097 with @option{pad_len}.
2100 Specify the duration of samples of silence to add. See
2101 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2102 for the accepted syntax. Used only if set to non-zero value.
2105 Specify the minimum total duration in the output audio stream. See
2106 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2107 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2108 the input audio length, silence is added to the end, until the value is reached.
2109 This option is mutually exclusive with @option{pad_dur}
2112 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2113 nor @option{whole_dur} option is set, the filter will add silence to the end of
2114 the input stream indefinitely.
2116 @subsection Examples
2120 Add 1024 samples of silence to the end of the input:
2126 Make sure the audio output will contain at least 10000 samples, pad
2127 the input with silence if required:
2129 apad=whole_len=10000
2133 Use @command{ffmpeg} to pad the audio input with silence, so that the
2134 video stream will always result the shortest and will be converted
2135 until the end in the output file when using the @option{shortest}
2138 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2143 Add a phasing effect to the input audio.
2145 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2146 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2148 A description of the accepted parameters follows.
2152 Set input gain. Default is 0.4.
2155 Set output gain. Default is 0.74
2158 Set delay in milliseconds. Default is 3.0.
2161 Set decay. Default is 0.4.
2164 Set modulation speed in Hz. Default is 0.5.
2167 Set modulation type. Default is triangular.
2169 It accepts the following values:
2176 @section aphaseshift
2177 Apply phase shift to input audio samples.
2179 The filter accepts the following options:
2183 Specify phase shift. Allowed range is from -1.0 to 1.0.
2184 Default value is 0.0.
2187 @subsection Commands
2189 This filter supports the above option as @ref{commands}.
2193 Audio pulsator is something between an autopanner and a tremolo.
2194 But it can produce funny stereo effects as well. Pulsator changes the volume
2195 of the left and right channel based on a LFO (low frequency oscillator) with
2196 different waveforms and shifted phases.
2197 This filter have the ability to define an offset between left and right
2198 channel. An offset of 0 means that both LFO shapes match each other.
2199 The left and right channel are altered equally - a conventional tremolo.
2200 An offset of 50% means that the shape of the right channel is exactly shifted
2201 in phase (or moved backwards about half of the frequency) - pulsator acts as
2202 an autopanner. At 1 both curves match again. Every setting in between moves the
2203 phase shift gapless between all stages and produces some "bypassing" sounds with
2204 sine and triangle waveforms. The more you set the offset near 1 (starting from
2205 the 0.5) the faster the signal passes from the left to the right speaker.
2207 The filter accepts the following options:
2211 Set input gain. By default it is 1. Range is [0.015625 - 64].
2214 Set output gain. By default it is 1. Range is [0.015625 - 64].
2217 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2218 sawup or sawdown. Default is sine.
2221 Set modulation. Define how much of original signal is affected by the LFO.
2224 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2227 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2230 Set pulse width. Default is 1. Allowed range is [0 - 2].
2233 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2236 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2240 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2244 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2245 if timing is set to hz.
2251 Resample the input audio to the specified parameters, using the
2252 libswresample library. If none are specified then the filter will
2253 automatically convert between its input and output.
2255 This filter is also able to stretch/squeeze the audio data to make it match
2256 the timestamps or to inject silence / cut out audio to make it match the
2257 timestamps, do a combination of both or do neither.
2259 The filter accepts the syntax
2260 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2261 expresses a sample rate and @var{resampler_options} is a list of
2262 @var{key}=@var{value} pairs, separated by ":". See the
2263 @ref{Resampler Options,,"Resampler Options" section in the
2264 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2265 for the complete list of supported options.
2267 @subsection Examples
2271 Resample the input audio to 44100Hz:
2277 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2278 samples per second compensation:
2280 aresample=async=1000
2286 Reverse an audio clip.
2288 Warning: This filter requires memory to buffer the entire clip, so trimming
2291 @subsection Examples
2295 Take the first 5 seconds of a clip, and reverse it.
2297 atrim=end=5,areverse
2303 Reduce noise from speech using Recurrent Neural Networks.
2305 This filter accepts the following options:
2309 Set train model file to load. This option is always required.
2312 Set how much to mix filtered samples into final output.
2313 Allowed range is from -1 to 1. Default value is 1.
2314 Negative values are special, they set how much to keep filtered noise
2315 in the final filter output. Set this option to -1 to hear actual
2316 noise removed from input signal.
2319 @section asetnsamples
2321 Set the number of samples per each output audio frame.
2323 The last output packet may contain a different number of samples, as
2324 the filter will flush all the remaining samples when the input audio
2327 The filter accepts the following options:
2331 @item nb_out_samples, n
2332 Set the number of frames per each output audio frame. The number is
2333 intended as the number of samples @emph{per each channel}.
2334 Default value is 1024.
2337 If set to 1, the filter will pad the last audio frame with zeroes, so
2338 that the last frame will contain the same number of samples as the
2339 previous ones. Default value is 1.
2342 For example, to set the number of per-frame samples to 1234 and
2343 disable padding for the last frame, use:
2345 asetnsamples=n=1234:p=0
2350 Set the sample rate without altering the PCM data.
2351 This will result in a change of speed and pitch.
2353 The filter accepts the following options:
2356 @item sample_rate, r
2357 Set the output sample rate. Default is 44100 Hz.
2362 Show a line containing various information for each input audio frame.
2363 The input audio is not modified.
2365 The shown line contains a sequence of key/value pairs of the form
2366 @var{key}:@var{value}.
2368 The following values are shown in the output:
2372 The (sequential) number of the input frame, starting from 0.
2375 The presentation timestamp of the input frame, in time base units; the time base
2376 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2379 The presentation timestamp of the input frame in seconds.
2382 position of the frame in the input stream, -1 if this information in
2383 unavailable and/or meaningless (for example in case of synthetic audio)
2392 The sample rate for the audio frame.
2395 The number of samples (per channel) in the frame.
2398 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2399 audio, the data is treated as if all the planes were concatenated.
2401 @item plane_checksums
2402 A list of Adler-32 checksums for each data plane.
2406 Apply audio soft clipping.
2408 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2409 along a smooth curve, rather than the abrupt shape of hard-clipping.
2411 This filter accepts the following options:
2415 Set type of soft-clipping.
2417 It accepts the following values:
2431 Set additional parameter which controls sigmoid function.
2434 Set oversampling factor.
2437 @subsection Commands
2439 This filter supports the all above options as @ref{commands}.
2442 Automatic Speech Recognition
2444 This filter uses PocketSphinx for speech recognition. To enable
2445 compilation of this filter, you need to configure FFmpeg with
2446 @code{--enable-pocketsphinx}.
2448 It accepts the following options:
2452 Set sampling rate of input audio. Defaults is @code{16000}.
2453 This need to match speech models, otherwise one will get poor results.
2456 Set dictionary containing acoustic model files.
2459 Set pronunciation dictionary.
2462 Set language model file.
2465 Set language model set.
2468 Set which language model to use.
2471 Set output for log messages.
2474 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2479 Display time domain statistical information about the audio channels.
2480 Statistics are calculated and displayed for each audio channel and,
2481 where applicable, an overall figure is also given.
2483 It accepts the following option:
2486 Short window length in seconds, used for peak and trough RMS measurement.
2487 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2491 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2492 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2495 Available keys for each channel are:
2541 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2542 this @code{lavfi.astats.Overall.Peak_count}.
2544 For description what each key means read below.
2547 Set number of frame after which stats are going to be recalculated.
2548 Default is disabled.
2550 @item measure_perchannel
2551 Select the entries which need to be measured per channel. The metadata keys can
2552 be used as flags, default is @option{all} which measures everything.
2553 @option{none} disables all per channel measurement.
2555 @item measure_overall
2556 Select the entries which need to be measured overall. The metadata keys can
2557 be used as flags, default is @option{all} which measures everything.
2558 @option{none} disables all overall measurement.
2562 A description of each shown parameter follows:
2566 Mean amplitude displacement from zero.
2569 Minimal sample level.
2572 Maximal sample level.
2574 @item Min difference
2575 Minimal difference between two consecutive samples.
2577 @item Max difference
2578 Maximal difference between two consecutive samples.
2580 @item Mean difference
2581 Mean difference between two consecutive samples.
2582 The average of each difference between two consecutive samples.
2584 @item RMS difference
2585 Root Mean Square difference between two consecutive samples.
2589 Standard peak and RMS level measured in dBFS.
2593 Peak and trough values for RMS level measured over a short window.
2596 Standard ratio of peak to RMS level (note: not in dB).
2599 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2600 (i.e. either @var{Min level} or @var{Max level}).
2603 Number of occasions (not the number of samples) that the signal attained either
2604 @var{Min level} or @var{Max level}.
2606 @item Noise floor dB
2607 Minimum local peak measured in dBFS over a short window.
2609 @item Noise floor count
2610 Number of occasions (not the number of samples) that the signal attained
2614 Overall bit depth of audio. Number of bits used for each sample.
2617 Measured dynamic range of audio in dB.
2619 @item Zero crossings
2620 Number of points where the waveform crosses the zero level axis.
2622 @item Zero crossings rate
2623 Rate of Zero crossings and number of audio samples.
2627 Boost subwoofer frequencies.
2629 The filter accepts the following options:
2633 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2634 Default value is 0.7.
2637 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2638 Default value is 0.7.
2641 Set delay line decay gain value. Allowed range is from 0 to 1.
2642 Default value is 0.7.
2645 Set delay line feedback gain value. Allowed range is from 0 to 1.
2646 Default value is 0.9.
2649 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2650 Default value is 100.
2653 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2654 Default value is 0.5.
2657 Set delay. Allowed range is from 1 to 100.
2658 Default value is 20.
2661 @subsection Commands
2663 This filter supports the all above options as @ref{commands}.
2666 Cut super frequencies.
2668 The filter accepts the following options:
2672 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2673 Default value is 20000.
2676 Set filter order. Available values are from 3 to 20.
2677 Default value is 10.
2680 @subsection Commands
2682 This filter supports the all above options as @ref{commands}.
2688 The filter accepts exactly one parameter, the audio tempo. If not
2689 specified then the filter will assume nominal 1.0 tempo. Tempo must
2690 be in the [0.5, 100.0] range.
2692 Note that tempo greater than 2 will skip some samples rather than
2693 blend them in. If for any reason this is a concern it is always
2694 possible to daisy-chain several instances of atempo to achieve the
2695 desired product tempo.
2697 @subsection Examples
2701 Slow down audio to 80% tempo:
2707 To speed up audio to 300% tempo:
2713 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2715 atempo=sqrt(3),atempo=sqrt(3)
2719 @subsection Commands
2721 This filter supports the following commands:
2724 Change filter tempo scale factor.
2725 Syntax for the command is : "@var{tempo}"
2730 Trim the input so that the output contains one continuous subpart of the input.
2732 It accepts the following parameters:
2735 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2736 sample with the timestamp @var{start} will be the first sample in the output.
2739 Specify time of the first audio sample that will be dropped, i.e. the
2740 audio sample immediately preceding the one with the timestamp @var{end} will be
2741 the last sample in the output.
2744 Same as @var{start}, except this option sets the start timestamp in samples
2748 Same as @var{end}, except this option sets the end timestamp in samples instead
2752 The maximum duration of the output in seconds.
2755 The number of the first sample that should be output.
2758 The number of the first sample that should be dropped.
2761 @option{start}, @option{end}, and @option{duration} are expressed as time
2762 duration specifications; see
2763 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2765 Note that the first two sets of the start/end options and the @option{duration}
2766 option look at the frame timestamp, while the _sample options simply count the
2767 samples that pass through the filter. So start/end_pts and start/end_sample will
2768 give different results when the timestamps are wrong, inexact or do not start at
2769 zero. Also note that this filter does not modify the timestamps. If you wish
2770 to have the output timestamps start at zero, insert the asetpts filter after the
2773 If multiple start or end options are set, this filter tries to be greedy and
2774 keep all samples that match at least one of the specified constraints. To keep
2775 only the part that matches all the constraints at once, chain multiple atrim
2778 The defaults are such that all the input is kept. So it is possible to set e.g.
2779 just the end values to keep everything before the specified time.
2784 Drop everything except the second minute of input:
2786 ffmpeg -i INPUT -af atrim=60:120
2790 Keep only the first 1000 samples:
2792 ffmpeg -i INPUT -af atrim=end_sample=1000
2797 @section axcorrelate
2798 Calculate normalized cross-correlation between two input audio streams.
2800 Resulted samples are always between -1 and 1 inclusive.
2801 If result is 1 it means two input samples are highly correlated in that selected segment.
2802 Result 0 means they are not correlated at all.
2803 If result is -1 it means two input samples are out of phase, which means they cancel each
2806 The filter accepts the following options:
2810 Set size of segment over which cross-correlation is calculated.
2811 Default is 256. Allowed range is from 2 to 131072.
2814 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2815 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2816 are always zero and thus need much less calculations to make.
2817 This is generally not true, but is valid for typical audio streams.
2820 @subsection Examples
2824 Calculate correlation between channels in stereo audio stream:
2826 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2832 Apply a two-pole Butterworth band-pass filter with central
2833 frequency @var{frequency}, and (3dB-point) band-width width.
2834 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2835 instead of the default: constant 0dB peak gain.
2836 The filter roll off at 6dB per octave (20dB per decade).
2838 The filter accepts the following options:
2842 Set the filter's central frequency. Default is @code{3000}.
2845 Constant skirt gain if set to 1. Defaults to 0.
2848 Set method to specify band-width of filter.
2863 Specify the band-width of a filter in width_type units.
2866 How much to use filtered signal in output. Default is 1.
2867 Range is between 0 and 1.
2870 Specify which channels to filter, by default all available are filtered.
2873 Normalize biquad coefficients, by default is disabled.
2874 Enabling it will normalize magnitude response at DC to 0dB.
2877 Set transform type of IIR filter.
2886 @subsection Commands
2888 This filter supports the following commands:
2891 Change bandpass frequency.
2892 Syntax for the command is : "@var{frequency}"
2895 Change bandpass width_type.
2896 Syntax for the command is : "@var{width_type}"
2899 Change bandpass width.
2900 Syntax for the command is : "@var{width}"
2903 Change bandpass mix.
2904 Syntax for the command is : "@var{mix}"
2909 Apply a two-pole Butterworth band-reject filter with central
2910 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
2911 The filter roll off at 6dB per octave (20dB per decade).
2913 The filter accepts the following options:
2917 Set the filter's central frequency. Default is @code{3000}.
2920 Set method to specify band-width of filter.
2935 Specify the band-width of a filter in width_type units.
2938 How much to use filtered signal in output. Default is 1.
2939 Range is between 0 and 1.
2942 Specify which channels to filter, by default all available are filtered.
2945 Normalize biquad coefficients, by default is disabled.
2946 Enabling it will normalize magnitude response at DC to 0dB.
2949 Set transform type of IIR filter.
2958 @subsection Commands
2960 This filter supports the following commands:
2963 Change bandreject frequency.
2964 Syntax for the command is : "@var{frequency}"
2967 Change bandreject width_type.
2968 Syntax for the command is : "@var{width_type}"
2971 Change bandreject width.
2972 Syntax for the command is : "@var{width}"
2975 Change bandreject mix.
2976 Syntax for the command is : "@var{mix}"
2979 @section bass, lowshelf
2981 Boost or cut the bass (lower) frequencies of the audio using a two-pole
2982 shelving filter with a response similar to that of a standard
2983 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
2985 The filter accepts the following options:
2989 Give the gain at 0 Hz. Its useful range is about -20
2990 (for a large cut) to +20 (for a large boost).
2991 Beware of clipping when using a positive gain.
2994 Set the filter's central frequency and so can be used
2995 to extend or reduce the frequency range to be boosted or cut.
2996 The default value is @code{100} Hz.
2999 Set method to specify band-width of filter.
3014 Determine how steep is the filter's shelf transition.
3017 How much to use filtered signal in output. Default is 1.
3018 Range is between 0 and 1.
3021 Specify which channels to filter, by default all available are filtered.
3024 Normalize biquad coefficients, by default is disabled.
3025 Enabling it will normalize magnitude response at DC to 0dB.
3028 Set transform type of IIR filter.
3037 @subsection Commands
3039 This filter supports the following commands:
3042 Change bass frequency.
3043 Syntax for the command is : "@var{frequency}"
3046 Change bass width_type.
3047 Syntax for the command is : "@var{width_type}"
3051 Syntax for the command is : "@var{width}"
3055 Syntax for the command is : "@var{gain}"
3059 Syntax for the command is : "@var{mix}"
3064 Apply a biquad IIR filter with the given coefficients.
3065 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3066 are the numerator and denominator coefficients respectively.
3067 and @var{channels}, @var{c} specify which channels to filter, by default all
3068 available are filtered.
3070 @subsection Commands
3072 This filter supports the following commands:
3080 Change biquad parameter.
3081 Syntax for the command is : "@var{value}"
3084 How much to use filtered signal in output. Default is 1.
3085 Range is between 0 and 1.
3088 Specify which channels to filter, by default all available are filtered.
3091 Normalize biquad coefficients, by default is disabled.
3092 Enabling it will normalize magnitude response at DC to 0dB.
3095 Set transform type of IIR filter.
3105 Bauer stereo to binaural transformation, which improves headphone listening of
3106 stereo audio records.
3108 To enable compilation of this filter you need to configure FFmpeg with
3109 @code{--enable-libbs2b}.
3111 It accepts the following parameters:
3115 Pre-defined crossfeed level.
3119 Default level (fcut=700, feed=50).
3122 Chu Moy circuit (fcut=700, feed=60).
3125 Jan Meier circuit (fcut=650, feed=95).
3130 Cut frequency (in Hz).
3139 Remap input channels to new locations.
3141 It accepts the following parameters:
3144 Map channels from input to output. The argument is a '|'-separated list of
3145 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3146 @var{in_channel} form. @var{in_channel} can be either the name of the input
3147 channel (e.g. FL for front left) or its index in the input channel layout.
3148 @var{out_channel} is the name of the output channel or its index in the output
3149 channel layout. If @var{out_channel} is not given then it is implicitly an
3150 index, starting with zero and increasing by one for each mapping.
3152 @item channel_layout
3153 The channel layout of the output stream.
3156 If no mapping is present, the filter will implicitly map input channels to
3157 output channels, preserving indices.
3159 @subsection Examples
3163 For example, assuming a 5.1+downmix input MOV file,
3165 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3167 will create an output WAV file tagged as stereo from the downmix channels of
3171 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3173 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3177 @section channelsplit
3179 Split each channel from an input audio stream into a separate output stream.
3181 It accepts the following parameters:
3183 @item channel_layout
3184 The channel layout of the input stream. The default is "stereo".
3186 A channel layout describing the channels to be extracted as separate output streams
3187 or "all" to extract each input channel as a separate stream. The default is "all".
3189 Choosing channels not present in channel layout in the input will result in an error.
3192 @subsection Examples
3196 For example, assuming a stereo input MP3 file,
3198 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3200 will create an output Matroska file with two audio streams, one containing only
3201 the left channel and the other the right channel.
3204 Split a 5.1 WAV file into per-channel files:
3206 ffmpeg -i in.wav -filter_complex
3207 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3208 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3209 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3214 Extract only LFE from a 5.1 WAV file:
3216 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3217 -map '[LFE]' lfe.wav
3222 Add a chorus effect to the audio.
3224 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3226 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3227 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3228 The modulation depth defines the range the modulated delay is played before or after
3229 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3230 sound tuned around the original one, like in a chorus where some vocals are slightly
3233 It accepts the following parameters:
3236 Set input gain. Default is 0.4.
3239 Set output gain. Default is 0.4.
3242 Set delays. A typical delay is around 40ms to 60ms.
3254 @subsection Examples
3260 chorus=0.7:0.9:55:0.4:0.25:2
3266 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3270 Fuller sounding chorus with three delays:
3272 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
3277 Compress or expand the audio's dynamic range.
3279 It accepts the following parameters:
3285 A list of times in seconds for each channel over which the instantaneous level
3286 of the input signal is averaged to determine its volume. @var{attacks} refers to
3287 increase of volume and @var{decays} refers to decrease of volume. For most
3288 situations, the attack time (response to the audio getting louder) should be
3289 shorter than the decay time, because the human ear is more sensitive to sudden
3290 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3291 a typical value for decay is 0.8 seconds.
3292 If specified number of attacks & decays is lower than number of channels, the last
3293 set attack/decay will be used for all remaining channels.
3296 A list of points for the transfer function, specified in dB relative to the
3297 maximum possible signal amplitude. Each key points list must be defined using
3298 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3299 @code{x0/y0 x1/y1 x2/y2 ....}
3301 The input values must be in strictly increasing order but the transfer function
3302 does not have to be monotonically rising. The point @code{0/0} is assumed but
3303 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3304 function are @code{-70/-70|-60/-20|1/0}.
3307 Set the curve radius in dB for all joints. It defaults to 0.01.
3310 Set the additional gain in dB to be applied at all points on the transfer
3311 function. This allows for easy adjustment of the overall gain.
3315 Set an initial volume, in dB, to be assumed for each channel when filtering
3316 starts. This permits the user to supply a nominal level initially, so that, for
3317 example, a very large gain is not applied to initial signal levels before the
3318 companding has begun to operate. A typical value for audio which is initially
3319 quiet is -90 dB. It defaults to 0.
3322 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3323 delayed before being fed to the volume adjuster. Specifying a delay
3324 approximately equal to the attack/decay times allows the filter to effectively
3325 operate in predictive rather than reactive mode. It defaults to 0.
3329 @subsection Examples
3333 Make music with both quiet and loud passages suitable for listening to in a
3336 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3339 Another example for audio with whisper and explosion parts:
3341 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3345 A noise gate for when the noise is at a lower level than the signal:
3347 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3351 Here is another noise gate, this time for when the noise is at a higher level
3352 than the signal (making it, in some ways, similar to squelch):
3354 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3358 2:1 compression starting at -6dB:
3360 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3364 2:1 compression starting at -9dB:
3366 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3370 2:1 compression starting at -12dB:
3372 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3376 2:1 compression starting at -18dB:
3378 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3382 3:1 compression starting at -15dB:
3384 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3390 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3396 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
3400 Hard limiter at -6dB:
3402 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3406 Hard limiter at -12dB:
3408 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3412 Hard noise gate at -35 dB:
3414 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3420 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3424 @section compensationdelay
3426 Compensation Delay Line is a metric based delay to compensate differing
3427 positions of microphones or speakers.
3429 For example, you have recorded guitar with two microphones placed in
3430 different locations. Because the front of sound wave has fixed speed in
3431 normal conditions, the phasing of microphones can vary and depends on
3432 their location and interposition. The best sound mix can be achieved when
3433 these microphones are in phase (synchronized). Note that a distance of
3434 ~30 cm between microphones makes one microphone capture the signal in
3435 antiphase to the other microphone. That makes the final mix sound moody.
3436 This filter helps to solve phasing problems by adding different delays
3437 to each microphone track and make them synchronized.
3439 The best result can be reached when you take one track as base and
3440 synchronize other tracks one by one with it.
3441 Remember that synchronization/delay tolerance depends on sample rate, too.
3442 Higher sample rates will give more tolerance.
3444 The filter accepts the following parameters:
3448 Set millimeters distance. This is compensation distance for fine tuning.
3452 Set cm distance. This is compensation distance for tightening distance setup.
3456 Set meters distance. This is compensation distance for hard distance setup.
3460 Set dry amount. Amount of unprocessed (dry) signal.
3464 Set wet amount. Amount of processed (wet) signal.
3468 Set temperature in degrees Celsius. This is the temperature of the environment.
3473 Apply headphone crossfeed filter.
3475 Crossfeed is the process of blending the left and right channels of stereo
3477 It is mainly used to reduce extreme stereo separation of low frequencies.
3479 The intent is to produce more speaker like sound to the listener.
3481 The filter accepts the following options:
3485 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3486 This sets gain of low shelf filter for side part of stereo image.
3487 Default is -6dB. Max allowed is -30db when strength is set to 1.
3490 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3491 This sets cut off frequency of low shelf filter. Default is cut off near
3492 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3495 Set curve slope of low shelf filter. Default is 0.5.
3496 Allowed range is from 0.01 to 1.
3499 Set input gain. Default is 0.9.
3502 Set output gain. Default is 1.
3505 @subsection Commands
3507 This filter supports the all above options as @ref{commands}.
3509 @section crystalizer
3510 Simple algorithm to expand audio dynamic range.
3512 The filter accepts the following options:
3516 Sets the intensity of effect (default: 2.0). Must be in range between 0.0
3517 (unchanged sound) to 10.0 (maximum effect).
3520 Enable clipping. By default is enabled.
3523 @subsection Commands
3525 This filter supports the all above options as @ref{commands}.
3528 Apply a DC shift to the audio.
3530 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3531 in the recording chain) from the audio. The effect of a DC offset is reduced
3532 headroom and hence volume. The @ref{astats} filter can be used to determine if
3533 a signal has a DC offset.
3537 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3541 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3542 used to prevent clipping.
3547 Apply de-essing to the audio samples.
3551 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3555 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3559 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3563 Set the output mode.
3565 It accepts the following values:
3568 Pass input unchanged.
3571 Pass ess filtered out.
3576 Default value is @var{o}.
3582 Measure audio dynamic range.
3584 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3585 is found in transition material. And anything less that 8 have very poor dynamics
3586 and is very compressed.
3588 The filter accepts the following options:
3592 Set window length in seconds used to split audio into segments of equal length.
3593 Default is 3 seconds.
3597 Dynamic Audio Normalizer.
3599 This filter applies a certain amount of gain to the input audio in order
3600 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3601 contrast to more "simple" normalization algorithms, the Dynamic Audio
3602 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3603 This allows for applying extra gain to the "quiet" sections of the audio
3604 while avoiding distortions or clipping the "loud" sections. In other words:
3605 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3606 sections, in the sense that the volume of each section is brought to the
3607 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3608 this goal *without* applying "dynamic range compressing". It will retain 100%
3609 of the dynamic range *within* each section of the audio file.
3613 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3614 Default is 500 milliseconds.
3615 The Dynamic Audio Normalizer processes the input audio in small chunks,
3616 referred to as frames. This is required, because a peak magnitude has no
3617 meaning for just a single sample value. Instead, we need to determine the
3618 peak magnitude for a contiguous sequence of sample values. While a "standard"
3619 normalizer would simply use the peak magnitude of the complete file, the
3620 Dynamic Audio Normalizer determines the peak magnitude individually for each
3621 frame. The length of a frame is specified in milliseconds. By default, the
3622 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3623 been found to give good results with most files.
3624 Note that the exact frame length, in number of samples, will be determined
3625 automatically, based on the sampling rate of the individual input audio file.
3628 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3629 number. Default is 31.
3630 Probably the most important parameter of the Dynamic Audio Normalizer is the
3631 @code{window size} of the Gaussian smoothing filter. The filter's window size
3632 is specified in frames, centered around the current frame. For the sake of
3633 simplicity, this must be an odd number. Consequently, the default value of 31
3634 takes into account the current frame, as well as the 15 preceding frames and
3635 the 15 subsequent frames. Using a larger window results in a stronger
3636 smoothing effect and thus in less gain variation, i.e. slower gain
3637 adaptation. Conversely, using a smaller window results in a weaker smoothing
3638 effect and thus in more gain variation, i.e. faster gain adaptation.
3639 In other words, the more you increase this value, the more the Dynamic Audio
3640 Normalizer will behave like a "traditional" normalization filter. On the
3641 contrary, the more you decrease this value, the more the Dynamic Audio
3642 Normalizer will behave like a dynamic range compressor.
3645 Set the target peak value. This specifies the highest permissible magnitude
3646 level for the normalized audio input. This filter will try to approach the
3647 target peak magnitude as closely as possible, but at the same time it also
3648 makes sure that the normalized signal will never exceed the peak magnitude.
3649 A frame's maximum local gain factor is imposed directly by the target peak
3650 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3651 It is not recommended to go above this value.
3654 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3655 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3656 factor for each input frame, i.e. the maximum gain factor that does not
3657 result in clipping or distortion. The maximum gain factor is determined by
3658 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3659 additionally bounds the frame's maximum gain factor by a predetermined
3660 (global) maximum gain factor. This is done in order to avoid excessive gain
3661 factors in "silent" or almost silent frames. By default, the maximum gain
3662 factor is 10.0, For most inputs the default value should be sufficient and
3663 it usually is not recommended to increase this value. Though, for input
3664 with an extremely low overall volume level, it may be necessary to allow even
3665 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3666 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3667 Instead, a "sigmoid" threshold function will be applied. This way, the
3668 gain factors will smoothly approach the threshold value, but never exceed that
3672 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3673 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3674 This means that the maximum local gain factor for each frame is defined
3675 (only) by the frame's highest magnitude sample. This way, the samples can
3676 be amplified as much as possible without exceeding the maximum signal
3677 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3678 Normalizer can also take into account the frame's root mean square,
3679 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3680 determine the power of a time-varying signal. It is therefore considered
3681 that the RMS is a better approximation of the "perceived loudness" than
3682 just looking at the signal's peak magnitude. Consequently, by adjusting all
3683 frames to a constant RMS value, a uniform "perceived loudness" can be
3684 established. If a target RMS value has been specified, a frame's local gain
3685 factor is defined as the factor that would result in exactly that RMS value.
3686 Note, however, that the maximum local gain factor is still restricted by the
3687 frame's highest magnitude sample, in order to prevent clipping.
3690 Enable channels coupling. By default is enabled.
3691 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3692 amount. This means the same gain factor will be applied to all channels, i.e.
3693 the maximum possible gain factor is determined by the "loudest" channel.
3694 However, in some recordings, it may happen that the volume of the different
3695 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3696 In this case, this option can be used to disable the channel coupling. This way,
3697 the gain factor will be determined independently for each channel, depending
3698 only on the individual channel's highest magnitude sample. This allows for
3699 harmonizing the volume of the different channels.
3702 Enable DC bias correction. By default is disabled.
3703 An audio signal (in the time domain) is a sequence of sample values.
3704 In the Dynamic Audio Normalizer these sample values are represented in the
3705 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3706 audio signal, or "waveform", should be centered around the zero point.
3707 That means if we calculate the mean value of all samples in a file, or in a
3708 single frame, then the result should be 0.0 or at least very close to that
3709 value. If, however, there is a significant deviation of the mean value from
3710 0.0, in either positive or negative direction, this is referred to as a
3711 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3712 Audio Normalizer provides optional DC bias correction.
3713 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3714 the mean value, or "DC correction" offset, of each input frame and subtract
3715 that value from all of the frame's sample values which ensures those samples
3716 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3717 boundaries, the DC correction offset values will be interpolated smoothly
3718 between neighbouring frames.
3720 @item altboundary, b
3721 Enable alternative boundary mode. By default is disabled.
3722 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3723 around each frame. This includes the preceding frames as well as the
3724 subsequent frames. However, for the "boundary" frames, located at the very
3725 beginning and at the very end of the audio file, not all neighbouring
3726 frames are available. In particular, for the first few frames in the audio
3727 file, the preceding frames are not known. And, similarly, for the last few
3728 frames in the audio file, the subsequent frames are not known. Thus, the
3729 question arises which gain factors should be assumed for the missing frames
3730 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3731 to deal with this situation. The default boundary mode assumes a gain factor
3732 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3733 "fade out" at the beginning and at the end of the input, respectively.
3736 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3737 By default, the Dynamic Audio Normalizer does not apply "traditional"
3738 compression. This means that signal peaks will not be pruned and thus the
3739 full dynamic range will be retained within each local neighbourhood. However,
3740 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3741 normalization algorithm with a more "traditional" compression.
3742 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3743 (thresholding) function. If (and only if) the compression feature is enabled,
3744 all input frames will be processed by a soft knee thresholding function prior
3745 to the actual normalization process. Put simply, the thresholding function is
3746 going to prune all samples whose magnitude exceeds a certain threshold value.
3747 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3748 value. Instead, the threshold value will be adjusted for each individual
3750 In general, smaller parameters result in stronger compression, and vice versa.
3751 Values below 3.0 are not recommended, because audible distortion may appear.
3754 Set the target threshold value. This specifies the lowest permissible
3755 magnitude level for the audio input which will be normalized.
3756 If input frame volume is above this value frame will be normalized.
3757 Otherwise frame may not be normalized at all. The default value is set
3758 to 0, which means all input frames will be normalized.
3759 This option is mostly useful if digital noise is not wanted to be amplified.
3762 @subsection Commands
3764 This filter supports the all above options as @ref{commands}.
3768 Make audio easier to listen to on headphones.
3770 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3771 so that when listened to on headphones the stereo image is moved from
3772 inside your head (standard for headphones) to outside and in front of
3773 the listener (standard for speakers).
3779 Apply a two-pole peaking equalisation (EQ) filter. With this
3780 filter, the signal-level at and around a selected frequency can
3781 be increased or decreased, whilst (unlike bandpass and bandreject
3782 filters) that at all other frequencies is unchanged.
3784 In order to produce complex equalisation curves, this filter can
3785 be given several times, each with a different central frequency.
3787 The filter accepts the following options:
3791 Set the filter's central frequency in Hz.
3794 Set method to specify band-width of filter.
3809 Specify the band-width of a filter in width_type units.
3812 Set the required gain or attenuation in dB.
3813 Beware of clipping when using a positive gain.
3816 How much to use filtered signal in output. Default is 1.
3817 Range is between 0 and 1.
3820 Specify which channels to filter, by default all available are filtered.
3823 Normalize biquad coefficients, by default is disabled.
3824 Enabling it will normalize magnitude response at DC to 0dB.
3827 Set transform type of IIR filter.
3836 @subsection Examples
3839 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
3841 equalizer=f=1000:t=h:width=200:g=-10
3845 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
3847 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
3851 @subsection Commands
3853 This filter supports the following commands:
3856 Change equalizer frequency.
3857 Syntax for the command is : "@var{frequency}"
3860 Change equalizer width_type.
3861 Syntax for the command is : "@var{width_type}"
3864 Change equalizer width.
3865 Syntax for the command is : "@var{width}"
3868 Change equalizer gain.
3869 Syntax for the command is : "@var{gain}"
3872 Change equalizer mix.
3873 Syntax for the command is : "@var{mix}"
3876 @section extrastereo
3878 Linearly increases the difference between left and right channels which
3879 adds some sort of "live" effect to playback.
3881 The filter accepts the following options:
3885 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
3886 (average of both channels), with 1.0 sound will be unchanged, with
3887 -1.0 left and right channels will be swapped.
3890 Enable clipping. By default is enabled.
3893 @subsection Commands
3895 This filter supports the all above options as @ref{commands}.
3897 @section firequalizer
3898 Apply FIR Equalization using arbitrary frequency response.
3900 The filter accepts the following option:
3904 Set gain curve equation (in dB). The expression can contain variables:
3907 the evaluated frequency
3911 channel number, set to 0 when multichannels evaluation is disabled
3913 channel id, see libavutil/channel_layout.h, set to the first channel id when
3914 multichannels evaluation is disabled
3918 channel_layout, see libavutil/channel_layout.h
3923 @item gain_interpolate(f)
3924 interpolate gain on frequency f based on gain_entry
3925 @item cubic_interpolate(f)
3926 same as gain_interpolate, but smoother
3928 This option is also available as command. Default is @code{gain_interpolate(f)}.
3931 Set gain entry for gain_interpolate function. The expression can
3935 store gain entry at frequency f with value g
3937 This option is also available as command.
3940 Set filter delay in seconds. Higher value means more accurate.
3941 Default is @code{0.01}.
3944 Set filter accuracy in Hz. Lower value means more accurate.
3945 Default is @code{5}.
3948 Set window function. Acceptable values are:
3951 rectangular window, useful when gain curve is already smooth
3953 hann window (default)
3959 3-terms continuous 1st derivative nuttall window
3961 minimum 3-terms discontinuous nuttall window
3963 4-terms continuous 1st derivative nuttall window
3965 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
3967 blackman-harris window
3973 If enabled, use fixed number of audio samples. This improves speed when
3974 filtering with large delay. Default is disabled.
3977 Enable multichannels evaluation on gain. Default is disabled.
3980 Enable zero phase mode by subtracting timestamp to compensate delay.
3981 Default is disabled.
3984 Set scale used by gain. Acceptable values are:
3987 linear frequency, linear gain
3989 linear frequency, logarithmic (in dB) gain (default)
3991 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
3993 logarithmic frequency, logarithmic gain
3997 Set file for dumping, suitable for gnuplot.
4000 Set scale for dumpfile. Acceptable values are same with scale option.
4004 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4005 Default is disabled.
4008 Enable minimum phase impulse response. Default is disabled.
4011 @subsection Examples
4016 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4019 lowpass at 1000 Hz with gain_entry:
4021 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4024 custom equalization:
4026 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4029 higher delay with zero phase to compensate delay:
4031 firequalizer=delay=0.1:fixed=on:zero_phase=on
4034 lowpass on left channel, highpass on right channel:
4036 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4037 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4042 Apply a flanging effect to the audio.
4044 The filter accepts the following options:
4048 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4051 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4054 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4058 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4059 Default value is 71.
4062 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4065 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4066 Default value is @var{sinusoidal}.
4069 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4070 Default value is 25.
4073 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4074 Default is @var{linear}.
4078 Apply Haas effect to audio.
4080 Note that this makes most sense to apply on mono signals.
4081 With this filter applied to mono signals it give some directionality and
4082 stretches its stereo image.
4084 The filter accepts the following options:
4088 Set input level. By default is @var{1}, or 0dB
4091 Set output level. By default is @var{1}, or 0dB.
4094 Set gain applied to side part of signal. By default is @var{1}.
4097 Set kind of middle source. Can be one of the following:
4107 Pick middle part signal of stereo image.
4110 Pick side part signal of stereo image.
4114 Change middle phase. By default is disabled.
4117 Set left channel delay. By default is @var{2.05} milliseconds.
4120 Set left channel balance. By default is @var{-1}.
4123 Set left channel gain. By default is @var{1}.
4126 Change left phase. By default is disabled.
4129 Set right channel delay. By defaults is @var{2.12} milliseconds.
4132 Set right channel balance. By default is @var{1}.
4135 Set right channel gain. By default is @var{1}.
4138 Change right phase. By default is enabled.
4143 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4144 embedded HDCD codes is expanded into a 20-bit PCM stream.
4146 The filter supports the Peak Extend and Low-level Gain Adjustment features
4147 of HDCD, and detects the Transient Filter flag.
4150 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4153 When using the filter with wav, note the default encoding for wav is 16-bit,
4154 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4155 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4157 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4158 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4161 The filter accepts the following options:
4164 @item disable_autoconvert
4165 Disable any automatic format conversion or resampling in the filter graph.
4167 @item process_stereo
4168 Process the stereo channels together. If target_gain does not match between
4169 channels, consider it invalid and use the last valid target_gain.
4172 Set the code detect timer period in ms.
4175 Always extend peaks above -3dBFS even if PE isn't signaled.
4178 Replace audio with a solid tone and adjust the amplitude to signal some
4179 specific aspect of the decoding process. The output file can be loaded in
4180 an audio editor alongside the original to aid analysis.
4182 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4189 Gain adjustment level at each sample
4191 Samples where peak extend occurs
4193 Samples where the code detect timer is active
4195 Samples where the target gain does not match between channels
4201 Apply head-related transfer functions (HRTFs) to create virtual
4202 loudspeakers around the user for binaural listening via headphones.
4203 The HRIRs are provided via additional streams, for each channel
4204 one stereo input stream is needed.
4206 The filter accepts the following options:
4210 Set mapping of input streams for convolution.
4211 The argument is a '|'-separated list of channel names in order as they
4212 are given as additional stream inputs for filter.
4213 This also specify number of input streams. Number of input streams
4214 must be not less than number of channels in first stream plus one.
4217 Set gain applied to audio. Value is in dB. Default is 0.
4220 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4221 processing audio in time domain which is slow.
4222 @var{freq} is processing audio in frequency domain which is fast.
4223 Default is @var{freq}.
4226 Set custom gain for LFE channels. Value is in dB. Default is 0.
4229 Set size of frame in number of samples which will be processed at once.
4230 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4233 Set format of hrir stream.
4234 Default value is @var{stereo}. Alternative value is @var{multich}.
4235 If value is set to @var{stereo}, number of additional streams should
4236 be greater or equal to number of input channels in first input stream.
4237 Also each additional stream should have stereo number of channels.
4238 If value is set to @var{multich}, number of additional streams should
4239 be exactly one. Also number of input channels of additional stream
4240 should be equal or greater than twice number of channels of first input
4244 @subsection Examples
4248 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4249 each amovie filter use stereo file with IR coefficients as input.
4250 The files give coefficients for each position of virtual loudspeaker:
4253 -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"
4258 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4259 but now in @var{multich} @var{hrir} format.
4261 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"
4268 Apply a high-pass filter with 3dB point frequency.
4269 The filter can be either single-pole, or double-pole (the default).
4270 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4272 The filter accepts the following options:
4276 Set frequency in Hz. Default is 3000.
4279 Set number of poles. Default is 2.
4282 Set method to specify band-width of filter.
4297 Specify the band-width of a filter in width_type units.
4298 Applies only to double-pole filter.
4299 The default is 0.707q and gives a Butterworth response.
4302 How much to use filtered signal in output. Default is 1.
4303 Range is between 0 and 1.
4306 Specify which channels to filter, by default all available are filtered.
4309 Normalize biquad coefficients, by default is disabled.
4310 Enabling it will normalize magnitude response at DC to 0dB.
4313 Set transform type of IIR filter.
4322 @subsection Commands
4324 This filter supports the following commands:
4327 Change highpass frequency.
4328 Syntax for the command is : "@var{frequency}"
4331 Change highpass width_type.
4332 Syntax for the command is : "@var{width_type}"
4335 Change highpass width.
4336 Syntax for the command is : "@var{width}"
4339 Change highpass mix.
4340 Syntax for the command is : "@var{mix}"
4345 Join multiple input streams into one multi-channel stream.
4347 It accepts the following parameters:
4351 The number of input streams. It defaults to 2.
4353 @item channel_layout
4354 The desired output channel layout. It defaults to stereo.
4357 Map channels from inputs to output. The argument is a '|'-separated list of
4358 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4359 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4360 can be either the name of the input channel (e.g. FL for front left) or its
4361 index in the specified input stream. @var{out_channel} is the name of the output
4365 The filter will attempt to guess the mappings when they are not specified
4366 explicitly. It does so by first trying to find an unused matching input channel
4367 and if that fails it picks the first unused input channel.
4369 Join 3 inputs (with properly set channel layouts):
4371 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4374 Build a 5.1 output from 6 single-channel streams:
4376 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4377 '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'
4383 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4385 To enable compilation of this filter you need to configure FFmpeg with
4386 @code{--enable-ladspa}.
4390 Specifies the name of LADSPA plugin library to load. If the environment
4391 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4392 each one of the directories specified by the colon separated list in
4393 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4394 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4395 @file{/usr/lib/ladspa/}.
4398 Specifies the plugin within the library. Some libraries contain only
4399 one plugin, but others contain many of them. If this is not set filter
4400 will list all available plugins within the specified library.
4403 Set the '|' separated list of controls which are zero or more floating point
4404 values that determine the behavior of the loaded plugin (for example delay,
4406 Controls need to be defined using the following syntax:
4407 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4408 @var{valuei} is the value set on the @var{i}-th control.
4409 Alternatively they can be also defined using the following syntax:
4410 @var{value0}|@var{value1}|@var{value2}|..., where
4411 @var{valuei} is the value set on the @var{i}-th control.
4412 If @option{controls} is set to @code{help}, all available controls and
4413 their valid ranges are printed.
4415 @item sample_rate, s
4416 Specify the sample rate, default to 44100. Only used if plugin have
4420 Set the number of samples per channel per each output frame, default
4421 is 1024. Only used if plugin have zero inputs.
4424 Set the minimum duration of the sourced audio. See
4425 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4426 for the accepted syntax.
4427 Note that the resulting duration may be greater than the specified duration,
4428 as the generated audio is always cut at the end of a complete frame.
4429 If not specified, or the expressed duration is negative, the audio is
4430 supposed to be generated forever.
4431 Only used if plugin have zero inputs.
4434 Enable latency compensation, by default is disabled.
4435 Only used if plugin have inputs.
4438 @subsection Examples
4442 List all available plugins within amp (LADSPA example plugin) library:
4448 List all available controls and their valid ranges for @code{vcf_notch}
4449 plugin from @code{VCF} library:
4451 ladspa=f=vcf:p=vcf_notch:c=help
4455 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4458 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4462 Add reverberation to the audio using TAP-plugins
4463 (Tom's Audio Processing plugins):
4465 ladspa=file=tap_reverb:tap_reverb
4469 Generate white noise, with 0.2 amplitude:
4471 ladspa=file=cmt:noise_source_white:c=c0=.2
4475 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4476 @code{C* Audio Plugin Suite} (CAPS) library:
4478 ladspa=file=caps:Click:c=c1=20'
4482 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4484 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4488 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4489 @code{SWH Plugins} collection:
4491 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4495 Attenuate low frequencies using Multiband EQ from Steve Harris
4496 @code{SWH Plugins} collection:
4498 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4502 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4505 ladspa=caps:Narrower
4509 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4511 ladspa=caps:White:.2
4515 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4517 ladspa=caps:Fractal:c=c1=1
4521 Dynamic volume normalization using @code{VLevel} plugin:
4523 ladspa=vlevel-ladspa:vlevel_mono
4527 @subsection Commands
4529 This filter supports the following commands:
4532 Modify the @var{N}-th control value.
4534 If the specified value is not valid, it is ignored and prior one is kept.
4539 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4540 Support for both single pass (livestreams, files) and double pass (files) modes.
4541 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4542 detect true peaks, the audio stream will be upsampled to 192 kHz.
4543 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4545 The filter accepts the following options:
4549 Set integrated loudness target.
4550 Range is -70.0 - -5.0. Default value is -24.0.
4553 Set loudness range target.
4554 Range is 1.0 - 20.0. Default value is 7.0.
4557 Set maximum true peak.
4558 Range is -9.0 - +0.0. Default value is -2.0.
4560 @item measured_I, measured_i
4561 Measured IL of input file.
4562 Range is -99.0 - +0.0.
4564 @item measured_LRA, measured_lra
4565 Measured LRA of input file.
4566 Range is 0.0 - 99.0.
4568 @item measured_TP, measured_tp
4569 Measured true peak of input file.
4570 Range is -99.0 - +99.0.
4572 @item measured_thresh
4573 Measured threshold of input file.
4574 Range is -99.0 - +0.0.
4577 Set offset gain. Gain is applied before the true-peak limiter.
4578 Range is -99.0 - +99.0. Default is +0.0.
4581 Normalize by linearly scaling the source audio.
4582 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4583 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4584 be lower than source LRA and the change in integrated loudness shouldn't
4585 result in a true peak which exceeds the target TP. If any of these
4586 conditions aren't met, normalization mode will revert to @var{dynamic}.
4587 Options are @code{true} or @code{false}. Default is @code{true}.
4590 Treat mono input files as "dual-mono". If a mono file is intended for playback
4591 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4592 If set to @code{true}, this option will compensate for this effect.
4593 Multi-channel input files are not affected by this option.
4594 Options are true or false. Default is false.
4597 Set print format for stats. Options are summary, json, or none.
4598 Default value is none.
4603 Apply a low-pass filter with 3dB point frequency.
4604 The filter can be either single-pole or double-pole (the default).
4605 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4607 The filter accepts the following options:
4611 Set frequency in Hz. Default is 500.
4614 Set number of poles. Default is 2.
4617 Set method to specify band-width of filter.
4632 Specify the band-width of a filter in width_type units.
4633 Applies only to double-pole filter.
4634 The default is 0.707q and gives a Butterworth response.
4637 How much to use filtered signal in output. Default is 1.
4638 Range is between 0 and 1.
4641 Specify which channels to filter, by default all available are filtered.
4644 Normalize biquad coefficients, by default is disabled.
4645 Enabling it will normalize magnitude response at DC to 0dB.
4648 Set transform type of IIR filter.
4657 @subsection Examples
4660 Lowpass only LFE channel, it LFE is not present it does nothing:
4666 @subsection Commands
4668 This filter supports the following commands:
4671 Change lowpass frequency.
4672 Syntax for the command is : "@var{frequency}"
4675 Change lowpass width_type.
4676 Syntax for the command is : "@var{width_type}"
4679 Change lowpass width.
4680 Syntax for the command is : "@var{width}"
4684 Syntax for the command is : "@var{mix}"
4689 Load a LV2 (LADSPA Version 2) plugin.
4691 To enable compilation of this filter you need to configure FFmpeg with
4692 @code{--enable-lv2}.
4696 Specifies the plugin URI. You may need to escape ':'.
4699 Set the '|' separated list of controls which are zero or more floating point
4700 values that determine the behavior of the loaded plugin (for example delay,
4702 If @option{controls} is set to @code{help}, all available controls and
4703 their valid ranges are printed.
4705 @item sample_rate, s
4706 Specify the sample rate, default to 44100. Only used if plugin have
4710 Set the number of samples per channel per each output frame, default
4711 is 1024. Only used if plugin have zero inputs.
4714 Set the minimum duration of the sourced audio. See
4715 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4716 for the accepted syntax.
4717 Note that the resulting duration may be greater than the specified duration,
4718 as the generated audio is always cut at the end of a complete frame.
4719 If not specified, or the expressed duration is negative, the audio is
4720 supposed to be generated forever.
4721 Only used if plugin have zero inputs.
4724 @subsection Examples
4728 Apply bass enhancer plugin from Calf:
4730 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4734 Apply vinyl plugin from Calf:
4736 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4740 Apply bit crusher plugin from ArtyFX:
4742 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4747 Multiband Compress or expand the audio's dynamic range.
4749 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4750 This is akin to the crossover of a loudspeaker, and results in flat frequency
4751 response when absent compander action.
4753 It accepts the following parameters:
4757 This option syntax is:
4758 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4759 For explanation of each item refer to compand filter documentation.
4765 Mix channels with specific gain levels. The filter accepts the output
4766 channel layout followed by a set of channels definitions.
4768 This filter is also designed to efficiently remap the channels of an audio
4771 The filter accepts parameters of the form:
4772 "@var{l}|@var{outdef}|@var{outdef}|..."
4776 output channel layout or number of channels
4779 output channel specification, of the form:
4780 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
4783 output channel to define, either a channel name (FL, FR, etc.) or a channel
4784 number (c0, c1, etc.)
4787 multiplicative coefficient for the channel, 1 leaving the volume unchanged
4790 input channel to use, see out_name for details; it is not possible to mix
4791 named and numbered input channels
4794 If the `=' in a channel specification is replaced by `<', then the gains for
4795 that specification will be renormalized so that the total is 1, thus
4796 avoiding clipping noise.
4798 @subsection Mixing examples
4800 For example, if you want to down-mix from stereo to mono, but with a bigger
4801 factor for the left channel:
4803 pan=1c|c0=0.9*c0+0.1*c1
4806 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
4807 7-channels surround:
4809 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
4812 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
4813 that should be preferred (see "-ac" option) unless you have very specific
4816 @subsection Remapping examples
4818 The channel remapping will be effective if, and only if:
4821 @item gain coefficients are zeroes or ones,
4822 @item only one input per channel output,
4825 If all these conditions are satisfied, the filter will notify the user ("Pure
4826 channel mapping detected"), and use an optimized and lossless method to do the
4829 For example, if you have a 5.1 source and want a stereo audio stream by
4830 dropping the extra channels:
4832 pan="stereo| c0=FL | c1=FR"
4835 Given the same source, you can also switch front left and front right channels
4836 and keep the input channel layout:
4838 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
4841 If the input is a stereo audio stream, you can mute the front left channel (and
4842 still keep the stereo channel layout) with:
4847 Still with a stereo audio stream input, you can copy the right channel in both
4848 front left and right:
4850 pan="stereo| c0=FR | c1=FR"
4855 ReplayGain scanner filter. This filter takes an audio stream as an input and
4856 outputs it unchanged.
4857 At end of filtering it displays @code{track_gain} and @code{track_peak}.
4861 Convert the audio sample format, sample rate and channel layout. It is
4862 not meant to be used directly.
4865 Apply time-stretching and pitch-shifting with librubberband.
4867 To enable compilation of this filter, you need to configure FFmpeg with
4868 @code{--enable-librubberband}.
4870 The filter accepts the following options:
4874 Set tempo scale factor.
4877 Set pitch scale factor.
4880 Set transients detector.
4881 Possible values are:
4890 Possible values are:
4899 Possible values are:
4906 Set processing window size.
4907 Possible values are:
4916 Possible values are:
4923 Enable formant preservation when shift pitching.
4924 Possible values are:
4932 Possible values are:
4941 Possible values are:
4948 @subsection Commands
4950 This filter supports the following commands:
4953 Change filter tempo scale factor.
4954 Syntax for the command is : "@var{tempo}"
4957 Change filter pitch scale factor.
4958 Syntax for the command is : "@var{pitch}"
4961 @section sidechaincompress
4963 This filter acts like normal compressor but has the ability to compress
4964 detected signal using second input signal.
4965 It needs two input streams and returns one output stream.
4966 First input stream will be processed depending on second stream signal.
4967 The filtered signal then can be filtered with other filters in later stages of
4968 processing. See @ref{pan} and @ref{amerge} filter.
4970 The filter accepts the following options:
4974 Set input gain. Default is 1. Range is between 0.015625 and 64.
4977 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
4978 Default is @code{downward}.
4981 If a signal of second stream raises above this level it will affect the gain
4982 reduction of first stream.
4983 By default is 0.125. Range is between 0.00097563 and 1.
4986 Set a ratio about which the signal is reduced. 1:2 means that if the level
4987 raised 4dB above the threshold, it will be only 2dB above after the reduction.
4988 Default is 2. Range is between 1 and 20.
4991 Amount of milliseconds the signal has to rise above the threshold before gain
4992 reduction starts. Default is 20. Range is between 0.01 and 2000.
4995 Amount of milliseconds the signal has to fall below the threshold before
4996 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
4999 Set the amount by how much signal will be amplified after processing.
5000 Default is 1. Range is from 1 to 64.
5003 Curve the sharp knee around the threshold to enter gain reduction more softly.
5004 Default is 2.82843. Range is between 1 and 8.
5007 Choose if the @code{average} level between all channels of side-chain stream
5008 or the louder(@code{maximum}) channel of side-chain stream affects the
5009 reduction. Default is @code{average}.
5012 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5013 of @code{rms}. Default is @code{rms} which is mainly smoother.
5016 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5019 How much to use compressed signal in output. Default is 1.
5020 Range is between 0 and 1.
5023 @subsection Commands
5025 This filter supports the all above options as @ref{commands}.
5027 @subsection Examples
5031 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5032 depending on the signal of 2nd input and later compressed signal to be
5033 merged with 2nd input:
5035 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5039 @section sidechaingate
5041 A sidechain gate acts like a normal (wideband) gate but has the ability to
5042 filter the detected signal before sending it to the gain reduction stage.
5043 Normally a gate uses the full range signal to detect a level above the
5045 For example: If you cut all lower frequencies from your sidechain signal
5046 the gate will decrease the volume of your track only if not enough highs
5047 appear. With this technique you are able to reduce the resonation of a
5048 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5050 It needs two input streams and returns one output stream.
5051 First input stream will be processed depending on second stream signal.
5053 The filter accepts the following options:
5057 Set input level before filtering.
5058 Default is 1. Allowed range is from 0.015625 to 64.
5061 Set the mode of operation. Can be @code{upward} or @code{downward}.
5062 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5063 will be amplified, expanding dynamic range in upward direction.
5064 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5067 Set the level of gain reduction when the signal is below the threshold.
5068 Default is 0.06125. Allowed range is from 0 to 1.
5069 Setting this to 0 disables reduction and then filter behaves like expander.
5072 If a signal rises above this level the gain reduction is released.
5073 Default is 0.125. Allowed range is from 0 to 1.
5076 Set a ratio about which the signal is reduced.
5077 Default is 2. Allowed range is from 1 to 9000.
5080 Amount of milliseconds the signal has to rise above the threshold before gain
5082 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5085 Amount of milliseconds the signal has to fall below the threshold before the
5086 reduction is increased again. Default is 250 milliseconds.
5087 Allowed range is from 0.01 to 9000.
5090 Set amount of amplification of signal after processing.
5091 Default is 1. Allowed range is from 1 to 64.
5094 Curve the sharp knee around the threshold to enter gain reduction more softly.
5095 Default is 2.828427125. Allowed range is from 1 to 8.
5098 Choose if exact signal should be taken for detection or an RMS like one.
5099 Default is rms. Can be peak or rms.
5102 Choose if the average level between all channels or the louder channel affects
5104 Default is average. Can be average or maximum.
5107 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5110 @subsection Commands
5112 This filter supports the all above options as @ref{commands}.
5114 @section silencedetect
5116 Detect silence in an audio stream.
5118 This filter logs a message when it detects that the input audio volume is less
5119 or equal to a noise tolerance value for a duration greater or equal to the
5120 minimum detected noise duration.
5122 The printed times and duration are expressed in seconds. The
5123 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5124 is set on the first frame whose timestamp equals or exceeds the detection
5125 duration and it contains the timestamp of the first frame of the silence.
5127 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5128 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5129 keys are set on the first frame after the silence. If @option{mono} is
5130 enabled, and each channel is evaluated separately, the @code{.X}
5131 suffixed keys are used, and @code{X} corresponds to the channel number.
5133 The filter accepts the following options:
5137 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5138 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5141 Set silence duration until notification (default is 2 seconds). See
5142 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5143 for the accepted syntax.
5146 Process each channel separately, instead of combined. By default is disabled.
5149 @subsection Examples
5153 Detect 5 seconds of silence with -50dB noise tolerance:
5155 silencedetect=n=-50dB:d=5
5159 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5160 tolerance in @file{silence.mp3}:
5162 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5166 @section silenceremove
5168 Remove silence from the beginning, middle or end of the audio.
5170 The filter accepts the following options:
5174 This value is used to indicate if audio should be trimmed at beginning of
5175 the audio. A value of zero indicates no silence should be trimmed from the
5176 beginning. When specifying a non-zero value, it trims audio up until it
5177 finds non-silence. Normally, when trimming silence from beginning of audio
5178 the @var{start_periods} will be @code{1} but it can be increased to higher
5179 values to trim all audio up to specific count of non-silence periods.
5180 Default value is @code{0}.
5182 @item start_duration
5183 Specify the amount of time that non-silence must be detected before it stops
5184 trimming audio. By increasing the duration, bursts of noises can be treated
5185 as silence and trimmed off. Default value is @code{0}.
5187 @item start_threshold
5188 This indicates what sample value should be treated as silence. For digital
5189 audio, a value of @code{0} may be fine but for audio recorded from analog,
5190 you may wish to increase the value to account for background noise.
5191 Can be specified in dB (in case "dB" is appended to the specified value)
5192 or amplitude ratio. Default value is @code{0}.
5195 Specify max duration of silence at beginning that will be kept after
5196 trimming. Default is 0, which is equal to trimming all samples detected
5200 Specify mode of detection of silence end in start of multi-channel audio.
5201 Can be @var{any} or @var{all}. Default is @var{any}.
5202 With @var{any}, any sample that is detected as non-silence will cause
5203 stopped trimming of silence.
5204 With @var{all}, only if all channels are detected as non-silence will cause
5205 stopped trimming of silence.
5208 Set the count for trimming silence from the end of audio.
5209 To remove silence from the middle of a file, specify a @var{stop_periods}
5210 that is negative. This value is then treated as a positive value and is
5211 used to indicate the effect should restart processing as specified by
5212 @var{start_periods}, making it suitable for removing periods of silence
5213 in the middle of the audio.
5214 Default value is @code{0}.
5217 Specify a duration of silence that must exist before audio is not copied any
5218 more. By specifying a higher duration, silence that is wanted can be left in
5220 Default value is @code{0}.
5222 @item stop_threshold
5223 This is the same as @option{start_threshold} but for trimming silence from
5225 Can be specified in dB (in case "dB" is appended to the specified value)
5226 or amplitude ratio. Default value is @code{0}.
5229 Specify max duration of silence at end that will be kept after
5230 trimming. Default is 0, which is equal to trimming all samples detected
5234 Specify mode of detection of silence start in end of multi-channel audio.
5235 Can be @var{any} or @var{all}. Default is @var{any}.
5236 With @var{any}, any sample that is detected as non-silence will cause
5237 stopped trimming of silence.
5238 With @var{all}, only if all channels are detected as non-silence will cause
5239 stopped trimming of silence.
5242 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5243 and works better with digital silence which is exactly 0.
5244 Default value is @code{rms}.
5247 Set duration in number of seconds used to calculate size of window in number
5248 of samples for detecting silence.
5249 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5252 @subsection Examples
5256 The following example shows how this filter can be used to start a recording
5257 that does not contain the delay at the start which usually occurs between
5258 pressing the record button and the start of the performance:
5260 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5264 Trim all silence encountered from beginning to end where there is more than 1
5265 second of silence in audio:
5267 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5271 Trim all digital silence samples, using peak detection, from beginning to end
5272 where there is more than 0 samples of digital silence in audio and digital
5273 silence is detected in all channels at same positions in stream:
5275 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5281 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5282 loudspeakers around the user for binaural listening via headphones (audio
5283 formats up to 9 channels supported).
5284 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5285 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5286 Austrian Academy of Sciences.
5288 To enable compilation of this filter you need to configure FFmpeg with
5289 @code{--enable-libmysofa}.
5291 The filter accepts the following options:
5295 Set the SOFA file used for rendering.
5298 Set gain applied to audio. Value is in dB. Default is 0.
5301 Set rotation of virtual loudspeakers in deg. Default is 0.
5304 Set elevation of virtual speakers in deg. Default is 0.
5307 Set distance in meters between loudspeakers and the listener with near-field
5308 HRTFs. Default is 1.
5311 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5312 processing audio in time domain which is slow.
5313 @var{freq} is processing audio in frequency domain which is fast.
5314 Default is @var{freq}.
5317 Set custom positions of virtual loudspeakers. Syntax for this option is:
5318 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5319 Each virtual loudspeaker is described with short channel name following with
5320 azimuth and elevation in degrees.
5321 Each virtual loudspeaker description is separated by '|'.
5322 For example to override front left and front right channel positions use:
5323 'speakers=FL 45 15|FR 345 15'.
5324 Descriptions with unrecognised channel names are ignored.
5327 Set custom gain for LFE channels. Value is in dB. Default is 0.
5330 Set custom frame size in number of samples. Default is 1024.
5331 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5332 is set to @var{freq}.
5335 Should all IRs be normalized upon importing SOFA file.
5336 By default is enabled.
5339 Should nearest IRs be interpolated with neighbor IRs if exact position
5340 does not match. By default is disabled.
5343 Minphase all IRs upon loading of SOFA file. By default is disabled.
5346 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5349 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5352 @subsection Examples
5356 Using ClubFritz6 sofa file:
5358 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5362 Using ClubFritz12 sofa file and bigger radius with small rotation:
5364 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5368 Similar as above but with custom speaker positions for front left, front right, back left and back right
5369 and also with custom gain:
5371 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5378 This filter expands or compresses each half-cycle of audio samples
5379 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5380 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5382 The filter accepts the following options:
5386 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5387 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5390 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5391 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5392 would be such that local peak value reaches target peak value but never to surpass it and that
5393 ratio between new and previous peak value does not surpass this option value.
5395 @item compression, c
5396 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5397 This option controls maximum local half-cycle of samples compression. This option is used
5398 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5399 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5400 that peak's half-cycle will be compressed by current compression factor.
5403 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5404 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5405 Any half-cycle samples with their local peak value below or same as this option value will be
5406 compressed by current compression factor, otherwise, if greater than threshold value they will be
5407 expanded with expansion factor so that it could reach peak target value but never surpass it.
5410 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5411 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5412 each new half-cycle until it reaches @option{expansion} value.
5413 Setting this options too high may lead to distortions.
5416 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5417 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5418 each new half-cycle until it reaches @option{compression} value.
5421 Specify which channels to filter, by default all available channels are filtered.
5424 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5425 option. When enabled any half-cycle of samples with their local peak value below or same as
5426 @option{threshold} option will be expanded otherwise it will be compressed.
5429 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5430 When disabled each filtered channel gain calculation is independent, otherwise when this option
5431 is enabled the minimum of all possible gains for each filtered channel is used.
5434 @subsection Commands
5436 This filter supports the all above options as @ref{commands}.
5438 @section stereotools
5440 This filter has some handy utilities to manage stereo signals, for converting
5441 M/S stereo recordings to L/R signal while having control over the parameters
5442 or spreading the stereo image of master track.
5444 The filter accepts the following options:
5448 Set input level before filtering for both channels. Defaults is 1.
5449 Allowed range is from 0.015625 to 64.
5452 Set output level after filtering for both channels. Defaults is 1.
5453 Allowed range is from 0.015625 to 64.
5456 Set input balance between both channels. Default is 0.
5457 Allowed range is from -1 to 1.
5460 Set output balance between both channels. Default is 0.
5461 Allowed range is from -1 to 1.
5464 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5465 clipping. Disabled by default.
5468 Mute the left channel. Disabled by default.
5471 Mute the right channel. Disabled by default.
5474 Change the phase of the left channel. Disabled by default.
5477 Change the phase of the right channel. Disabled by default.
5480 Set stereo mode. Available values are:
5484 Left/Right to Left/Right, this is default.
5487 Left/Right to Mid/Side.
5490 Mid/Side to Left/Right.
5493 Left/Right to Left/Left.
5496 Left/Right to Right/Right.
5499 Left/Right to Left + Right.
5502 Left/Right to Right/Left.
5505 Mid/Side to Left/Left.
5508 Mid/Side to Right/Right.
5511 Mid/Side to Right/Left.
5514 Left/Right to Left - Right.
5518 Set level of side signal. Default is 1.
5519 Allowed range is from 0.015625 to 64.
5522 Set balance of side signal. Default is 0.
5523 Allowed range is from -1 to 1.
5526 Set level of the middle signal. Default is 1.
5527 Allowed range is from 0.015625 to 64.
5530 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5533 Set stereo base between mono and inversed channels. Default is 0.
5534 Allowed range is from -1 to 1.
5537 Set delay in milliseconds how much to delay left from right channel and
5538 vice versa. Default is 0. Allowed range is from -20 to 20.
5541 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5544 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5546 @item bmode_in, bmode_out
5547 Set balance mode for balance_in/balance_out option.
5549 Can be one of the following:
5553 Classic balance mode. Attenuate one channel at time.
5554 Gain is raised up to 1.
5557 Similar as classic mode above but gain is raised up to 2.
5560 Equal power distribution, from -6dB to +6dB range.
5564 @subsection Commands
5566 This filter supports the all above options as @ref{commands}.
5568 @subsection Examples
5572 Apply karaoke like effect:
5574 stereotools=mlev=0.015625
5578 Convert M/S signal to L/R:
5580 "stereotools=mode=ms>lr"
5584 @section stereowiden
5586 This filter enhance the stereo effect by suppressing signal common to both
5587 channels and by delaying the signal of left into right and vice versa,
5588 thereby widening the stereo effect.
5590 The filter accepts the following options:
5594 Time in milliseconds of the delay of left signal into right and vice versa.
5595 Default is 20 milliseconds.
5598 Amount of gain in delayed signal into right and vice versa. Gives a delay
5599 effect of left signal in right output and vice versa which gives widening
5600 effect. Default is 0.3.
5603 Cross feed of left into right with inverted phase. This helps in suppressing
5604 the mono. If the value is 1 it will cancel all the signal common to both
5605 channels. Default is 0.3.
5608 Set level of input signal of original channel. Default is 0.8.
5611 @subsection Commands
5613 This filter supports the all above options except @code{delay} as @ref{commands}.
5615 @section superequalizer
5616 Apply 18 band equalizer.
5618 The filter accepts the following options:
5625 Set 131Hz band gain.
5627 Set 185Hz band gain.
5629 Set 262Hz band gain.
5631 Set 370Hz band gain.
5633 Set 523Hz band gain.
5635 Set 740Hz band gain.
5637 Set 1047Hz band gain.
5639 Set 1480Hz band gain.
5641 Set 2093Hz band gain.
5643 Set 2960Hz band gain.
5645 Set 4186Hz band gain.
5647 Set 5920Hz band gain.
5649 Set 8372Hz band gain.
5651 Set 11840Hz band gain.
5653 Set 16744Hz band gain.
5655 Set 20000Hz band gain.
5659 Apply audio surround upmix filter.
5661 This filter allows to produce multichannel output from audio stream.
5663 The filter accepts the following options:
5667 Set output channel layout. By default, this is @var{5.1}.
5669 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5670 for the required syntax.
5673 Set input channel layout. By default, this is @var{stereo}.
5675 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5676 for the required syntax.
5679 Set input volume level. By default, this is @var{1}.
5682 Set output volume level. By default, this is @var{1}.
5685 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5688 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5691 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5694 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5695 In @var{add} mode, LFE channel is created from input audio and added to output.
5696 In @var{sub} mode, LFE channel is created from input audio and added to output but
5697 also all non-LFE output channels are subtracted with output LFE channel.
5700 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5701 Default is @var{90}.
5704 Set front center input volume. By default, this is @var{1}.
5707 Set front center output volume. By default, this is @var{1}.
5710 Set front left input volume. By default, this is @var{1}.
5713 Set front left output volume. By default, this is @var{1}.
5716 Set front right input volume. By default, this is @var{1}.
5719 Set front right output volume. By default, this is @var{1}.
5722 Set side left input volume. By default, this is @var{1}.
5725 Set side left output volume. By default, this is @var{1}.
5728 Set side right input volume. By default, this is @var{1}.
5731 Set side right output volume. By default, this is @var{1}.
5734 Set back left input volume. By default, this is @var{1}.
5737 Set back left output volume. By default, this is @var{1}.
5740 Set back right input volume. By default, this is @var{1}.
5743 Set back right output volume. By default, this is @var{1}.
5746 Set back center input volume. By default, this is @var{1}.
5749 Set back center output volume. By default, this is @var{1}.
5752 Set LFE input volume. By default, this is @var{1}.
5755 Set LFE output volume. By default, this is @var{1}.
5758 Set spread usage of stereo image across X axis for all channels.
5761 Set spread usage of stereo image across Y axis for all channels.
5763 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5764 Set spread usage of stereo image across X axis for each channel.
5766 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5767 Set spread usage of stereo image across Y axis for each channel.
5770 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5773 Set window function.
5775 It accepts the following values:
5798 Default is @code{hann}.
5801 Set window overlap. If set to 1, the recommended overlap for selected
5802 window function will be picked. Default is @code{0.5}.
5805 @section treble, highshelf
5807 Boost or cut treble (upper) frequencies of the audio using a two-pole
5808 shelving filter with a response similar to that of a standard
5809 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
5811 The filter accepts the following options:
5815 Give the gain at whichever is the lower of ~22 kHz and the
5816 Nyquist frequency. Its useful range is about -20 (for a large cut)
5817 to +20 (for a large boost). Beware of clipping when using a positive gain.
5820 Set the filter's central frequency and so can be used
5821 to extend or reduce the frequency range to be boosted or cut.
5822 The default value is @code{3000} Hz.
5825 Set method to specify band-width of filter.
5840 Determine how steep is the filter's shelf transition.
5843 How much to use filtered signal in output. Default is 1.
5844 Range is between 0 and 1.
5847 Specify which channels to filter, by default all available are filtered.
5850 Normalize biquad coefficients, by default is disabled.
5851 Enabling it will normalize magnitude response at DC to 0dB.
5854 Set transform type of IIR filter.
5863 @subsection Commands
5865 This filter supports the following commands:
5868 Change treble frequency.
5869 Syntax for the command is : "@var{frequency}"
5872 Change treble width_type.
5873 Syntax for the command is : "@var{width_type}"
5876 Change treble width.
5877 Syntax for the command is : "@var{width}"
5881 Syntax for the command is : "@var{gain}"
5885 Syntax for the command is : "@var{mix}"
5890 Sinusoidal amplitude modulation.
5892 The filter accepts the following options:
5896 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
5897 (20 Hz or lower) will result in a tremolo effect.
5898 This filter may also be used as a ring modulator by specifying
5899 a modulation frequency higher than 20 Hz.
5900 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5903 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5904 Default value is 0.5.
5909 Sinusoidal phase modulation.
5911 The filter accepts the following options:
5915 Modulation frequency in Hertz.
5916 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
5919 Depth of modulation as a percentage. Range is 0.0 - 1.0.
5920 Default value is 0.5.
5925 Adjust the input audio volume.
5927 It accepts the following parameters:
5931 Set audio volume expression.
5933 Output values are clipped to the maximum value.
5935 The output audio volume is given by the relation:
5937 @var{output_volume} = @var{volume} * @var{input_volume}
5940 The default value for @var{volume} is "1.0".
5943 This parameter represents the mathematical precision.
5945 It determines which input sample formats will be allowed, which affects the
5946 precision of the volume scaling.
5950 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
5952 32-bit floating-point; this limits input sample format to FLT. (default)
5954 64-bit floating-point; this limits input sample format to DBL.
5958 Choose the behaviour on encountering ReplayGain side data in input frames.
5962 Remove ReplayGain side data, ignoring its contents (the default).
5965 Ignore ReplayGain side data, but leave it in the frame.
5968 Prefer the track gain, if present.
5971 Prefer the album gain, if present.
5974 @item replaygain_preamp
5975 Pre-amplification gain in dB to apply to the selected replaygain gain.
5977 Default value for @var{replaygain_preamp} is 0.0.
5979 @item replaygain_noclip
5980 Prevent clipping by limiting the gain applied.
5982 Default value for @var{replaygain_noclip} is 1.
5985 Set when the volume expression is evaluated.
5987 It accepts the following values:
5990 only evaluate expression once during the filter initialization, or
5991 when the @samp{volume} command is sent
5994 evaluate expression for each incoming frame
5997 Default value is @samp{once}.
6000 The volume expression can contain the following parameters.
6004 frame number (starting at zero)
6007 @item nb_consumed_samples
6008 number of samples consumed by the filter
6010 number of samples in the current frame
6012 original frame position in the file
6018 PTS at start of stream
6020 time at start of stream
6026 last set volume value
6029 Note that when @option{eval} is set to @samp{once} only the
6030 @var{sample_rate} and @var{tb} variables are available, all other
6031 variables will evaluate to NAN.
6033 @subsection Commands
6035 This filter supports the following commands:
6038 Modify the volume expression.
6039 The command accepts the same syntax of the corresponding option.
6041 If the specified expression is not valid, it is kept at its current
6045 @subsection Examples
6049 Halve the input audio volume:
6053 volume=volume=-6.0206dB
6056 In all the above example the named key for @option{volume} can be
6057 omitted, for example like in:
6063 Increase input audio power by 6 decibels using fixed-point precision:
6065 volume=volume=6dB:precision=fixed
6069 Fade volume after time 10 with an annihilation period of 5 seconds:
6071 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6075 @section volumedetect
6077 Detect the volume of the input video.
6079 The filter has no parameters. The input is not modified. Statistics about
6080 the volume will be printed in the log when the input stream end is reached.
6082 In particular it will show the mean volume (root mean square), maximum
6083 volume (on a per-sample basis), and the beginning of a histogram of the
6084 registered volume values (from the maximum value to a cumulated 1/1000 of
6087 All volumes are in decibels relative to the maximum PCM value.
6089 @subsection Examples
6091 Here is an excerpt of the output:
6093 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6094 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6095 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6096 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6097 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6098 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6099 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6100 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6101 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6107 The mean square energy is approximately -27 dB, or 10^-2.7.
6109 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6111 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6114 In other words, raising the volume by +4 dB does not cause any clipping,
6115 raising it by +5 dB causes clipping for 6 samples, etc.
6117 @c man end AUDIO FILTERS
6119 @chapter Audio Sources
6120 @c man begin AUDIO SOURCES
6122 Below is a description of the currently available audio sources.
6126 Buffer audio frames, and make them available to the filter chain.
6128 This source is mainly intended for a programmatic use, in particular
6129 through the interface defined in @file{libavfilter/buffersrc.h}.
6131 It accepts the following parameters:
6135 The timebase which will be used for timestamps of submitted frames. It must be
6136 either a floating-point number or in @var{numerator}/@var{denominator} form.
6139 The sample rate of the incoming audio buffers.
6142 The sample format of the incoming audio buffers.
6143 Either a sample format name or its corresponding integer representation from
6144 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6146 @item channel_layout
6147 The channel layout of the incoming audio buffers.
6148 Either a channel layout name from channel_layout_map in
6149 @file{libavutil/channel_layout.c} or its corresponding integer representation
6150 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6153 The number of channels of the incoming audio buffers.
6154 If both @var{channels} and @var{channel_layout} are specified, then they
6159 @subsection Examples
6162 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6165 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6166 Since the sample format with name "s16p" corresponds to the number
6167 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6170 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6175 Generate an audio signal specified by an expression.
6177 This source accepts in input one or more expressions (one for each
6178 channel), which are evaluated and used to generate a corresponding
6181 This source accepts the following options:
6185 Set the '|'-separated expressions list for each separate channel. In case the
6186 @option{channel_layout} option is not specified, the selected channel layout
6187 depends on the number of provided expressions. Otherwise the last
6188 specified expression is applied to the remaining output channels.
6190 @item channel_layout, c
6191 Set the channel layout. The number of channels in the specified layout
6192 must be equal to the number of specified expressions.
6195 Set the minimum duration of the sourced audio. See
6196 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6197 for the accepted syntax.
6198 Note that the resulting duration may be greater than the specified
6199 duration, as the generated audio is always cut at the end of a
6202 If not specified, or the expressed duration is negative, the audio is
6203 supposed to be generated forever.
6206 Set the number of samples per channel per each output frame,
6209 @item sample_rate, s
6210 Specify the sample rate, default to 44100.
6213 Each expression in @var{exprs} can contain the following constants:
6217 number of the evaluated sample, starting from 0
6220 time of the evaluated sample expressed in seconds, starting from 0
6227 @subsection Examples
6237 Generate a sin signal with frequency of 440 Hz, set sample rate to
6240 aevalsrc="sin(440*2*PI*t):s=8000"
6244 Generate a two channels signal, specify the channel layout (Front
6245 Center + Back Center) explicitly:
6247 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6251 Generate white noise:
6253 aevalsrc="-2+random(0)"
6257 Generate an amplitude modulated signal:
6259 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6263 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6265 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6272 Generate a FIR coefficients using frequency sampling method.
6274 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6276 The filter accepts the following options:
6280 Set number of filter coefficents in output audio stream.
6281 Default value is 1025.
6284 Set frequency points from where magnitude and phase are set.
6285 This must be in non decreasing order, and first element must be 0, while last element
6286 must be 1. Elements are separated by white spaces.
6289 Set magnitude value for every frequency point set by @option{frequency}.
6290 Number of values must be same as number of frequency points.
6291 Values are separated by white spaces.
6294 Set phase value for every frequency point set by @option{frequency}.
6295 Number of values must be same as number of frequency points.
6296 Values are separated by white spaces.
6298 @item sample_rate, r
6299 Set sample rate, default is 44100.
6302 Set number of samples per each frame. Default is 1024.
6305 Set window function. Default is blackman.
6310 The null audio source, return unprocessed audio frames. It is mainly useful
6311 as a template and to be employed in analysis / debugging tools, or as
6312 the source for filters which ignore the input data (for example the sox
6315 This source accepts the following options:
6319 @item channel_layout, cl
6321 Specifies the channel layout, and can be either an integer or a string
6322 representing a channel layout. The default value of @var{channel_layout}
6325 Check the channel_layout_map definition in
6326 @file{libavutil/channel_layout.c} for the mapping between strings and
6327 channel layout values.
6329 @item sample_rate, r
6330 Specifies the sample rate, and defaults to 44100.
6333 Set the number of samples per requested frames.
6336 Set the duration of the sourced audio. See
6337 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6338 for the accepted syntax.
6340 If not specified, or the expressed duration is negative, the audio is
6341 supposed to be generated forever.
6344 @subsection Examples
6348 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6350 anullsrc=r=48000:cl=4
6354 Do the same operation with a more obvious syntax:
6356 anullsrc=r=48000:cl=mono
6360 All the parameters need to be explicitly defined.
6364 Synthesize a voice utterance using the libflite library.
6366 To enable compilation of this filter you need to configure FFmpeg with
6367 @code{--enable-libflite}.
6369 Note that versions of the flite library prior to 2.0 are not thread-safe.
6371 The filter accepts the following options:
6376 If set to 1, list the names of the available voices and exit
6377 immediately. Default value is 0.
6380 Set the maximum number of samples per frame. Default value is 512.
6383 Set the filename containing the text to speak.
6386 Set the text to speak.
6389 Set the voice to use for the speech synthesis. Default value is
6390 @code{kal}. See also the @var{list_voices} option.
6393 @subsection Examples
6397 Read from file @file{speech.txt}, and synthesize the text using the
6398 standard flite voice:
6400 flite=textfile=speech.txt
6404 Read the specified text selecting the @code{slt} voice:
6406 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6410 Input text to ffmpeg:
6412 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6416 Make @file{ffplay} speak the specified text, using @code{flite} and
6417 the @code{lavfi} device:
6419 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6423 For more information about libflite, check:
6424 @url{http://www.festvox.org/flite/}
6428 Generate a noise audio signal.
6430 The filter accepts the following options:
6433 @item sample_rate, r
6434 Specify the sample rate. Default value is 48000 Hz.
6437 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6441 Specify the duration of the generated audio stream. Not specifying this option
6442 results in noise with an infinite length.
6444 @item color, colour, c
6445 Specify the color of noise. Available noise colors are white, pink, brown,
6446 blue, violet and velvet. Default color is white.
6449 Specify a value used to seed the PRNG.
6452 Set the number of samples per each output frame, default is 1024.
6455 @subsection Examples
6460 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6462 anoisesrc=d=60:c=pink:r=44100:a=0.5
6468 Generate odd-tap Hilbert transform FIR coefficients.
6470 The resulting stream can be used with @ref{afir} filter for phase-shifting
6471 the signal by 90 degrees.
6473 This is used in many matrix coding schemes and for analytic signal generation.
6474 The process is often written as a multiplication by i (or j), the imaginary unit.
6476 The filter accepts the following options:
6480 @item sample_rate, s
6481 Set sample rate, default is 44100.
6484 Set length of FIR filter, default is 22051.
6487 Set number of samples per each frame.
6490 Set window function to be used when generating FIR coefficients.
6495 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6497 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6499 The filter accepts the following options:
6502 @item sample_rate, r
6503 Set sample rate, default is 44100.
6506 Set number of samples per each frame. Default is 1024.
6509 Set high-pass frequency. Default is 0.
6512 Set low-pass frequency. Default is 0.
6513 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6514 is higher than 0 then filter will create band-pass filter coefficients,
6515 otherwise band-reject filter coefficients.
6518 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6521 Set Kaiser window beta.
6524 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6527 Enable rounding, by default is disabled.
6530 Set number of taps for high-pass filter.
6533 Set number of taps for low-pass filter.
6538 Generate an audio signal made of a sine wave with amplitude 1/8.
6540 The audio signal is bit-exact.
6542 The filter accepts the following options:
6547 Set the carrier frequency. Default is 440 Hz.
6549 @item beep_factor, b
6550 Enable a periodic beep every second with frequency @var{beep_factor} times
6551 the carrier frequency. Default is 0, meaning the beep is disabled.
6553 @item sample_rate, r
6554 Specify the sample rate, default is 44100.
6557 Specify the duration of the generated audio stream.
6559 @item samples_per_frame
6560 Set the number of samples per output frame.
6562 The expression can contain the following constants:
6566 The (sequential) number of the output audio frame, starting from 0.
6569 The PTS (Presentation TimeStamp) of the output audio frame,
6570 expressed in @var{TB} units.
6573 The PTS of the output audio frame, expressed in seconds.
6576 The timebase of the output audio frames.
6579 Default is @code{1024}.
6582 @subsection Examples
6587 Generate a simple 440 Hz sine wave:
6593 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6597 sine=frequency=220:beep_factor=4:duration=5
6601 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6604 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6608 @c man end AUDIO SOURCES
6610 @chapter Audio Sinks
6611 @c man begin AUDIO SINKS
6613 Below is a description of the currently available audio sinks.
6615 @section abuffersink
6617 Buffer audio frames, and make them available to the end of filter chain.
6619 This sink is mainly intended for programmatic use, in particular
6620 through the interface defined in @file{libavfilter/buffersink.h}
6621 or the options system.
6623 It accepts a pointer to an AVABufferSinkContext structure, which
6624 defines the incoming buffers' formats, to be passed as the opaque
6625 parameter to @code{avfilter_init_filter} for initialization.
6628 Null audio sink; do absolutely nothing with the input audio. It is
6629 mainly useful as a template and for use in analysis / debugging
6632 @c man end AUDIO SINKS
6634 @chapter Video Filters
6635 @c man begin VIDEO FILTERS
6637 When you configure your FFmpeg build, you can disable any of the
6638 existing filters using @code{--disable-filters}.
6639 The configure output will show the video filters included in your
6642 Below is a description of the currently available video filters.
6646 Mark a region of interest in a video frame.
6648 The frame data is passed through unchanged, but metadata is attached
6649 to the frame indicating regions of interest which can affect the
6650 behaviour of later encoding. Multiple regions can be marked by
6651 applying the filter multiple times.
6655 Region distance in pixels from the left edge of the frame.
6657 Region distance in pixels from the top edge of the frame.
6659 Region width in pixels.
6661 Region height in pixels.
6663 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6664 and may contain the following variables:
6667 Width of the input frame.
6669 Height of the input frame.
6673 Quantisation offset to apply within the region.
6675 This must be a real value in the range -1 to +1. A value of zero
6676 indicates no quality change. A negative value asks for better quality
6677 (less quantisation), while a positive value asks for worse quality
6678 (greater quantisation).
6680 The range is calibrated so that the extreme values indicate the
6681 largest possible offset - if the rest of the frame is encoded with the
6682 worst possible quality, an offset of -1 indicates that this region
6683 should be encoded with the best possible quality anyway. Intermediate
6684 values are then interpolated in some codec-dependent way.
6686 For example, in 10-bit H.264 the quantisation parameter varies between
6687 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6688 this region should be encoded with a QP around one-tenth of the full
6689 range better than the rest of the frame. So, if most of the frame
6690 were to be encoded with a QP of around 30, this region would get a QP
6691 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6692 An extreme value of -1 would indicate that this region should be
6693 encoded with the best possible quality regardless of the treatment of
6694 the rest of the frame - that is, should be encoded at a QP of -12.
6696 If set to true, remove any existing regions of interest marked on the
6697 frame before adding the new one.
6700 @subsection Examples
6704 Mark the centre quarter of the frame as interesting.
6706 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6709 Mark the 100-pixel-wide region on the left edge of the frame as very
6710 uninteresting (to be encoded at much lower quality than the rest of
6713 addroi=0:0:100:ih:+1/5
6717 @section alphaextract
6719 Extract the alpha component from the input as a grayscale video. This
6720 is especially useful with the @var{alphamerge} filter.
6724 Add or replace the alpha component of the primary input with the
6725 grayscale value of a second input. This is intended for use with
6726 @var{alphaextract} to allow the transmission or storage of frame
6727 sequences that have alpha in a format that doesn't support an alpha
6730 For example, to reconstruct full frames from a normal YUV-encoded video
6731 and a separate video created with @var{alphaextract}, you might use:
6733 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6738 Amplify differences between current pixel and pixels of adjacent frames in
6739 same pixel location.
6741 This filter accepts the following options:
6745 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6746 For example radius of 3 will instruct filter to calculate average of 7 frames.
6749 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6752 Set threshold for difference amplification. Any difference greater or equal to
6753 this value will not alter source pixel. Default is 10.
6754 Allowed range is from 0 to 65535.
6757 Set tolerance for difference amplification. Any difference lower to
6758 this value will not alter source pixel. Default is 0.
6759 Allowed range is from 0 to 65535.
6762 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6763 This option controls maximum possible value that will decrease source pixel value.
6766 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
6767 This option controls maximum possible value that will increase source pixel value.
6770 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
6773 @subsection Commands
6775 This filter supports the following @ref{commands} that corresponds to option of same name:
6787 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
6788 and libavformat to work. On the other hand, it is limited to ASS (Advanced
6789 Substation Alpha) subtitles files.
6791 This filter accepts the following option in addition to the common options from
6792 the @ref{subtitles} filter:
6796 Set the shaping engine
6798 Available values are:
6801 The default libass shaping engine, which is the best available.
6803 Fast, font-agnostic shaper that can do only substitutions
6805 Slower shaper using OpenType for substitutions and positioning
6808 The default is @code{auto}.
6812 Apply an Adaptive Temporal Averaging Denoiser to the video input.
6814 The filter accepts the following options:
6818 Set threshold A for 1st plane. Default is 0.02.
6819 Valid range is 0 to 0.3.
6822 Set threshold B for 1st plane. Default is 0.04.
6823 Valid range is 0 to 5.
6826 Set threshold A for 2nd plane. Default is 0.02.
6827 Valid range is 0 to 0.3.
6830 Set threshold B for 2nd plane. Default is 0.04.
6831 Valid range is 0 to 5.
6834 Set threshold A for 3rd plane. Default is 0.02.
6835 Valid range is 0 to 0.3.
6838 Set threshold B for 3rd plane. Default is 0.04.
6839 Valid range is 0 to 5.
6841 Threshold A is designed to react on abrupt changes in the input signal and
6842 threshold B is designed to react on continuous changes in the input signal.
6845 Set number of frames filter will use for averaging. Default is 9. Must be odd
6846 number in range [5, 129].
6849 Set what planes of frame filter will use for averaging. Default is all.
6852 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
6853 Alternatively can be set to @code{s} serial.
6855 Parallel can be faster then serial, while other way around is never true.
6856 Parallel will abort early on first change being greater then thresholds, while serial
6857 will continue processing other side of frames if they are equal or below thresholds.
6860 @subsection Commands
6861 This filter supports same @ref{commands} as options except option @code{s}.
6862 The command accepts the same syntax of the corresponding option.
6866 Apply average blur filter.
6868 The filter accepts the following options:
6872 Set horizontal radius size.
6875 Set which planes to filter. By default all planes are filtered.
6878 Set vertical radius size, if zero it will be same as @code{sizeX}.
6879 Default is @code{0}.
6882 @subsection Commands
6883 This filter supports same commands as options.
6884 The command accepts the same syntax of the corresponding option.
6886 If the specified expression is not valid, it is kept at its current
6891 Compute the bounding box for the non-black pixels in the input frame
6894 This filter computes the bounding box containing all the pixels with a
6895 luminance value greater than the minimum allowed value.
6896 The parameters describing the bounding box are printed on the filter
6899 The filter accepts the following option:
6903 Set the minimal luminance value. Default is @code{16}.
6907 Apply bilateral filter, spatial smoothing while preserving edges.
6909 The filter accepts the following options:
6912 Set sigma of gaussian function to calculate spatial weight.
6913 Allowed range is 0 to 512. Default is 0.1.
6916 Set sigma of gaussian function to calculate range weight.
6917 Allowed range is 0 to 1. Default is 0.1.
6920 Set planes to filter. Default is first only.
6923 @section bitplanenoise
6925 Show and measure bit plane noise.
6927 The filter accepts the following options:
6931 Set which plane to analyze. Default is @code{1}.
6934 Filter out noisy pixels from @code{bitplane} set above.
6935 Default is disabled.
6938 @section blackdetect
6940 Detect video intervals that are (almost) completely black. Can be
6941 useful to detect chapter transitions, commercials, or invalid
6944 The filter outputs its detection analysis to both the log as well as
6945 frame metadata. If a black segment of at least the specified minimum
6946 duration is found, a line with the start and end timestamps as well
6947 as duration is printed to the log with level @code{info}. In addition,
6948 a log line with level @code{debug} is printed per frame showing the
6949 black amount detected for that frame.
6951 The filter also attaches metadata to the first frame of a black
6952 segment with key @code{lavfi.black_start} and to the first frame
6953 after the black segment ends with key @code{lavfi.black_end}. The
6954 value is the frame's timestamp. This metadata is added regardless
6955 of the minimum duration specified.
6957 The filter accepts the following options:
6960 @item black_min_duration, d
6961 Set the minimum detected black duration expressed in seconds. It must
6962 be a non-negative floating point number.
6964 Default value is 2.0.
6966 @item picture_black_ratio_th, pic_th
6967 Set the threshold for considering a picture "black".
6968 Express the minimum value for the ratio:
6970 @var{nb_black_pixels} / @var{nb_pixels}
6973 for which a picture is considered black.
6974 Default value is 0.98.
6976 @item pixel_black_th, pix_th
6977 Set the threshold for considering a pixel "black".
6979 The threshold expresses the maximum pixel luminance value for which a
6980 pixel is considered "black". The provided value is scaled according to
6981 the following equation:
6983 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
6986 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
6987 the input video format, the range is [0-255] for YUV full-range
6988 formats and [16-235] for YUV non full-range formats.
6990 Default value is 0.10.
6993 The following example sets the maximum pixel threshold to the minimum
6994 value, and detects only black intervals of 2 or more seconds:
6996 blackdetect=d=2:pix_th=0.00
7001 Detect frames that are (almost) completely black. Can be useful to
7002 detect chapter transitions or commercials. Output lines consist of
7003 the frame number of the detected frame, the percentage of blackness,
7004 the position in the file if known or -1 and the timestamp in seconds.
7006 In order to display the output lines, you need to set the loglevel at
7007 least to the AV_LOG_INFO value.
7009 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7010 The value represents the percentage of pixels in the picture that
7011 are below the threshold value.
7013 It accepts the following parameters:
7018 The percentage of the pixels that have to be below the threshold; it defaults to
7021 @item threshold, thresh
7022 The threshold below which a pixel value is considered black; it defaults to
7030 Blend two video frames into each other.
7032 The @code{blend} filter takes two input streams and outputs one
7033 stream, the first input is the "top" layer and second input is
7034 "bottom" layer. By default, the output terminates when the longest input terminates.
7036 The @code{tblend} (time blend) filter takes two consecutive frames
7037 from one single stream, and outputs the result obtained by blending
7038 the new frame on top of the old frame.
7040 A description of the accepted options follows.
7048 Set blend mode for specific pixel component or all pixel components in case
7049 of @var{all_mode}. Default value is @code{normal}.
7051 Available values for component modes are:
7093 Set blend opacity for specific pixel component or all pixel components in case
7094 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7101 Set blend expression for specific pixel component or all pixel components in case
7102 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7104 The expressions can use the following variables:
7108 The sequential number of the filtered frame, starting from @code{0}.
7112 the coordinates of the current sample
7116 the width and height of currently filtered plane
7120 Width and height scale for the plane being filtered. It is the
7121 ratio between the dimensions of the current plane to the luma plane,
7122 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7123 the luma plane and @code{0.5,0.5} for the chroma planes.
7126 Time of the current frame, expressed in seconds.
7129 Value of pixel component at current location for first video frame (top layer).
7132 Value of pixel component at current location for second video frame (bottom layer).
7136 The @code{blend} filter also supports the @ref{framesync} options.
7138 @subsection Examples
7142 Apply transition from bottom layer to top layer in first 10 seconds:
7144 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7148 Apply linear horizontal transition from top layer to bottom layer:
7150 blend=all_expr='A*(X/W)+B*(1-X/W)'
7154 Apply 1x1 checkerboard effect:
7156 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7160 Apply uncover left effect:
7162 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7166 Apply uncover down effect:
7168 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7172 Apply uncover up-left effect:
7174 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7178 Split diagonally video and shows top and bottom layer on each side:
7180 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7184 Display differences between the current and the previous frame:
7186 tblend=all_mode=grainextract
7192 Denoise frames using Block-Matching 3D algorithm.
7194 The filter accepts the following options.
7198 Set denoising strength. Default value is 1.
7199 Allowed range is from 0 to 999.9.
7200 The denoising algorithm is very sensitive to sigma, so adjust it
7201 according to the source.
7204 Set local patch size. This sets dimensions in 2D.
7207 Set sliding step for processing blocks. Default value is 4.
7208 Allowed range is from 1 to 64.
7209 Smaller values allows processing more reference blocks and is slower.
7212 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7213 When set to 1, no block matching is done. Larger values allows more blocks
7215 Allowed range is from 1 to 256.
7218 Set radius for search block matching. Default is 9.
7219 Allowed range is from 1 to INT32_MAX.
7222 Set step between two search locations for block matching. Default is 1.
7223 Allowed range is from 1 to 64. Smaller is slower.
7226 Set threshold of mean square error for block matching. Valid range is 0 to
7230 Set thresholding parameter for hard thresholding in 3D transformed domain.
7231 Larger values results in stronger hard-thresholding filtering in frequency
7235 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7236 Default is @code{basic}.
7239 If enabled, filter will use 2nd stream for block matching.
7240 Default is disabled for @code{basic} value of @var{estim} option,
7241 and always enabled if value of @var{estim} is @code{final}.
7244 Set planes to filter. Default is all available except alpha.
7247 @subsection Examples
7251 Basic filtering with bm3d:
7253 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7257 Same as above, but filtering only luma:
7259 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7263 Same as above, but with both estimation modes:
7265 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
7269 Same as above, but prefilter with @ref{nlmeans} filter instead:
7271 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
7277 Apply a boxblur algorithm to the input video.
7279 It accepts the following parameters:
7283 @item luma_radius, lr
7284 @item luma_power, lp
7285 @item chroma_radius, cr
7286 @item chroma_power, cp
7287 @item alpha_radius, ar
7288 @item alpha_power, ap
7292 A description of the accepted options follows.
7295 @item luma_radius, lr
7296 @item chroma_radius, cr
7297 @item alpha_radius, ar
7298 Set an expression for the box radius in pixels used for blurring the
7299 corresponding input plane.
7301 The radius value must be a non-negative number, and must not be
7302 greater than the value of the expression @code{min(w,h)/2} for the
7303 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7306 Default value for @option{luma_radius} is "2". If not specified,
7307 @option{chroma_radius} and @option{alpha_radius} default to the
7308 corresponding value set for @option{luma_radius}.
7310 The expressions can contain the following constants:
7314 The input width and height in pixels.
7318 The input chroma image width and height in pixels.
7322 The horizontal and vertical chroma subsample values. For example, for the
7323 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7326 @item luma_power, lp
7327 @item chroma_power, cp
7328 @item alpha_power, ap
7329 Specify how many times the boxblur filter is applied to the
7330 corresponding plane.
7332 Default value for @option{luma_power} is 2. If not specified,
7333 @option{chroma_power} and @option{alpha_power} default to the
7334 corresponding value set for @option{luma_power}.
7336 A value of 0 will disable the effect.
7339 @subsection Examples
7343 Apply a boxblur filter with the luma, chroma, and alpha radii
7346 boxblur=luma_radius=2:luma_power=1
7351 Set the luma radius to 2, and alpha and chroma radius to 0:
7353 boxblur=2:1:cr=0:ar=0
7357 Set the luma and chroma radii to a fraction of the video dimension:
7359 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7365 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7366 Deinterlacing Filter").
7368 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7369 interpolation algorithms.
7370 It accepts the following parameters:
7374 The interlacing mode to adopt. It accepts one of the following values:
7378 Output one frame for each frame.
7380 Output one frame for each field.
7383 The default value is @code{send_field}.
7386 The picture field parity assumed for the input interlaced video. It accepts one
7387 of the following values:
7391 Assume the top field is first.
7393 Assume the bottom field is first.
7395 Enable automatic detection of field parity.
7398 The default value is @code{auto}.
7399 If the interlacing is unknown or the decoder does not export this information,
7400 top field first will be assumed.
7403 Specify which frames to deinterlace. Accepts one of the following
7408 Deinterlace all frames.
7410 Only deinterlace frames marked as interlaced.
7413 The default value is @code{all}.
7418 Apply Contrast Adaptive Sharpen filter to video stream.
7420 The filter accepts the following options:
7424 Set the sharpening strength. Default value is 0.
7427 Set planes to filter. Default value is to filter all
7428 planes except alpha plane.
7432 Remove all color information for all colors except for certain one.
7434 The filter accepts the following options:
7438 The color which will not be replaced with neutral chroma.
7441 Similarity percentage with the above color.
7442 0.01 matches only the exact key color, while 1.0 matches everything.
7446 0.0 makes pixels either fully gray, or not gray at all.
7447 Higher values result in more preserved color.
7450 Signals that the color passed is already in YUV instead of RGB.
7452 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7453 This can be used to pass exact YUV values as hexadecimal numbers.
7456 @subsection Commands
7457 This filter supports same @ref{commands} as options.
7458 The command accepts the same syntax of the corresponding option.
7460 If the specified expression is not valid, it is kept at its current
7464 YUV colorspace color/chroma keying.
7466 The filter accepts the following options:
7470 The color which will be replaced with transparency.
7473 Similarity percentage with the key color.
7475 0.01 matches only the exact key color, while 1.0 matches everything.
7480 0.0 makes pixels either fully transparent, or not transparent at all.
7482 Higher values result in semi-transparent pixels, with a higher transparency
7483 the more similar the pixels color is to the key color.
7486 Signals that the color passed is already in YUV instead of RGB.
7488 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7489 This can be used to pass exact YUV values as hexadecimal numbers.
7492 @subsection Commands
7493 This filter supports same @ref{commands} as options.
7494 The command accepts the same syntax of the corresponding option.
7496 If the specified expression is not valid, it is kept at its current
7499 @subsection Examples
7503 Make every green pixel in the input image transparent:
7505 ffmpeg -i input.png -vf chromakey=green out.png
7509 Overlay a greenscreen-video on top of a static black background.
7511 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
7516 Reduce chrominance noise.
7518 The filter accepts the following options:
7522 Set threshold for averaging chrominance values.
7523 Sum of absolute difference of U and V pixel components or current
7524 pixel and neighbour pixels lower than this threshold will be used in
7525 averaging. Luma component is left unchanged and is copied to output.
7526 Default value is 30. Allowed range is from 1 to 200.
7529 Set horizontal radius of rectangle used for averaging.
7530 Allowed range is from 1 to 100. Default value is 5.
7533 Set vertical radius of rectangle used for averaging.
7534 Allowed range is from 1 to 100. Default value is 5.
7537 Set horizontal step when averaging. Default value is 1.
7538 Allowed range is from 1 to 50.
7539 Mostly useful to speed-up filtering.
7542 Set vertical step when averaging. Default value is 1.
7543 Allowed range is from 1 to 50.
7544 Mostly useful to speed-up filtering.
7547 @subsection Commands
7548 This filter supports same @ref{commands} as options.
7549 The command accepts the same syntax of the corresponding option.
7551 @section chromashift
7552 Shift chroma pixels horizontally and/or vertically.
7554 The filter accepts the following options:
7557 Set amount to shift chroma-blue horizontally.
7559 Set amount to shift chroma-blue vertically.
7561 Set amount to shift chroma-red horizontally.
7563 Set amount to shift chroma-red vertically.
7565 Set edge mode, can be @var{smear}, default, or @var{warp}.
7568 @subsection Commands
7570 This filter supports the all above options as @ref{commands}.
7574 Display CIE color diagram with pixels overlaid onto it.
7576 The filter accepts the following options:
7591 @item uhdtv, rec2020
7605 Set what gamuts to draw.
7607 See @code{system} option for available values.
7610 Set ciescope size, by default set to 512.
7613 Set intensity used to map input pixel values to CIE diagram.
7616 Set contrast used to draw tongue colors that are out of active color system gamut.
7619 Correct gamma displayed on scope, by default enabled.
7622 Show white point on CIE diagram, by default disabled.
7625 Set input gamma. Used only with XYZ input color space.
7630 Visualize information exported by some codecs.
7632 Some codecs can export information through frames using side-data or other
7633 means. For example, some MPEG based codecs export motion vectors through the
7634 @var{export_mvs} flag in the codec @option{flags2} option.
7636 The filter accepts the following option:
7640 Set motion vectors to visualize.
7642 Available flags for @var{mv} are:
7646 forward predicted MVs of P-frames
7648 forward predicted MVs of B-frames
7650 backward predicted MVs of B-frames
7654 Display quantization parameters using the chroma planes.
7657 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7659 Available flags for @var{mv_type} are:
7663 forward predicted MVs
7665 backward predicted MVs
7668 @item frame_type, ft
7669 Set frame type to visualize motion vectors of.
7671 Available flags for @var{frame_type} are:
7675 intra-coded frames (I-frames)
7677 predicted frames (P-frames)
7679 bi-directionally predicted frames (B-frames)
7683 @subsection Examples
7687 Visualize forward predicted MVs of all frames using @command{ffplay}:
7689 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7693 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7695 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7699 @section colorbalance
7700 Modify intensity of primary colors (red, green and blue) of input frames.
7702 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7703 regions for the red-cyan, green-magenta or blue-yellow balance.
7705 A positive adjustment value shifts the balance towards the primary color, a negative
7706 value towards the complementary color.
7708 The filter accepts the following options:
7714 Adjust red, green and blue shadows (darkest pixels).
7719 Adjust red, green and blue midtones (medium pixels).
7724 Adjust red, green and blue highlights (brightest pixels).
7726 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7729 Preserve lightness when changing color balance. Default is disabled.
7732 @subsection Examples
7736 Add red color cast to shadows:
7742 @subsection Commands
7744 This filter supports the all above options as @ref{commands}.
7746 @section colorchannelmixer
7748 Adjust video input frames by re-mixing color channels.
7750 This filter modifies a color channel by adding the values associated to
7751 the other channels of the same pixels. For example if the value to
7752 modify is red, the output value will be:
7754 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
7757 The filter accepts the following options:
7764 Adjust contribution of input red, green, blue and alpha channels for output red channel.
7765 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
7771 Adjust contribution of input red, green, blue and alpha channels for output green channel.
7772 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
7778 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
7779 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
7785 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
7786 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
7788 Allowed ranges for options are @code{[-2.0, 2.0]}.
7791 @subsection Examples
7795 Convert source to grayscale:
7797 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
7800 Simulate sepia tones:
7802 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
7806 @subsection Commands
7808 This filter supports the all above options as @ref{commands}.
7811 RGB colorspace color keying.
7813 The filter accepts the following options:
7817 The color which will be replaced with transparency.
7820 Similarity percentage with the key color.
7822 0.01 matches only the exact key color, while 1.0 matches everything.
7827 0.0 makes pixels either fully transparent, or not transparent at all.
7829 Higher values result in semi-transparent pixels, with a higher transparency
7830 the more similar the pixels color is to the key color.
7833 @subsection Examples
7837 Make every green pixel in the input image transparent:
7839 ffmpeg -i input.png -vf colorkey=green out.png
7843 Overlay a greenscreen-video on top of a static background image.
7845 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
7849 @subsection Commands
7850 This filter supports same @ref{commands} as options.
7851 The command accepts the same syntax of the corresponding option.
7853 If the specified expression is not valid, it is kept at its current
7857 Remove all color information for all RGB colors except for certain one.
7859 The filter accepts the following options:
7863 The color which will not be replaced with neutral gray.
7866 Similarity percentage with the above color.
7867 0.01 matches only the exact key color, while 1.0 matches everything.
7870 Blend percentage. 0.0 makes pixels fully gray.
7871 Higher values result in more preserved color.
7874 @subsection Commands
7875 This filter supports same @ref{commands} as options.
7876 The command accepts the same syntax of the corresponding option.
7878 If the specified expression is not valid, it is kept at its current
7881 @section colorlevels
7883 Adjust video input frames using levels.
7885 The filter accepts the following options:
7892 Adjust red, green, blue and alpha input black point.
7893 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7899 Adjust red, green, blue and alpha input white point.
7900 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
7902 Input levels are used to lighten highlights (bright tones), darken shadows
7903 (dark tones), change the balance of bright and dark tones.
7909 Adjust red, green, blue and alpha output black point.
7910 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
7916 Adjust red, green, blue and alpha output white point.
7917 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
7919 Output levels allows manual selection of a constrained output level range.
7922 @subsection Examples
7926 Make video output darker:
7928 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
7934 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
7938 Make video output lighter:
7940 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
7944 Increase brightness:
7946 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
7950 @subsection Commands
7952 This filter supports the all above options as @ref{commands}.
7954 @section colormatrix
7956 Convert color matrix.
7958 The filter accepts the following options:
7963 Specify the source and destination color matrix. Both values must be
7966 The accepted values are:
7994 For example to convert from BT.601 to SMPTE-240M, use the command:
7996 colormatrix=bt601:smpte240m
8001 Convert colorspace, transfer characteristics or color primaries.
8002 Input video needs to have an even size.
8004 The filter accepts the following options:
8009 Specify all color properties at once.
8011 The accepted values are:
8041 Specify output colorspace.
8043 The accepted values are:
8052 BT.470BG or BT.601-6 625
8055 SMPTE-170M or BT.601-6 525
8064 BT.2020 with non-constant luminance
8070 Specify output transfer characteristics.
8072 The accepted values are:
8084 Constant gamma of 2.2
8087 Constant gamma of 2.8
8090 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8108 BT.2020 for 10-bits content
8111 BT.2020 for 12-bits content
8117 Specify output color primaries.
8119 The accepted values are:
8128 BT.470BG or BT.601-6 625
8131 SMPTE-170M or BT.601-6 525
8155 Specify output color range.
8157 The accepted values are:
8160 TV (restricted) range
8163 MPEG (restricted) range
8174 Specify output color format.
8176 The accepted values are:
8179 YUV 4:2:0 planar 8-bits
8182 YUV 4:2:0 planar 10-bits
8185 YUV 4:2:0 planar 12-bits
8188 YUV 4:2:2 planar 8-bits
8191 YUV 4:2:2 planar 10-bits
8194 YUV 4:2:2 planar 12-bits
8197 YUV 4:4:4 planar 8-bits
8200 YUV 4:4:4 planar 10-bits
8203 YUV 4:4:4 planar 12-bits
8208 Do a fast conversion, which skips gamma/primary correction. This will take
8209 significantly less CPU, but will be mathematically incorrect. To get output
8210 compatible with that produced by the colormatrix filter, use fast=1.
8213 Specify dithering mode.
8215 The accepted values are:
8221 Floyd-Steinberg dithering
8225 Whitepoint adaptation mode.
8227 The accepted values are:
8230 Bradford whitepoint adaptation
8233 von Kries whitepoint adaptation
8236 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8240 Override all input properties at once. Same accepted values as @ref{all}.
8243 Override input colorspace. Same accepted values as @ref{space}.
8246 Override input color primaries. Same accepted values as @ref{primaries}.
8249 Override input transfer characteristics. Same accepted values as @ref{trc}.
8252 Override input color range. Same accepted values as @ref{range}.
8256 The filter converts the transfer characteristics, color space and color
8257 primaries to the specified user values. The output value, if not specified,
8258 is set to a default value based on the "all" property. If that property is
8259 also not specified, the filter will log an error. The output color range and
8260 format default to the same value as the input color range and format. The
8261 input transfer characteristics, color space, color primaries and color range
8262 should be set on the input data. If any of these are missing, the filter will
8263 log an error and no conversion will take place.
8265 For example to convert the input to SMPTE-240M, use the command:
8267 colorspace=smpte240m
8270 @section convolution
8272 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8274 The filter accepts the following options:
8281 Set matrix for each plane.
8282 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8283 and from 1 to 49 odd number of signed integers in @var{row} mode.
8289 Set multiplier for calculated value for each plane.
8290 If unset or 0, it will be sum of all matrix elements.
8296 Set bias for each plane. This value is added to the result of the multiplication.
8297 Useful for making the overall image brighter or darker. Default is 0.0.
8303 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8304 Default is @var{square}.
8307 @subsection Examples
8313 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"
8319 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"
8325 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"
8331 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"
8335 Apply laplacian edge detector which includes diagonals:
8337 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"
8343 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"
8349 Apply 2D convolution of video stream in frequency domain using second stream
8352 The filter accepts the following options:
8356 Set which planes to process.
8359 Set which impulse video frames will be processed, can be @var{first}
8360 or @var{all}. Default is @var{all}.
8363 The @code{convolve} filter also supports the @ref{framesync} options.
8367 Copy the input video source unchanged to the output. This is mainly useful for
8372 Video filtering on GPU using Apple's CoreImage API on OSX.
8374 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8375 processed by video hardware. However, software-based OpenGL implementations
8376 exist which means there is no guarantee for hardware processing. It depends on
8379 There are many filters and image generators provided by Apple that come with a
8380 large variety of options. The filter has to be referenced by its name along
8383 The coreimage filter accepts the following options:
8386 List all available filters and generators along with all their respective
8387 options as well as possible minimum and maximum values along with the default
8394 Specify all filters by their respective name and options.
8395 Use @var{list_filters} to determine all valid filter names and options.
8396 Numerical options are specified by a float value and are automatically clamped
8397 to their respective value range. Vector and color options have to be specified
8398 by a list of space separated float values. Character escaping has to be done.
8399 A special option name @code{default} is available to use default options for a
8402 It is required to specify either @code{default} or at least one of the filter options.
8403 All omitted options are used with their default values.
8404 The syntax of the filter string is as follows:
8406 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8410 Specify a rectangle where the output of the filter chain is copied into the
8411 input image. It is given by a list of space separated float values:
8413 output_rect=x\ y\ width\ height
8415 If not given, the output rectangle equals the dimensions of the input image.
8416 The output rectangle is automatically cropped at the borders of the input
8417 image. Negative values are valid for each component.
8419 output_rect=25\ 25\ 100\ 100
8423 Several filters can be chained for successive processing without GPU-HOST
8424 transfers allowing for fast processing of complex filter chains.
8425 Currently, only filters with zero (generators) or exactly one (filters) input
8426 image and one output image are supported. Also, transition filters are not yet
8429 Some filters generate output images with additional padding depending on the
8430 respective filter kernel. The padding is automatically removed to ensure the
8431 filter output has the same size as the input image.
8433 For image generators, the size of the output image is determined by the
8434 previous output image of the filter chain or the input image of the whole
8435 filterchain, respectively. The generators do not use the pixel information of
8436 this image to generate their output. However, the generated output is
8437 blended onto this image, resulting in partial or complete coverage of the
8440 The @ref{coreimagesrc} video source can be used for generating input images
8441 which are directly fed into the filter chain. By using it, providing input
8442 images by another video source or an input video is not required.
8444 @subsection Examples
8449 List all filters available:
8451 coreimage=list_filters=true
8455 Use the CIBoxBlur filter with default options to blur an image:
8457 coreimage=filter=CIBoxBlur@@default
8461 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8462 its center at 100x100 and a radius of 50 pixels:
8464 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8468 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8469 given as complete and escaped command-line for Apple's standard bash shell:
8471 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8477 Cover a rectangular object
8479 It accepts the following options:
8483 Filepath of the optional cover image, needs to be in yuv420.
8488 It accepts the following values:
8491 cover it by the supplied image
8493 cover it by interpolating the surrounding pixels
8496 Default value is @var{blur}.
8499 @subsection Examples
8503 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8505 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8511 Crop the input video to given dimensions.
8513 It accepts the following parameters:
8517 The width of the output video. It defaults to @code{iw}.
8518 This expression is evaluated only once during the filter
8519 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8522 The height of the output video. It defaults to @code{ih}.
8523 This expression is evaluated only once during the filter
8524 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8527 The horizontal position, in the input video, of the left edge of the output
8528 video. It defaults to @code{(in_w-out_w)/2}.
8529 This expression is evaluated per-frame.
8532 The vertical position, in the input video, of the top edge of the output video.
8533 It defaults to @code{(in_h-out_h)/2}.
8534 This expression is evaluated per-frame.
8537 If set to 1 will force the output display aspect ratio
8538 to be the same of the input, by changing the output sample aspect
8539 ratio. It defaults to 0.
8542 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8543 width/height/x/y as specified and will not be rounded to nearest smaller value.
8547 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8548 expressions containing the following constants:
8553 The computed values for @var{x} and @var{y}. They are evaluated for
8558 The input width and height.
8562 These are the same as @var{in_w} and @var{in_h}.
8566 The output (cropped) width and height.
8570 These are the same as @var{out_w} and @var{out_h}.
8573 same as @var{iw} / @var{ih}
8576 input sample aspect ratio
8579 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8583 horizontal and vertical chroma subsample values. For example for the
8584 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8587 The number of the input frame, starting from 0.
8590 the position in the file of the input frame, NAN if unknown
8593 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8597 The expression for @var{out_w} may depend on the value of @var{out_h},
8598 and the expression for @var{out_h} may depend on @var{out_w}, but they
8599 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8600 evaluated after @var{out_w} and @var{out_h}.
8602 The @var{x} and @var{y} parameters specify the expressions for the
8603 position of the top-left corner of the output (non-cropped) area. They
8604 are evaluated for each frame. If the evaluated value is not valid, it
8605 is approximated to the nearest valid value.
8607 The expression for @var{x} may depend on @var{y}, and the expression
8608 for @var{y} may depend on @var{x}.
8610 @subsection Examples
8614 Crop area with size 100x100 at position (12,34).
8619 Using named options, the example above becomes:
8621 crop=w=100:h=100:x=12:y=34
8625 Crop the central input area with size 100x100:
8631 Crop the central input area with size 2/3 of the input video:
8633 crop=2/3*in_w:2/3*in_h
8637 Crop the input video central square:
8644 Delimit the rectangle with the top-left corner placed at position
8645 100:100 and the right-bottom corner corresponding to the right-bottom
8646 corner of the input image.
8648 crop=in_w-100:in_h-100:100:100
8652 Crop 10 pixels from the left and right borders, and 20 pixels from
8653 the top and bottom borders
8655 crop=in_w-2*10:in_h-2*20
8659 Keep only the bottom right quarter of the input image:
8661 crop=in_w/2:in_h/2:in_w/2:in_h/2
8665 Crop height for getting Greek harmony:
8667 crop=in_w:1/PHI*in_w
8671 Apply trembling effect:
8673 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)
8677 Apply erratic camera effect depending on timestamp:
8679 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)"
8683 Set x depending on the value of y:
8685 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8689 @subsection Commands
8691 This filter supports the following commands:
8697 Set width/height of the output video and the horizontal/vertical position
8699 The command accepts the same syntax of the corresponding option.
8701 If the specified expression is not valid, it is kept at its current
8707 Auto-detect the crop size.
8709 It calculates the necessary cropping parameters and prints the
8710 recommended parameters via the logging system. The detected dimensions
8711 correspond to the non-black area of the input video.
8713 It accepts the following parameters:
8718 Set higher black value threshold, which can be optionally specified
8719 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8720 value greater to the set value is considered non-black. It defaults to 24.
8721 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8722 on the bitdepth of the pixel format.
8725 The value which the width/height should be divisible by. It defaults to
8726 16. The offset is automatically adjusted to center the video. Use 2 to
8727 get only even dimensions (needed for 4:2:2 video). 16 is best when
8728 encoding to most video codecs.
8730 @item reset_count, reset
8731 Set the counter that determines after how many frames cropdetect will
8732 reset the previously detected largest video area and start over to
8733 detect the current optimal crop area. Default value is 0.
8735 This can be useful when channel logos distort the video area. 0
8736 indicates 'never reset', and returns the largest area encountered during
8743 Delay video filtering until a given wallclock timestamp. The filter first
8744 passes on @option{preroll} amount of frames, then it buffers at most
8745 @option{buffer} amount of frames and waits for the cue. After reaching the cue
8746 it forwards the buffered frames and also any subsequent frames coming in its
8749 The filter can be used synchronize the output of multiple ffmpeg processes for
8750 realtime output devices like decklink. By putting the delay in the filtering
8751 chain and pre-buffering frames the process can pass on data to output almost
8752 immediately after the target wallclock timestamp is reached.
8754 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
8760 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
8763 The duration of content to pass on as preroll expressed in seconds. Default is 0.
8766 The maximum duration of content to buffer before waiting for the cue expressed
8767 in seconds. Default is 0.
8774 Apply color adjustments using curves.
8776 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
8777 component (red, green and blue) has its values defined by @var{N} key points
8778 tied from each other using a smooth curve. The x-axis represents the pixel
8779 values from the input frame, and the y-axis the new pixel values to be set for
8782 By default, a component curve is defined by the two points @var{(0;0)} and
8783 @var{(1;1)}. This creates a straight line where each original pixel value is
8784 "adjusted" to its own value, which means no change to the image.
8786 The filter allows you to redefine these two points and add some more. A new
8787 curve (using a natural cubic spline interpolation) will be define to pass
8788 smoothly through all these new coordinates. The new defined points needs to be
8789 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
8790 be in the @var{[0;1]} interval. If the computed curves happened to go outside
8791 the vector spaces, the values will be clipped accordingly.
8793 The filter accepts the following options:
8797 Select one of the available color presets. This option can be used in addition
8798 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
8799 options takes priority on the preset values.
8800 Available presets are:
8803 @item color_negative
8806 @item increase_contrast
8808 @item linear_contrast
8809 @item medium_contrast
8811 @item strong_contrast
8814 Default is @code{none}.
8816 Set the master key points. These points will define a second pass mapping. It
8817 is sometimes called a "luminance" or "value" mapping. It can be used with
8818 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
8819 post-processing LUT.
8821 Set the key points for the red component.
8823 Set the key points for the green component.
8825 Set the key points for the blue component.
8827 Set the key points for all components (not including master).
8828 Can be used in addition to the other key points component
8829 options. In this case, the unset component(s) will fallback on this
8830 @option{all} setting.
8832 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
8834 Save Gnuplot script of the curves in specified file.
8837 To avoid some filtergraph syntax conflicts, each key points list need to be
8838 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
8840 @subsection Examples
8844 Increase slightly the middle level of blue:
8846 curves=blue='0/0 0.5/0.58 1/1'
8852 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'
8854 Here we obtain the following coordinates for each components:
8857 @code{(0;0.11) (0.42;0.51) (1;0.95)}
8859 @code{(0;0) (0.50;0.48) (1;1)}
8861 @code{(0;0.22) (0.49;0.44) (1;0.80)}
8865 The previous example can also be achieved with the associated built-in preset:
8867 curves=preset=vintage
8877 Use a Photoshop preset and redefine the points of the green component:
8879 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
8883 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
8884 and @command{gnuplot}:
8886 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
8887 gnuplot -p /tmp/curves.plt
8893 Video data analysis filter.
8895 This filter shows hexadecimal pixel values of part of video.
8897 The filter accepts the following options:
8901 Set output video size.
8904 Set x offset from where to pick pixels.
8907 Set y offset from where to pick pixels.
8910 Set scope mode, can be one of the following:
8913 Draw hexadecimal pixel values with white color on black background.
8916 Draw hexadecimal pixel values with input video pixel color on black
8920 Draw hexadecimal pixel values on color background picked from input video,
8921 the text color is picked in such way so its always visible.
8925 Draw rows and columns numbers on left and top of video.
8928 Set background opacity.
8931 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
8935 Apply Directional blur filter.
8937 The filter accepts the following options:
8941 Set angle of directional blur. Default is @code{45}.
8944 Set radius of directional blur. Default is @code{5}.
8947 Set which planes to filter. By default all planes are filtered.
8950 @subsection Commands
8951 This filter supports same @ref{commands} as options.
8952 The command accepts the same syntax of the corresponding option.
8954 If the specified expression is not valid, it is kept at its current
8959 Denoise frames using 2D DCT (frequency domain filtering).
8961 This filter is not designed for real time.
8963 The filter accepts the following options:
8967 Set the noise sigma constant.
8969 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
8970 coefficient (absolute value) below this threshold with be dropped.
8972 If you need a more advanced filtering, see @option{expr}.
8974 Default is @code{0}.
8977 Set number overlapping pixels for each block. Since the filter can be slow, you
8978 may want to reduce this value, at the cost of a less effective filter and the
8979 risk of various artefacts.
8981 If the overlapping value doesn't permit processing the whole input width or
8982 height, a warning will be displayed and according borders won't be denoised.
8984 Default value is @var{blocksize}-1, which is the best possible setting.
8987 Set the coefficient factor expression.
8989 For each coefficient of a DCT block, this expression will be evaluated as a
8990 multiplier value for the coefficient.
8992 If this is option is set, the @option{sigma} option will be ignored.
8994 The absolute value of the coefficient can be accessed through the @var{c}
8998 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
8999 @var{blocksize}, which is the width and height of the processed blocks.
9001 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9002 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9003 on the speed processing. Also, a larger block size does not necessarily means a
9007 @subsection Examples
9009 Apply a denoise with a @option{sigma} of @code{4.5}:
9014 The same operation can be achieved using the expression system:
9016 dctdnoiz=e='gte(c, 4.5*3)'
9019 Violent denoise using a block size of @code{16x16}:
9026 Remove banding artifacts from input video.
9027 It works by replacing banded pixels with average value of referenced pixels.
9029 The filter accepts the following options:
9036 Set banding detection threshold for each plane. Default is 0.02.
9037 Valid range is 0.00003 to 0.5.
9038 If difference between current pixel and reference pixel is less than threshold,
9039 it will be considered as banded.
9042 Banding detection range in pixels. Default is 16. If positive, random number
9043 in range 0 to set value will be used. If negative, exact absolute value
9045 The range defines square of four pixels around current pixel.
9048 Set direction in radians from which four pixel will be compared. If positive,
9049 random direction from 0 to set direction will be picked. If negative, exact of
9050 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9051 will pick only pixels on same row and -PI/2 will pick only pixels on same
9055 If enabled, current pixel is compared with average value of all four
9056 surrounding pixels. The default is enabled. If disabled current pixel is
9057 compared with all four surrounding pixels. The pixel is considered banded
9058 if only all four differences with surrounding pixels are less than threshold.
9061 If enabled, current pixel is changed if and only if all pixel components are banded,
9062 e.g. banding detection threshold is triggered for all color components.
9063 The default is disabled.
9068 Remove blocking artifacts from input video.
9070 The filter accepts the following options:
9074 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9075 This controls what kind of deblocking is applied.
9078 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9084 Set blocking detection thresholds. Allowed range is 0 to 1.
9085 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9086 Using higher threshold gives more deblocking strength.
9087 Setting @var{alpha} controls threshold detection at exact edge of block.
9088 Remaining options controls threshold detection near the edge. Each one for
9089 below/above or left/right. Setting any of those to @var{0} disables
9093 Set planes to filter. Default is to filter all available planes.
9096 @subsection Examples
9100 Deblock using weak filter and block size of 4 pixels.
9102 deblock=filter=weak:block=4
9106 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9107 deblocking more edges.
9109 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9113 Similar as above, but filter only first plane.
9115 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9119 Similar as above, but filter only second and third plane.
9121 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9128 Drop duplicated frames at regular intervals.
9130 The filter accepts the following options:
9134 Set the number of frames from which one will be dropped. Setting this to
9135 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9136 Default is @code{5}.
9139 Set the threshold for duplicate detection. If the difference metric for a frame
9140 is less than or equal to this value, then it is declared as duplicate. Default
9144 Set scene change threshold. Default is @code{15}.
9148 Set the size of the x and y-axis blocks used during metric calculations.
9149 Larger blocks give better noise suppression, but also give worse detection of
9150 small movements. Must be a power of two. Default is @code{32}.
9153 Mark main input as a pre-processed input and activate clean source input
9154 stream. This allows the input to be pre-processed with various filters to help
9155 the metrics calculation while keeping the frame selection lossless. When set to
9156 @code{1}, the first stream is for the pre-processed input, and the second
9157 stream is the clean source from where the kept frames are chosen. Default is
9161 Set whether or not chroma is considered in the metric calculations. Default is
9167 Apply 2D deconvolution of video stream in frequency domain using second stream
9170 The filter accepts the following options:
9174 Set which planes to process.
9177 Set which impulse video frames will be processed, can be @var{first}
9178 or @var{all}. Default is @var{all}.
9181 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9182 and height are not same and not power of 2 or if stream prior to convolving
9186 The @code{deconvolve} filter also supports the @ref{framesync} options.
9190 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9192 It accepts the following options:
9196 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9197 @var{rainbows} for cross-color reduction.
9200 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9203 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9206 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9209 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9214 Apply deflate effect to the video.
9216 This filter replaces the pixel by the local(3x3) average by taking into account
9217 only values lower than the pixel.
9219 It accepts the following options:
9226 Limit the maximum change for each plane, default is 65535.
9227 If 0, plane will remain unchanged.
9230 @subsection Commands
9232 This filter supports the all above options as @ref{commands}.
9236 Remove temporal frame luminance variations.
9238 It accepts the following options:
9242 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9245 Set averaging mode to smooth temporal luminance variations.
9247 Available values are:
9272 Do not actually modify frame. Useful when one only wants metadata.
9277 Remove judder produced by partially interlaced telecined content.
9279 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9280 source was partially telecined content then the output of @code{pullup,dejudder}
9281 will have a variable frame rate. May change the recorded frame rate of the
9282 container. Aside from that change, this filter will not affect constant frame
9285 The option available in this filter is:
9289 Specify the length of the window over which the judder repeats.
9291 Accepts any integer greater than 1. Useful values are:
9295 If the original was telecined from 24 to 30 fps (Film to NTSC).
9298 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9301 If a mixture of the two.
9304 The default is @samp{4}.
9309 Suppress a TV station logo by a simple interpolation of the surrounding
9310 pixels. Just set a rectangle covering the logo and watch it disappear
9311 (and sometimes something even uglier appear - your mileage may vary).
9313 It accepts the following parameters:
9318 Specify the top left corner coordinates of the logo. They must be
9323 Specify the width and height of the logo to clear. They must be
9327 Specify the thickness of the fuzzy edge of the rectangle (added to
9328 @var{w} and @var{h}). The default value is 1. This option is
9329 deprecated, setting higher values should no longer be necessary and
9333 When set to 1, a green rectangle is drawn on the screen to simplify
9334 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9335 The default value is 0.
9337 The rectangle is drawn on the outermost pixels which will be (partly)
9338 replaced with interpolated values. The values of the next pixels
9339 immediately outside this rectangle in each direction will be used to
9340 compute the interpolated pixel values inside the rectangle.
9344 @subsection Examples
9348 Set a rectangle covering the area with top left corner coordinates 0,0
9349 and size 100x77, and a band of size 10:
9351 delogo=x=0:y=0:w=100:h=77:band=10
9359 Remove the rain in the input image/video by applying the derain methods based on
9360 convolutional neural networks. Supported models:
9364 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9365 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9368 Training as well as model generation scripts are provided in
9369 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9371 Native model files (.model) can be generated from TensorFlow model
9372 files (.pb) by using tools/python/convert.py
9374 The filter accepts the following options:
9378 Specify which filter to use. This option accepts the following values:
9382 Derain filter. To conduct derain filter, you need to use a derain model.
9385 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9387 Default value is @samp{derain}.
9390 Specify which DNN backend to use for model loading and execution. This option accepts
9391 the following values:
9395 Native implementation of DNN loading and execution.
9398 TensorFlow backend. To enable this backend you
9399 need to install the TensorFlow for C library (see
9400 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9401 @code{--enable-libtensorflow}
9403 Default value is @samp{native}.
9406 Set path to model file specifying network architecture and its parameters.
9407 Note that different backends use different file formats. TensorFlow and native
9408 backend can load files for only its format.
9411 It can also be finished with @ref{dnn_processing} filter.
9415 Attempt to fix small changes in horizontal and/or vertical shift. This
9416 filter helps remove camera shake from hand-holding a camera, bumping a
9417 tripod, moving on a vehicle, etc.
9419 The filter accepts the following options:
9427 Specify a rectangular area where to limit the search for motion
9429 If desired the search for motion vectors can be limited to a
9430 rectangular area of the frame defined by its top left corner, width
9431 and height. These parameters have the same meaning as the drawbox
9432 filter which can be used to visualise the position of the bounding
9435 This is useful when simultaneous movement of subjects within the frame
9436 might be confused for camera motion by the motion vector search.
9438 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9439 then the full frame is used. This allows later options to be set
9440 without specifying the bounding box for the motion vector search.
9442 Default - search the whole frame.
9446 Specify the maximum extent of movement in x and y directions in the
9447 range 0-64 pixels. Default 16.
9450 Specify how to generate pixels to fill blanks at the edge of the
9451 frame. Available values are:
9454 Fill zeroes at blank locations
9456 Original image at blank locations
9458 Extruded edge value at blank locations
9460 Mirrored edge at blank locations
9462 Default value is @samp{mirror}.
9465 Specify the blocksize to use for motion search. Range 4-128 pixels,
9469 Specify the contrast threshold for blocks. Only blocks with more than
9470 the specified contrast (difference between darkest and lightest
9471 pixels) will be considered. Range 1-255, default 125.
9474 Specify the search strategy. Available values are:
9477 Set exhaustive search
9479 Set less exhaustive search.
9481 Default value is @samp{exhaustive}.
9484 If set then a detailed log of the motion search is written to the
9491 Remove unwanted contamination of foreground colors, caused by reflected color of
9492 greenscreen or bluescreen.
9494 This filter accepts the following options:
9498 Set what type of despill to use.
9501 Set how spillmap will be generated.
9504 Set how much to get rid of still remaining spill.
9507 Controls amount of red in spill area.
9510 Controls amount of green in spill area.
9511 Should be -1 for greenscreen.
9514 Controls amount of blue in spill area.
9515 Should be -1 for bluescreen.
9518 Controls brightness of spill area, preserving colors.
9521 Modify alpha from generated spillmap.
9524 @subsection Commands
9526 This filter supports the all above options as @ref{commands}.
9530 Apply an exact inverse of the telecine operation. It requires a predefined
9531 pattern specified using the pattern option which must be the same as that passed
9532 to the telecine filter.
9534 This filter accepts the following options:
9543 The default value is @code{top}.
9547 A string of numbers representing the pulldown pattern you wish to apply.
9548 The default value is @code{23}.
9551 A number representing position of the first frame with respect to the telecine
9552 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9557 Apply dilation effect to the video.
9559 This filter replaces the pixel by the local(3x3) maximum.
9561 It accepts the following options:
9568 Limit the maximum change for each plane, default is 65535.
9569 If 0, plane will remain unchanged.
9572 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9575 Flags to local 3x3 coordinates maps like this:
9582 @subsection Commands
9584 This filter supports the all above options as @ref{commands}.
9588 Displace pixels as indicated by second and third input stream.
9590 It takes three input streams and outputs one stream, the first input is the
9591 source, and second and third input are displacement maps.
9593 The second input specifies how much to displace pixels along the
9594 x-axis, while the third input specifies how much to displace pixels
9596 If one of displacement map streams terminates, last frame from that
9597 displacement map will be used.
9599 Note that once generated, displacements maps can be reused over and over again.
9601 A description of the accepted options follows.
9605 Set displace behavior for pixels that are out of range.
9607 Available values are:
9610 Missing pixels are replaced by black pixels.
9613 Adjacent pixels will spread out to replace missing pixels.
9616 Out of range pixels are wrapped so they point to pixels of other side.
9619 Out of range pixels will be replaced with mirrored pixels.
9621 Default is @samp{smear}.
9625 @subsection Examples
9629 Add ripple effect to rgb input of video size hd720:
9631 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
9635 Add wave effect to rgb input of video size hd720:
9637 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
9641 @anchor{dnn_processing}
9642 @section dnn_processing
9644 Do image processing with deep neural networks. It works together with another filter
9645 which converts the pixel format of the Frame to what the dnn network requires.
9647 The filter accepts the following options:
9651 Specify which DNN backend to use for model loading and execution. This option accepts
9652 the following values:
9656 Native implementation of DNN loading and execution.
9659 TensorFlow backend. To enable this backend you
9660 need to install the TensorFlow for C library (see
9661 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9662 @code{--enable-libtensorflow}
9665 OpenVINO backend. To enable this backend you
9666 need to build and install the OpenVINO for C library (see
9667 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9668 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9669 be needed if the header files and libraries are not installed into system path)
9673 Default value is @samp{native}.
9676 Set path to model file specifying network architecture and its parameters.
9677 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9678 backend can load files for only its format.
9680 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9683 Set the input name of the dnn network.
9686 Set the output name of the dnn network.
9690 @subsection Examples
9694 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9696 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9700 Halve the pixel value of the frame with format gray32f:
9702 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
9706 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9708 ./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
9712 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9714 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9721 Draw a colored box on the input image.
9723 It accepts the following parameters:
9728 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
9732 The expressions which specify the width and height of the box; if 0 they are interpreted as
9733 the input width and height. It defaults to 0.
9736 Specify the color of the box to write. For the general syntax of this option,
9737 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9738 value @code{invert} is used, the box edge color is the same as the
9739 video with inverted luma.
9742 The expression which sets the thickness of the box edge.
9743 A value of @code{fill} will create a filled box. Default value is @code{3}.
9745 See below for the list of accepted constants.
9748 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
9749 will overwrite the video's color and alpha pixels.
9750 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
9753 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9754 following constants:
9758 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9762 horizontal and vertical chroma subsample values. For example for the
9763 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9767 The input width and height.
9770 The input sample aspect ratio.
9774 The x and y offset coordinates where the box is drawn.
9778 The width and height of the drawn box.
9781 The thickness of the drawn box.
9783 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
9784 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
9788 @subsection Examples
9792 Draw a black box around the edge of the input image:
9798 Draw a box with color red and an opacity of 50%:
9800 drawbox=10:20:200:60:red@@0.5
9803 The previous example can be specified as:
9805 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
9809 Fill the box with pink color:
9811 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
9815 Draw a 2-pixel red 2.40:1 mask:
9817 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
9821 @subsection Commands
9822 This filter supports same commands as options.
9823 The command accepts the same syntax of the corresponding option.
9825 If the specified expression is not valid, it is kept at its current
9830 Draw a graph using input video metadata.
9832 It accepts the following parameters:
9836 Set 1st frame metadata key from which metadata values will be used to draw a graph.
9839 Set 1st foreground color expression.
9842 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
9845 Set 2nd foreground color expression.
9848 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
9851 Set 3rd foreground color expression.
9854 Set 4th frame metadata key from which metadata values will be used to draw a graph.
9857 Set 4th foreground color expression.
9860 Set minimal value of metadata value.
9863 Set maximal value of metadata value.
9866 Set graph background color. Default is white.
9871 Available values for mode is:
9878 Default is @code{line}.
9883 Available values for slide is:
9886 Draw new frame when right border is reached.
9889 Replace old columns with new ones.
9892 Scroll from right to left.
9895 Scroll from left to right.
9898 Draw single picture.
9901 Default is @code{frame}.
9904 Set size of graph video. For the syntax of this option, check the
9905 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
9906 The default value is @code{900x256}.
9909 Set the output frame rate. Default value is @code{25}.
9911 The foreground color expressions can use the following variables:
9914 Minimal value of metadata value.
9917 Maximal value of metadata value.
9920 Current metadata key value.
9923 The color is defined as 0xAABBGGRR.
9926 Example using metadata from @ref{signalstats} filter:
9928 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
9931 Example using metadata from @ref{ebur128} filter:
9933 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
9938 Draw a grid on the input image.
9940 It accepts the following parameters:
9945 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
9949 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
9950 input width and height, respectively, minus @code{thickness}, so image gets
9951 framed. Default to 0.
9954 Specify the color of the grid. For the general syntax of this option,
9955 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
9956 value @code{invert} is used, the grid color is the same as the
9957 video with inverted luma.
9960 The expression which sets the thickness of the grid line. Default value is @code{1}.
9962 See below for the list of accepted constants.
9965 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
9966 will overwrite the video's color and alpha pixels.
9967 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
9970 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
9971 following constants:
9975 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
9979 horizontal and vertical chroma subsample values. For example for the
9980 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
9984 The input grid cell width and height.
9987 The input sample aspect ratio.
9991 The x and y coordinates of some point of grid intersection (meant to configure offset).
9995 The width and height of the drawn cell.
9998 The thickness of the drawn cell.
10000 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10001 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10005 @subsection Examples
10009 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10011 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10015 Draw a white 3x3 grid with an opacity of 50%:
10017 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10021 @subsection Commands
10022 This filter supports same commands as options.
10023 The command accepts the same syntax of the corresponding option.
10025 If the specified expression is not valid, it is kept at its current
10031 Draw a text string or text from a specified file on top of a video, using the
10032 libfreetype library.
10034 To enable compilation of this filter, you need to configure FFmpeg with
10035 @code{--enable-libfreetype}.
10036 To enable default font fallback and the @var{font} option you need to
10037 configure FFmpeg with @code{--enable-libfontconfig}.
10038 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10039 @code{--enable-libfribidi}.
10043 It accepts the following parameters:
10048 Used to draw a box around text using the background color.
10049 The value must be either 1 (enable) or 0 (disable).
10050 The default value of @var{box} is 0.
10053 Set the width of the border to be drawn around the box using @var{boxcolor}.
10054 The default value of @var{boxborderw} is 0.
10057 The color to be used for drawing box around text. For the syntax of this
10058 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10060 The default value of @var{boxcolor} is "white".
10063 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10064 The default value of @var{line_spacing} is 0.
10067 Set the width of the border to be drawn around the text using @var{bordercolor}.
10068 The default value of @var{borderw} is 0.
10071 Set the color to be used for drawing border around text. For the syntax of this
10072 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10074 The default value of @var{bordercolor} is "black".
10077 Select how the @var{text} is expanded. Can be either @code{none},
10078 @code{strftime} (deprecated) or
10079 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10083 Set a start time for the count. Value is in microseconds. Only applied
10084 in the deprecated strftime expansion mode. To emulate in normal expansion
10085 mode use the @code{pts} function, supplying the start time (in seconds)
10086 as the second argument.
10089 If true, check and fix text coords to avoid clipping.
10092 The color to be used for drawing fonts. For the syntax of this option, check
10093 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10095 The default value of @var{fontcolor} is "black".
10097 @item fontcolor_expr
10098 String which is expanded the same way as @var{text} to obtain dynamic
10099 @var{fontcolor} value. By default this option has empty value and is not
10100 processed. When this option is set, it overrides @var{fontcolor} option.
10103 The font family to be used for drawing text. By default Sans.
10106 The font file to be used for drawing text. The path must be included.
10107 This parameter is mandatory if the fontconfig support is disabled.
10110 Draw the text applying alpha blending. The value can
10111 be a number between 0.0 and 1.0.
10112 The expression accepts the same variables @var{x, y} as well.
10113 The default value is 1.
10114 Please see @var{fontcolor_expr}.
10117 The font size to be used for drawing text.
10118 The default value of @var{fontsize} is 16.
10121 If set to 1, attempt to shape the text (for example, reverse the order of
10122 right-to-left text and join Arabic characters) before drawing it.
10123 Otherwise, just draw the text exactly as given.
10124 By default 1 (if supported).
10126 @item ft_load_flags
10127 The flags to be used for loading the fonts.
10129 The flags map the corresponding flags supported by libfreetype, and are
10130 a combination of the following values:
10137 @item vertical_layout
10138 @item force_autohint
10141 @item ignore_global_advance_width
10143 @item ignore_transform
10145 @item linear_design
10149 Default value is "default".
10151 For more information consult the documentation for the FT_LOAD_*
10155 The color to be used for drawing a shadow behind the drawn text. For the
10156 syntax of this option, check the @ref{color syntax,,"Color" section in the
10157 ffmpeg-utils manual,ffmpeg-utils}.
10159 The default value of @var{shadowcolor} is "black".
10163 The x and y offsets for the text shadow position with respect to the
10164 position of the text. They can be either positive or negative
10165 values. The default value for both is "0".
10168 The starting frame number for the n/frame_num variable. The default value
10172 The size in number of spaces to use for rendering the tab.
10173 Default value is 4.
10176 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10177 format. It can be used with or without text parameter. @var{timecode_rate}
10178 option must be specified.
10180 @item timecode_rate, rate, r
10181 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10182 integer. Minimum value is "1".
10183 Drop-frame timecode is supported for frame rates 30 & 60.
10186 If set to 1, the output of the timecode option will wrap around at 24 hours.
10187 Default is 0 (disabled).
10190 The text string to be drawn. The text must be a sequence of UTF-8
10191 encoded characters.
10192 This parameter is mandatory if no file is specified with the parameter
10196 A text file containing text to be drawn. The text must be a sequence
10197 of UTF-8 encoded characters.
10199 This parameter is mandatory if no text string is specified with the
10200 parameter @var{text}.
10202 If both @var{text} and @var{textfile} are specified, an error is thrown.
10205 If set to 1, the @var{textfile} will be reloaded before each frame.
10206 Be sure to update it atomically, or it may be read partially, or even fail.
10210 The expressions which specify the offsets where text will be drawn
10211 within the video frame. They are relative to the top/left border of the
10214 The default value of @var{x} and @var{y} is "0".
10216 See below for the list of accepted constants and functions.
10219 The parameters for @var{x} and @var{y} are expressions containing the
10220 following constants and functions:
10224 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10228 horizontal and vertical chroma subsample values. For example for the
10229 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10232 the height of each text line
10240 @item max_glyph_a, ascent
10241 the maximum distance from the baseline to the highest/upper grid
10242 coordinate used to place a glyph outline point, for all the rendered
10244 It is a positive value, due to the grid's orientation with the Y axis
10247 @item max_glyph_d, descent
10248 the maximum distance from the baseline to the lowest grid coordinate
10249 used to place a glyph outline point, for all the rendered glyphs.
10250 This is a negative value, due to the grid's orientation, with the Y axis
10254 maximum glyph height, that is the maximum height for all the glyphs
10255 contained in the rendered text, it is equivalent to @var{ascent} -
10259 maximum glyph width, that is the maximum width for all the glyphs
10260 contained in the rendered text
10263 the number of input frame, starting from 0
10265 @item rand(min, max)
10266 return a random number included between @var{min} and @var{max}
10269 The input sample aspect ratio.
10272 timestamp expressed in seconds, NAN if the input timestamp is unknown
10275 the height of the rendered text
10278 the width of the rendered text
10282 the x and y offset coordinates where the text is drawn.
10284 These parameters allow the @var{x} and @var{y} expressions to refer
10285 to each other, so you can for example specify @code{y=x/dar}.
10288 A one character description of the current frame's picture type.
10291 The current packet's position in the input file or stream
10292 (in bytes, from the start of the input). A value of -1 indicates
10293 this info is not available.
10296 The current packet's duration, in seconds.
10299 The current packet's size (in bytes).
10302 @anchor{drawtext_expansion}
10303 @subsection Text expansion
10305 If @option{expansion} is set to @code{strftime},
10306 the filter recognizes strftime() sequences in the provided text and
10307 expands them accordingly. Check the documentation of strftime(). This
10308 feature is deprecated.
10310 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10312 If @option{expansion} is set to @code{normal} (which is the default),
10313 the following expansion mechanism is used.
10315 The backslash character @samp{\}, followed by any character, always expands to
10316 the second character.
10318 Sequences of the form @code{%@{...@}} are expanded. The text between the
10319 braces is a function name, possibly followed by arguments separated by ':'.
10320 If the arguments contain special characters or delimiters (':' or '@}'),
10321 they should be escaped.
10323 Note that they probably must also be escaped as the value for the
10324 @option{text} option in the filter argument string and as the filter
10325 argument in the filtergraph description, and possibly also for the shell,
10326 that makes up to four levels of escaping; using a text file avoids these
10329 The following functions are available:
10334 The expression evaluation result.
10336 It must take one argument specifying the expression to be evaluated,
10337 which accepts the same constants and functions as the @var{x} and
10338 @var{y} values. Note that not all constants should be used, for
10339 example the text size is not known when evaluating the expression, so
10340 the constants @var{text_w} and @var{text_h} will have an undefined
10343 @item expr_int_format, eif
10344 Evaluate the expression's value and output as formatted integer.
10346 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10347 The second argument specifies the output format. Allowed values are @samp{x},
10348 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10349 @code{printf} function.
10350 The third parameter is optional and sets the number of positions taken by the output.
10351 It can be used to add padding with zeros from the left.
10354 The time at which the filter is running, expressed in UTC.
10355 It can accept an argument: a strftime() format string.
10358 The time at which the filter is running, expressed in the local time zone.
10359 It can accept an argument: a strftime() format string.
10362 Frame metadata. Takes one or two arguments.
10364 The first argument is mandatory and specifies the metadata key.
10366 The second argument is optional and specifies a default value, used when the
10367 metadata key is not found or empty.
10369 Available metadata can be identified by inspecting entries
10370 starting with TAG included within each frame section
10371 printed by running @code{ffprobe -show_frames}.
10373 String metadata generated in filters leading to
10374 the drawtext filter are also available.
10377 The frame number, starting from 0.
10380 A one character description of the current picture type.
10383 The timestamp of the current frame.
10384 It can take up to three arguments.
10386 The first argument is the format of the timestamp; it defaults to @code{flt}
10387 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10388 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10389 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10390 @code{localtime} stands for the timestamp of the frame formatted as
10391 local time zone time.
10393 The second argument is an offset added to the timestamp.
10395 If the format is set to @code{hms}, a third argument @code{24HH} may be
10396 supplied to present the hour part of the formatted timestamp in 24h format
10399 If the format is set to @code{localtime} or @code{gmtime},
10400 a third argument may be supplied: a strftime() format string.
10401 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10404 @subsection Commands
10406 This filter supports altering parameters via commands:
10409 Alter existing filter parameters.
10411 Syntax for the argument is the same as for filter invocation, e.g.
10414 fontsize=56:fontcolor=green:text='Hello World'
10417 Full filter invocation with sendcmd would look like this:
10420 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10424 If the entire argument can't be parsed or applied as valid values then the filter will
10425 continue with its existing parameters.
10427 @subsection Examples
10431 Draw "Test Text" with font FreeSerif, using the default values for the
10432 optional parameters.
10435 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10439 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10440 and y=50 (counting from the top-left corner of the screen), text is
10441 yellow with a red box around it. Both the text and the box have an
10445 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10446 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10449 Note that the double quotes are not necessary if spaces are not used
10450 within the parameter list.
10453 Show the text at the center of the video frame:
10455 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10459 Show the text at a random position, switching to a new position every 30 seconds:
10461 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)"
10465 Show a text line sliding from right to left in the last row of the video
10466 frame. The file @file{LONG_LINE} is assumed to contain a single line
10469 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10473 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10475 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10479 Draw a single green letter "g", at the center of the input video.
10480 The glyph baseline is placed at half screen height.
10482 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10486 Show text for 1 second every 3 seconds:
10488 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10492 Use fontconfig to set the font. Note that the colons need to be escaped.
10494 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10498 Draw "Test Text" with font size dependent on height of the video.
10500 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10504 Print the date of a real-time encoding (see strftime(3)):
10506 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10510 Show text fading in and out (appearing/disappearing):
10513 DS=1.0 # display start
10514 DE=10.0 # display end
10515 FID=1.5 # fade in duration
10516 FOD=5 # fade out duration
10517 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 @}"
10521 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10522 and the @option{fontsize} value are included in the @option{y} offset.
10524 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10525 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10529 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10530 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10531 must have option @option{-export_path_metadata 1} for the special metadata fields
10532 to be available for filters.
10534 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10539 For more information about libfreetype, check:
10540 @url{http://www.freetype.org/}.
10542 For more information about fontconfig, check:
10543 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10545 For more information about libfribidi, check:
10546 @url{http://fribidi.org/}.
10548 @section edgedetect
10550 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10552 The filter accepts the following options:
10557 Set low and high threshold values used by the Canny thresholding
10560 The high threshold selects the "strong" edge pixels, which are then
10561 connected through 8-connectivity with the "weak" edge pixels selected
10562 by the low threshold.
10564 @var{low} and @var{high} threshold values must be chosen in the range
10565 [0,1], and @var{low} should be lesser or equal to @var{high}.
10567 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10571 Define the drawing mode.
10575 Draw white/gray wires on black background.
10578 Mix the colors to create a paint/cartoon effect.
10581 Apply Canny edge detector on all selected planes.
10583 Default value is @var{wires}.
10586 Select planes for filtering. By default all available planes are filtered.
10589 @subsection Examples
10593 Standard edge detection with custom values for the hysteresis thresholding:
10595 edgedetect=low=0.1:high=0.4
10599 Painting effect without thresholding:
10601 edgedetect=mode=colormix:high=0
10607 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10609 For each input image, the filter will compute the optimal mapping from
10610 the input to the output given the codebook length, that is the number
10611 of distinct output colors.
10613 This filter accepts the following options.
10616 @item codebook_length, l
10617 Set codebook length. The value must be a positive integer, and
10618 represents the number of distinct output colors. Default value is 256.
10621 Set the maximum number of iterations to apply for computing the optimal
10622 mapping. The higher the value the better the result and the higher the
10623 computation time. Default value is 1.
10626 Set a random seed, must be an integer included between 0 and
10627 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10628 will try to use a good random seed on a best effort basis.
10631 Set pal8 output pixel format. This option does not work with codebook
10632 length greater than 256.
10637 Measure graylevel entropy in histogram of color channels of video frames.
10639 It accepts the following parameters:
10643 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10645 @var{diff} mode measures entropy of histogram delta values, absolute differences
10646 between neighbour histogram values.
10650 Set brightness, contrast, saturation and approximate gamma adjustment.
10652 The filter accepts the following options:
10656 Set the contrast expression. The value must be a float value in range
10657 @code{-1000.0} to @code{1000.0}. The default value is "1".
10660 Set the brightness expression. The value must be a float value in
10661 range @code{-1.0} to @code{1.0}. The default value is "0".
10664 Set the saturation expression. The value must be a float in
10665 range @code{0.0} to @code{3.0}. The default value is "1".
10668 Set the gamma expression. The value must be a float in range
10669 @code{0.1} to @code{10.0}. The default value is "1".
10672 Set the gamma expression for red. The value must be a float in
10673 range @code{0.1} to @code{10.0}. The default value is "1".
10676 Set the gamma expression for green. The value must be a float in range
10677 @code{0.1} to @code{10.0}. The default value is "1".
10680 Set the gamma expression for blue. The value must be a float in range
10681 @code{0.1} to @code{10.0}. The default value is "1".
10684 Set the gamma weight expression. It can be used to reduce the effect
10685 of a high gamma value on bright image areas, e.g. keep them from
10686 getting overamplified and just plain white. The value must be a float
10687 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10688 gamma correction all the way down while @code{1.0} leaves it at its
10689 full strength. Default is "1".
10692 Set when the expressions for brightness, contrast, saturation and
10693 gamma expressions are evaluated.
10695 It accepts the following values:
10698 only evaluate expressions once during the filter initialization or
10699 when a command is processed
10702 evaluate expressions for each incoming frame
10705 Default value is @samp{init}.
10708 The expressions accept the following parameters:
10711 frame count of the input frame starting from 0
10714 byte position of the corresponding packet in the input file, NAN if
10718 frame rate of the input video, NAN if the input frame rate is unknown
10721 timestamp expressed in seconds, NAN if the input timestamp is unknown
10724 @subsection Commands
10725 The filter supports the following commands:
10729 Set the contrast expression.
10732 Set the brightness expression.
10735 Set the saturation expression.
10738 Set the gamma expression.
10741 Set the gamma_r expression.
10744 Set gamma_g expression.
10747 Set gamma_b expression.
10750 Set gamma_weight expression.
10752 The command accepts the same syntax of the corresponding option.
10754 If the specified expression is not valid, it is kept at its current
10761 Apply erosion effect to the video.
10763 This filter replaces the pixel by the local(3x3) minimum.
10765 It accepts the following options:
10772 Limit the maximum change for each plane, default is 65535.
10773 If 0, plane will remain unchanged.
10776 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
10779 Flags to local 3x3 coordinates maps like this:
10786 @subsection Commands
10788 This filter supports the all above options as @ref{commands}.
10790 @section extractplanes
10792 Extract color channel components from input video stream into
10793 separate grayscale video streams.
10795 The filter accepts the following option:
10799 Set plane(s) to extract.
10801 Available values for planes are:
10812 Choosing planes not available in the input will result in an error.
10813 That means you cannot select @code{r}, @code{g}, @code{b} planes
10814 with @code{y}, @code{u}, @code{v} planes at same time.
10817 @subsection Examples
10821 Extract luma, u and v color channel component from input video frame
10822 into 3 grayscale outputs:
10824 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
10830 Apply a fade-in/out effect to the input video.
10832 It accepts the following parameters:
10836 The effect type can be either "in" for a fade-in, or "out" for a fade-out
10838 Default is @code{in}.
10840 @item start_frame, s
10841 Specify the number of the frame to start applying the fade
10842 effect at. Default is 0.
10845 The number of frames that the fade effect lasts. At the end of the
10846 fade-in effect, the output video will have the same intensity as the input video.
10847 At the end of the fade-out transition, the output video will be filled with the
10848 selected @option{color}.
10852 If set to 1, fade only alpha channel, if one exists on the input.
10853 Default value is 0.
10855 @item start_time, st
10856 Specify the timestamp (in seconds) of the frame to start to apply the fade
10857 effect. If both start_frame and start_time are specified, the fade will start at
10858 whichever comes last. Default is 0.
10861 The number of seconds for which the fade effect has to last. At the end of the
10862 fade-in effect the output video will have the same intensity as the input video,
10863 at the end of the fade-out transition the output video will be filled with the
10864 selected @option{color}.
10865 If both duration and nb_frames are specified, duration is used. Default is 0
10866 (nb_frames is used by default).
10869 Specify the color of the fade. Default is "black".
10872 @subsection Examples
10876 Fade in the first 30 frames of video:
10881 The command above is equivalent to:
10887 Fade out the last 45 frames of a 200-frame video:
10890 fade=type=out:start_frame=155:nb_frames=45
10894 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
10896 fade=in:0:25, fade=out:975:25
10900 Make the first 5 frames yellow, then fade in from frame 5-24:
10902 fade=in:5:20:color=yellow
10906 Fade in alpha over first 25 frames of video:
10908 fade=in:0:25:alpha=1
10912 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
10914 fade=t=in:st=5.5:d=0.5
10920 Denoise frames using 3D FFT (frequency domain filtering).
10922 The filter accepts the following options:
10926 Set the noise sigma constant. This sets denoising strength.
10927 Default value is 1. Allowed range is from 0 to 30.
10928 Using very high sigma with low overlap may give blocking artifacts.
10931 Set amount of denoising. By default all detected noise is reduced.
10932 Default value is 1. Allowed range is from 0 to 1.
10935 Set size of block, Default is 4, can be 3, 4, 5 or 6.
10936 Actual size of block in pixels is 2 to power of @var{block}, so by default
10937 block size in pixels is 2^4 which is 16.
10940 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
10943 Set number of previous frames to use for denoising. By default is set to 0.
10946 Set number of next frames to to use for denoising. By default is set to 0.
10949 Set planes which will be filtered, by default are all available filtered
10954 Apply arbitrary expressions to samples in frequency domain
10958 Adjust the dc value (gain) of the luma plane of the image. The filter
10959 accepts an integer value in range @code{0} to @code{1000}. The default
10960 value is set to @code{0}.
10963 Adjust the dc value (gain) of the 1st chroma plane of the image. The
10964 filter accepts an integer value in range @code{0} to @code{1000}. The
10965 default value is set to @code{0}.
10968 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
10969 filter accepts an integer value in range @code{0} to @code{1000}. The
10970 default value is set to @code{0}.
10973 Set the frequency domain weight expression for the luma plane.
10976 Set the frequency domain weight expression for the 1st chroma plane.
10979 Set the frequency domain weight expression for the 2nd chroma plane.
10982 Set when the expressions are evaluated.
10984 It accepts the following values:
10987 Only evaluate expressions once during the filter initialization.
10990 Evaluate expressions for each incoming frame.
10993 Default value is @samp{init}.
10995 The filter accepts the following variables:
10998 The coordinates of the current sample.
11002 The width and height of the image.
11005 The number of input frame, starting from 0.
11008 @subsection Examples
11014 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11020 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11026 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11032 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11039 Extract a single field from an interlaced image using stride
11040 arithmetic to avoid wasting CPU time. The output frames are marked as
11043 The filter accepts the following options:
11047 Specify whether to extract the top (if the value is @code{0} or
11048 @code{top}) or the bottom field (if the value is @code{1} or
11054 Create new frames by copying the top and bottom fields from surrounding frames
11055 supplied as numbers by the hint file.
11059 Set file containing hints: absolute/relative frame numbers.
11061 There must be one line for each frame in a clip. Each line must contain two
11062 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11063 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11064 is current frame number for @code{absolute} mode or out of [-1, 1] range
11065 for @code{relative} mode. First number tells from which frame to pick up top
11066 field and second number tells from which frame to pick up bottom field.
11068 If optionally followed by @code{+} output frame will be marked as interlaced,
11069 else if followed by @code{-} output frame will be marked as progressive, else
11070 it will be marked same as input frame.
11071 If optionally followed by @code{t} output frame will use only top field, or in
11072 case of @code{b} it will use only bottom field.
11073 If line starts with @code{#} or @code{;} that line is skipped.
11076 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11079 Example of first several lines of @code{hint} file for @code{relative} mode:
11081 0,0 - # first frame
11082 1,0 - # second frame, use third's frame top field and second's frame bottom field
11083 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11098 @section fieldmatch
11100 Field matching filter for inverse telecine. It is meant to reconstruct the
11101 progressive frames from a telecined stream. The filter does not drop duplicated
11102 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11103 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11105 The separation of the field matching and the decimation is notably motivated by
11106 the possibility of inserting a de-interlacing filter fallback between the two.
11107 If the source has mixed telecined and real interlaced content,
11108 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11109 But these remaining combed frames will be marked as interlaced, and thus can be
11110 de-interlaced by a later filter such as @ref{yadif} before decimation.
11112 In addition to the various configuration options, @code{fieldmatch} can take an
11113 optional second stream, activated through the @option{ppsrc} option. If
11114 enabled, the frames reconstruction will be based on the fields and frames from
11115 this second stream. This allows the first input to be pre-processed in order to
11116 help the various algorithms of the filter, while keeping the output lossless
11117 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11118 or brightness/contrast adjustments can help.
11120 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11121 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11122 which @code{fieldmatch} is based on. While the semantic and usage are very
11123 close, some behaviour and options names can differ.
11125 The @ref{decimate} filter currently only works for constant frame rate input.
11126 If your input has mixed telecined (30fps) and progressive content with a lower
11127 framerate like 24fps use the following filterchain to produce the necessary cfr
11128 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11130 The filter accepts the following options:
11134 Specify the assumed field order of the input stream. Available values are:
11138 Auto detect parity (use FFmpeg's internal parity value).
11140 Assume bottom field first.
11142 Assume top field first.
11145 Note that it is sometimes recommended not to trust the parity announced by the
11148 Default value is @var{auto}.
11151 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11152 sense that it won't risk creating jerkiness due to duplicate frames when
11153 possible, but if there are bad edits or blended fields it will end up
11154 outputting combed frames when a good match might actually exist. On the other
11155 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11156 but will almost always find a good frame if there is one. The other values are
11157 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11158 jerkiness and creating duplicate frames versus finding good matches in sections
11159 with bad edits, orphaned fields, blended fields, etc.
11161 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11163 Available values are:
11167 2-way matching (p/c)
11169 2-way matching, and trying 3rd match if still combed (p/c + n)
11171 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11173 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11174 still combed (p/c + n + u/b)
11176 3-way matching (p/c/n)
11178 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11179 detected as combed (p/c/n + u/b)
11182 The parenthesis at the end indicate the matches that would be used for that
11183 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11186 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11189 Default value is @var{pc_n}.
11192 Mark the main input stream as a pre-processed input, and enable the secondary
11193 input stream as the clean source to pick the fields from. See the filter
11194 introduction for more details. It is similar to the @option{clip2} feature from
11197 Default value is @code{0} (disabled).
11200 Set the field to match from. It is recommended to set this to the same value as
11201 @option{order} unless you experience matching failures with that setting. In
11202 certain circumstances changing the field that is used to match from can have a
11203 large impact on matching performance. Available values are:
11207 Automatic (same value as @option{order}).
11209 Match from the bottom field.
11211 Match from the top field.
11214 Default value is @var{auto}.
11217 Set whether or not chroma is included during the match comparisons. In most
11218 cases it is recommended to leave this enabled. You should set this to @code{0}
11219 only if your clip has bad chroma problems such as heavy rainbowing or other
11220 artifacts. Setting this to @code{0} could also be used to speed things up at
11221 the cost of some accuracy.
11223 Default value is @code{1}.
11227 These define an exclusion band which excludes the lines between @option{y0} and
11228 @option{y1} from being included in the field matching decision. An exclusion
11229 band can be used to ignore subtitles, a logo, or other things that may
11230 interfere with the matching. @option{y0} sets the starting scan line and
11231 @option{y1} sets the ending line; all lines in between @option{y0} and
11232 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11233 @option{y0} and @option{y1} to the same value will disable the feature.
11234 @option{y0} and @option{y1} defaults to @code{0}.
11237 Set the scene change detection threshold as a percentage of maximum change on
11238 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11239 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11240 @option{scthresh} is @code{[0.0, 100.0]}.
11242 Default value is @code{12.0}.
11245 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11246 account the combed scores of matches when deciding what match to use as the
11247 final match. Available values are:
11251 No final matching based on combed scores.
11253 Combed scores are only used when a scene change is detected.
11255 Use combed scores all the time.
11258 Default is @var{sc}.
11261 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11262 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11263 Available values are:
11267 No forced calculation.
11269 Force p/c/n calculations.
11271 Force p/c/n/u/b calculations.
11274 Default value is @var{none}.
11277 This is the area combing threshold used for combed frame detection. This
11278 essentially controls how "strong" or "visible" combing must be to be detected.
11279 Larger values mean combing must be more visible and smaller values mean combing
11280 can be less visible or strong and still be detected. Valid settings are from
11281 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11282 be detected as combed). This is basically a pixel difference value. A good
11283 range is @code{[8, 12]}.
11285 Default value is @code{9}.
11288 Sets whether or not chroma is considered in the combed frame decision. Only
11289 disable this if your source has chroma problems (rainbowing, etc.) that are
11290 causing problems for the combed frame detection with chroma enabled. Actually,
11291 using @option{chroma}=@var{0} is usually more reliable, except for the case
11292 where there is chroma only combing in the source.
11294 Default value is @code{0}.
11298 Respectively set the x-axis and y-axis size of the window used during combed
11299 frame detection. This has to do with the size of the area in which
11300 @option{combpel} pixels are required to be detected as combed for a frame to be
11301 declared combed. See the @option{combpel} parameter description for more info.
11302 Possible values are any number that is a power of 2 starting at 4 and going up
11305 Default value is @code{16}.
11308 The number of combed pixels inside any of the @option{blocky} by
11309 @option{blockx} size blocks on the frame for the frame to be detected as
11310 combed. While @option{cthresh} controls how "visible" the combing must be, this
11311 setting controls "how much" combing there must be in any localized area (a
11312 window defined by the @option{blockx} and @option{blocky} settings) on the
11313 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11314 which point no frames will ever be detected as combed). This setting is known
11315 as @option{MI} in TFM/VFM vocabulary.
11317 Default value is @code{80}.
11320 @anchor{p/c/n/u/b meaning}
11321 @subsection p/c/n/u/b meaning
11323 @subsubsection p/c/n
11325 We assume the following telecined stream:
11328 Top fields: 1 2 2 3 4
11329 Bottom fields: 1 2 3 4 4
11332 The numbers correspond to the progressive frame the fields relate to. Here, the
11333 first two frames are progressive, the 3rd and 4th are combed, and so on.
11335 When @code{fieldmatch} is configured to run a matching from bottom
11336 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11341 B 1 2 3 4 4 <-- matching reference
11350 As a result of the field matching, we can see that some frames get duplicated.
11351 To perform a complete inverse telecine, you need to rely on a decimation filter
11352 after this operation. See for instance the @ref{decimate} filter.
11354 The same operation now matching from top fields (@option{field}=@var{top})
11359 T 1 2 2 3 4 <-- matching reference
11369 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11370 basically, they refer to the frame and field of the opposite parity:
11373 @item @var{p} matches the field of the opposite parity in the previous frame
11374 @item @var{c} matches the field of the opposite parity in the current frame
11375 @item @var{n} matches the field of the opposite parity in the next frame
11380 The @var{u} and @var{b} matching are a bit special in the sense that they match
11381 from the opposite parity flag. In the following examples, we assume that we are
11382 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11383 'x' is placed above and below each matched fields.
11385 With bottom matching (@option{field}=@var{bottom}):
11390 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11391 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11399 With top matching (@option{field}=@var{top}):
11404 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11405 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11413 @subsection Examples
11415 Simple IVTC of a top field first telecined stream:
11417 fieldmatch=order=tff:combmatch=none, decimate
11420 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11422 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11425 @section fieldorder
11427 Transform the field order of the input video.
11429 It accepts the following parameters:
11434 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11435 for bottom field first.
11438 The default value is @samp{tff}.
11440 The transformation is done by shifting the picture content up or down
11441 by one line, and filling the remaining line with appropriate picture content.
11442 This method is consistent with most broadcast field order converters.
11444 If the input video is not flagged as being interlaced, or it is already
11445 flagged as being of the required output field order, then this filter does
11446 not alter the incoming video.
11448 It is very useful when converting to or from PAL DV material,
11449 which is bottom field first.
11453 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11456 @section fifo, afifo
11458 Buffer input images and send them when they are requested.
11460 It is mainly useful when auto-inserted by the libavfilter
11463 It does not take parameters.
11465 @section fillborders
11467 Fill borders of the input video, without changing video stream dimensions.
11468 Sometimes video can have garbage at the four edges and you may not want to
11469 crop video input to keep size multiple of some number.
11471 This filter accepts the following options:
11475 Number of pixels to fill from left border.
11478 Number of pixels to fill from right border.
11481 Number of pixels to fill from top border.
11484 Number of pixels to fill from bottom border.
11489 It accepts the following values:
11492 fill pixels using outermost pixels
11495 fill pixels using mirroring
11498 fill pixels with constant value
11501 Default is @var{smear}.
11504 Set color for pixels in fixed mode. Default is @var{black}.
11507 @subsection Commands
11508 This filter supports same @ref{commands} as options.
11509 The command accepts the same syntax of the corresponding option.
11511 If the specified expression is not valid, it is kept at its current
11516 Find a rectangular object
11518 It accepts the following options:
11522 Filepath of the object image, needs to be in gray8.
11525 Detection threshold, default is 0.5.
11528 Number of mipmaps, default is 3.
11530 @item xmin, ymin, xmax, ymax
11531 Specifies the rectangle in which to search.
11534 @subsection Examples
11538 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11540 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11546 Flood area with values of same pixel components with another values.
11548 It accepts the following options:
11551 Set pixel x coordinate.
11554 Set pixel y coordinate.
11557 Set source #0 component value.
11560 Set source #1 component value.
11563 Set source #2 component value.
11566 Set source #3 component value.
11569 Set destination #0 component value.
11572 Set destination #1 component value.
11575 Set destination #2 component value.
11578 Set destination #3 component value.
11584 Convert the input video to one of the specified pixel formats.
11585 Libavfilter will try to pick one that is suitable as input to
11588 It accepts the following parameters:
11592 A '|'-separated list of pixel format names, such as
11593 "pix_fmts=yuv420p|monow|rgb24".
11597 @subsection Examples
11601 Convert the input video to the @var{yuv420p} format
11603 format=pix_fmts=yuv420p
11606 Convert the input video to any of the formats in the list
11608 format=pix_fmts=yuv420p|yuv444p|yuv410p
11615 Convert the video to specified constant frame rate by duplicating or dropping
11616 frames as necessary.
11618 It accepts the following parameters:
11622 The desired output frame rate. The default is @code{25}.
11625 Assume the first PTS should be the given value, in seconds. This allows for
11626 padding/trimming at the start of stream. By default, no assumption is made
11627 about the first frame's expected PTS, so no padding or trimming is done.
11628 For example, this could be set to 0 to pad the beginning with duplicates of
11629 the first frame if a video stream starts after the audio stream or to trim any
11630 frames with a negative PTS.
11633 Timestamp (PTS) rounding method.
11635 Possible values are:
11642 round towards -infinity
11644 round towards +infinity
11648 The default is @code{near}.
11651 Action performed when reading the last frame.
11653 Possible values are:
11656 Use same timestamp rounding method as used for other frames.
11658 Pass through last frame if input duration has not been reached yet.
11660 The default is @code{round}.
11664 Alternatively, the options can be specified as a flat string:
11665 @var{fps}[:@var{start_time}[:@var{round}]].
11667 See also the @ref{setpts} filter.
11669 @subsection Examples
11673 A typical usage in order to set the fps to 25:
11679 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
11681 fps=fps=film:round=near
11687 Pack two different video streams into a stereoscopic video, setting proper
11688 metadata on supported codecs. The two views should have the same size and
11689 framerate and processing will stop when the shorter video ends. Please note
11690 that you may conveniently adjust view properties with the @ref{scale} and
11693 It accepts the following parameters:
11697 The desired packing format. Supported values are:
11702 The views are next to each other (default).
11705 The views are on top of each other.
11708 The views are packed by line.
11711 The views are packed by column.
11714 The views are temporally interleaved.
11723 # Convert left and right views into a frame-sequential video
11724 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
11726 # Convert views into a side-by-side video with the same output resolution as the input
11727 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
11732 Change the frame rate by interpolating new video output frames from the source
11735 This filter is not designed to function correctly with interlaced media. If
11736 you wish to change the frame rate of interlaced media then you are required
11737 to deinterlace before this filter and re-interlace after this filter.
11739 A description of the accepted options follows.
11743 Specify the output frames per second. This option can also be specified
11744 as a value alone. The default is @code{50}.
11747 Specify the start of a range where the output frame will be created as a
11748 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11749 the default is @code{15}.
11752 Specify the end of a range where the output frame will be created as a
11753 linear interpolation of two frames. The range is [@code{0}-@code{255}],
11754 the default is @code{240}.
11757 Specify the level at which a scene change is detected as a value between
11758 0 and 100 to indicate a new scene; a low value reflects a low
11759 probability for the current frame to introduce a new scene, while a higher
11760 value means the current frame is more likely to be one.
11761 The default is @code{8.2}.
11764 Specify flags influencing the filter process.
11766 Available value for @var{flags} is:
11769 @item scene_change_detect, scd
11770 Enable scene change detection using the value of the option @var{scene}.
11771 This flag is enabled by default.
11777 Select one frame every N-th frame.
11779 This filter accepts the following option:
11782 Select frame after every @code{step} frames.
11783 Allowed values are positive integers higher than 0. Default value is @code{1}.
11786 @section freezedetect
11788 Detect frozen video.
11790 This filter logs a message and sets frame metadata when it detects that the
11791 input video has no significant change in content during a specified duration.
11792 Video freeze detection calculates the mean average absolute difference of all
11793 the components of video frames and compares it to a noise floor.
11795 The printed times and duration are expressed in seconds. The
11796 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
11797 whose timestamp equals or exceeds the detection duration and it contains the
11798 timestamp of the first frame of the freeze. The
11799 @code{lavfi.freezedetect.freeze_duration} and
11800 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
11803 The filter accepts the following options:
11807 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
11808 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
11812 Set freeze duration until notification (default is 2 seconds).
11815 @section freezeframes
11817 Freeze video frames.
11819 This filter freezes video frames using frame from 2nd input.
11821 The filter accepts the following options:
11825 Set number of first frame from which to start freeze.
11828 Set number of last frame from which to end freeze.
11831 Set number of frame from 2nd input which will be used instead of replaced frames.
11837 Apply a frei0r effect to the input video.
11839 To enable the compilation of this filter, you need to install the frei0r
11840 header and configure FFmpeg with @code{--enable-frei0r}.
11842 It accepts the following parameters:
11847 The name of the frei0r effect to load. If the environment variable
11848 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
11849 directories specified by the colon-separated list in @env{FREI0R_PATH}.
11850 Otherwise, the standard frei0r paths are searched, in this order:
11851 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
11852 @file{/usr/lib/frei0r-1/}.
11854 @item filter_params
11855 A '|'-separated list of parameters to pass to the frei0r effect.
11859 A frei0r effect parameter can be a boolean (its value is either
11860 "y" or "n"), a double, a color (specified as
11861 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
11862 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
11863 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
11864 a position (specified as @var{X}/@var{Y}, where
11865 @var{X} and @var{Y} are floating point numbers) and/or a string.
11867 The number and types of parameters depend on the loaded effect. If an
11868 effect parameter is not specified, the default value is set.
11870 @subsection Examples
11874 Apply the distort0r effect, setting the first two double parameters:
11876 frei0r=filter_name=distort0r:filter_params=0.5|0.01
11880 Apply the colordistance effect, taking a color as the first parameter:
11882 frei0r=colordistance:0.2/0.3/0.4
11883 frei0r=colordistance:violet
11884 frei0r=colordistance:0x112233
11888 Apply the perspective effect, specifying the top left and top right image
11891 frei0r=perspective:0.2/0.2|0.8/0.2
11895 For more information, see
11896 @url{http://frei0r.dyne.org}
11898 @subsection Commands
11900 This filter supports the @option{filter_params} option as @ref{commands}.
11904 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
11906 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
11907 processing filter, one of them is performed once per block, not per pixel.
11908 This allows for much higher speed.
11910 The filter accepts the following options:
11914 Set quality. This option defines the number of levels for averaging. It accepts
11915 an integer in the range 4-5. Default value is @code{4}.
11918 Force a constant quantization parameter. It accepts an integer in range 0-63.
11919 If not set, the filter will use the QP from the video stream (if available).
11922 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
11923 more details but also more artifacts, while higher values make the image smoother
11924 but also blurrier. Default value is @code{0} − PSNR optimal.
11926 @item use_bframe_qp
11927 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
11928 option may cause flicker since the B-Frames have often larger QP. Default is
11929 @code{0} (not enabled).
11935 Apply Gaussian blur filter.
11937 The filter accepts the following options:
11941 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
11944 Set number of steps for Gaussian approximation. Default is @code{1}.
11947 Set which planes to filter. By default all planes are filtered.
11950 Set vertical sigma, if negative it will be same as @code{sigma}.
11951 Default is @code{-1}.
11954 @subsection Commands
11955 This filter supports same commands as options.
11956 The command accepts the same syntax of the corresponding option.
11958 If the specified expression is not valid, it is kept at its current
11963 Apply generic equation to each pixel.
11965 The filter accepts the following options:
11968 @item lum_expr, lum
11969 Set the luminance expression.
11971 Set the chrominance blue expression.
11973 Set the chrominance red expression.
11974 @item alpha_expr, a
11975 Set the alpha expression.
11977 Set the red expression.
11978 @item green_expr, g
11979 Set the green expression.
11981 Set the blue expression.
11984 The colorspace is selected according to the specified options. If one
11985 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
11986 options is specified, the filter will automatically select a YCbCr
11987 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
11988 @option{blue_expr} options is specified, it will select an RGB
11991 If one of the chrominance expression is not defined, it falls back on the other
11992 one. If no alpha expression is specified it will evaluate to opaque value.
11993 If none of chrominance expressions are specified, they will evaluate
11994 to the luminance expression.
11996 The expressions can use the following variables and functions:
12000 The sequential number of the filtered frame, starting from @code{0}.
12004 The coordinates of the current sample.
12008 The width and height of the image.
12012 Width and height scale depending on the currently filtered plane. It is the
12013 ratio between the corresponding luma plane number of pixels and the current
12014 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12015 @code{0.5,0.5} for chroma planes.
12018 Time of the current frame, expressed in seconds.
12021 Return the value of the pixel at location (@var{x},@var{y}) of the current
12025 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12029 Return the value of the pixel at location (@var{x},@var{y}) of the
12030 blue-difference chroma plane. Return 0 if there is no such plane.
12033 Return the value of the pixel at location (@var{x},@var{y}) of the
12034 red-difference chroma plane. Return 0 if there is no such plane.
12039 Return the value of the pixel at location (@var{x},@var{y}) of the
12040 red/green/blue component. Return 0 if there is no such component.
12043 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12044 plane. Return 0 if there is no such plane.
12046 @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)
12047 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12048 sums of samples within a rectangle. See the functions without the sum postfix.
12050 @item interpolation
12051 Set one of interpolation methods:
12056 Default is bilinear.
12059 For functions, if @var{x} and @var{y} are outside the area, the value will be
12060 automatically clipped to the closer edge.
12062 Please note that this filter can use multiple threads in which case each slice
12063 will have its own expression state. If you want to use only a single expression
12064 state because your expressions depend on previous state then you should limit
12065 the number of filter threads to 1.
12067 @subsection Examples
12071 Flip the image horizontally:
12077 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12078 wavelength of 100 pixels:
12080 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12084 Generate a fancy enigmatic moving light:
12086 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
12090 Generate a quick emboss effect:
12092 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12096 Modify RGB components depending on pixel position:
12098 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12102 Create a radial gradient that is the same size as the input (also see
12103 the @ref{vignette} filter):
12105 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12111 Fix the banding artifacts that are sometimes introduced into nearly flat
12112 regions by truncation to 8-bit color depth.
12113 Interpolate the gradients that should go where the bands are, and
12116 It is designed for playback only. Do not use it prior to
12117 lossy compression, because compression tends to lose the dither and
12118 bring back the bands.
12120 It accepts the following parameters:
12125 The maximum amount by which the filter will change any one pixel. This is also
12126 the threshold for detecting nearly flat regions. Acceptable values range from
12127 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12131 The neighborhood to fit the gradient to. A larger radius makes for smoother
12132 gradients, but also prevents the filter from modifying the pixels near detailed
12133 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12134 values will be clipped to the valid range.
12138 Alternatively, the options can be specified as a flat string:
12139 @var{strength}[:@var{radius}]
12141 @subsection Examples
12145 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12151 Specify radius, omitting the strength (which will fall-back to the default
12159 @anchor{graphmonitor}
12160 @section graphmonitor
12161 Show various filtergraph stats.
12163 With this filter one can debug complete filtergraph.
12164 Especially issues with links filling with queued frames.
12166 The filter accepts the following options:
12170 Set video output size. Default is @var{hd720}.
12173 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12176 Set output mode, can be @var{fulll} or @var{compact}.
12177 In @var{compact} mode only filters with some queued frames have displayed stats.
12180 Set flags which enable which stats are shown in video.
12182 Available values for flags are:
12185 Display number of queued frames in each link.
12187 @item frame_count_in
12188 Display number of frames taken from filter.
12190 @item frame_count_out
12191 Display number of frames given out from filter.
12194 Display current filtered frame pts.
12197 Display current filtered frame time.
12200 Display time base for filter link.
12203 Display used format for filter link.
12206 Display video size or number of audio channels in case of audio used by filter link.
12209 Display video frame rate or sample rate in case of audio used by filter link.
12212 Display link output status.
12216 Set upper limit for video rate of output stream, Default value is @var{25}.
12217 This guarantee that output video frame rate will not be higher than this value.
12221 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12222 and corrects the scene colors accordingly.
12224 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12226 The filter accepts the following options:
12230 The order of differentiation to be applied on the scene. Must be chosen in the range
12231 [0,2] and default value is 1.
12234 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12235 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12236 max value instead of calculating Minkowski distance.
12239 The standard deviation of Gaussian blur to be applied on the scene. Must be
12240 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12241 can't be equal to 0 if @var{difford} is greater than 0.
12244 @subsection Examples
12250 greyedge=difford=1:minknorm=5:sigma=2
12256 greyedge=difford=1:minknorm=0:sigma=2
12264 Apply a Hald CLUT to a video stream.
12266 First input is the video stream to process, and second one is the Hald CLUT.
12267 The Hald CLUT input can be a simple picture or a complete video stream.
12269 The filter accepts the following options:
12273 Force termination when the shortest input terminates. Default is @code{0}.
12275 Continue applying the last CLUT after the end of the stream. A value of
12276 @code{0} disable the filter after the last frame of the CLUT is reached.
12277 Default is @code{1}.
12280 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12281 filters share the same internals).
12283 This filter also supports the @ref{framesync} options.
12285 More information about the Hald CLUT can be found on Eskil Steenberg's website
12286 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12288 @subsection Workflow examples
12290 @subsubsection Hald CLUT video stream
12292 Generate an identity Hald CLUT stream altered with various effects:
12294 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
12297 Note: make sure you use a lossless codec.
12299 Then use it with @code{haldclut} to apply it on some random stream:
12301 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12304 The Hald CLUT will be applied to the 10 first seconds (duration of
12305 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12306 to the remaining frames of the @code{mandelbrot} stream.
12308 @subsubsection Hald CLUT with preview
12310 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12311 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12312 biggest possible square starting at the top left of the picture. The remaining
12313 padding pixels (bottom or right) will be ignored. This area can be used to add
12314 a preview of the Hald CLUT.
12316 Typically, the following generated Hald CLUT will be supported by the
12317 @code{haldclut} filter:
12320 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12321 pad=iw+320 [padded_clut];
12322 smptebars=s=320x256, split [a][b];
12323 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12324 [main][b] overlay=W-320" -frames:v 1 clut.png
12327 It contains the original and a preview of the effect of the CLUT: SMPTE color
12328 bars are displayed on the right-top, and below the same color bars processed by
12331 Then, the effect of this Hald CLUT can be visualized with:
12333 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12338 Flip the input video horizontally.
12340 For example, to horizontally flip the input video with @command{ffmpeg}:
12342 ffmpeg -i in.avi -vf "hflip" out.avi
12346 This filter applies a global color histogram equalization on a
12349 It can be used to correct video that has a compressed range of pixel
12350 intensities. The filter redistributes the pixel intensities to
12351 equalize their distribution across the intensity range. It may be
12352 viewed as an "automatically adjusting contrast filter". This filter is
12353 useful only for correcting degraded or poorly captured source
12356 The filter accepts the following options:
12360 Determine the amount of equalization to be applied. As the strength
12361 is reduced, the distribution of pixel intensities more-and-more
12362 approaches that of the input frame. The value must be a float number
12363 in the range [0,1] and defaults to 0.200.
12366 Set the maximum intensity that can generated and scale the output
12367 values appropriately. The strength should be set as desired and then
12368 the intensity can be limited if needed to avoid washing-out. The value
12369 must be a float number in the range [0,1] and defaults to 0.210.
12372 Set the antibanding level. If enabled the filter will randomly vary
12373 the luminance of output pixels by a small amount to avoid banding of
12374 the histogram. Possible values are @code{none}, @code{weak} or
12375 @code{strong}. It defaults to @code{none}.
12381 Compute and draw a color distribution histogram for the input video.
12383 The computed histogram is a representation of the color component
12384 distribution in an image.
12386 Standard histogram displays the color components distribution in an image.
12387 Displays color graph for each color component. Shows distribution of
12388 the Y, U, V, A or R, G, B components, depending on input format, in the
12389 current frame. Below each graph a color component scale meter is shown.
12391 The filter accepts the following options:
12395 Set height of level. Default value is @code{200}.
12396 Allowed range is [50, 2048].
12399 Set height of color scale. Default value is @code{12}.
12400 Allowed range is [0, 40].
12404 It accepts the following values:
12407 Per color component graphs are placed below each other.
12410 Per color component graphs are placed side by side.
12413 Presents information identical to that in the @code{parade}, except
12414 that the graphs representing color components are superimposed directly
12417 Default is @code{stack}.
12420 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12421 Default is @code{linear}.
12424 Set what color components to display.
12425 Default is @code{7}.
12428 Set foreground opacity. Default is @code{0.7}.
12431 Set background opacity. Default is @code{0.5}.
12434 @subsection Examples
12439 Calculate and draw histogram:
12441 ffplay -i input -vf histogram
12449 This is a high precision/quality 3d denoise filter. It aims to reduce
12450 image noise, producing smooth images and making still images really
12451 still. It should enhance compressibility.
12453 It accepts the following optional parameters:
12457 A non-negative floating point number which specifies spatial luma strength.
12458 It defaults to 4.0.
12460 @item chroma_spatial
12461 A non-negative floating point number which specifies spatial chroma strength.
12462 It defaults to 3.0*@var{luma_spatial}/4.0.
12465 A floating point number which specifies luma temporal strength. It defaults to
12466 6.0*@var{luma_spatial}/4.0.
12469 A floating point number which specifies chroma temporal strength. It defaults to
12470 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12473 @subsection Commands
12474 This filter supports same @ref{commands} as options.
12475 The command accepts the same syntax of the corresponding option.
12477 If the specified expression is not valid, it is kept at its current
12480 @anchor{hwdownload}
12481 @section hwdownload
12483 Download hardware frames to system memory.
12485 The input must be in hardware frames, and the output a non-hardware format.
12486 Not all formats will be supported on the output - it may be necessary to insert
12487 an additional @option{format} filter immediately following in the graph to get
12488 the output in a supported format.
12492 Map hardware frames to system memory or to another device.
12494 This filter has several different modes of operation; which one is used depends
12495 on the input and output formats:
12498 Hardware frame input, normal frame output
12500 Map the input frames to system memory and pass them to the output. If the
12501 original hardware frame is later required (for example, after overlaying
12502 something else on part of it), the @option{hwmap} filter can be used again
12503 in the next mode to retrieve it.
12505 Normal frame input, hardware frame output
12507 If the input is actually a software-mapped hardware frame, then unmap it -
12508 that is, return the original hardware frame.
12510 Otherwise, a device must be provided. Create new hardware surfaces on that
12511 device for the output, then map them back to the software format at the input
12512 and give those frames to the preceding filter. This will then act like the
12513 @option{hwupload} filter, but may be able to avoid an additional copy when
12514 the input is already in a compatible format.
12516 Hardware frame input and output
12518 A device must be supplied for the output, either directly or with the
12519 @option{derive_device} option. The input and output devices must be of
12520 different types and compatible - the exact meaning of this is
12521 system-dependent, but typically it means that they must refer to the same
12522 underlying hardware context (for example, refer to the same graphics card).
12524 If the input frames were originally created on the output device, then unmap
12525 to retrieve the original frames.
12527 Otherwise, map the frames to the output device - create new hardware frames
12528 on the output corresponding to the frames on the input.
12531 The following additional parameters are accepted:
12535 Set the frame mapping mode. Some combination of:
12538 The mapped frame should be readable.
12540 The mapped frame should be writeable.
12542 The mapping will always overwrite the entire frame.
12544 This may improve performance in some cases, as the original contents of the
12545 frame need not be loaded.
12547 The mapping must not involve any copying.
12549 Indirect mappings to copies of frames are created in some cases where either
12550 direct mapping is not possible or it would have unexpected properties.
12551 Setting this flag ensures that the mapping is direct and will fail if that is
12554 Defaults to @var{read+write} if not specified.
12556 @item derive_device @var{type}
12557 Rather than using the device supplied at initialisation, instead derive a new
12558 device of type @var{type} from the device the input frames exist on.
12561 In a hardware to hardware mapping, map in reverse - create frames in the sink
12562 and map them back to the source. This may be necessary in some cases where
12563 a mapping in one direction is required but only the opposite direction is
12564 supported by the devices being used.
12566 This option is dangerous - it may break the preceding filter in undefined
12567 ways if there are any additional constraints on that filter's output.
12568 Do not use it without fully understanding the implications of its use.
12574 Upload system memory frames to hardware surfaces.
12576 The device to upload to must be supplied when the filter is initialised. If
12577 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12578 option or with the @option{derive_device} option. The input and output devices
12579 must be of different types and compatible - the exact meaning of this is
12580 system-dependent, but typically it means that they must refer to the same
12581 underlying hardware context (for example, refer to the same graphics card).
12583 The following additional parameters are accepted:
12586 @item derive_device @var{type}
12587 Rather than using the device supplied at initialisation, instead derive a new
12588 device of type @var{type} from the device the input frames exist on.
12591 @anchor{hwupload_cuda}
12592 @section hwupload_cuda
12594 Upload system memory frames to a CUDA device.
12596 It accepts the following optional parameters:
12600 The number of the CUDA device to use
12605 Apply a high-quality magnification filter designed for pixel art. This filter
12606 was originally created by Maxim Stepin.
12608 It accepts the following option:
12612 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12613 @code{hq3x} and @code{4} for @code{hq4x}.
12614 Default is @code{3}.
12618 Stack input videos horizontally.
12620 All streams must be of same pixel format and of same height.
12622 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12623 to create same output.
12625 The filter accepts the following option:
12629 Set number of input streams. Default is 2.
12632 If set to 1, force the output to terminate when the shortest input
12633 terminates. Default value is 0.
12638 Modify the hue and/or the saturation of the input.
12640 It accepts the following parameters:
12644 Specify the hue angle as a number of degrees. It accepts an expression,
12645 and defaults to "0".
12648 Specify the saturation in the [-10,10] range. It accepts an expression and
12652 Specify the hue angle as a number of radians. It accepts an
12653 expression, and defaults to "0".
12656 Specify the brightness in the [-10,10] range. It accepts an expression and
12660 @option{h} and @option{H} are mutually exclusive, and can't be
12661 specified at the same time.
12663 The @option{b}, @option{h}, @option{H} and @option{s} option values are
12664 expressions containing the following constants:
12668 frame count of the input frame starting from 0
12671 presentation timestamp of the input frame expressed in time base units
12674 frame rate of the input video, NAN if the input frame rate is unknown
12677 timestamp expressed in seconds, NAN if the input timestamp is unknown
12680 time base of the input video
12683 @subsection Examples
12687 Set the hue to 90 degrees and the saturation to 1.0:
12693 Same command but expressing the hue in radians:
12699 Rotate hue and make the saturation swing between 0
12700 and 2 over a period of 1 second:
12702 hue="H=2*PI*t: s=sin(2*PI*t)+1"
12706 Apply a 3 seconds saturation fade-in effect starting at 0:
12708 hue="s=min(t/3\,1)"
12711 The general fade-in expression can be written as:
12713 hue="s=min(0\, max((t-START)/DURATION\, 1))"
12717 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
12719 hue="s=max(0\, min(1\, (8-t)/3))"
12722 The general fade-out expression can be written as:
12724 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
12729 @subsection Commands
12731 This filter supports the following commands:
12737 Modify the hue and/or the saturation and/or brightness of the input video.
12738 The command accepts the same syntax of the corresponding option.
12740 If the specified expression is not valid, it is kept at its current
12744 @section hysteresis
12746 Grow first stream into second stream by connecting components.
12747 This makes it possible to build more robust edge masks.
12749 This filter accepts the following options:
12753 Set which planes will be processed as bitmap, unprocessed planes will be
12754 copied from first stream.
12755 By default value 0xf, all planes will be processed.
12758 Set threshold which is used in filtering. If pixel component value is higher than
12759 this value filter algorithm for connecting components is activated.
12760 By default value is 0.
12763 The @code{hysteresis} filter also supports the @ref{framesync} options.
12767 Detect video interlacing type.
12769 This filter tries to detect if the input frames are interlaced, progressive,
12770 top or bottom field first. It will also try to detect fields that are
12771 repeated between adjacent frames (a sign of telecine).
12773 Single frame detection considers only immediately adjacent frames when classifying each frame.
12774 Multiple frame detection incorporates the classification history of previous frames.
12776 The filter will log these metadata values:
12779 @item single.current_frame
12780 Detected type of current frame using single-frame detection. One of:
12781 ``tff'' (top field first), ``bff'' (bottom field first),
12782 ``progressive'', or ``undetermined''
12785 Cumulative number of frames detected as top field first using single-frame detection.
12788 Cumulative number of frames detected as top field first using multiple-frame detection.
12791 Cumulative number of frames detected as bottom field first using single-frame detection.
12793 @item multiple.current_frame
12794 Detected type of current frame using multiple-frame detection. One of:
12795 ``tff'' (top field first), ``bff'' (bottom field first),
12796 ``progressive'', or ``undetermined''
12799 Cumulative number of frames detected as bottom field first using multiple-frame detection.
12801 @item single.progressive
12802 Cumulative number of frames detected as progressive using single-frame detection.
12804 @item multiple.progressive
12805 Cumulative number of frames detected as progressive using multiple-frame detection.
12807 @item single.undetermined
12808 Cumulative number of frames that could not be classified using single-frame detection.
12810 @item multiple.undetermined
12811 Cumulative number of frames that could not be classified using multiple-frame detection.
12813 @item repeated.current_frame
12814 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
12816 @item repeated.neither
12817 Cumulative number of frames with no repeated field.
12820 Cumulative number of frames with the top field repeated from the previous frame's top field.
12822 @item repeated.bottom
12823 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
12826 The filter accepts the following options:
12830 Set interlacing threshold.
12832 Set progressive threshold.
12834 Threshold for repeated field detection.
12836 Number of frames after which a given frame's contribution to the
12837 statistics is halved (i.e., it contributes only 0.5 to its
12838 classification). The default of 0 means that all frames seen are given
12839 full weight of 1.0 forever.
12840 @item analyze_interlaced_flag
12841 When this is not 0 then idet will use the specified number of frames to determine
12842 if the interlaced flag is accurate, it will not count undetermined frames.
12843 If the flag is found to be accurate it will be used without any further
12844 computations, if it is found to be inaccurate it will be cleared without any
12845 further computations. This allows inserting the idet filter as a low computational
12846 method to clean up the interlaced flag
12851 Deinterleave or interleave fields.
12853 This filter allows one to process interlaced images fields without
12854 deinterlacing them. Deinterleaving splits the input frame into 2
12855 fields (so called half pictures). Odd lines are moved to the top
12856 half of the output image, even lines to the bottom half.
12857 You can process (filter) them independently and then re-interleave them.
12859 The filter accepts the following options:
12863 @item chroma_mode, c
12864 @item alpha_mode, a
12865 Available values for @var{luma_mode}, @var{chroma_mode} and
12866 @var{alpha_mode} are:
12872 @item deinterleave, d
12873 Deinterleave fields, placing one above the other.
12875 @item interleave, i
12876 Interleave fields. Reverse the effect of deinterleaving.
12878 Default value is @code{none}.
12880 @item luma_swap, ls
12881 @item chroma_swap, cs
12882 @item alpha_swap, as
12883 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
12886 @subsection Commands
12888 This filter supports the all above options as @ref{commands}.
12892 Apply inflate effect to the video.
12894 This filter replaces the pixel by the local(3x3) average by taking into account
12895 only values higher than the pixel.
12897 It accepts the following options:
12904 Limit the maximum change for each plane, default is 65535.
12905 If 0, plane will remain unchanged.
12908 @subsection Commands
12910 This filter supports the all above options as @ref{commands}.
12914 Simple interlacing filter from progressive contents. This interleaves upper (or
12915 lower) lines from odd frames with lower (or upper) lines from even frames,
12916 halving the frame rate and preserving image height.
12919 Original Original New Frame
12920 Frame 'j' Frame 'j+1' (tff)
12921 ========== =========== ==================
12922 Line 0 --------------------> Frame 'j' Line 0
12923 Line 1 Line 1 ----> Frame 'j+1' Line 1
12924 Line 2 ---------------------> Frame 'j' Line 2
12925 Line 3 Line 3 ----> Frame 'j+1' Line 3
12927 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
12930 It accepts the following optional parameters:
12934 This determines whether the interlaced frame is taken from the even
12935 (tff - default) or odd (bff) lines of the progressive frame.
12938 Vertical lowpass filter to avoid twitter interlacing and
12939 reduce moire patterns.
12943 Disable vertical lowpass filter
12946 Enable linear filter (default)
12949 Enable complex filter. This will slightly less reduce twitter and moire
12950 but better retain detail and subjective sharpness impression.
12957 Deinterlace input video by applying Donald Graft's adaptive kernel
12958 deinterling. Work on interlaced parts of a video to produce
12959 progressive frames.
12961 The description of the accepted parameters follows.
12965 Set the threshold which affects the filter's tolerance when
12966 determining if a pixel line must be processed. It must be an integer
12967 in the range [0,255] and defaults to 10. A value of 0 will result in
12968 applying the process on every pixels.
12971 Paint pixels exceeding the threshold value to white if set to 1.
12975 Set the fields order. Swap fields if set to 1, leave fields alone if
12979 Enable additional sharpening if set to 1. Default is 0.
12982 Enable twoway sharpening if set to 1. Default is 0.
12985 @subsection Examples
12989 Apply default values:
12991 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
12995 Enable additional sharpening:
13001 Paint processed pixels in white:
13009 Slowly update darker pixels.
13011 This filter makes short flashes of light appear longer.
13012 This filter accepts the following options:
13016 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13019 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13022 @section lenscorrection
13024 Correct radial lens distortion
13026 This filter can be used to correct for radial distortion as can result from the use
13027 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13028 one can use tools available for example as part of opencv or simply trial-and-error.
13029 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13030 and extract the k1 and k2 coefficients from the resulting matrix.
13032 Note that effectively the same filter is available in the open-source tools Krita and
13033 Digikam from the KDE project.
13035 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13036 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13037 brightness distribution, so you may want to use both filters together in certain
13038 cases, though you will have to take care of ordering, i.e. whether vignetting should
13039 be applied before or after lens correction.
13041 @subsection Options
13043 The filter accepts the following options:
13047 Relative x-coordinate of the focal point of the image, and thereby the center of the
13048 distortion. This value has a range [0,1] and is expressed as fractions of the image
13049 width. Default is 0.5.
13051 Relative y-coordinate of the focal point of the image, and thereby the center of the
13052 distortion. This value has a range [0,1] and is expressed as fractions of the image
13053 height. Default is 0.5.
13055 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13056 no correction. Default is 0.
13058 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13059 0 means no correction. Default is 0.
13062 The formula that generates the correction is:
13064 @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)
13066 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13067 distances from the focal point in the source and target images, respectively.
13071 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13073 The @code{lensfun} filter requires the camera make, camera model, and lens model
13074 to apply the lens correction. The filter will load the lensfun database and
13075 query it to find the corresponding camera and lens entries in the database. As
13076 long as these entries can be found with the given options, the filter can
13077 perform corrections on frames. Note that incomplete strings will result in the
13078 filter choosing the best match with the given options, and the filter will
13079 output the chosen camera and lens models (logged with level "info"). You must
13080 provide the make, camera model, and lens model as they are required.
13082 The filter accepts the following options:
13086 The make of the camera (for example, "Canon"). This option is required.
13089 The model of the camera (for example, "Canon EOS 100D"). This option is
13093 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13094 option is required.
13097 The type of correction to apply. The following values are valid options:
13101 Enables fixing lens vignetting.
13104 Enables fixing lens geometry. This is the default.
13107 Enables fixing chromatic aberrations.
13110 Enables fixing lens vignetting and lens geometry.
13113 Enables fixing lens vignetting and chromatic aberrations.
13116 Enables fixing both lens geometry and chromatic aberrations.
13119 Enables all possible corrections.
13123 The focal length of the image/video (zoom; expected constant for video). For
13124 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13125 range should be chosen when using that lens. Default 18.
13128 The aperture of the image/video (expected constant for video). Note that
13129 aperture is only used for vignetting correction. Default 3.5.
13131 @item focus_distance
13132 The focus distance of the image/video (expected constant for video). Note that
13133 focus distance is only used for vignetting and only slightly affects the
13134 vignetting correction process. If unknown, leave it at the default value (which
13138 The scale factor which is applied after transformation. After correction the
13139 video is no longer necessarily rectangular. This parameter controls how much of
13140 the resulting image is visible. The value 0 means that a value will be chosen
13141 automatically such that there is little or no unmapped area in the output
13142 image. 1.0 means that no additional scaling is done. Lower values may result
13143 in more of the corrected image being visible, while higher values may avoid
13144 unmapped areas in the output.
13146 @item target_geometry
13147 The target geometry of the output image/video. The following values are valid
13151 @item rectilinear (default)
13154 @item equirectangular
13155 @item fisheye_orthographic
13156 @item fisheye_stereographic
13157 @item fisheye_equisolid
13158 @item fisheye_thoby
13161 Apply the reverse of image correction (instead of correcting distortion, apply
13164 @item interpolation
13165 The type of interpolation used when correcting distortion. The following values
13170 @item linear (default)
13175 @subsection Examples
13179 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13180 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13184 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
13188 Apply the same as before, but only for the first 5 seconds of video.
13191 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
13198 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13199 score between two input videos.
13201 The obtained VMAF score is printed through the logging system.
13203 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13204 After installing the library it can be enabled using:
13205 @code{./configure --enable-libvmaf}.
13206 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13208 The filter has following options:
13212 Set the model path which is to be used for SVM.
13213 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13216 Set the file path to be used to store logs.
13219 Set the format of the log file (csv, json or xml).
13221 @item enable_transform
13222 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13223 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13224 Default value: @code{false}
13227 Invokes the phone model which will generate VMAF scores higher than in the
13228 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13229 Default value: @code{false}
13232 Enables computing psnr along with vmaf.
13233 Default value: @code{false}
13236 Enables computing ssim along with vmaf.
13237 Default value: @code{false}
13240 Enables computing ms_ssim along with vmaf.
13241 Default value: @code{false}
13244 Set the pool method to be used for computing vmaf.
13245 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13248 Set number of threads to be used when computing vmaf.
13249 Default value: @code{0}, which makes use of all available logical processors.
13252 Set interval for frame subsampling used when computing vmaf.
13253 Default value: @code{1}
13255 @item enable_conf_interval
13256 Enables confidence interval.
13257 Default value: @code{false}
13260 This filter also supports the @ref{framesync} options.
13262 @subsection Examples
13265 On the below examples the input file @file{main.mpg} being processed is
13266 compared with the reference file @file{ref.mpg}.
13269 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13273 Example with options:
13275 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13279 Example with options and different containers:
13281 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 -
13287 Limits the pixel components values to the specified range [min, max].
13289 The filter accepts the following options:
13293 Lower bound. Defaults to the lowest allowed value for the input.
13296 Upper bound. Defaults to the highest allowed value for the input.
13299 Specify which planes will be processed. Defaults to all available.
13306 The filter accepts the following options:
13310 Set the number of loops. Setting this value to -1 will result in infinite loops.
13314 Set maximal size in number of frames. Default is 0.
13317 Set first frame of loop. Default is 0.
13320 @subsection Examples
13324 Loop single first frame infinitely:
13326 loop=loop=-1:size=1:start=0
13330 Loop single first frame 10 times:
13332 loop=loop=10:size=1:start=0
13336 Loop 10 first frames 5 times:
13338 loop=loop=5:size=10:start=0
13344 Apply a 1D LUT to an input video.
13346 The filter accepts the following options:
13350 Set the 1D LUT file name.
13352 Currently supported formats:
13361 Select interpolation mode.
13363 Available values are:
13367 Use values from the nearest defined point.
13369 Interpolate values using the linear interpolation.
13371 Interpolate values using the cosine interpolation.
13373 Interpolate values using the cubic interpolation.
13375 Interpolate values using the spline interpolation.
13382 Apply a 3D LUT to an input video.
13384 The filter accepts the following options:
13388 Set the 3D LUT file name.
13390 Currently supported formats:
13404 Select interpolation mode.
13406 Available values are:
13410 Use values from the nearest defined point.
13412 Interpolate values using the 8 points defining a cube.
13414 Interpolate values using a tetrahedron.
13420 Turn certain luma values into transparency.
13422 The filter accepts the following options:
13426 Set the luma which will be used as base for transparency.
13427 Default value is @code{0}.
13430 Set the range of luma values to be keyed out.
13431 Default value is @code{0.01}.
13434 Set the range of softness. Default value is @code{0}.
13435 Use this to control gradual transition from zero to full transparency.
13438 @subsection Commands
13439 This filter supports same @ref{commands} as options.
13440 The command accepts the same syntax of the corresponding option.
13442 If the specified expression is not valid, it is kept at its current
13445 @section lut, lutrgb, lutyuv
13447 Compute a look-up table for binding each pixel component input value
13448 to an output value, and apply it to the input video.
13450 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13451 to an RGB input video.
13453 These filters accept the following parameters:
13456 set first pixel component expression
13458 set second pixel component expression
13460 set third pixel component expression
13462 set fourth pixel component expression, corresponds to the alpha component
13465 set red component expression
13467 set green component expression
13469 set blue component expression
13471 alpha component expression
13474 set Y/luminance component expression
13476 set U/Cb component expression
13478 set V/Cr component expression
13481 Each of them specifies the expression to use for computing the lookup table for
13482 the corresponding pixel component values.
13484 The exact component associated to each of the @var{c*} options depends on the
13487 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13488 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13490 The expressions can contain the following constants and functions:
13495 The input width and height.
13498 The input value for the pixel component.
13501 The input value, clipped to the @var{minval}-@var{maxval} range.
13504 The maximum value for the pixel component.
13507 The minimum value for the pixel component.
13510 The negated value for the pixel component value, clipped to the
13511 @var{minval}-@var{maxval} range; it corresponds to the expression
13512 "maxval-clipval+minval".
13515 The computed value in @var{val}, clipped to the
13516 @var{minval}-@var{maxval} range.
13518 @item gammaval(gamma)
13519 The computed gamma correction value of the pixel component value,
13520 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13522 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13526 All expressions default to "val".
13528 @subsection Examples
13532 Negate input video:
13534 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13535 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13538 The above is the same as:
13540 lutrgb="r=negval:g=negval:b=negval"
13541 lutyuv="y=negval:u=negval:v=negval"
13551 Remove chroma components, turning the video into a graytone image:
13553 lutyuv="u=128:v=128"
13557 Apply a luma burning effect:
13563 Remove green and blue components:
13569 Set a constant alpha channel value on input:
13571 format=rgba,lutrgb=a="maxval-minval/2"
13575 Correct luminance gamma by a factor of 0.5:
13577 lutyuv=y=gammaval(0.5)
13581 Discard least significant bits of luma:
13583 lutyuv=y='bitand(val, 128+64+32)'
13587 Technicolor like effect:
13589 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13593 @section lut2, tlut2
13595 The @code{lut2} filter takes two input streams and outputs one
13598 The @code{tlut2} (time lut2) filter takes two consecutive frames
13599 from one single stream.
13601 This filter accepts the following parameters:
13604 set first pixel component expression
13606 set second pixel component expression
13608 set third pixel component expression
13610 set fourth pixel component expression, corresponds to the alpha component
13613 set output bit depth, only available for @code{lut2} filter. By default is 0,
13614 which means bit depth is automatically picked from first input format.
13617 The @code{lut2} filter also supports the @ref{framesync} options.
13619 Each of them specifies the expression to use for computing the lookup table for
13620 the corresponding pixel component values.
13622 The exact component associated to each of the @var{c*} options depends on the
13625 The expressions can contain the following constants:
13630 The input width and height.
13633 The first input value for the pixel component.
13636 The second input value for the pixel component.
13639 The first input video bit depth.
13642 The second input video bit depth.
13645 All expressions default to "x".
13647 @subsection Examples
13651 Highlight differences between two RGB video streams:
13653 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)'
13657 Highlight differences between two YUV video streams:
13659 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)'
13663 Show max difference between two video streams:
13665 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)))'
13669 @section maskedclamp
13671 Clamp the first input stream with the second input and third input stream.
13673 Returns the value of first stream to be between second input
13674 stream - @code{undershoot} and third input stream + @code{overshoot}.
13676 This filter accepts the following options:
13679 Default value is @code{0}.
13682 Default value is @code{0}.
13685 Set which planes will be processed as bitmap, unprocessed planes will be
13686 copied from first stream.
13687 By default value 0xf, all planes will be processed.
13692 Merge the second and third input stream into output stream using absolute differences
13693 between second input stream and first input stream and absolute difference between
13694 third input stream and first input stream. The picked value will be from second input
13695 stream if second absolute difference is greater than first one or from third input stream
13698 This filter accepts the following options:
13701 Set which planes will be processed as bitmap, unprocessed planes will be
13702 copied from first stream.
13703 By default value 0xf, all planes will be processed.
13706 @section maskedmerge
13708 Merge the first input stream with the second input stream using per pixel
13709 weights in the third input stream.
13711 A value of 0 in the third stream pixel component means that pixel component
13712 from first stream is returned unchanged, while maximum value (eg. 255 for
13713 8-bit videos) means that pixel component from second stream is returned
13714 unchanged. Intermediate values define the amount of merging between both
13715 input stream's pixel components.
13717 This filter accepts the following options:
13720 Set which planes will be processed as bitmap, unprocessed planes will be
13721 copied from first stream.
13722 By default value 0xf, all planes will be processed.
13727 Merge the second and third input stream into output stream using absolute differences
13728 between second input stream and first input stream and absolute difference between
13729 third input stream and first input stream. The picked value will be from second input
13730 stream if second absolute difference is less than first one or from third input stream
13733 This filter accepts the following options:
13736 Set which planes will be processed as bitmap, unprocessed planes will be
13737 copied from first stream.
13738 By default value 0xf, all planes will be processed.
13741 @section maskedthreshold
13742 Pick pixels comparing absolute difference of two video streams with fixed
13745 If absolute difference between pixel component of first and second video
13746 stream is equal or lower than user supplied threshold than pixel component
13747 from first video stream is picked, otherwise pixel component from second
13748 video stream is picked.
13750 This filter accepts the following options:
13753 Set threshold used when picking pixels from absolute difference from two input
13757 Set which planes will be processed as bitmap, unprocessed planes will be
13758 copied from second stream.
13759 By default value 0xf, all planes will be processed.
13763 Create mask from input video.
13765 For example it is useful to create motion masks after @code{tblend} filter.
13767 This filter accepts the following options:
13771 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
13774 Set high threshold. Any pixel component higher than this value will be set to max value
13775 allowed for current pixel format.
13778 Set planes to filter, by default all available planes are filtered.
13781 Fill all frame pixels with this value.
13784 Set max average pixel value for frame. If sum of all pixel components is higher that this
13785 average, output frame will be completely filled with value set by @var{fill} option.
13786 Typically useful for scene changes when used in combination with @code{tblend} filter.
13791 Apply motion-compensation deinterlacing.
13793 It needs one field per frame as input and must thus be used together
13794 with yadif=1/3 or equivalent.
13796 This filter accepts the following options:
13799 Set the deinterlacing mode.
13801 It accepts one of the following values:
13806 use iterative motion estimation
13808 like @samp{slow}, but use multiple reference frames.
13810 Default value is @samp{fast}.
13813 Set the picture field parity assumed for the input video. It must be
13814 one of the following values:
13818 assume top field first
13820 assume bottom field first
13823 Default value is @samp{bff}.
13826 Set per-block quantization parameter (QP) used by the internal
13829 Higher values should result in a smoother motion vector field but less
13830 optimal individual vectors. Default value is 1.
13835 Pick median pixel from certain rectangle defined by radius.
13837 This filter accepts the following options:
13841 Set horizontal radius size. Default value is @code{1}.
13842 Allowed range is integer from 1 to 127.
13845 Set which planes to process. Default is @code{15}, which is all available planes.
13848 Set vertical radius size. Default value is @code{0}.
13849 Allowed range is integer from 0 to 127.
13850 If it is 0, value will be picked from horizontal @code{radius} option.
13853 Set median percentile. Default value is @code{0.5}.
13854 Default value of @code{0.5} will pick always median values, while @code{0} will pick
13855 minimum values, and @code{1} maximum values.
13858 @subsection Commands
13859 This filter supports same @ref{commands} as options.
13860 The command accepts the same syntax of the corresponding option.
13862 If the specified expression is not valid, it is kept at its current
13865 @section mergeplanes
13867 Merge color channel components from several video streams.
13869 The filter accepts up to 4 input streams, and merge selected input
13870 planes to the output video.
13872 This filter accepts the following options:
13875 Set input to output plane mapping. Default is @code{0}.
13877 The mappings is specified as a bitmap. It should be specified as a
13878 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
13879 mapping for the first plane of the output stream. 'A' sets the number of
13880 the input stream to use (from 0 to 3), and 'a' the plane number of the
13881 corresponding input to use (from 0 to 3). The rest of the mappings is
13882 similar, 'Bb' describes the mapping for the output stream second
13883 plane, 'Cc' describes the mapping for the output stream third plane and
13884 'Dd' describes the mapping for the output stream fourth plane.
13887 Set output pixel format. Default is @code{yuva444p}.
13890 @subsection Examples
13894 Merge three gray video streams of same width and height into single video stream:
13896 [a0][a1][a2]mergeplanes=0x001020:yuv444p
13900 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
13902 [a0][a1]mergeplanes=0x00010210:yuva444p
13906 Swap Y and A plane in yuva444p stream:
13908 format=yuva444p,mergeplanes=0x03010200:yuva444p
13912 Swap U and V plane in yuv420p stream:
13914 format=yuv420p,mergeplanes=0x000201:yuv420p
13918 Cast a rgb24 clip to yuv444p:
13920 format=rgb24,mergeplanes=0x000102:yuv444p
13926 Estimate and export motion vectors using block matching algorithms.
13927 Motion vectors are stored in frame side data to be used by other filters.
13929 This filter accepts the following options:
13932 Specify the motion estimation method. Accepts one of the following values:
13936 Exhaustive search algorithm.
13938 Three step search algorithm.
13940 Two dimensional logarithmic search algorithm.
13942 New three step search algorithm.
13944 Four step search algorithm.
13946 Diamond search algorithm.
13948 Hexagon-based search algorithm.
13950 Enhanced predictive zonal search algorithm.
13952 Uneven multi-hexagon search algorithm.
13954 Default value is @samp{esa}.
13957 Macroblock size. Default @code{16}.
13960 Search parameter. Default @code{7}.
13963 @section midequalizer
13965 Apply Midway Image Equalization effect using two video streams.
13967 Midway Image Equalization adjusts a pair of images to have the same
13968 histogram, while maintaining their dynamics as much as possible. It's
13969 useful for e.g. matching exposures from a pair of stereo cameras.
13971 This filter has two inputs and one output, which must be of same pixel format, but
13972 may be of different sizes. The output of filter is first input adjusted with
13973 midway histogram of both inputs.
13975 This filter accepts the following option:
13979 Set which planes to process. Default is @code{15}, which is all available planes.
13982 @section minterpolate
13984 Convert the video to specified frame rate using motion interpolation.
13986 This filter accepts the following options:
13989 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}.
13992 Motion interpolation mode. Following values are accepted:
13995 Duplicate previous or next frame for interpolating new ones.
13997 Blend source frames. Interpolated frame is mean of previous and next frames.
13999 Motion compensated interpolation. Following options are effective when this mode is selected:
14003 Motion compensation mode. Following values are accepted:
14006 Overlapped block motion compensation.
14008 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14010 Default mode is @samp{obmc}.
14013 Motion estimation mode. Following values are accepted:
14016 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14018 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14020 Default mode is @samp{bilat}.
14023 The algorithm to be used for motion estimation. Following values are accepted:
14026 Exhaustive search algorithm.
14028 Three step search algorithm.
14030 Two dimensional logarithmic search algorithm.
14032 New three step search algorithm.
14034 Four step search algorithm.
14036 Diamond search algorithm.
14038 Hexagon-based search algorithm.
14040 Enhanced predictive zonal search algorithm.
14042 Uneven multi-hexagon search algorithm.
14044 Default algorithm is @samp{epzs}.
14047 Macroblock size. Default @code{16}.
14050 Motion estimation search parameter. Default @code{32}.
14053 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).
14058 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:
14061 Disable scene change detection.
14063 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14065 Default method is @samp{fdiff}.
14067 @item scd_threshold
14068 Scene change detection threshold. Default is @code{10.}.
14073 Mix several video input streams into one video stream.
14075 A description of the accepted options follows.
14079 The number of inputs. If unspecified, it defaults to 2.
14082 Specify weight of each input video stream as sequence.
14083 Each weight is separated by space. If number of weights
14084 is smaller than number of @var{frames} last specified
14085 weight will be used for all remaining unset weights.
14088 Specify scale, if it is set it will be multiplied with sum
14089 of each weight multiplied with pixel values to give final destination
14090 pixel value. By default @var{scale} is auto scaled to sum of weights.
14093 Specify how end of stream is determined.
14096 The duration of the longest input. (default)
14099 The duration of the shortest input.
14102 The duration of the first input.
14106 @section mpdecimate
14108 Drop frames that do not differ greatly from the previous frame in
14109 order to reduce frame rate.
14111 The main use of this filter is for very-low-bitrate encoding
14112 (e.g. streaming over dialup modem), but it could in theory be used for
14113 fixing movies that were inverse-telecined incorrectly.
14115 A description of the accepted options follows.
14119 Set the maximum number of consecutive frames which can be dropped (if
14120 positive), or the minimum interval between dropped frames (if
14121 negative). If the value is 0, the frame is dropped disregarding the
14122 number of previous sequentially dropped frames.
14124 Default value is 0.
14129 Set the dropping threshold values.
14131 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14132 represent actual pixel value differences, so a threshold of 64
14133 corresponds to 1 unit of difference for each pixel, or the same spread
14134 out differently over the block.
14136 A frame is a candidate for dropping if no 8x8 blocks differ by more
14137 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14138 meaning the whole image) differ by more than a threshold of @option{lo}.
14140 Default value for @option{hi} is 64*12, default value for @option{lo} is
14141 64*5, and default value for @option{frac} is 0.33.
14147 Negate (invert) the input video.
14149 It accepts the following option:
14154 With value 1, it negates the alpha component, if present. Default value is 0.
14160 Denoise frames using Non-Local Means algorithm.
14162 Each pixel is adjusted by looking for other pixels with similar contexts. This
14163 context similarity is defined by comparing their surrounding patches of size
14164 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14167 Note that the research area defines centers for patches, which means some
14168 patches will be made of pixels outside that research area.
14170 The filter accepts the following options.
14174 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14177 Set patch size. Default is 7. Must be odd number in range [0, 99].
14180 Same as @option{p} but for chroma planes.
14182 The default value is @var{0} and means automatic.
14185 Set research size. Default is 15. Must be odd number in range [0, 99].
14188 Same as @option{r} but for chroma planes.
14190 The default value is @var{0} and means automatic.
14195 Deinterlace video using neural network edge directed interpolation.
14197 This filter accepts the following options:
14201 Mandatory option, without binary file filter can not work.
14202 Currently file can be found here:
14203 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14206 Set which frames to deinterlace, by default it is @code{all}.
14207 Can be @code{all} or @code{interlaced}.
14210 Set mode of operation.
14212 Can be one of the following:
14216 Use frame flags, both fields.
14218 Use frame flags, single field.
14220 Use top field only.
14222 Use bottom field only.
14224 Use both fields, top first.
14226 Use both fields, bottom first.
14230 Set which planes to process, by default filter process all frames.
14233 Set size of local neighborhood around each pixel, used by the predictor neural
14236 Can be one of the following:
14249 Set the number of neurons in predictor neural network.
14250 Can be one of the following:
14261 Controls the number of different neural network predictions that are blended
14262 together to compute the final output value. Can be @code{fast}, default or
14266 Set which set of weights to use in the predictor.
14267 Can be one of the following:
14271 weights trained to minimize absolute error
14273 weights trained to minimize squared error
14277 Controls whether or not the prescreener neural network is used to decide
14278 which pixels should be processed by the predictor neural network and which
14279 can be handled by simple cubic interpolation.
14280 The prescreener is trained to know whether cubic interpolation will be
14281 sufficient for a pixel or whether it should be predicted by the predictor nn.
14282 The computational complexity of the prescreener nn is much less than that of
14283 the predictor nn. Since most pixels can be handled by cubic interpolation,
14284 using the prescreener generally results in much faster processing.
14285 The prescreener is pretty accurate, so the difference between using it and not
14286 using it is almost always unnoticeable.
14288 Can be one of the following:
14296 Default is @code{new}.
14299 Set various debugging flags.
14304 Force libavfilter not to use any of the specified pixel formats for the
14305 input to the next filter.
14307 It accepts the following parameters:
14311 A '|'-separated list of pixel format names, such as
14312 pix_fmts=yuv420p|monow|rgb24".
14316 @subsection Examples
14320 Force libavfilter to use a format different from @var{yuv420p} for the
14321 input to the vflip filter:
14323 noformat=pix_fmts=yuv420p,vflip
14327 Convert the input video to any of the formats not contained in the list:
14329 noformat=yuv420p|yuv444p|yuv410p
14335 Add noise on video input frame.
14337 The filter accepts the following options:
14345 Set noise seed for specific pixel component or all pixel components in case
14346 of @var{all_seed}. Default value is @code{123457}.
14348 @item all_strength, alls
14349 @item c0_strength, c0s
14350 @item c1_strength, c1s
14351 @item c2_strength, c2s
14352 @item c3_strength, c3s
14353 Set noise strength for specific pixel component or all pixel components in case
14354 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14356 @item all_flags, allf
14357 @item c0_flags, c0f
14358 @item c1_flags, c1f
14359 @item c2_flags, c2f
14360 @item c3_flags, c3f
14361 Set pixel component flags or set flags for all components if @var{all_flags}.
14362 Available values for component flags are:
14365 averaged temporal noise (smoother)
14367 mix random noise with a (semi)regular pattern
14369 temporal noise (noise pattern changes between frames)
14371 uniform noise (gaussian otherwise)
14375 @subsection Examples
14377 Add temporal and uniform noise to input video:
14379 noise=alls=20:allf=t+u
14384 Normalize RGB video (aka histogram stretching, contrast stretching).
14385 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14387 For each channel of each frame, the filter computes the input range and maps
14388 it linearly to the user-specified output range. The output range defaults
14389 to the full dynamic range from pure black to pure white.
14391 Temporal smoothing can be used on the input range to reduce flickering (rapid
14392 changes in brightness) caused when small dark or bright objects enter or leave
14393 the scene. This is similar to the auto-exposure (automatic gain control) on a
14394 video camera, and, like a video camera, it may cause a period of over- or
14395 under-exposure of the video.
14397 The R,G,B channels can be normalized independently, which may cause some
14398 color shifting, or linked together as a single channel, which prevents
14399 color shifting. Linked normalization preserves hue. Independent normalization
14400 does not, so it can be used to remove some color casts. Independent and linked
14401 normalization can be combined in any ratio.
14403 The normalize filter accepts the following options:
14408 Colors which define the output range. The minimum input value is mapped to
14409 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14410 The defaults are black and white respectively. Specifying white for
14411 @var{blackpt} and black for @var{whitept} will give color-inverted,
14412 normalized video. Shades of grey can be used to reduce the dynamic range
14413 (contrast). Specifying saturated colors here can create some interesting
14417 The number of previous frames to use for temporal smoothing. The input range
14418 of each channel is smoothed using a rolling average over the current frame
14419 and the @var{smoothing} previous frames. The default is 0 (no temporal
14423 Controls the ratio of independent (color shifting) channel normalization to
14424 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14425 independent. Defaults to 1.0 (fully independent).
14428 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14429 expensive no-op. Defaults to 1.0 (full strength).
14433 @subsection Commands
14434 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14435 The command accepts the same syntax of the corresponding option.
14437 If the specified expression is not valid, it is kept at its current
14440 @subsection Examples
14442 Stretch video contrast to use the full dynamic range, with no temporal
14443 smoothing; may flicker depending on the source content:
14445 normalize=blackpt=black:whitept=white:smoothing=0
14448 As above, but with 50 frames of temporal smoothing; flicker should be
14449 reduced, depending on the source content:
14451 normalize=blackpt=black:whitept=white:smoothing=50
14454 As above, but with hue-preserving linked channel normalization:
14456 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14459 As above, but with half strength:
14461 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14464 Map the darkest input color to red, the brightest input color to cyan:
14466 normalize=blackpt=red:whitept=cyan
14471 Pass the video source unchanged to the output.
14474 Optical Character Recognition
14476 This filter uses Tesseract for optical character recognition. To enable
14477 compilation of this filter, you need to configure FFmpeg with
14478 @code{--enable-libtesseract}.
14480 It accepts the following options:
14484 Set datapath to tesseract data. Default is to use whatever was
14485 set at installation.
14488 Set language, default is "eng".
14491 Set character whitelist.
14494 Set character blacklist.
14497 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14498 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14502 Apply a video transform using libopencv.
14504 To enable this filter, install the libopencv library and headers and
14505 configure FFmpeg with @code{--enable-libopencv}.
14507 It accepts the following parameters:
14512 The name of the libopencv filter to apply.
14514 @item filter_params
14515 The parameters to pass to the libopencv filter. If not specified, the default
14516 values are assumed.
14520 Refer to the official libopencv documentation for more precise
14522 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14524 Several libopencv filters are supported; see the following subsections.
14529 Dilate an image by using a specific structuring element.
14530 It corresponds to the libopencv function @code{cvDilate}.
14532 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14534 @var{struct_el} represents a structuring element, and has the syntax:
14535 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14537 @var{cols} and @var{rows} represent the number of columns and rows of
14538 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14539 point, and @var{shape} the shape for the structuring element. @var{shape}
14540 must be "rect", "cross", "ellipse", or "custom".
14542 If the value for @var{shape} is "custom", it must be followed by a
14543 string of the form "=@var{filename}". The file with name
14544 @var{filename} is assumed to represent a binary image, with each
14545 printable character corresponding to a bright pixel. When a custom
14546 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14547 or columns and rows of the read file are assumed instead.
14549 The default value for @var{struct_el} is "3x3+0x0/rect".
14551 @var{nb_iterations} specifies the number of times the transform is
14552 applied to the image, and defaults to 1.
14556 # Use the default values
14559 # Dilate using a structuring element with a 5x5 cross, iterating two times
14560 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14562 # Read the shape from the file diamond.shape, iterating two times.
14563 # The file diamond.shape may contain a pattern of characters like this
14569 # The specified columns and rows are ignored
14570 # but the anchor point coordinates are not
14571 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14576 Erode an image by using a specific structuring element.
14577 It corresponds to the libopencv function @code{cvErode}.
14579 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14580 with the same syntax and semantics as the @ref{dilate} filter.
14584 Smooth the input video.
14586 The filter takes the following parameters:
14587 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14589 @var{type} is the type of smooth filter to apply, and must be one of
14590 the following values: "blur", "blur_no_scale", "median", "gaussian",
14591 or "bilateral". The default value is "gaussian".
14593 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14594 depends on the smooth type. @var{param1} and
14595 @var{param2} accept integer positive values or 0. @var{param3} and
14596 @var{param4} accept floating point values.
14598 The default value for @var{param1} is 3. The default value for the
14599 other parameters is 0.
14601 These parameters correspond to the parameters assigned to the
14602 libopencv function @code{cvSmooth}.
14604 @section oscilloscope
14606 2D Video Oscilloscope.
14608 Useful to measure spatial impulse, step responses, chroma delays, etc.
14610 It accepts the following parameters:
14614 Set scope center x position.
14617 Set scope center y position.
14620 Set scope size, relative to frame diagonal.
14623 Set scope tilt/rotation.
14629 Set trace center x position.
14632 Set trace center y position.
14635 Set trace width, relative to width of frame.
14638 Set trace height, relative to height of frame.
14641 Set which components to trace. By default it traces first three components.
14644 Draw trace grid. By default is enabled.
14647 Draw some statistics. By default is enabled.
14650 Draw scope. By default is enabled.
14653 @subsection Commands
14654 This filter supports same @ref{commands} as options.
14655 The command accepts the same syntax of the corresponding option.
14657 If the specified expression is not valid, it is kept at its current
14660 @subsection Examples
14664 Inspect full first row of video frame.
14666 oscilloscope=x=0.5:y=0:s=1
14670 Inspect full last row of video frame.
14672 oscilloscope=x=0.5:y=1:s=1
14676 Inspect full 5th line of video frame of height 1080.
14678 oscilloscope=x=0.5:y=5/1080:s=1
14682 Inspect full last column of video frame.
14684 oscilloscope=x=1:y=0.5:s=1:t=1
14692 Overlay one video on top of another.
14694 It takes two inputs and has one output. The first input is the "main"
14695 video on which the second input is overlaid.
14697 It accepts the following parameters:
14699 A description of the accepted options follows.
14704 Set the expression for the x and y coordinates of the overlaid video
14705 on the main video. Default value is "0" for both expressions. In case
14706 the expression is invalid, it is set to a huge value (meaning that the
14707 overlay will not be displayed within the output visible area).
14710 See @ref{framesync}.
14713 Set when the expressions for @option{x}, and @option{y} are evaluated.
14715 It accepts the following values:
14718 only evaluate expressions once during the filter initialization or
14719 when a command is processed
14722 evaluate expressions for each incoming frame
14725 Default value is @samp{frame}.
14728 See @ref{framesync}.
14731 Set the format for the output video.
14733 It accepts the following values:
14736 force YUV420 output
14739 force YUV420p10 output
14742 force YUV422 output
14745 force YUV422p10 output
14748 force YUV444 output
14751 force packed RGB output
14754 force planar RGB output
14757 automatically pick format
14760 Default value is @samp{yuv420}.
14763 See @ref{framesync}.
14766 Set format of alpha of the overlaid video, it can be @var{straight} or
14767 @var{premultiplied}. Default is @var{straight}.
14770 The @option{x}, and @option{y} expressions can contain the following
14776 The main input width and height.
14780 The overlay input width and height.
14784 The computed values for @var{x} and @var{y}. They are evaluated for
14789 horizontal and vertical chroma subsample values of the output
14790 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
14794 the number of input frame, starting from 0
14797 the position in the file of the input frame, NAN if unknown
14800 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
14804 This filter also supports the @ref{framesync} options.
14806 Note that the @var{n}, @var{pos}, @var{t} variables are available only
14807 when evaluation is done @emph{per frame}, and will evaluate to NAN
14808 when @option{eval} is set to @samp{init}.
14810 Be aware that frames are taken from each input video in timestamp
14811 order, hence, if their initial timestamps differ, it is a good idea
14812 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
14813 have them begin in the same zero timestamp, as the example for
14814 the @var{movie} filter does.
14816 You can chain together more overlays but you should test the
14817 efficiency of such approach.
14819 @subsection Commands
14821 This filter supports the following commands:
14825 Modify the x and y of the overlay input.
14826 The command accepts the same syntax of the corresponding option.
14828 If the specified expression is not valid, it is kept at its current
14832 @subsection Examples
14836 Draw the overlay at 10 pixels from the bottom right corner of the main
14839 overlay=main_w-overlay_w-10:main_h-overlay_h-10
14842 Using named options the example above becomes:
14844 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
14848 Insert a transparent PNG logo in the bottom left corner of the input,
14849 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
14851 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
14855 Insert 2 different transparent PNG logos (second logo on bottom
14856 right corner) using the @command{ffmpeg} tool:
14858 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
14862 Add a transparent color layer on top of the main video; @code{WxH}
14863 must specify the size of the main input to the overlay filter:
14865 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
14869 Play an original video and a filtered version (here with the deshake
14870 filter) side by side using the @command{ffplay} tool:
14872 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
14875 The above command is the same as:
14877 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
14881 Make a sliding overlay appearing from the left to the right top part of the
14882 screen starting since time 2:
14884 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
14888 Compose output by putting two input videos side to side:
14890 ffmpeg -i left.avi -i right.avi -filter_complex "
14891 nullsrc=size=200x100 [background];
14892 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
14893 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
14894 [background][left] overlay=shortest=1 [background+left];
14895 [background+left][right] overlay=shortest=1:x=100 [left+right]
14900 Mask 10-20 seconds of a video by applying the delogo filter to a section
14902 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
14903 -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]'
14908 Chain several overlays in cascade:
14910 nullsrc=s=200x200 [bg];
14911 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
14912 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
14913 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
14914 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
14915 [in3] null, [mid2] overlay=100:100 [out0]
14920 @anchor{overlay_cuda}
14921 @section overlay_cuda
14923 Overlay one video on top of another.
14925 This is the CUDA variant of the @ref{overlay} filter.
14926 It only accepts CUDA frames. The underlying input pixel formats have to match.
14928 It takes two inputs and has one output. The first input is the "main"
14929 video on which the second input is overlaid.
14931 It accepts the following parameters:
14936 Set the x and y coordinates of the overlaid video on the main video.
14937 Default value is "0" for both expressions.
14940 See @ref{framesync}.
14943 See @ref{framesync}.
14946 See @ref{framesync}.
14950 This filter also supports the @ref{framesync} options.
14954 Apply Overcomplete Wavelet denoiser.
14956 The filter accepts the following options:
14962 Larger depth values will denoise lower frequency components more, but
14963 slow down filtering.
14965 Must be an int in the range 8-16, default is @code{8}.
14967 @item luma_strength, ls
14970 Must be a double value in the range 0-1000, default is @code{1.0}.
14972 @item chroma_strength, cs
14973 Set chroma strength.
14975 Must be a double value in the range 0-1000, default is @code{1.0}.
14981 Add paddings to the input image, and place the original input at the
14982 provided @var{x}, @var{y} coordinates.
14984 It accepts the following parameters:
14989 Specify an expression for the size of the output image with the
14990 paddings added. If the value for @var{width} or @var{height} is 0, the
14991 corresponding input size is used for the output.
14993 The @var{width} expression can reference the value set by the
14994 @var{height} expression, and vice versa.
14996 The default value of @var{width} and @var{height} is 0.
15000 Specify the offsets to place the input image at within the padded area,
15001 with respect to the top/left border of the output image.
15003 The @var{x} expression can reference the value set by the @var{y}
15004 expression, and vice versa.
15006 The default value of @var{x} and @var{y} is 0.
15008 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15009 so the input image is centered on the padded area.
15012 Specify the color of the padded area. For the syntax of this option,
15013 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15014 manual,ffmpeg-utils}.
15016 The default value of @var{color} is "black".
15019 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15021 It accepts the following values:
15025 Only evaluate expressions once during the filter initialization or when
15026 a command is processed.
15029 Evaluate expressions for each incoming frame.
15033 Default value is @samp{init}.
15036 Pad to aspect instead to a resolution.
15040 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15041 options are expressions containing the following constants:
15046 The input video width and height.
15050 These are the same as @var{in_w} and @var{in_h}.
15054 The output width and height (the size of the padded area), as
15055 specified by the @var{width} and @var{height} expressions.
15059 These are the same as @var{out_w} and @var{out_h}.
15063 The x and y offsets as specified by the @var{x} and @var{y}
15064 expressions, or NAN if not yet specified.
15067 same as @var{iw} / @var{ih}
15070 input sample aspect ratio
15073 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15077 The horizontal and vertical chroma subsample values. For example for the
15078 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15081 @subsection Examples
15085 Add paddings with the color "violet" to the input video. The output video
15086 size is 640x480, and the top-left corner of the input video is placed at
15089 pad=640:480:0:40:violet
15092 The example above is equivalent to the following command:
15094 pad=width=640:height=480:x=0:y=40:color=violet
15098 Pad the input to get an output with dimensions increased by 3/2,
15099 and put the input video at the center of the padded area:
15101 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15105 Pad the input to get a squared output with size equal to the maximum
15106 value between the input width and height, and put the input video at
15107 the center of the padded area:
15109 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15113 Pad the input to get a final w/h ratio of 16:9:
15115 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15119 In case of anamorphic video, in order to set the output display aspect
15120 correctly, it is necessary to use @var{sar} in the expression,
15121 according to the relation:
15123 (ih * X / ih) * sar = output_dar
15124 X = output_dar / sar
15127 Thus the previous example needs to be modified to:
15129 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15133 Double the output size and put the input video in the bottom-right
15134 corner of the output padded area:
15136 pad="2*iw:2*ih:ow-iw:oh-ih"
15140 @anchor{palettegen}
15141 @section palettegen
15143 Generate one palette for a whole video stream.
15145 It accepts the following options:
15149 Set the maximum number of colors to quantize in the palette.
15150 Note: the palette will still contain 256 colors; the unused palette entries
15153 @item reserve_transparent
15154 Create a palette of 255 colors maximum and reserve the last one for
15155 transparency. Reserving the transparency color is useful for GIF optimization.
15156 If not set, the maximum of colors in the palette will be 256. You probably want
15157 to disable this option for a standalone image.
15160 @item transparency_color
15161 Set the color that will be used as background for transparency.
15164 Set statistics mode.
15166 It accepts the following values:
15169 Compute full frame histograms.
15171 Compute histograms only for the part that differs from previous frame. This
15172 might be relevant to give more importance to the moving part of your input if
15173 the background is static.
15175 Compute new histogram for each frame.
15178 Default value is @var{full}.
15181 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15182 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15183 color quantization of the palette. This information is also visible at
15184 @var{info} logging level.
15186 @subsection Examples
15190 Generate a representative palette of a given video using @command{ffmpeg}:
15192 ffmpeg -i input.mkv -vf palettegen palette.png
15196 @section paletteuse
15198 Use a palette to downsample an input video stream.
15200 The filter takes two inputs: one video stream and a palette. The palette must
15201 be a 256 pixels image.
15203 It accepts the following options:
15207 Select dithering mode. Available algorithms are:
15210 Ordered 8x8 bayer dithering (deterministic)
15212 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15213 Note: this dithering is sometimes considered "wrong" and is included as a
15215 @item floyd_steinberg
15216 Floyd and Steingberg dithering (error diffusion)
15218 Frankie Sierra dithering v2 (error diffusion)
15220 Frankie Sierra dithering v2 "Lite" (error diffusion)
15223 Default is @var{sierra2_4a}.
15226 When @var{bayer} dithering is selected, this option defines the scale of the
15227 pattern (how much the crosshatch pattern is visible). A low value means more
15228 visible pattern for less banding, and higher value means less visible pattern
15229 at the cost of more banding.
15231 The option must be an integer value in the range [0,5]. Default is @var{2}.
15234 If set, define the zone to process
15238 Only the changing rectangle will be reprocessed. This is similar to GIF
15239 cropping/offsetting compression mechanism. This option can be useful for speed
15240 if only a part of the image is changing, and has use cases such as limiting the
15241 scope of the error diffusal @option{dither} to the rectangle that bounds the
15242 moving scene (it leads to more deterministic output if the scene doesn't change
15243 much, and as a result less moving noise and better GIF compression).
15246 Default is @var{none}.
15249 Take new palette for each output frame.
15251 @item alpha_threshold
15252 Sets the alpha threshold for transparency. Alpha values above this threshold
15253 will be treated as completely opaque, and values below this threshold will be
15254 treated as completely transparent.
15256 The option must be an integer value in the range [0,255]. Default is @var{128}.
15259 @subsection Examples
15263 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15264 using @command{ffmpeg}:
15266 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15270 @section perspective
15272 Correct perspective of video not recorded perpendicular to the screen.
15274 A description of the accepted parameters follows.
15285 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15286 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15287 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15288 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15289 then the corners of the source will be sent to the specified coordinates.
15291 The expressions can use the following variables:
15296 the width and height of video frame.
15300 Output frame count.
15303 @item interpolation
15304 Set interpolation for perspective correction.
15306 It accepts the following values:
15312 Default value is @samp{linear}.
15315 Set interpretation of coordinate options.
15317 It accepts the following values:
15321 Send point in the source specified by the given coordinates to
15322 the corners of the destination.
15324 @item 1, destination
15326 Send the corners of the source to the point in the destination specified
15327 by the given coordinates.
15329 Default value is @samp{source}.
15333 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15335 It accepts the following values:
15338 only evaluate expressions once during the filter initialization or
15339 when a command is processed
15342 evaluate expressions for each incoming frame
15345 Default value is @samp{init}.
15350 Delay interlaced video by one field time so that the field order changes.
15352 The intended use is to fix PAL movies that have been captured with the
15353 opposite field order to the film-to-video transfer.
15355 A description of the accepted parameters follows.
15361 It accepts the following values:
15364 Capture field order top-first, transfer bottom-first.
15365 Filter will delay the bottom field.
15368 Capture field order bottom-first, transfer top-first.
15369 Filter will delay the top field.
15372 Capture and transfer with the same field order. This mode only exists
15373 for the documentation of the other options to refer to, but if you
15374 actually select it, the filter will faithfully do nothing.
15377 Capture field order determined automatically by field flags, transfer
15379 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15380 basis using field flags. If no field information is available,
15381 then this works just like @samp{u}.
15384 Capture unknown or varying, transfer opposite.
15385 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15386 analyzing the images and selecting the alternative that produces best
15387 match between the fields.
15390 Capture top-first, transfer unknown or varying.
15391 Filter selects among @samp{t} and @samp{p} using image analysis.
15394 Capture bottom-first, transfer unknown or varying.
15395 Filter selects among @samp{b} and @samp{p} using image analysis.
15398 Capture determined by field flags, transfer unknown or varying.
15399 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15400 image analysis. If no field information is available, then this works just
15401 like @samp{U}. This is the default mode.
15404 Both capture and transfer unknown or varying.
15405 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15409 @section photosensitivity
15410 Reduce various flashes in video, so to help users with epilepsy.
15412 It accepts the following options:
15415 Set how many frames to use when filtering. Default is 30.
15418 Set detection threshold factor. Default is 1.
15422 Set how many pixels to skip when sampling frames. Default is 1.
15423 Allowed range is from 1 to 1024.
15426 Leave frames unchanged. Default is disabled.
15429 @section pixdesctest
15431 Pixel format descriptor test filter, mainly useful for internal
15432 testing. The output video should be equal to the input video.
15436 format=monow, pixdesctest
15439 can be used to test the monowhite pixel format descriptor definition.
15443 Display sample values of color channels. Mainly useful for checking color
15444 and levels. Minimum supported resolution is 640x480.
15446 The filters accept the following options:
15450 Set scope X position, relative offset on X axis.
15453 Set scope Y position, relative offset on Y axis.
15462 Set window opacity. This window also holds statistics about pixel area.
15465 Set window X position, relative offset on X axis.
15468 Set window Y position, relative offset on Y axis.
15473 Enable the specified chain of postprocessing subfilters using libpostproc. This
15474 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15475 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15476 Each subfilter and some options have a short and a long name that can be used
15477 interchangeably, i.e. dr/dering are the same.
15479 The filters accept the following options:
15483 Set postprocessing subfilters string.
15486 All subfilters share common options to determine their scope:
15490 Honor the quality commands for this subfilter.
15493 Do chrominance filtering, too (default).
15496 Do luminance filtering only (no chrominance).
15499 Do chrominance filtering only (no luminance).
15502 These options can be appended after the subfilter name, separated by a '|'.
15504 Available subfilters are:
15507 @item hb/hdeblock[|difference[|flatness]]
15508 Horizontal deblocking filter
15511 Difference factor where higher values mean more deblocking (default: @code{32}).
15513 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15516 @item vb/vdeblock[|difference[|flatness]]
15517 Vertical deblocking filter
15520 Difference factor where higher values mean more deblocking (default: @code{32}).
15522 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15525 @item ha/hadeblock[|difference[|flatness]]
15526 Accurate horizontal deblocking filter
15529 Difference factor where higher values mean more deblocking (default: @code{32}).
15531 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15534 @item va/vadeblock[|difference[|flatness]]
15535 Accurate vertical deblocking filter
15538 Difference factor where higher values mean more deblocking (default: @code{32}).
15540 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15544 The horizontal and vertical deblocking filters share the difference and
15545 flatness values so you cannot set different horizontal and vertical
15549 @item h1/x1hdeblock
15550 Experimental horizontal deblocking filter
15552 @item v1/x1vdeblock
15553 Experimental vertical deblocking filter
15558 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15561 larger -> stronger filtering
15563 larger -> stronger filtering
15565 larger -> stronger filtering
15568 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15571 Stretch luminance to @code{0-255}.
15574 @item lb/linblenddeint
15575 Linear blend deinterlacing filter that deinterlaces the given block by
15576 filtering all lines with a @code{(1 2 1)} filter.
15578 @item li/linipoldeint
15579 Linear interpolating deinterlacing filter that deinterlaces the given block by
15580 linearly interpolating every second line.
15582 @item ci/cubicipoldeint
15583 Cubic interpolating deinterlacing filter deinterlaces the given block by
15584 cubically interpolating every second line.
15586 @item md/mediandeint
15587 Median deinterlacing filter that deinterlaces the given block by applying a
15588 median filter to every second line.
15590 @item fd/ffmpegdeint
15591 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15592 second line with a @code{(-1 4 2 4 -1)} filter.
15595 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15596 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15598 @item fq/forceQuant[|quantizer]
15599 Overrides the quantizer table from the input with the constant quantizer you
15607 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15610 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15613 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15616 @subsection Examples
15620 Apply horizontal and vertical deblocking, deringing and automatic
15621 brightness/contrast:
15627 Apply default filters without brightness/contrast correction:
15633 Apply default filters and temporal denoiser:
15635 pp=default/tmpnoise|1|2|3
15639 Apply deblocking on luminance only, and switch vertical deblocking on or off
15640 automatically depending on available CPU time:
15647 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
15648 similar to spp = 6 with 7 point DCT, where only the center sample is
15651 The filter accepts the following options:
15655 Force a constant quantization parameter. It accepts an integer in range
15656 0 to 63. If not set, the filter will use the QP from the video stream
15660 Set thresholding mode. Available modes are:
15664 Set hard thresholding.
15666 Set soft thresholding (better de-ringing effect, but likely blurrier).
15668 Set medium thresholding (good results, default).
15672 @section premultiply
15673 Apply alpha premultiply effect to input video stream using first plane
15674 of second stream as alpha.
15676 Both streams must have same dimensions and same pixel format.
15678 The filter accepts the following option:
15682 Set which planes will be processed, unprocessed planes will be copied.
15683 By default value 0xf, all planes will be processed.
15686 Do not require 2nd input for processing, instead use alpha plane from input stream.
15690 Apply prewitt operator to input video stream.
15692 The filter accepts the following option:
15696 Set which planes will be processed, unprocessed planes will be copied.
15697 By default value 0xf, all planes will be processed.
15700 Set value which will be multiplied with filtered result.
15703 Set value which will be added to filtered result.
15706 @section pseudocolor
15708 Alter frame colors in video with pseudocolors.
15710 This filter accepts the following options:
15714 set pixel first component expression
15717 set pixel second component expression
15720 set pixel third component expression
15723 set pixel fourth component expression, corresponds to the alpha component
15726 set component to use as base for altering colors
15729 Each of them specifies the expression to use for computing the lookup table for
15730 the corresponding pixel component values.
15732 The expressions can contain the following constants and functions:
15737 The input width and height.
15740 The input value for the pixel component.
15742 @item ymin, umin, vmin, amin
15743 The minimum allowed component value.
15745 @item ymax, umax, vmax, amax
15746 The maximum allowed component value.
15749 All expressions default to "val".
15751 @subsection Examples
15755 Change too high luma values to gradient:
15757 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'"
15763 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
15764 Ratio) between two input videos.
15766 This filter takes in input two input videos, the first input is
15767 considered the "main" source and is passed unchanged to the
15768 output. The second input is used as a "reference" video for computing
15771 Both video inputs must have the same resolution and pixel format for
15772 this filter to work correctly. Also it assumes that both inputs
15773 have the same number of frames, which are compared one by one.
15775 The obtained average PSNR is printed through the logging system.
15777 The filter stores the accumulated MSE (mean squared error) of each
15778 frame, and at the end of the processing it is averaged across all frames
15779 equally, and the following formula is applied to obtain the PSNR:
15782 PSNR = 10*log10(MAX^2/MSE)
15785 Where MAX is the average of the maximum values of each component of the
15788 The description of the accepted parameters follows.
15791 @item stats_file, f
15792 If specified the filter will use the named file to save the PSNR of
15793 each individual frame. When filename equals "-" the data is sent to
15796 @item stats_version
15797 Specifies which version of the stats file format to use. Details of
15798 each format are written below.
15799 Default value is 1.
15801 @item stats_add_max
15802 Determines whether the max value is output to the stats log.
15803 Default value is 0.
15804 Requires stats_version >= 2. If this is set and stats_version < 2,
15805 the filter will return an error.
15808 This filter also supports the @ref{framesync} options.
15810 The file printed if @var{stats_file} is selected, contains a sequence of
15811 key/value pairs of the form @var{key}:@var{value} for each compared
15814 If a @var{stats_version} greater than 1 is specified, a header line precedes
15815 the list of per-frame-pair stats, with key value pairs following the frame
15816 format with the following parameters:
15819 @item psnr_log_version
15820 The version of the log file format. Will match @var{stats_version}.
15823 A comma separated list of the per-frame-pair parameters included in
15827 A description of each shown per-frame-pair parameter follows:
15831 sequential number of the input frame, starting from 1
15834 Mean Square Error pixel-by-pixel average difference of the compared
15835 frames, averaged over all the image components.
15837 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
15838 Mean Square Error pixel-by-pixel average difference of the compared
15839 frames for the component specified by the suffix.
15841 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
15842 Peak Signal to Noise ratio of the compared frames for the component
15843 specified by the suffix.
15845 @item max_avg, max_y, max_u, max_v
15846 Maximum allowed value for each channel, and average over all
15850 @subsection Examples
15855 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
15856 [main][ref] psnr="stats_file=stats.log" [out]
15859 On this example the input file being processed is compared with the
15860 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
15861 is stored in @file{stats.log}.
15864 Another example with different containers:
15866 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 -
15873 Pulldown reversal (inverse telecine) filter, capable of handling mixed
15874 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
15877 The pullup filter is designed to take advantage of future context in making
15878 its decisions. This filter is stateless in the sense that it does not lock
15879 onto a pattern to follow, but it instead looks forward to the following
15880 fields in order to identify matches and rebuild progressive frames.
15882 To produce content with an even framerate, insert the fps filter after
15883 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
15884 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
15886 The filter accepts the following options:
15893 These options set the amount of "junk" to ignore at the left, right, top, and
15894 bottom of the image, respectively. Left and right are in units of 8 pixels,
15895 while top and bottom are in units of 2 lines.
15896 The default is 8 pixels on each side.
15899 Set the strict breaks. Setting this option to 1 will reduce the chances of
15900 filter generating an occasional mismatched frame, but it may also cause an
15901 excessive number of frames to be dropped during high motion sequences.
15902 Conversely, setting it to -1 will make filter match fields more easily.
15903 This may help processing of video where there is slight blurring between
15904 the fields, but may also cause there to be interlaced frames in the output.
15905 Default value is @code{0}.
15908 Set the metric plane to use. It accepts the following values:
15914 Use chroma blue plane.
15917 Use chroma red plane.
15920 This option may be set to use chroma plane instead of the default luma plane
15921 for doing filter's computations. This may improve accuracy on very clean
15922 source material, but more likely will decrease accuracy, especially if there
15923 is chroma noise (rainbow effect) or any grayscale video.
15924 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
15925 load and make pullup usable in realtime on slow machines.
15928 For best results (without duplicated frames in the output file) it is
15929 necessary to change the output frame rate. For example, to inverse
15930 telecine NTSC input:
15932 ffmpeg -i input -vf pullup -r 24000/1001 ...
15937 Change video quantization parameters (QP).
15939 The filter accepts the following option:
15943 Set expression for quantization parameter.
15946 The expression is evaluated through the eval API and can contain, among others,
15947 the following constants:
15951 1 if index is not 129, 0 otherwise.
15954 Sequential index starting from -129 to 128.
15957 @subsection Examples
15961 Some equation like:
15969 Flush video frames from internal cache of frames into a random order.
15970 No frame is discarded.
15971 Inspired by @ref{frei0r} nervous filter.
15975 Set size in number of frames of internal cache, in range from @code{2} to
15976 @code{512}. Default is @code{30}.
15979 Set seed for random number generator, must be an integer included between
15980 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
15981 less than @code{0}, the filter will try to use a good random seed on a
15985 @section readeia608
15987 Read closed captioning (EIA-608) information from the top lines of a video frame.
15989 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
15990 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
15991 with EIA-608 data (starting from 0). A description of each metadata value follows:
15994 @item lavfi.readeia608.X.cc
15995 The two bytes stored as EIA-608 data (printed in hexadecimal).
15997 @item lavfi.readeia608.X.line
15998 The number of the line on which the EIA-608 data was identified and read.
16001 This filter accepts the following options:
16005 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16008 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16011 Set the ratio of width reserved for sync code detection.
16012 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16015 Enable checking the parity bit. In the event of a parity error, the filter will output
16016 @code{0x00} for that character. Default is false.
16019 Lowpass lines prior to further processing. Default is enabled.
16022 @subsection Commands
16024 This filter supports the all above options as @ref{commands}.
16026 @subsection Examples
16030 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16032 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
16038 Read vertical interval timecode (VITC) information from the top lines of a
16041 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16042 timecode value, if a valid timecode has been detected. Further metadata key
16043 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16044 timecode data has been found or not.
16046 This filter accepts the following options:
16050 Set the maximum number of lines to scan for VITC data. If the value is set to
16051 @code{-1} the full video frame is scanned. Default is @code{45}.
16054 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16055 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16058 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16059 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16062 @subsection Examples
16066 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16067 draw @code{--:--:--:--} as a placeholder:
16069 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16075 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16077 Destination pixel at position (X, Y) will be picked from source (x, y) position
16078 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16079 value for pixel will be used for destination pixel.
16081 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16082 will have Xmap/Ymap video stream dimensions.
16083 Xmap and Ymap input video streams are 16bit depth, single channel.
16087 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16088 Default is @code{color}.
16091 Specify the color of the unmapped pixels. For the syntax of this option,
16092 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16093 manual,ffmpeg-utils}. Default color is @code{black}.
16096 @section removegrain
16098 The removegrain filter is a spatial denoiser for progressive video.
16102 Set mode for the first plane.
16105 Set mode for the second plane.
16108 Set mode for the third plane.
16111 Set mode for the fourth plane.
16114 Range of mode is from 0 to 24. Description of each mode follows:
16118 Leave input plane unchanged. Default.
16121 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16124 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16127 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16130 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16131 This is equivalent to a median filter.
16134 Line-sensitive clipping giving the minimal change.
16137 Line-sensitive clipping, intermediate.
16140 Line-sensitive clipping, intermediate.
16143 Line-sensitive clipping, intermediate.
16146 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16149 Replaces the target pixel with the closest neighbour.
16152 [1 2 1] horizontal and vertical kernel blur.
16158 Bob mode, interpolates top field from the line where the neighbours
16159 pixels are the closest.
16162 Bob mode, interpolates bottom field from the line where the neighbours
16163 pixels are the closest.
16166 Bob mode, interpolates top field. Same as 13 but with a more complicated
16167 interpolation formula.
16170 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16171 interpolation formula.
16174 Clips the pixel with the minimum and maximum of respectively the maximum and
16175 minimum of each pair of opposite neighbour pixels.
16178 Line-sensitive clipping using opposite neighbours whose greatest distance from
16179 the current pixel is minimal.
16182 Replaces the pixel with the average of its 8 neighbours.
16185 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16188 Clips pixels using the averages of opposite neighbour.
16191 Same as mode 21 but simpler and faster.
16194 Small edge and halo removal, but reputed useless.
16200 @section removelogo
16202 Suppress a TV station logo, using an image file to determine which
16203 pixels comprise the logo. It works by filling in the pixels that
16204 comprise the logo with neighboring pixels.
16206 The filter accepts the following options:
16210 Set the filter bitmap file, which can be any image format supported by
16211 libavformat. The width and height of the image file must match those of the
16212 video stream being processed.
16215 Pixels in the provided bitmap image with a value of zero are not
16216 considered part of the logo, non-zero pixels are considered part of
16217 the logo. If you use white (255) for the logo and black (0) for the
16218 rest, you will be safe. For making the filter bitmap, it is
16219 recommended to take a screen capture of a black frame with the logo
16220 visible, and then using a threshold filter followed by the erode
16221 filter once or twice.
16223 If needed, little splotches can be fixed manually. Remember that if
16224 logo pixels are not covered, the filter quality will be much
16225 reduced. Marking too many pixels as part of the logo does not hurt as
16226 much, but it will increase the amount of blurring needed to cover over
16227 the image and will destroy more information than necessary, and extra
16228 pixels will slow things down on a large logo.
16230 @section repeatfields
16232 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16233 fields based on its value.
16237 Reverse a video clip.
16239 Warning: This filter requires memory to buffer the entire clip, so trimming
16242 @subsection Examples
16246 Take the first 5 seconds of a clip, and reverse it.
16253 Shift R/G/B/A pixels horizontally and/or vertically.
16255 The filter accepts the following options:
16258 Set amount to shift red horizontally.
16260 Set amount to shift red vertically.
16262 Set amount to shift green horizontally.
16264 Set amount to shift green vertically.
16266 Set amount to shift blue horizontally.
16268 Set amount to shift blue vertically.
16270 Set amount to shift alpha horizontally.
16272 Set amount to shift alpha vertically.
16274 Set edge mode, can be @var{smear}, default, or @var{warp}.
16277 @subsection Commands
16279 This filter supports the all above options as @ref{commands}.
16282 Apply roberts cross operator to input video stream.
16284 The filter accepts the following option:
16288 Set which planes will be processed, unprocessed planes will be copied.
16289 By default value 0xf, all planes will be processed.
16292 Set value which will be multiplied with filtered result.
16295 Set value which will be added to filtered result.
16300 Rotate video by an arbitrary angle expressed in radians.
16302 The filter accepts the following options:
16304 A description of the optional parameters follows.
16307 Set an expression for the angle by which to rotate the input video
16308 clockwise, expressed as a number of radians. A negative value will
16309 result in a counter-clockwise rotation. By default it is set to "0".
16311 This expression is evaluated for each frame.
16314 Set the output width expression, default value is "iw".
16315 This expression is evaluated just once during configuration.
16318 Set the output height expression, default value is "ih".
16319 This expression is evaluated just once during configuration.
16322 Enable bilinear interpolation if set to 1, a value of 0 disables
16323 it. Default value is 1.
16326 Set the color used to fill the output area not covered by the rotated
16327 image. For the general syntax of this option, check the
16328 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16329 If the special value "none" is selected then no
16330 background is printed (useful for example if the background is never shown).
16332 Default value is "black".
16335 The expressions for the angle and the output size can contain the
16336 following constants and functions:
16340 sequential number of the input frame, starting from 0. It is always NAN
16341 before the first frame is filtered.
16344 time in seconds of the input frame, it is set to 0 when the filter is
16345 configured. It is always NAN before the first frame is filtered.
16349 horizontal and vertical chroma subsample values. For example for the
16350 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16354 the input video width and height
16358 the output width and height, that is the size of the padded area as
16359 specified by the @var{width} and @var{height} expressions
16363 the minimal width/height required for completely containing the input
16364 video rotated by @var{a} radians.
16366 These are only available when computing the @option{out_w} and
16367 @option{out_h} expressions.
16370 @subsection Examples
16374 Rotate the input by PI/6 radians clockwise:
16380 Rotate the input by PI/6 radians counter-clockwise:
16386 Rotate the input by 45 degrees clockwise:
16392 Apply a constant rotation with period T, starting from an angle of PI/3:
16394 rotate=PI/3+2*PI*t/T
16398 Make the input video rotation oscillating with a period of T
16399 seconds and an amplitude of A radians:
16401 rotate=A*sin(2*PI/T*t)
16405 Rotate the video, output size is chosen so that the whole rotating
16406 input video is always completely contained in the output:
16408 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16412 Rotate the video, reduce the output size so that no background is ever
16415 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16419 @subsection Commands
16421 The filter supports the following commands:
16425 Set the angle expression.
16426 The command accepts the same syntax of the corresponding option.
16428 If the specified expression is not valid, it is kept at its current
16434 Apply Shape Adaptive Blur.
16436 The filter accepts the following options:
16439 @item luma_radius, lr
16440 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16441 value is 1.0. A greater value will result in a more blurred image, and
16442 in slower processing.
16444 @item luma_pre_filter_radius, lpfr
16445 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16448 @item luma_strength, ls
16449 Set luma maximum difference between pixels to still be considered, must
16450 be a value in the 0.1-100.0 range, default value is 1.0.
16452 @item chroma_radius, cr
16453 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16454 greater value will result in a more blurred image, and in slower
16457 @item chroma_pre_filter_radius, cpfr
16458 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16460 @item chroma_strength, cs
16461 Set chroma maximum difference between pixels to still be considered,
16462 must be a value in the -0.9-100.0 range.
16465 Each chroma option value, if not explicitly specified, is set to the
16466 corresponding luma option value.
16471 Scale (resize) the input video, using the libswscale library.
16473 The scale filter forces the output display aspect ratio to be the same
16474 of the input, by changing the output sample aspect ratio.
16476 If the input image format is different from the format requested by
16477 the next filter, the scale filter will convert the input to the
16480 @subsection Options
16481 The filter accepts the following options, or any of the options
16482 supported by the libswscale scaler.
16484 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16485 the complete list of scaler options.
16490 Set the output video dimension expression. Default value is the input
16493 If the @var{width} or @var{w} value is 0, the input width is used for
16494 the output. If the @var{height} or @var{h} value is 0, the input height
16495 is used for the output.
16497 If one and only one of the values is -n with n >= 1, the scale filter
16498 will use a value that maintains the aspect ratio of the input image,
16499 calculated from the other specified dimension. After that it will,
16500 however, make sure that the calculated dimension is divisible by n and
16501 adjust the value if necessary.
16503 If both values are -n with n >= 1, the behavior will be identical to
16504 both values being set to 0 as previously detailed.
16506 See below for the list of accepted constants for use in the dimension
16510 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16514 Only evaluate expressions once during the filter initialization or when a command is processed.
16517 Evaluate expressions for each incoming frame.
16521 Default value is @samp{init}.
16525 Set the interlacing mode. It accepts the following values:
16529 Force interlaced aware scaling.
16532 Do not apply interlaced scaling.
16535 Select interlaced aware scaling depending on whether the source frames
16536 are flagged as interlaced or not.
16539 Default value is @samp{0}.
16542 Set libswscale scaling flags. See
16543 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16544 complete list of values. If not explicitly specified the filter applies
16548 @item param0, param1
16549 Set libswscale input parameters for scaling algorithms that need them. See
16550 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16551 complete documentation. If not explicitly specified the filter applies
16557 Set the video size. For the syntax of this option, check the
16558 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16560 @item in_color_matrix
16561 @item out_color_matrix
16562 Set in/output YCbCr color space type.
16564 This allows the autodetected value to be overridden as well as allows forcing
16565 a specific value used for the output and encoder.
16567 If not specified, the color space type depends on the pixel format.
16573 Choose automatically.
16576 Format conforming to International Telecommunication Union (ITU)
16577 Recommendation BT.709.
16580 Set color space conforming to the United States Federal Communications
16581 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16586 Set color space conforming to:
16590 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16593 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16596 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16601 Set color space conforming to SMPTE ST 240:1999.
16604 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16609 Set in/output YCbCr sample range.
16611 This allows the autodetected value to be overridden as well as allows forcing
16612 a specific value used for the output and encoder. If not specified, the
16613 range depends on the pixel format. Possible values:
16617 Choose automatically.
16620 Set full range (0-255 in case of 8-bit luma).
16622 @item mpeg/limited/tv
16623 Set "MPEG" range (16-235 in case of 8-bit luma).
16626 @item force_original_aspect_ratio
16627 Enable decreasing or increasing output video width or height if necessary to
16628 keep the original aspect ratio. Possible values:
16632 Scale the video as specified and disable this feature.
16635 The output video dimensions will automatically be decreased if needed.
16638 The output video dimensions will automatically be increased if needed.
16642 One useful instance of this option is that when you know a specific device's
16643 maximum allowed resolution, you can use this to limit the output video to
16644 that, while retaining the aspect ratio. For example, device A allows
16645 1280x720 playback, and your video is 1920x800. Using this option (set it to
16646 decrease) and specifying 1280x720 to the command line makes the output
16649 Please note that this is a different thing than specifying -1 for @option{w}
16650 or @option{h}, you still need to specify the output resolution for this option
16653 @item force_divisible_by
16654 Ensures that both the output dimensions, width and height, are divisible by the
16655 given integer when used together with @option{force_original_aspect_ratio}. This
16656 works similar to using @code{-n} in the @option{w} and @option{h} options.
16658 This option respects the value set for @option{force_original_aspect_ratio},
16659 increasing or decreasing the resolution accordingly. The video's aspect ratio
16660 may be slightly modified.
16662 This option can be handy if you need to have a video fit within or exceed
16663 a defined resolution using @option{force_original_aspect_ratio} but also have
16664 encoder restrictions on width or height divisibility.
16668 The values of the @option{w} and @option{h} options are expressions
16669 containing the following constants:
16674 The input width and height
16678 These are the same as @var{in_w} and @var{in_h}.
16682 The output (scaled) width and height
16686 These are the same as @var{out_w} and @var{out_h}
16689 The same as @var{iw} / @var{ih}
16692 input sample aspect ratio
16695 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
16699 horizontal and vertical input chroma subsample values. For example for the
16700 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16704 horizontal and vertical output chroma subsample values. For example for the
16705 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16708 The (sequential) number of the input frame, starting from 0.
16709 Only available with @code{eval=frame}.
16712 The presentation timestamp of the input frame, expressed as a number of
16713 seconds. Only available with @code{eval=frame}.
16716 The position (byte offset) of the frame in the input stream, or NaN if
16717 this information is unavailable and/or meaningless (for example in case of synthetic video).
16718 Only available with @code{eval=frame}.
16721 @subsection Examples
16725 Scale the input video to a size of 200x100
16730 This is equivalent to:
16741 Specify a size abbreviation for the output size:
16746 which can also be written as:
16752 Scale the input to 2x:
16754 scale=w=2*iw:h=2*ih
16758 The above is the same as:
16760 scale=2*in_w:2*in_h
16764 Scale the input to 2x with forced interlaced scaling:
16766 scale=2*iw:2*ih:interl=1
16770 Scale the input to half size:
16772 scale=w=iw/2:h=ih/2
16776 Increase the width, and set the height to the same size:
16782 Seek Greek harmony:
16789 Increase the height, and set the width to 3/2 of the height:
16791 scale=w=3/2*oh:h=3/5*ih
16795 Increase the size, making the size a multiple of the chroma
16798 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
16802 Increase the width to a maximum of 500 pixels,
16803 keeping the same aspect ratio as the input:
16805 scale=w='min(500\, iw*3/2):h=-1'
16809 Make pixels square by combining scale and setsar:
16811 scale='trunc(ih*dar):ih',setsar=1/1
16815 Make pixels square by combining scale and setsar,
16816 making sure the resulting resolution is even (required by some codecs):
16818 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
16822 @subsection Commands
16824 This filter supports the following commands:
16828 Set the output video dimension expression.
16829 The command accepts the same syntax of the corresponding option.
16831 If the specified expression is not valid, it is kept at its current
16837 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
16838 format conversion on CUDA video frames. Setting the output width and height
16839 works in the same way as for the @var{scale} filter.
16841 The following additional options are accepted:
16844 The pixel format of the output CUDA frames. If set to the string "same" (the
16845 default), the input format will be kept. Note that automatic format negotiation
16846 and conversion is not yet supported for hardware frames
16849 The interpolation algorithm used for resizing. One of the following:
16856 @item cubic2p_bspline
16857 2-parameter cubic (B=1, C=0)
16859 @item cubic2p_catmullrom
16860 2-parameter cubic (B=0, C=1/2)
16862 @item cubic2p_b05c03
16863 2-parameter cubic (B=1/2, C=3/10)
16871 @item force_original_aspect_ratio
16872 Enable decreasing or increasing output video width or height if necessary to
16873 keep the original aspect ratio. Possible values:
16877 Scale the video as specified and disable this feature.
16880 The output video dimensions will automatically be decreased if needed.
16883 The output video dimensions will automatically be increased if needed.
16887 One useful instance of this option is that when you know a specific device's
16888 maximum allowed resolution, you can use this to limit the output video to
16889 that, while retaining the aspect ratio. For example, device A allows
16890 1280x720 playback, and your video is 1920x800. Using this option (set it to
16891 decrease) and specifying 1280x720 to the command line makes the output
16894 Please note that this is a different thing than specifying -1 for @option{w}
16895 or @option{h}, you still need to specify the output resolution for this option
16898 @item force_divisible_by
16899 Ensures that both the output dimensions, width and height, are divisible by the
16900 given integer when used together with @option{force_original_aspect_ratio}. This
16901 works similar to using @code{-n} in the @option{w} and @option{h} options.
16903 This option respects the value set for @option{force_original_aspect_ratio},
16904 increasing or decreasing the resolution accordingly. The video's aspect ratio
16905 may be slightly modified.
16907 This option can be handy if you need to have a video fit within or exceed
16908 a defined resolution using @option{force_original_aspect_ratio} but also have
16909 encoder restrictions on width or height divisibility.
16915 Scale (resize) the input video, based on a reference video.
16917 See the scale filter for available options, scale2ref supports the same but
16918 uses the reference video instead of the main input as basis. scale2ref also
16919 supports the following additional constants for the @option{w} and
16920 @option{h} options:
16925 The main input video's width and height
16928 The same as @var{main_w} / @var{main_h}
16931 The main input video's sample aspect ratio
16933 @item main_dar, mdar
16934 The main input video's display aspect ratio. Calculated from
16935 @code{(main_w / main_h) * main_sar}.
16939 The main input video's horizontal and vertical chroma subsample values.
16940 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
16944 The (sequential) number of the main input frame, starting from 0.
16945 Only available with @code{eval=frame}.
16948 The presentation timestamp of the main input frame, expressed as a number of
16949 seconds. Only available with @code{eval=frame}.
16952 The position (byte offset) of the frame in the main input stream, or NaN if
16953 this information is unavailable and/or meaningless (for example in case of synthetic video).
16954 Only available with @code{eval=frame}.
16957 @subsection Examples
16961 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
16963 'scale2ref[b][a];[a][b]overlay'
16967 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
16969 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
16973 @subsection Commands
16975 This filter supports the following commands:
16979 Set the output video dimension expression.
16980 The command accepts the same syntax of the corresponding option.
16982 If the specified expression is not valid, it is kept at its current
16987 Scroll input video horizontally and/or vertically by constant speed.
16989 The filter accepts the following options:
16991 @item horizontal, h
16992 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
16993 Negative values changes scrolling direction.
16996 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
16997 Negative values changes scrolling direction.
17000 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17003 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17006 @subsection Commands
17008 This filter supports the following @ref{commands}:
17010 @item horizontal, h
17011 Set the horizontal scrolling speed.
17013 Set the vertical scrolling speed.
17019 Detect video scene change.
17021 This filter sets frame metadata with mafd between frame, the scene score, and
17022 forward the frame to the next filter, so they can use these metadata to detect
17023 scene change or others.
17025 In addition, this filter logs a message and sets frame metadata when it detects
17026 a scene change by @option{threshold}.
17028 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17030 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17031 to detect scene change.
17033 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17034 detect scene change with @option{threshold}.
17036 The filter accepts the following options:
17040 Set the scene change detection threshold as a percentage of maximum change. Good
17041 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17044 Default value is @code{10.}.
17047 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17048 You can enable it if you want to get snapshot of scene change frames only.
17051 @anchor{selectivecolor}
17052 @section selectivecolor
17054 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17055 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17056 by the "purity" of the color (that is, how saturated it already is).
17058 This filter is similar to the Adobe Photoshop Selective Color tool.
17060 The filter accepts the following options:
17063 @item correction_method
17064 Select color correction method.
17066 Available values are:
17069 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17072 Specified adjustments are relative to the original component value.
17074 Default is @code{absolute}.
17076 Adjustments for red pixels (pixels where the red component is the maximum)
17078 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17080 Adjustments for green pixels (pixels where the green component is the maximum)
17082 Adjustments for cyan pixels (pixels where the red component is the minimum)
17084 Adjustments for blue pixels (pixels where the blue component is the maximum)
17086 Adjustments for magenta pixels (pixels where the green component is the minimum)
17088 Adjustments for white pixels (pixels where all components are greater than 128)
17090 Adjustments for all pixels except pure black and pure white
17092 Adjustments for black pixels (pixels where all components are lesser than 128)
17094 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17097 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17098 4 space separated floating point adjustment values in the [-1,1] range,
17099 respectively to adjust the amount of cyan, magenta, yellow and black for the
17100 pixels of its range.
17102 @subsection Examples
17106 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17107 increase magenta by 27% in blue areas:
17109 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17113 Use a Photoshop selective color preset:
17115 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17119 @anchor{separatefields}
17120 @section separatefields
17122 The @code{separatefields} takes a frame-based video input and splits
17123 each frame into its components fields, producing a new half height clip
17124 with twice the frame rate and twice the frame count.
17126 This filter use field-dominance information in frame to decide which
17127 of each pair of fields to place first in the output.
17128 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17130 @section setdar, setsar
17132 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17135 This is done by changing the specified Sample (aka Pixel) Aspect
17136 Ratio, according to the following equation:
17138 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17141 Keep in mind that the @code{setdar} filter does not modify the pixel
17142 dimensions of the video frame. Also, the display aspect ratio set by
17143 this filter may be changed by later filters in the filterchain,
17144 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17147 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17148 the filter output video.
17150 Note that as a consequence of the application of this filter, the
17151 output display aspect ratio will change according to the equation
17154 Keep in mind that the sample aspect ratio set by the @code{setsar}
17155 filter may be changed by later filters in the filterchain, e.g. if
17156 another "setsar" or a "setdar" filter is applied.
17158 It accepts the following parameters:
17161 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17162 Set the aspect ratio used by the filter.
17164 The parameter can be a floating point number string, an expression, or
17165 a string of the form @var{num}:@var{den}, where @var{num} and
17166 @var{den} are the numerator and denominator of the aspect ratio. If
17167 the parameter is not specified, it is assumed the value "0".
17168 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17172 Set the maximum integer value to use for expressing numerator and
17173 denominator when reducing the expressed aspect ratio to a rational.
17174 Default value is @code{100}.
17178 The parameter @var{sar} is an expression containing
17179 the following constants:
17183 These are approximated values for the mathematical constants e
17184 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17187 The input width and height.
17190 These are the same as @var{w} / @var{h}.
17193 The input sample aspect ratio.
17196 The input display aspect ratio. It is the same as
17197 (@var{w} / @var{h}) * @var{sar}.
17200 Horizontal and vertical chroma subsample values. For example, for the
17201 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17204 @subsection Examples
17209 To change the display aspect ratio to 16:9, specify one of the following:
17216 To change the sample aspect ratio to 10:11, specify:
17222 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17223 1000 in the aspect ratio reduction, use the command:
17225 setdar=ratio=16/9:max=1000
17233 Force field for the output video frame.
17235 The @code{setfield} filter marks the interlace type field for the
17236 output frames. It does not change the input frame, but only sets the
17237 corresponding property, which affects how the frame is treated by
17238 following filters (e.g. @code{fieldorder} or @code{yadif}).
17240 The filter accepts the following options:
17245 Available values are:
17249 Keep the same field property.
17252 Mark the frame as bottom-field-first.
17255 Mark the frame as top-field-first.
17258 Mark the frame as progressive.
17265 Force frame parameter for the output video frame.
17267 The @code{setparams} filter marks interlace and color range for the
17268 output frames. It does not change the input frame, but only sets the
17269 corresponding property, which affects how the frame is treated by
17274 Available values are:
17278 Keep the same field property (default).
17281 Mark the frame as bottom-field-first.
17284 Mark the frame as top-field-first.
17287 Mark the frame as progressive.
17291 Available values are:
17295 Keep the same color range property (default).
17297 @item unspecified, unknown
17298 Mark the frame as unspecified color range.
17300 @item limited, tv, mpeg
17301 Mark the frame as limited range.
17303 @item full, pc, jpeg
17304 Mark the frame as full range.
17307 @item color_primaries
17308 Set the color primaries.
17309 Available values are:
17313 Keep the same color primaries property (default).
17330 Set the color transfer.
17331 Available values are:
17335 Keep the same color trc property (default).
17357 Set the colorspace.
17358 Available values are:
17362 Keep the same colorspace property (default).
17375 @item chroma-derived-nc
17376 @item chroma-derived-c
17383 Show a line containing various information for each input video frame.
17384 The input video is not modified.
17386 This filter supports the following options:
17390 Calculate checksums of each plane. By default enabled.
17393 The shown line contains a sequence of key/value pairs of the form
17394 @var{key}:@var{value}.
17396 The following values are shown in the output:
17400 The (sequential) number of the input frame, starting from 0.
17403 The Presentation TimeStamp of the input frame, expressed as a number of
17404 time base units. The time base unit depends on the filter input pad.
17407 The Presentation TimeStamp of the input frame, expressed as a number of
17411 The position of the frame in the input stream, or -1 if this information is
17412 unavailable and/or meaningless (for example in case of synthetic video).
17415 The pixel format name.
17418 The sample aspect ratio of the input frame, expressed in the form
17419 @var{num}/@var{den}.
17422 The size of the input frame. For the syntax of this option, check the
17423 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17426 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17427 for bottom field first).
17430 This is 1 if the frame is a key frame, 0 otherwise.
17433 The picture type of the input frame ("I" for an I-frame, "P" for a
17434 P-frame, "B" for a B-frame, or "?" for an unknown type).
17435 Also refer to the documentation of the @code{AVPictureType} enum and of
17436 the @code{av_get_picture_type_char} function defined in
17437 @file{libavutil/avutil.h}.
17440 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17442 @item plane_checksum
17443 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17444 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17447 The mean value of pixels in each plane of the input frame, expressed in the form
17448 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17451 The standard deviation of pixel values in each plane of the input frame, expressed
17452 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17456 @section showpalette
17458 Displays the 256 colors palette of each frame. This filter is only relevant for
17459 @var{pal8} pixel format frames.
17461 It accepts the following option:
17465 Set the size of the box used to represent one palette color entry. Default is
17466 @code{30} (for a @code{30x30} pixel box).
17469 @section shuffleframes
17471 Reorder and/or duplicate and/or drop video frames.
17473 It accepts the following parameters:
17477 Set the destination indexes of input frames.
17478 This is space or '|' separated list of indexes that maps input frames to output
17479 frames. Number of indexes also sets maximal value that each index may have.
17480 '-1' index have special meaning and that is to drop frame.
17483 The first frame has the index 0. The default is to keep the input unchanged.
17485 @subsection Examples
17489 Swap second and third frame of every three frames of the input:
17491 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17495 Swap 10th and 1st frame of every ten frames of the input:
17497 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17501 @section shuffleplanes
17503 Reorder and/or duplicate video planes.
17505 It accepts the following parameters:
17510 The index of the input plane to be used as the first output plane.
17513 The index of the input plane to be used as the second output plane.
17516 The index of the input plane to be used as the third output plane.
17519 The index of the input plane to be used as the fourth output plane.
17523 The first plane has the index 0. The default is to keep the input unchanged.
17525 @subsection Examples
17529 Swap the second and third planes of the input:
17531 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17535 @anchor{signalstats}
17536 @section signalstats
17537 Evaluate various visual metrics that assist in determining issues associated
17538 with the digitization of analog video media.
17540 By default the filter will log these metadata values:
17544 Display the minimal Y value contained within the input frame. Expressed in
17548 Display the Y value at the 10% percentile within the input frame. Expressed in
17552 Display the average Y value within the input frame. Expressed in range of
17556 Display the Y value at the 90% percentile within the input frame. Expressed in
17560 Display the maximum Y value contained within the input frame. Expressed in
17564 Display the minimal U value contained within the input frame. Expressed in
17568 Display the U value at the 10% percentile within the input frame. Expressed in
17572 Display the average U value within the input frame. Expressed in range of
17576 Display the U value at the 90% percentile within the input frame. Expressed in
17580 Display the maximum U value contained within the input frame. Expressed in
17584 Display the minimal V value contained within the input frame. Expressed in
17588 Display the V value at the 10% percentile within the input frame. Expressed in
17592 Display the average V value within the input frame. Expressed in range of
17596 Display the V value at the 90% percentile within the input frame. Expressed in
17600 Display the maximum V value contained within the input frame. Expressed in
17604 Display the minimal saturation value contained within the input frame.
17605 Expressed in range of [0-~181.02].
17608 Display the saturation value at the 10% percentile within the input frame.
17609 Expressed in range of [0-~181.02].
17612 Display the average saturation value within the input frame. Expressed in range
17616 Display the saturation value at the 90% percentile within the input frame.
17617 Expressed in range of [0-~181.02].
17620 Display the maximum saturation value contained within the input frame.
17621 Expressed in range of [0-~181.02].
17624 Display the median value for hue within the input frame. Expressed in range of
17628 Display the average value for hue within the input frame. Expressed in range of
17632 Display the average of sample value difference between all values of the Y
17633 plane in the current frame and corresponding values of the previous input frame.
17634 Expressed in range of [0-255].
17637 Display the average of sample value difference between all values of the U
17638 plane in the current frame and corresponding values of the previous input frame.
17639 Expressed in range of [0-255].
17642 Display the average of sample value difference between all values of the V
17643 plane in the current frame and corresponding values of the previous input frame.
17644 Expressed in range of [0-255].
17647 Display bit depth of Y plane in current frame.
17648 Expressed in range of [0-16].
17651 Display bit depth of U plane in current frame.
17652 Expressed in range of [0-16].
17655 Display bit depth of V plane in current frame.
17656 Expressed in range of [0-16].
17659 The filter accepts the following options:
17665 @option{stat} specify an additional form of image analysis.
17666 @option{out} output video with the specified type of pixel highlighted.
17668 Both options accept the following values:
17672 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
17673 unlike the neighboring pixels of the same field. Examples of temporal outliers
17674 include the results of video dropouts, head clogs, or tape tracking issues.
17677 Identify @var{vertical line repetition}. Vertical line repetition includes
17678 similar rows of pixels within a frame. In born-digital video vertical line
17679 repetition is common, but this pattern is uncommon in video digitized from an
17680 analog source. When it occurs in video that results from the digitization of an
17681 analog source it can indicate concealment from a dropout compensator.
17684 Identify pixels that fall outside of legal broadcast range.
17688 Set the highlight color for the @option{out} option. The default color is
17692 @subsection Examples
17696 Output data of various video metrics:
17698 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
17702 Output specific data about the minimum and maximum values of the Y plane per frame:
17704 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
17708 Playback video while highlighting pixels that are outside of broadcast range in red.
17710 ffplay example.mov -vf signalstats="out=brng:color=red"
17714 Playback video with signalstats metadata drawn over the frame.
17716 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
17719 The contents of signalstat_drawtext.txt used in the command are:
17722 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
17723 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
17724 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
17725 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
17733 Calculates the MPEG-7 Video Signature. The filter can handle more than one
17734 input. In this case the matching between the inputs can be calculated additionally.
17735 The filter always passes through the first input. The signature of each stream can
17736 be written into a file.
17738 It accepts the following options:
17742 Enable or disable the matching process.
17744 Available values are:
17748 Disable the calculation of a matching (default).
17750 Calculate the matching for the whole video and output whether the whole video
17751 matches or only parts.
17753 Calculate only until a matching is found or the video ends. Should be faster in
17758 Set the number of inputs. The option value must be a non negative integer.
17759 Default value is 1.
17762 Set the path to which the output is written. If there is more than one input,
17763 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
17764 integer), that will be replaced with the input number. If no filename is
17765 specified, no output will be written. This is the default.
17768 Choose the output format.
17770 Available values are:
17774 Use the specified binary representation (default).
17776 Use the specified xml representation.
17780 Set threshold to detect one word as similar. The option value must be an integer
17781 greater than zero. The default value is 9000.
17784 Set threshold to detect all words as similar. The option value must be an integer
17785 greater than zero. The default value is 60000.
17788 Set threshold to detect frames as similar. The option value must be an integer
17789 greater than zero. The default value is 116.
17792 Set the minimum length of a sequence in frames to recognize it as matching
17793 sequence. The option value must be a non negative integer value.
17794 The default value is 0.
17797 Set the minimum relation, that matching frames to all frames must have.
17798 The option value must be a double value between 0 and 1. The default value is 0.5.
17801 @subsection Examples
17805 To calculate the signature of an input video and store it in signature.bin:
17807 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
17811 To detect whether two videos match and store the signatures in XML format in
17812 signature0.xml and signature1.xml:
17814 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 -
17822 Blur the input video without impacting the outlines.
17824 It accepts the following options:
17827 @item luma_radius, lr
17828 Set the luma radius. The option value must be a float number in
17829 the range [0.1,5.0] that specifies the variance of the gaussian filter
17830 used to blur the image (slower if larger). Default value is 1.0.
17832 @item luma_strength, ls
17833 Set the luma strength. The option value must be a float number
17834 in the range [-1.0,1.0] that configures the blurring. A value included
17835 in [0.0,1.0] will blur the image whereas a value included in
17836 [-1.0,0.0] will sharpen the image. Default value is 1.0.
17838 @item luma_threshold, lt
17839 Set the luma threshold used as a coefficient to determine
17840 whether a pixel should be blurred or not. The option value must be an
17841 integer in the range [-30,30]. A value of 0 will filter all the image,
17842 a value included in [0,30] will filter flat areas and a value included
17843 in [-30,0] will filter edges. Default value is 0.
17845 @item chroma_radius, cr
17846 Set the chroma radius. The option value must be a float number in
17847 the range [0.1,5.0] that specifies the variance of the gaussian filter
17848 used to blur the image (slower if larger). Default value is @option{luma_radius}.
17850 @item chroma_strength, cs
17851 Set the chroma strength. The option value must be a float number
17852 in the range [-1.0,1.0] that configures the blurring. A value included
17853 in [0.0,1.0] will blur the image whereas a value included in
17854 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
17856 @item chroma_threshold, ct
17857 Set the chroma threshold used as a coefficient to determine
17858 whether a pixel should be blurred or not. The option value must be an
17859 integer in the range [-30,30]. A value of 0 will filter all the image,
17860 a value included in [0,30] will filter flat areas and a value included
17861 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
17864 If a chroma option is not explicitly set, the corresponding luma value
17868 Apply sobel operator to input video stream.
17870 The filter accepts the following option:
17874 Set which planes will be processed, unprocessed planes will be copied.
17875 By default value 0xf, all planes will be processed.
17878 Set value which will be multiplied with filtered result.
17881 Set value which will be added to filtered result.
17887 Apply a simple postprocessing filter that compresses and decompresses the image
17888 at several (or - in the case of @option{quality} level @code{6} - all) shifts
17889 and average the results.
17891 The filter accepts the following options:
17895 Set quality. This option defines the number of levels for averaging. It accepts
17896 an integer in the range 0-6. If set to @code{0}, the filter will have no
17897 effect. A value of @code{6} means the higher quality. For each increment of
17898 that value the speed drops by a factor of approximately 2. Default value is
17902 Force a constant quantization parameter. If not set, the filter will use the QP
17903 from the video stream (if available).
17906 Set thresholding mode. Available modes are:
17910 Set hard thresholding (default).
17912 Set soft thresholding (better de-ringing effect, but likely blurrier).
17915 @item use_bframe_qp
17916 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
17917 option may cause flicker since the B-Frames have often larger QP. Default is
17918 @code{0} (not enabled).
17921 @subsection Commands
17923 This filter supports the following commands:
17925 @item quality, level
17926 Set quality level. The value @code{max} can be used to set the maximum level,
17927 currently @code{6}.
17933 Scale the input by applying one of the super-resolution methods based on
17934 convolutional neural networks. Supported models:
17938 Super-Resolution Convolutional Neural Network model (SRCNN).
17939 See @url{https://arxiv.org/abs/1501.00092}.
17942 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
17943 See @url{https://arxiv.org/abs/1609.05158}.
17946 Training scripts as well as scripts for model file (.pb) saving can be found at
17947 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
17948 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
17950 Native model files (.model) can be generated from TensorFlow model
17951 files (.pb) by using tools/python/convert.py
17953 The filter accepts the following options:
17957 Specify which DNN backend to use for model loading and execution. This option accepts
17958 the following values:
17962 Native implementation of DNN loading and execution.
17965 TensorFlow backend. To enable this backend you
17966 need to install the TensorFlow for C library (see
17967 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
17968 @code{--enable-libtensorflow}
17971 Default value is @samp{native}.
17974 Set path to model file specifying network architecture and its parameters.
17975 Note that different backends use different file formats. TensorFlow backend
17976 can load files for both formats, while native backend can load files for only
17980 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
17981 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
17982 input upscaled using bicubic upscaling with proper scale factor.
17985 This feature can also be finished with @ref{dnn_processing} filter.
17989 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
17991 This filter takes in input two input videos, the first input is
17992 considered the "main" source and is passed unchanged to the
17993 output. The second input is used as a "reference" video for computing
17996 Both video inputs must have the same resolution and pixel format for
17997 this filter to work correctly. Also it assumes that both inputs
17998 have the same number of frames, which are compared one by one.
18000 The filter stores the calculated SSIM of each frame.
18002 The description of the accepted parameters follows.
18005 @item stats_file, f
18006 If specified the filter will use the named file to save the SSIM of
18007 each individual frame. When filename equals "-" the data is sent to
18011 The file printed if @var{stats_file} is selected, contains a sequence of
18012 key/value pairs of the form @var{key}:@var{value} for each compared
18015 A description of each shown parameter follows:
18019 sequential number of the input frame, starting from 1
18021 @item Y, U, V, R, G, B
18022 SSIM of the compared frames for the component specified by the suffix.
18025 SSIM of the compared frames for the whole frame.
18028 Same as above but in dB representation.
18031 This filter also supports the @ref{framesync} options.
18033 @subsection Examples
18038 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18039 [main][ref] ssim="stats_file=stats.log" [out]
18042 On this example the input file being processed is compared with the
18043 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18044 is stored in @file{stats.log}.
18047 Another example with both psnr and ssim at same time:
18049 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18053 Another example with different containers:
18055 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 -
18061 Convert between different stereoscopic image formats.
18063 The filters accept the following options:
18067 Set stereoscopic image format of input.
18069 Available values for input image formats are:
18072 side by side parallel (left eye left, right eye right)
18075 side by side crosseye (right eye left, left eye right)
18078 side by side parallel with half width resolution
18079 (left eye left, right eye right)
18082 side by side crosseye with half width resolution
18083 (right eye left, left eye right)
18087 above-below (left eye above, right eye below)
18091 above-below (right eye above, left eye below)
18095 above-below with half height resolution
18096 (left eye above, right eye below)
18100 above-below with half height resolution
18101 (right eye above, left eye below)
18104 alternating frames (left eye first, right eye second)
18107 alternating frames (right eye first, left eye second)
18110 interleaved rows (left eye has top row, right eye starts on next row)
18113 interleaved rows (right eye has top row, left eye starts on next row)
18116 interleaved columns, left eye first
18119 interleaved columns, right eye first
18121 Default value is @samp{sbsl}.
18125 Set stereoscopic image format of output.
18129 side by side parallel (left eye left, right eye right)
18132 side by side crosseye (right eye left, left eye right)
18135 side by side parallel with half width resolution
18136 (left eye left, right eye right)
18139 side by side crosseye with half width resolution
18140 (right eye left, left eye right)
18144 above-below (left eye above, right eye below)
18148 above-below (right eye above, left eye below)
18152 above-below with half height resolution
18153 (left eye above, right eye below)
18157 above-below with half height resolution
18158 (right eye above, left eye below)
18161 alternating frames (left eye first, right eye second)
18164 alternating frames (right eye first, left eye second)
18167 interleaved rows (left eye has top row, right eye starts on next row)
18170 interleaved rows (right eye has top row, left eye starts on next row)
18173 anaglyph red/blue gray
18174 (red filter on left eye, blue filter on right eye)
18177 anaglyph red/green gray
18178 (red filter on left eye, green filter on right eye)
18181 anaglyph red/cyan gray
18182 (red filter on left eye, cyan filter on right eye)
18185 anaglyph red/cyan half colored
18186 (red filter on left eye, cyan filter on right eye)
18189 anaglyph red/cyan color
18190 (red filter on left eye, cyan filter on right eye)
18193 anaglyph red/cyan color optimized with the least squares projection of dubois
18194 (red filter on left eye, cyan filter on right eye)
18197 anaglyph green/magenta gray
18198 (green filter on left eye, magenta filter on right eye)
18201 anaglyph green/magenta half colored
18202 (green filter on left eye, magenta filter on right eye)
18205 anaglyph green/magenta colored
18206 (green filter on left eye, magenta filter on right eye)
18209 anaglyph green/magenta color optimized with the least squares projection of dubois
18210 (green filter on left eye, magenta filter on right eye)
18213 anaglyph yellow/blue gray
18214 (yellow filter on left eye, blue filter on right eye)
18217 anaglyph yellow/blue half colored
18218 (yellow filter on left eye, blue filter on right eye)
18221 anaglyph yellow/blue colored
18222 (yellow filter on left eye, blue filter on right eye)
18225 anaglyph yellow/blue color optimized with the least squares projection of dubois
18226 (yellow filter on left eye, blue filter on right eye)
18229 mono output (left eye only)
18232 mono output (right eye only)
18235 checkerboard, left eye first
18238 checkerboard, right eye first
18241 interleaved columns, left eye first
18244 interleaved columns, right eye first
18250 Default value is @samp{arcd}.
18253 @subsection Examples
18257 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18263 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18269 @section streamselect, astreamselect
18270 Select video or audio streams.
18272 The filter accepts the following options:
18276 Set number of inputs. Default is 2.
18279 Set input indexes to remap to outputs.
18282 @subsection Commands
18284 The @code{streamselect} and @code{astreamselect} filter supports the following
18289 Set input indexes to remap to outputs.
18292 @subsection Examples
18296 Select first 5 seconds 1st stream and rest of time 2nd stream:
18298 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18302 Same as above, but for audio:
18304 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18311 Draw subtitles on top of input video using the libass library.
18313 To enable compilation of this filter you need to configure FFmpeg with
18314 @code{--enable-libass}. This filter also requires a build with libavcodec and
18315 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18316 Alpha) subtitles format.
18318 The filter accepts the following options:
18322 Set the filename of the subtitle file to read. It must be specified.
18324 @item original_size
18325 Specify the size of the original video, the video for which the ASS file
18326 was composed. For the syntax of this option, check the
18327 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18328 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18329 correctly scale the fonts if the aspect ratio has been changed.
18332 Set a directory path containing fonts that can be used by the filter.
18333 These fonts will be used in addition to whatever the font provider uses.
18336 Process alpha channel, by default alpha channel is untouched.
18339 Set subtitles input character encoding. @code{subtitles} filter only. Only
18340 useful if not UTF-8.
18342 @item stream_index, si
18343 Set subtitles stream index. @code{subtitles} filter only.
18346 Override default style or script info parameters of the subtitles. It accepts a
18347 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18350 If the first key is not specified, it is assumed that the first value
18351 specifies the @option{filename}.
18353 For example, to render the file @file{sub.srt} on top of the input
18354 video, use the command:
18359 which is equivalent to:
18361 subtitles=filename=sub.srt
18364 To render the default subtitles stream from file @file{video.mkv}, use:
18366 subtitles=video.mkv
18369 To render the second subtitles stream from that file, use:
18371 subtitles=video.mkv:si=1
18374 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18375 @code{DejaVu Serif}, use:
18377 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18380 @section super2xsai
18382 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18383 Interpolate) pixel art scaling algorithm.
18385 Useful for enlarging pixel art images without reducing sharpness.
18389 Swap two rectangular objects in video.
18391 This filter accepts the following options:
18401 Set 1st rect x coordinate.
18404 Set 1st rect y coordinate.
18407 Set 2nd rect x coordinate.
18410 Set 2nd rect y coordinate.
18412 All expressions are evaluated once for each frame.
18415 The all options are expressions containing the following constants:
18420 The input width and height.
18423 same as @var{w} / @var{h}
18426 input sample aspect ratio
18429 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18432 The number of the input frame, starting from 0.
18435 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18438 the position in the file of the input frame, NAN if unknown
18445 Blend successive video frames.
18451 Apply telecine process to the video.
18453 This filter accepts the following options:
18462 The default value is @code{top}.
18466 A string of numbers representing the pulldown pattern you wish to apply.
18467 The default value is @code{23}.
18471 Some typical patterns:
18476 24p: 2332 (preferred)
18483 24p: 222222222223 ("Euro pulldown")
18488 @section thistogram
18490 Compute and draw a color distribution histogram for the input video across time.
18492 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18493 at certain time, this filter shows also past histograms of number of frames defined
18494 by @code{width} option.
18496 The computed histogram is a representation of the color component
18497 distribution in an image.
18499 The filter accepts the following options:
18503 Set width of single color component output. Default value is @code{0}.
18504 Value of @code{0} means width will be picked from input video.
18505 This also set number of passed histograms to keep.
18506 Allowed range is [0, 8192].
18508 @item display_mode, d
18510 It accepts the following values:
18513 Per color component graphs are placed below each other.
18516 Per color component graphs are placed side by side.
18519 Presents information identical to that in the @code{parade}, except
18520 that the graphs representing color components are superimposed directly
18523 Default is @code{stack}.
18525 @item levels_mode, m
18526 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18527 Default is @code{linear}.
18529 @item components, c
18530 Set what color components to display.
18531 Default is @code{7}.
18534 Set background opacity. Default is @code{0.9}.
18537 Show envelope. Default is disabled.
18540 Set envelope color. Default is @code{gold}.
18545 Available values for slide is:
18548 Draw new frame when right border is reached.
18551 Replace old columns with new ones.
18554 Scroll from right to left.
18557 Scroll from left to right.
18560 Draw single picture.
18563 Default is @code{replace}.
18568 Apply threshold effect to video stream.
18570 This filter needs four video streams to perform thresholding.
18571 First stream is stream we are filtering.
18572 Second stream is holding threshold values, third stream is holding min values,
18573 and last, fourth stream is holding max values.
18575 The filter accepts the following option:
18579 Set which planes will be processed, unprocessed planes will be copied.
18580 By default value 0xf, all planes will be processed.
18583 For example if first stream pixel's component value is less then threshold value
18584 of pixel component from 2nd threshold stream, third stream value will picked,
18585 otherwise fourth stream pixel component value will be picked.
18587 Using color source filter one can perform various types of thresholding:
18589 @subsection Examples
18593 Binary threshold, using gray color as threshold:
18595 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
18599 Inverted binary threshold, using gray color as threshold:
18601 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
18605 Truncate binary threshold, using gray color as threshold:
18607 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
18611 Threshold to zero, using gray color as threshold:
18613 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
18617 Inverted threshold to zero, using gray color as threshold:
18619 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
18624 Select the most representative frame in a given sequence of consecutive frames.
18626 The filter accepts the following options:
18630 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
18631 will pick one of them, and then handle the next batch of @var{n} frames until
18632 the end. Default is @code{100}.
18635 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
18636 value will result in a higher memory usage, so a high value is not recommended.
18638 @subsection Examples
18642 Extract one picture each 50 frames:
18648 Complete example of a thumbnail creation with @command{ffmpeg}:
18650 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
18657 Tile several successive frames together.
18659 The @ref{untile} filter can do the reverse.
18661 The filter accepts the following options:
18666 Set the grid size (i.e. the number of lines and columns). For the syntax of
18667 this option, check the
18668 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18671 Set the maximum number of frames to render in the given area. It must be less
18672 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
18673 the area will be used.
18676 Set the outer border margin in pixels.
18679 Set the inner border thickness (i.e. the number of pixels between frames). For
18680 more advanced padding options (such as having different values for the edges),
18681 refer to the pad video filter.
18684 Specify the color of the unused area. For the syntax of this option, check the
18685 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
18686 The default value of @var{color} is "black".
18689 Set the number of frames to overlap when tiling several successive frames together.
18690 The value must be between @code{0} and @var{nb_frames - 1}.
18693 Set the number of frames to initially be empty before displaying first output frame.
18694 This controls how soon will one get first output frame.
18695 The value must be between @code{0} and @var{nb_frames - 1}.
18698 @subsection Examples
18702 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
18704 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
18706 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
18707 duplicating each output frame to accommodate the originally detected frame
18711 Display @code{5} pictures in an area of @code{3x2} frames,
18712 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
18713 mixed flat and named options:
18715 tile=3x2:nb_frames=5:padding=7:margin=2
18719 @section tinterlace
18721 Perform various types of temporal field interlacing.
18723 Frames are counted starting from 1, so the first input frame is
18726 The filter accepts the following options:
18731 Specify the mode of the interlacing. This option can also be specified
18732 as a value alone. See below for a list of values for this option.
18734 Available values are:
18738 Move odd frames into the upper field, even into the lower field,
18739 generating a double height frame at half frame rate.
18743 Frame 1 Frame 2 Frame 3 Frame 4
18745 11111 22222 33333 44444
18746 11111 22222 33333 44444
18747 11111 22222 33333 44444
18748 11111 22222 33333 44444
18762 Only output odd frames, even frames are dropped, generating a frame with
18763 unchanged height at half frame rate.
18768 Frame 1 Frame 2 Frame 3 Frame 4
18770 11111 22222 33333 44444
18771 11111 22222 33333 44444
18772 11111 22222 33333 44444
18773 11111 22222 33333 44444
18783 Only output even frames, odd frames are dropped, generating a frame with
18784 unchanged height at half frame rate.
18789 Frame 1 Frame 2 Frame 3 Frame 4
18791 11111 22222 33333 44444
18792 11111 22222 33333 44444
18793 11111 22222 33333 44444
18794 11111 22222 33333 44444
18804 Expand each frame to full height, but pad alternate lines with black,
18805 generating a frame with double height at the same input frame rate.
18810 Frame 1 Frame 2 Frame 3 Frame 4
18812 11111 22222 33333 44444
18813 11111 22222 33333 44444
18814 11111 22222 33333 44444
18815 11111 22222 33333 44444
18818 11111 ..... 33333 .....
18819 ..... 22222 ..... 44444
18820 11111 ..... 33333 .....
18821 ..... 22222 ..... 44444
18822 11111 ..... 33333 .....
18823 ..... 22222 ..... 44444
18824 11111 ..... 33333 .....
18825 ..... 22222 ..... 44444
18829 @item interleave_top, 4
18830 Interleave the upper field from odd frames with the lower field from
18831 even frames, generating a frame with unchanged height at half frame rate.
18836 Frame 1 Frame 2 Frame 3 Frame 4
18838 11111<- 22222 33333<- 44444
18839 11111 22222<- 33333 44444<-
18840 11111<- 22222 33333<- 44444
18841 11111 22222<- 33333 44444<-
18851 @item interleave_bottom, 5
18852 Interleave the lower field from odd frames with the upper field from
18853 even frames, generating a frame with unchanged height at half frame rate.
18858 Frame 1 Frame 2 Frame 3 Frame 4
18860 11111 22222<- 33333 44444<-
18861 11111<- 22222 33333<- 44444
18862 11111 22222<- 33333 44444<-
18863 11111<- 22222 33333<- 44444
18873 @item interlacex2, 6
18874 Double frame rate with unchanged height. Frames are inserted each
18875 containing the second temporal field from the previous input frame and
18876 the first temporal field from the next input frame. This mode relies on
18877 the top_field_first flag. Useful for interlaced video displays with no
18878 field synchronisation.
18883 Frame 1 Frame 2 Frame 3 Frame 4
18885 11111 22222 33333 44444
18886 11111 22222 33333 44444
18887 11111 22222 33333 44444
18888 11111 22222 33333 44444
18891 11111 22222 22222 33333 33333 44444 44444
18892 11111 11111 22222 22222 33333 33333 44444
18893 11111 22222 22222 33333 33333 44444 44444
18894 11111 11111 22222 22222 33333 33333 44444
18899 Move odd frames into the upper field, even into the lower field,
18900 generating a double height frame at same frame rate.
18905 Frame 1 Frame 2 Frame 3 Frame 4
18907 11111 22222 33333 44444
18908 11111 22222 33333 44444
18909 11111 22222 33333 44444
18910 11111 22222 33333 44444
18913 11111 33333 33333 55555
18914 22222 22222 44444 44444
18915 11111 33333 33333 55555
18916 22222 22222 44444 44444
18917 11111 33333 33333 55555
18918 22222 22222 44444 44444
18919 11111 33333 33333 55555
18920 22222 22222 44444 44444
18925 Numeric values are deprecated but are accepted for backward
18926 compatibility reasons.
18928 Default mode is @code{merge}.
18931 Specify flags influencing the filter process.
18933 Available value for @var{flags} is:
18936 @item low_pass_filter, vlpf
18937 Enable linear vertical low-pass filtering in the filter.
18938 Vertical low-pass filtering is required when creating an interlaced
18939 destination from a progressive source which contains high-frequency
18940 vertical detail. Filtering will reduce interlace 'twitter' and Moire
18943 @item complex_filter, cvlpf
18944 Enable complex vertical low-pass filtering.
18945 This will slightly less reduce interlace 'twitter' and Moire
18946 patterning but better retain detail and subjective sharpness impression.
18949 Bypass already interlaced frames, only adjust the frame rate.
18952 Vertical low-pass filtering and bypassing already interlaced frames can only be
18953 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
18958 Pick median pixels from several successive input video frames.
18960 The filter accepts the following options:
18964 Set radius of median filter.
18965 Default is 1. Allowed range is from 1 to 127.
18968 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
18971 Set median percentile. Default value is @code{0.5}.
18972 Default value of @code{0.5} will pick always median values, while @code{0} will pick
18973 minimum values, and @code{1} maximum values.
18978 Mix successive video frames.
18980 A description of the accepted options follows.
18984 The number of successive frames to mix. If unspecified, it defaults to 3.
18987 Specify weight of each input video frame.
18988 Each weight is separated by space. If number of weights is smaller than
18989 number of @var{frames} last specified weight will be used for all remaining
18993 Specify scale, if it is set it will be multiplied with sum
18994 of each weight multiplied with pixel values to give final destination
18995 pixel value. By default @var{scale} is auto scaled to sum of weights.
18998 @subsection Examples
19002 Average 7 successive frames:
19004 tmix=frames=7:weights="1 1 1 1 1 1 1"
19008 Apply simple temporal convolution:
19010 tmix=frames=3:weights="-1 3 -1"
19014 Similar as above but only showing temporal differences:
19016 tmix=frames=3:weights="-1 2 -1":scale=1
19022 Tone map colors from different dynamic ranges.
19024 This filter expects data in single precision floating point, as it needs to
19025 operate on (and can output) out-of-range values. Another filter, such as
19026 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19028 The tonemapping algorithms implemented only work on linear light, so input
19029 data should be linearized beforehand (and possibly correctly tagged).
19032 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19035 @subsection Options
19036 The filter accepts the following options.
19040 Set the tone map algorithm to use.
19042 Possible values are:
19045 Do not apply any tone map, only desaturate overbright pixels.
19048 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19049 in-range values, while distorting out-of-range values.
19052 Stretch the entire reference gamut to a linear multiple of the display.
19055 Fit a logarithmic transfer between the tone curves.
19058 Preserve overall image brightness with a simple curve, using nonlinear
19059 contrast, which results in flattening details and degrading color accuracy.
19062 Preserve both dark and bright details better than @var{reinhard}, at the cost
19063 of slightly darkening everything. Use it when detail preservation is more
19064 important than color and brightness accuracy.
19067 Smoothly map out-of-range values, while retaining contrast and colors for
19068 in-range material as much as possible. Use it when color accuracy is more
19069 important than detail preservation.
19075 Tune the tone mapping algorithm.
19077 This affects the following algorithms:
19083 Specifies the scale factor to use while stretching.
19087 Specifies the exponent of the function.
19091 Specify an extra linear coefficient to multiply into the signal before clipping.
19095 Specify the local contrast coefficient at the display peak.
19096 Default to 0.5, which means that in-gamut values will be about half as bright
19103 Specify the transition point from linear to mobius transform. Every value
19104 below this point is guaranteed to be mapped 1:1. The higher the value, the
19105 more accurate the result will be, at the cost of losing bright details.
19106 Default to 0.3, which due to the steep initial slope still preserves in-range
19107 colors fairly accurately.
19111 Apply desaturation for highlights that exceed this level of brightness. The
19112 higher the parameter, the more color information will be preserved. This
19113 setting helps prevent unnaturally blown-out colors for super-highlights, by
19114 (smoothly) turning into white instead. This makes images feel more natural,
19115 at the cost of reducing information about out-of-range colors.
19117 The default of 2.0 is somewhat conservative and will mostly just apply to
19118 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19120 This option works only if the input frame has a supported color tag.
19123 Override signal/nominal/reference peak with this value. Useful when the
19124 embedded peak information in display metadata is not reliable or when tone
19125 mapping from a lower range to a higher range.
19130 Temporarily pad video frames.
19132 The filter accepts the following options:
19136 Specify number of delay frames before input video stream. Default is 0.
19139 Specify number of padding frames after input video stream.
19140 Set to -1 to pad indefinitely. Default is 0.
19143 Set kind of frames added to beginning of stream.
19144 Can be either @var{add} or @var{clone}.
19145 With @var{add} frames of solid-color are added.
19146 With @var{clone} frames are clones of first frame.
19147 Default is @var{add}.
19150 Set kind of frames added to end of stream.
19151 Can be either @var{add} or @var{clone}.
19152 With @var{add} frames of solid-color are added.
19153 With @var{clone} frames are clones of last frame.
19154 Default is @var{add}.
19156 @item start_duration, stop_duration
19157 Specify the duration of the start/stop delay. See
19158 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19159 for the accepted syntax.
19160 These options override @var{start} and @var{stop}. Default is 0.
19163 Specify the color of the padded area. For the syntax of this option,
19164 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19165 manual,ffmpeg-utils}.
19167 The default value of @var{color} is "black".
19173 Transpose rows with columns in the input video and optionally flip it.
19175 It accepts the following parameters:
19180 Specify the transposition direction.
19182 Can assume the following values:
19184 @item 0, 4, cclock_flip
19185 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19193 Rotate by 90 degrees clockwise, that is:
19201 Rotate by 90 degrees counterclockwise, that is:
19208 @item 3, 7, clock_flip
19209 Rotate by 90 degrees clockwise and vertically flip, that is:
19217 For values between 4-7, the transposition is only done if the input
19218 video geometry is portrait and not landscape. These values are
19219 deprecated, the @code{passthrough} option should be used instead.
19221 Numerical values are deprecated, and should be dropped in favor of
19222 symbolic constants.
19225 Do not apply the transposition if the input geometry matches the one
19226 specified by the specified value. It accepts the following values:
19229 Always apply transposition.
19231 Preserve portrait geometry (when @var{height} >= @var{width}).
19233 Preserve landscape geometry (when @var{width} >= @var{height}).
19236 Default value is @code{none}.
19239 For example to rotate by 90 degrees clockwise and preserve portrait
19242 transpose=dir=1:passthrough=portrait
19245 The command above can also be specified as:
19247 transpose=1:portrait
19250 @section transpose_npp
19252 Transpose rows with columns in the input video and optionally flip it.
19253 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19255 It accepts the following parameters:
19260 Specify the transposition direction.
19262 Can assume the following values:
19265 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19268 Rotate by 90 degrees clockwise.
19271 Rotate by 90 degrees counterclockwise.
19274 Rotate by 90 degrees clockwise and vertically flip.
19278 Do not apply the transposition if the input geometry matches the one
19279 specified by the specified value. It accepts the following values:
19282 Always apply transposition. (default)
19284 Preserve portrait geometry (when @var{height} >= @var{width}).
19286 Preserve landscape geometry (when @var{width} >= @var{height}).
19292 Trim the input so that the output contains one continuous subpart of the input.
19294 It accepts the following parameters:
19297 Specify the time of the start of the kept section, i.e. the frame with the
19298 timestamp @var{start} will be the first frame in the output.
19301 Specify the time of the first frame that will be dropped, i.e. the frame
19302 immediately preceding the one with the timestamp @var{end} will be the last
19303 frame in the output.
19306 This is the same as @var{start}, except this option sets the start timestamp
19307 in timebase units instead of seconds.
19310 This is the same as @var{end}, except this option sets the end timestamp
19311 in timebase units instead of seconds.
19314 The maximum duration of the output in seconds.
19317 The number of the first frame that should be passed to the output.
19320 The number of the first frame that should be dropped.
19323 @option{start}, @option{end}, and @option{duration} are expressed as time
19324 duration specifications; see
19325 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19326 for the accepted syntax.
19328 Note that the first two sets of the start/end options and the @option{duration}
19329 option look at the frame timestamp, while the _frame variants simply count the
19330 frames that pass through the filter. Also note that this filter does not modify
19331 the timestamps. If you wish for the output timestamps to start at zero, insert a
19332 setpts filter after the trim filter.
19334 If multiple start or end options are set, this filter tries to be greedy and
19335 keep all the frames that match at least one of the specified constraints. To keep
19336 only the part that matches all the constraints at once, chain multiple trim
19339 The defaults are such that all the input is kept. So it is possible to set e.g.
19340 just the end values to keep everything before the specified time.
19345 Drop everything except the second minute of input:
19347 ffmpeg -i INPUT -vf trim=60:120
19351 Keep only the first second:
19353 ffmpeg -i INPUT -vf trim=duration=1
19358 @section unpremultiply
19359 Apply alpha unpremultiply effect to input video stream using first plane
19360 of second stream as alpha.
19362 Both streams must have same dimensions and same pixel format.
19364 The filter accepts the following option:
19368 Set which planes will be processed, unprocessed planes will be copied.
19369 By default value 0xf, all planes will be processed.
19371 If the format has 1 or 2 components, then luma is bit 0.
19372 If the format has 3 or 4 components:
19373 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19374 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19375 If present, the alpha channel is always the last bit.
19378 Do not require 2nd input for processing, instead use alpha plane from input stream.
19384 Sharpen or blur the input video.
19386 It accepts the following parameters:
19389 @item luma_msize_x, lx
19390 Set the luma matrix horizontal size. It must be an odd integer between
19391 3 and 23. The default value is 5.
19393 @item luma_msize_y, ly
19394 Set the luma matrix vertical size. It must be an odd integer between 3
19395 and 23. The default value is 5.
19397 @item luma_amount, la
19398 Set the luma effect strength. It must be a floating point number, reasonable
19399 values lay between -1.5 and 1.5.
19401 Negative values will blur the input video, while positive values will
19402 sharpen it, a value of zero will disable the effect.
19404 Default value is 1.0.
19406 @item chroma_msize_x, cx
19407 Set the chroma matrix horizontal size. It must be an odd integer
19408 between 3 and 23. The default value is 5.
19410 @item chroma_msize_y, cy
19411 Set the chroma matrix vertical size. It must be an odd integer
19412 between 3 and 23. The default value is 5.
19414 @item chroma_amount, ca
19415 Set the chroma effect strength. It must be a floating point number, reasonable
19416 values lay between -1.5 and 1.5.
19418 Negative values will blur the input video, while positive values will
19419 sharpen it, a value of zero will disable the effect.
19421 Default value is 0.0.
19425 All parameters are optional and default to the equivalent of the
19426 string '5:5:1.0:5:5:0.0'.
19428 @subsection Examples
19432 Apply strong luma sharpen effect:
19434 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19438 Apply a strong blur of both luma and chroma parameters:
19440 unsharp=7:7:-2:7:7:-2
19447 Decompose a video made of tiled images into the individual images.
19449 The frame rate of the output video is the frame rate of the input video
19450 multiplied by the number of tiles.
19452 This filter does the reverse of @ref{tile}.
19454 The filter accepts the following options:
19459 Set the grid size (i.e. the number of lines and columns). For the syntax of
19460 this option, check the
19461 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19464 @subsection Examples
19468 Produce a 1-second video from a still image file made of 25 frames stacked
19469 vertically, like an analogic film reel:
19471 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19477 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19478 the image at several (or - in the case of @option{quality} level @code{8} - all)
19479 shifts and average the results.
19481 The way this differs from the behavior of spp is that uspp actually encodes &
19482 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19483 DCT similar to MJPEG.
19485 The filter accepts the following options:
19489 Set quality. This option defines the number of levels for averaging. It accepts
19490 an integer in the range 0-8. If set to @code{0}, the filter will have no
19491 effect. A value of @code{8} means the higher quality. For each increment of
19492 that value the speed drops by a factor of approximately 2. Default value is
19496 Force a constant quantization parameter. If not set, the filter will use the QP
19497 from the video stream (if available).
19502 Convert 360 videos between various formats.
19504 The filter accepts the following options:
19510 Set format of the input/output video.
19518 Equirectangular projection.
19523 Cubemap with 3x2/6x1/1x6 layout.
19525 Format specific options:
19530 Set padding proportion for the input/output cubemap. Values in decimals.
19537 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)
19540 Default value is @b{@samp{0}}.
19541 Maximum value is @b{@samp{0.1}}.
19545 Set fixed padding for the input/output cubemap. Values in pixels.
19547 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19551 Set order of faces for the input/output cubemap. Choose one direction for each position.
19553 Designation of directions:
19569 Default value is @b{@samp{rludfb}}.
19573 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
19575 Designation of angles:
19578 0 degrees clockwise
19580 90 degrees clockwise
19582 180 degrees clockwise
19584 270 degrees clockwise
19587 Default value is @b{@samp{000000}}.
19591 Equi-Angular Cubemap.
19598 Format specific options:
19603 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19605 If diagonal field of view is set it overrides horizontal and vertical field of view.
19610 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19612 If diagonal field of view is set it overrides horizontal and vertical field of view.
19618 Format specific options:
19623 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19625 If diagonal field of view is set it overrides horizontal and vertical field of view.
19630 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19632 If diagonal field of view is set it overrides horizontal and vertical field of view.
19638 Facebook's 360 formats.
19641 Stereographic format.
19643 Format specific options:
19648 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19650 If diagonal field of view is set it overrides horizontal and vertical field of view.
19655 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19657 If diagonal field of view is set it overrides horizontal and vertical field of view.
19664 Ball format, gives significant distortion toward the back.
19667 Hammer-Aitoff map projection format.
19670 Sinusoidal map projection format.
19673 Fisheye projection.
19675 Format specific options:
19680 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19682 If diagonal field of view is set it overrides horizontal and vertical field of view.
19687 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19689 If diagonal field of view is set it overrides horizontal and vertical field of view.
19693 Pannini projection.
19695 Format specific options:
19698 Set output pannini parameter.
19701 Set input pannini parameter.
19705 Cylindrical projection.
19707 Format specific options:
19712 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19714 If diagonal field of view is set it overrides horizontal and vertical field of view.
19719 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19721 If diagonal field of view is set it overrides horizontal and vertical field of view.
19725 Perspective projection. @i{(output only)}
19727 Format specific options:
19730 Set perspective parameter.
19734 Tetrahedron projection.
19737 Truncated square pyramid projection.
19741 Half equirectangular projection.
19746 Format specific options:
19751 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19753 If diagonal field of view is set it overrides horizontal and vertical field of view.
19758 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19760 If diagonal field of view is set it overrides horizontal and vertical field of view.
19764 Orthographic format.
19766 Format specific options:
19771 Set output horizontal/vertical/diagonal field of view. Values in degrees.
19773 If diagonal field of view is set it overrides horizontal and vertical field of view.
19778 Set input horizontal/vertical/diagonal field of view. Values in degrees.
19780 If diagonal field of view is set it overrides horizontal and vertical field of view.
19784 Octahedron projection.
19788 Set interpolation method.@*
19789 @i{Note: more complex interpolation methods require much more memory to run.}
19799 Bilinear interpolation.
19801 Lagrange9 interpolation.
19804 Bicubic interpolation.
19807 Lanczos interpolation.
19810 Spline16 interpolation.
19813 Gaussian interpolation.
19815 Mitchell interpolation.
19818 Default value is @b{@samp{line}}.
19822 Set the output video resolution.
19824 Default resolution depends on formats.
19828 Set the input/output stereo format.
19839 Default value is @b{@samp{2d}} for input and output format.
19844 Set rotation for the output video. Values in degrees.
19847 Set rotation order for the output video. Choose one item for each position.
19858 Default value is @b{@samp{ypr}}.
19863 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
19867 Set if input video is flipped horizontally/vertically. Boolean values.
19870 Set if input video is transposed. Boolean value, by default disabled.
19873 Set if output video needs to be transposed. Boolean value, by default disabled.
19876 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
19879 @subsection Examples
19883 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
19885 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
19888 Extract back view of Equi-Angular Cubemap:
19890 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
19893 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
19895 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
19899 @subsection Commands
19901 This filter supports subset of above options as @ref{commands}.
19903 @section vaguedenoiser
19905 Apply a wavelet based denoiser.
19907 It transforms each frame from the video input into the wavelet domain,
19908 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
19909 the obtained coefficients. It does an inverse wavelet transform after.
19910 Due to wavelet properties, it should give a nice smoothed result, and
19911 reduced noise, without blurring picture features.
19913 This filter accepts the following options:
19917 The filtering strength. The higher, the more filtered the video will be.
19918 Hard thresholding can use a higher threshold than soft thresholding
19919 before the video looks overfiltered. Default value is 2.
19922 The filtering method the filter will use.
19924 It accepts the following values:
19927 All values under the threshold will be zeroed.
19930 All values under the threshold will be zeroed. All values above will be
19931 reduced by the threshold.
19934 Scales or nullifies coefficients - intermediary between (more) soft and
19935 (less) hard thresholding.
19938 Default is garrote.
19941 Number of times, the wavelet will decompose the picture. Picture can't
19942 be decomposed beyond a particular point (typically, 8 for a 640x480
19943 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
19946 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
19949 A list of the planes to process. By default all planes are processed.
19952 The threshold type the filter will use.
19954 It accepts the following values:
19957 Threshold used is same for all decompositions.
19960 Threshold used depends also on each decomposition coefficients.
19963 Default is universal.
19966 @section vectorscope
19968 Display 2 color component values in the two dimensional graph (which is called
19971 This filter accepts the following options:
19975 Set vectorscope mode.
19977 It accepts the following values:
19981 Gray values are displayed on graph, higher brightness means more pixels have
19982 same component color value on location in graph. This is the default mode.
19985 Gray values are displayed on graph. Surrounding pixels values which are not
19986 present in video frame are drawn in gradient of 2 color components which are
19987 set by option @code{x} and @code{y}. The 3rd color component is static.
19990 Actual color components values present in video frame are displayed on graph.
19993 Similar as color2 but higher frequency of same values @code{x} and @code{y}
19994 on graph increases value of another color component, which is luminance by
19995 default values of @code{x} and @code{y}.
19998 Actual colors present in video frame are displayed on graph. If two different
19999 colors map to same position on graph then color with higher value of component
20000 not present in graph is picked.
20003 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20004 component picked from radial gradient.
20008 Set which color component will be represented on X-axis. Default is @code{1}.
20011 Set which color component will be represented on Y-axis. Default is @code{2}.
20014 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20015 of color component which represents frequency of (X, Y) location in graph.
20020 No envelope, this is default.
20023 Instant envelope, even darkest single pixel will be clearly highlighted.
20026 Hold maximum and minimum values presented in graph over time. This way you
20027 can still spot out of range values without constantly looking at vectorscope.
20030 Peak and instant envelope combined together.
20034 Set what kind of graticule to draw.
20043 Set graticule opacity.
20046 Set graticule flags.
20050 Draw graticule for white point.
20053 Draw graticule for black point.
20056 Draw color points short names.
20060 Set background opacity.
20062 @item lthreshold, l
20063 Set low threshold for color component not represented on X or Y axis.
20064 Values lower than this value will be ignored. Default is 0.
20065 Note this value is multiplied with actual max possible value one pixel component
20066 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20069 @item hthreshold, h
20070 Set high threshold for color component not represented on X or Y axis.
20071 Values higher than this value will be ignored. Default is 1.
20072 Note this value is multiplied with actual max possible value one pixel component
20073 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20074 is 0.9 * 255 = 230.
20076 @item colorspace, c
20077 Set what kind of colorspace to use when drawing graticule.
20087 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20088 This means no tint, and output will remain gray.
20091 @anchor{vidstabdetect}
20092 @section vidstabdetect
20094 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20095 @ref{vidstabtransform} for pass 2.
20097 This filter generates a file with relative translation and rotation
20098 transform information about subsequent frames, which is then used by
20099 the @ref{vidstabtransform} filter.
20101 To enable compilation of this filter you need to configure FFmpeg with
20102 @code{--enable-libvidstab}.
20104 This filter accepts the following options:
20108 Set the path to the file used to write the transforms information.
20109 Default value is @file{transforms.trf}.
20112 Set how shaky the video is and how quick the camera is. It accepts an
20113 integer in the range 1-10, a value of 1 means little shakiness, a
20114 value of 10 means strong shakiness. Default value is 5.
20117 Set the accuracy of the detection process. It must be a value in the
20118 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20119 accuracy. Default value is 15.
20122 Set stepsize of the search process. The region around minimum is
20123 scanned with 1 pixel resolution. Default value is 6.
20126 Set minimum contrast. Below this value a local measurement field is
20127 discarded. Must be a floating point value in the range 0-1. Default
20131 Set reference frame number for tripod mode.
20133 If enabled, the motion of the frames is compared to a reference frame
20134 in the filtered stream, identified by the specified number. The idea
20135 is to compensate all movements in a more-or-less static scene and keep
20136 the camera view absolutely still.
20138 If set to 0, it is disabled. The frames are counted starting from 1.
20141 Show fields and transforms in the resulting frames. It accepts an
20142 integer in the range 0-2. Default value is 0, which disables any
20146 @subsection Examples
20150 Use default values:
20156 Analyze strongly shaky movie and put the results in file
20157 @file{mytransforms.trf}:
20159 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20163 Visualize the result of internal transformations in the resulting
20166 vidstabdetect=show=1
20170 Analyze a video with medium shakiness using @command{ffmpeg}:
20172 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20176 @anchor{vidstabtransform}
20177 @section vidstabtransform
20179 Video stabilization/deshaking: pass 2 of 2,
20180 see @ref{vidstabdetect} for pass 1.
20182 Read a file with transform information for each frame and
20183 apply/compensate them. Together with the @ref{vidstabdetect}
20184 filter this can be used to deshake videos. See also
20185 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20186 the @ref{unsharp} filter, see below.
20188 To enable compilation of this filter you need to configure FFmpeg with
20189 @code{--enable-libvidstab}.
20191 @subsection Options
20195 Set path to the file used to read the transforms. Default value is
20196 @file{transforms.trf}.
20199 Set the number of frames (value*2 + 1) used for lowpass filtering the
20200 camera movements. Default value is 10.
20202 For example a number of 10 means that 21 frames are used (10 in the
20203 past and 10 in the future) to smoothen the motion in the video. A
20204 larger value leads to a smoother video, but limits the acceleration of
20205 the camera (pan/tilt movements). 0 is a special case where a static
20206 camera is simulated.
20209 Set the camera path optimization algorithm.
20211 Accepted values are:
20214 gaussian kernel low-pass filter on camera motion (default)
20216 averaging on transformations
20220 Set maximal number of pixels to translate frames. Default value is -1,
20224 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20225 value is -1, meaning no limit.
20228 Specify how to deal with borders that may be visible due to movement
20231 Available values are:
20234 keep image information from previous frame (default)
20236 fill the border black
20240 Invert transforms if set to 1. Default value is 0.
20243 Consider transforms as relative to previous frame if set to 1,
20244 absolute if set to 0. Default value is 0.
20247 Set percentage to zoom. A positive value will result in a zoom-in
20248 effect, a negative value in a zoom-out effect. Default value is 0 (no
20252 Set optimal zooming to avoid borders.
20254 Accepted values are:
20259 optimal static zoom value is determined (only very strong movements
20260 will lead to visible borders) (default)
20262 optimal adaptive zoom value is determined (no borders will be
20263 visible), see @option{zoomspeed}
20266 Note that the value given at zoom is added to the one calculated here.
20269 Set percent to zoom maximally each frame (enabled when
20270 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20274 Specify type of interpolation.
20276 Available values are:
20281 linear only horizontal
20283 linear in both directions (default)
20285 cubic in both directions (slow)
20289 Enable virtual tripod mode if set to 1, which is equivalent to
20290 @code{relative=0:smoothing=0}. Default value is 0.
20292 Use also @code{tripod} option of @ref{vidstabdetect}.
20295 Increase log verbosity if set to 1. Also the detected global motions
20296 are written to the temporary file @file{global_motions.trf}. Default
20300 @subsection Examples
20304 Use @command{ffmpeg} for a typical stabilization with default values:
20306 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20309 Note the use of the @ref{unsharp} filter which is always recommended.
20312 Zoom in a bit more and load transform data from a given file:
20314 vidstabtransform=zoom=5:input="mytransforms.trf"
20318 Smoothen the video even more:
20320 vidstabtransform=smoothing=30
20326 Flip the input video vertically.
20328 For example, to vertically flip a video with @command{ffmpeg}:
20330 ffmpeg -i in.avi -vf "vflip" out.avi
20335 Detect variable frame rate video.
20337 This filter tries to detect if the input is variable or constant frame rate.
20339 At end it will output number of frames detected as having variable delta pts,
20340 and ones with constant delta pts.
20341 If there was frames with variable delta, than it will also show min, max and
20342 average delta encountered.
20346 Boost or alter saturation.
20348 The filter accepts the following options:
20351 Set strength of boost if positive value or strength of alter if negative value.
20352 Default is 0. Allowed range is from -2 to 2.
20355 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20358 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20361 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20364 Set the red luma coefficient.
20367 Set the green luma coefficient.
20370 Set the blue luma coefficient.
20373 If @code{intensity} is negative and this is set to 1, colors will change,
20374 otherwise colors will be less saturated, more towards gray.
20377 @subsection Commands
20379 This filter supports the all above options as @ref{commands}.
20384 Make or reverse a natural vignetting effect.
20386 The filter accepts the following options:
20390 Set lens angle expression as a number of radians.
20392 The value is clipped in the @code{[0,PI/2]} range.
20394 Default value: @code{"PI/5"}
20398 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20402 Set forward/backward mode.
20404 Available modes are:
20407 The larger the distance from the central point, the darker the image becomes.
20410 The larger the distance from the central point, the brighter the image becomes.
20411 This can be used to reverse a vignette effect, though there is no automatic
20412 detection to extract the lens @option{angle} and other settings (yet). It can
20413 also be used to create a burning effect.
20416 Default value is @samp{forward}.
20419 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20421 It accepts the following values:
20424 Evaluate expressions only once during the filter initialization.
20427 Evaluate expressions for each incoming frame. This is way slower than the
20428 @samp{init} mode since it requires all the scalers to be re-computed, but it
20429 allows advanced dynamic expressions.
20432 Default value is @samp{init}.
20435 Set dithering to reduce the circular banding effects. Default is @code{1}
20439 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20440 Setting this value to the SAR of the input will make a rectangular vignetting
20441 following the dimensions of the video.
20443 Default is @code{1/1}.
20446 @subsection Expressions
20448 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20449 following parameters.
20454 input width and height
20457 the number of input frame, starting from 0
20460 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20461 @var{TB} units, NAN if undefined
20464 frame rate of the input video, NAN if the input frame rate is unknown
20467 the PTS (Presentation TimeStamp) of the filtered video frame,
20468 expressed in seconds, NAN if undefined
20471 time base of the input video
20475 @subsection Examples
20479 Apply simple strong vignetting effect:
20485 Make a flickering vignetting:
20487 vignette='PI/4+random(1)*PI/50':eval=frame
20492 @section vmafmotion
20494 Obtain the average VMAF motion score of a video.
20495 It is one of the component metrics of VMAF.
20497 The obtained average motion score is printed through the logging system.
20499 The filter accepts the following options:
20503 If specified, the filter will use the named file to save the motion score of
20504 each frame with respect to the previous frame.
20505 When filename equals "-" the data is sent to standard output.
20510 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20514 Stack input videos vertically.
20516 All streams must be of same pixel format and of same width.
20518 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20519 to create same output.
20521 The filter accepts the following options:
20525 Set number of input streams. Default is 2.
20528 If set to 1, force the output to terminate when the shortest input
20529 terminates. Default value is 0.
20534 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20535 Deinterlacing Filter").
20537 Based on the process described by Martin Weston for BBC R&D, and
20538 implemented based on the de-interlace algorithm written by Jim
20539 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20540 uses filter coefficients calculated by BBC R&D.
20542 This filter uses field-dominance information in frame to decide which
20543 of each pair of fields to place first in the output.
20544 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20546 There are two sets of filter coefficients, so called "simple"
20547 and "complex". Which set of filter coefficients is used can
20548 be set by passing an optional parameter:
20552 Set the interlacing filter coefficients. Accepts one of the following values:
20556 Simple filter coefficient set.
20558 More-complex filter coefficient set.
20560 Default value is @samp{complex}.
20563 Specify which frames to deinterlace. Accepts one of the following values:
20567 Deinterlace all frames,
20569 Only deinterlace frames marked as interlaced.
20572 Default value is @samp{all}.
20576 Video waveform monitor.
20578 The waveform monitor plots color component intensity. By default luminance
20579 only. Each column of the waveform corresponds to a column of pixels in the
20582 It accepts the following options:
20586 Can be either @code{row}, or @code{column}. Default is @code{column}.
20587 In row mode, the graph on the left side represents color component value 0 and
20588 the right side represents value = 255. In column mode, the top side represents
20589 color component value = 0 and bottom side represents value = 255.
20592 Set intensity. Smaller values are useful to find out how many values of the same
20593 luminance are distributed across input rows/columns.
20594 Default value is @code{0.04}. Allowed range is [0, 1].
20597 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
20598 In mirrored mode, higher values will be represented on the left
20599 side for @code{row} mode and at the top for @code{column} mode. Default is
20600 @code{1} (mirrored).
20604 It accepts the following values:
20607 Presents information identical to that in the @code{parade}, except
20608 that the graphs representing color components are superimposed directly
20611 This display mode makes it easier to spot relative differences or similarities
20612 in overlapping areas of the color components that are supposed to be identical,
20613 such as neutral whites, grays, or blacks.
20616 Display separate graph for the color components side by side in
20617 @code{row} mode or one below the other in @code{column} mode.
20620 Display separate graph for the color components side by side in
20621 @code{column} mode or one below the other in @code{row} mode.
20623 Using this display mode makes it easy to spot color casts in the highlights
20624 and shadows of an image, by comparing the contours of the top and the bottom
20625 graphs of each waveform. Since whites, grays, and blacks are characterized
20626 by exactly equal amounts of red, green, and blue, neutral areas of the picture
20627 should display three waveforms of roughly equal width/height. If not, the
20628 correction is easy to perform by making level adjustments the three waveforms.
20630 Default is @code{stack}.
20632 @item components, c
20633 Set which color components to display. Default is 1, which means only luminance
20634 or red color component if input is in RGB colorspace. If is set for example to
20635 7 it will display all 3 (if) available color components.
20640 No envelope, this is default.
20643 Instant envelope, minimum and maximum values presented in graph will be easily
20644 visible even with small @code{step} value.
20647 Hold minimum and maximum values presented in graph across time. This way you
20648 can still spot out of range values without constantly looking at waveforms.
20651 Peak and instant envelope combined together.
20657 No filtering, this is default.
20660 Luma and chroma combined together.
20663 Similar as above, but shows difference between blue and red chroma.
20666 Similar as above, but use different colors.
20669 Similar as above, but again with different colors.
20672 Displays only chroma.
20675 Displays actual color value on waveform.
20678 Similar as above, but with luma showing frequency of chroma values.
20682 Set which graticule to display.
20686 Do not display graticule.
20689 Display green graticule showing legal broadcast ranges.
20692 Display orange graticule showing legal broadcast ranges.
20695 Display invert graticule showing legal broadcast ranges.
20699 Set graticule opacity.
20702 Set graticule flags.
20706 Draw numbers above lines. By default enabled.
20709 Draw dots instead of lines.
20713 Set scale used for displaying graticule.
20720 Default is digital.
20723 Set background opacity.
20727 Set tint for output.
20728 Only used with lowpass filter and when display is not overlay and input
20729 pixel formats are not RGB.
20732 @section weave, doubleweave
20734 The @code{weave} takes a field-based video input and join
20735 each two sequential fields into single frame, producing a new double
20736 height clip with half the frame rate and half the frame count.
20738 The @code{doubleweave} works same as @code{weave} but without
20739 halving frame rate and frame count.
20741 It accepts the following option:
20745 Set first field. Available values are:
20749 Set the frame as top-field-first.
20752 Set the frame as bottom-field-first.
20756 @subsection Examples
20760 Interlace video using @ref{select} and @ref{separatefields} filter:
20762 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
20767 Apply the xBR high-quality magnification filter which is designed for pixel
20768 art. It follows a set of edge-detection rules, see
20769 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
20771 It accepts the following option:
20775 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
20776 @code{3xBR} and @code{4} for @code{4xBR}.
20777 Default is @code{3}.
20782 Apply cross fade from one input video stream to another input video stream.
20783 The cross fade is applied for specified duration.
20785 The filter accepts the following options:
20789 Set one of available transition effects:
20837 Default transition effect is fade.
20840 Set cross fade duration in seconds.
20841 Default duration is 1 second.
20844 Set cross fade start relative to first input stream in seconds.
20845 Default offset is 0.
20848 Set expression for custom transition effect.
20850 The expressions can use the following variables and functions:
20855 The coordinates of the current sample.
20859 The width and height of the image.
20862 Progress of transition effect.
20865 Currently processed plane.
20868 Return value of first input at current location and plane.
20871 Return value of second input at current location and plane.
20877 Return the value of the pixel at location (@var{x},@var{y}) of the
20878 first/second/third/fourth component of first input.
20884 Return the value of the pixel at location (@var{x},@var{y}) of the
20885 first/second/third/fourth component of second input.
20889 @subsection Examples
20893 Cross fade from one input video to another input video, with fade transition and duration of transition
20894 of 2 seconds starting at offset of 5 seconds:
20896 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
20901 Pick median pixels from several input videos.
20903 The filter accepts the following options:
20907 Set number of inputs.
20908 Default is 3. Allowed range is from 3 to 255.
20909 If number of inputs is even number, than result will be mean value between two median values.
20912 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
20915 Set median percentile. Default value is @code{0.5}.
20916 Default value of @code{0.5} will pick always median values, while @code{0} will pick
20917 minimum values, and @code{1} maximum values.
20921 Stack video inputs into custom layout.
20923 All streams must be of same pixel format.
20925 The filter accepts the following options:
20929 Set number of input streams. Default is 2.
20932 Specify layout of inputs.
20933 This option requires the desired layout configuration to be explicitly set by the user.
20934 This sets position of each video input in output. Each input
20935 is separated by '|'.
20936 The first number represents the column, and the second number represents the row.
20937 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
20938 where X is video input from which to take width or height.
20939 Multiple values can be used when separated by '+'. In such
20940 case values are summed together.
20942 Note that if inputs are of different sizes gaps may appear, as not all of
20943 the output video frame will be filled. Similarly, videos can overlap each
20944 other if their position doesn't leave enough space for the full frame of
20947 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
20948 a layout must be set by the user.
20951 If set to 1, force the output to terminate when the shortest input
20952 terminates. Default value is 0.
20955 If set to valid color, all unused pixels will be filled with that color.
20956 By default fill is set to none, so it is disabled.
20959 @subsection Examples
20963 Display 4 inputs into 2x2 grid.
20967 input1(0, 0) | input3(w0, 0)
20968 input2(0, h0) | input4(w0, h0)
20972 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
20975 Note that if inputs are of different sizes, gaps or overlaps may occur.
20978 Display 4 inputs into 1x4 grid.
20985 input4(0, h0+h1+h2)
20989 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
20992 Note that if inputs are of different widths, unused space will appear.
20995 Display 9 inputs into 3x3 grid.
20999 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21000 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21001 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21005 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
21008 Note that if inputs are of different sizes, gaps or overlaps may occur.
21011 Display 16 inputs into 4x4 grid.
21015 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21016 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21017 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21018 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21022 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|
21023 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
21026 Note that if inputs are of different sizes, gaps or overlaps may occur.
21033 Deinterlace the input video ("yadif" means "yet another deinterlacing
21036 It accepts the following parameters:
21042 The interlacing mode to adopt. It accepts one of the following values:
21045 @item 0, send_frame
21046 Output one frame for each frame.
21047 @item 1, send_field
21048 Output one frame for each field.
21049 @item 2, send_frame_nospatial
21050 Like @code{send_frame}, but it skips the spatial interlacing check.
21051 @item 3, send_field_nospatial
21052 Like @code{send_field}, but it skips the spatial interlacing check.
21055 The default value is @code{send_frame}.
21058 The picture field parity assumed for the input interlaced video. It accepts one
21059 of the following values:
21063 Assume the top field is first.
21065 Assume the bottom field is first.
21067 Enable automatic detection of field parity.
21070 The default value is @code{auto}.
21071 If the interlacing is unknown or the decoder does not export this information,
21072 top field first will be assumed.
21075 Specify which frames to deinterlace. Accepts one of the following
21080 Deinterlace all frames.
21081 @item 1, interlaced
21082 Only deinterlace frames marked as interlaced.
21085 The default value is @code{all}.
21088 @section yadif_cuda
21090 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21091 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21094 It accepts the following parameters:
21100 The interlacing mode to adopt. It accepts one of the following values:
21103 @item 0, send_frame
21104 Output one frame for each frame.
21105 @item 1, send_field
21106 Output one frame for each field.
21107 @item 2, send_frame_nospatial
21108 Like @code{send_frame}, but it skips the spatial interlacing check.
21109 @item 3, send_field_nospatial
21110 Like @code{send_field}, but it skips the spatial interlacing check.
21113 The default value is @code{send_frame}.
21116 The picture field parity assumed for the input interlaced video. It accepts one
21117 of the following values:
21121 Assume the top field is first.
21123 Assume the bottom field is first.
21125 Enable automatic detection of field parity.
21128 The default value is @code{auto}.
21129 If the interlacing is unknown or the decoder does not export this information,
21130 top field first will be assumed.
21133 Specify which frames to deinterlace. Accepts one of the following
21138 Deinterlace all frames.
21139 @item 1, interlaced
21140 Only deinterlace frames marked as interlaced.
21143 The default value is @code{all}.
21148 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21149 The algorithm is described in
21150 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21152 It accepts the following parameters:
21156 Set the window radius. Default value is 3.
21159 Set which planes to filter. Default is only the first plane.
21162 Set blur strength. Default value is 128.
21165 @subsection Commands
21166 This filter supports same @ref{commands} as options.
21170 Apply Zoom & Pan effect.
21172 This filter accepts the following options:
21176 Set the zoom expression. Range is 1-10. Default is 1.
21180 Set the x and y expression. Default is 0.
21183 Set the duration expression in number of frames.
21184 This sets for how many number of frames effect will last for
21185 single input image.
21188 Set the output image size, default is 'hd720'.
21191 Set the output frame rate, default is '25'.
21194 Each expression can contain the following constants:
21213 Output frame count.
21216 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21218 @item out_time, time, ot
21219 The output timestamp expressed in seconds.
21223 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21224 for current input frame.
21228 'x' and 'y' of last output frame of previous input frame or 0 when there was
21229 not yet such frame (first input frame).
21232 Last calculated zoom from 'z' expression for current input frame.
21235 Last calculated zoom of last output frame of previous input frame.
21238 Number of output frames for current input frame. Calculated from 'd' expression
21239 for each input frame.
21242 number of output frames created for previous input frame
21245 Rational number: input width / input height
21248 sample aspect ratio
21251 display aspect ratio
21255 @subsection Examples
21259 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21261 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
21265 Zoom in up to 1.5x and pan always at center of picture:
21267 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21271 Same as above but without pausing:
21273 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21277 Zoom in 2x into center of picture only for the first second of the input video:
21279 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21286 Scale (resize) the input video, using the z.lib library:
21287 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21288 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21290 The zscale filter forces the output display aspect ratio to be the same
21291 as the input, by changing the output sample aspect ratio.
21293 If the input image format is different from the format requested by
21294 the next filter, the zscale filter will convert the input to the
21297 @subsection Options
21298 The filter accepts the following options.
21303 Set the output video dimension expression. Default value is the input
21306 If the @var{width} or @var{w} value is 0, the input width is used for
21307 the output. If the @var{height} or @var{h} value is 0, the input height
21308 is used for the output.
21310 If one and only one of the values is -n with n >= 1, the zscale filter
21311 will use a value that maintains the aspect ratio of the input image,
21312 calculated from the other specified dimension. After that it will,
21313 however, make sure that the calculated dimension is divisible by n and
21314 adjust the value if necessary.
21316 If both values are -n with n >= 1, the behavior will be identical to
21317 both values being set to 0 as previously detailed.
21319 See below for the list of accepted constants for use in the dimension
21323 Set the video size. For the syntax of this option, check the
21324 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21327 Set the dither type.
21329 Possible values are:
21334 @item error_diffusion
21340 Set the resize filter type.
21342 Possible values are:
21352 Default is bilinear.
21355 Set the color range.
21357 Possible values are:
21364 Default is same as input.
21367 Set the color primaries.
21369 Possible values are:
21379 Default is same as input.
21382 Set the transfer characteristics.
21384 Possible values are:
21398 Default is same as input.
21401 Set the colorspace matrix.
21403 Possible value are:
21414 Default is same as input.
21417 Set the input color range.
21419 Possible values are:
21426 Default is same as input.
21428 @item primariesin, pin
21429 Set the input color primaries.
21431 Possible values are:
21441 Default is same as input.
21443 @item transferin, tin
21444 Set the input transfer characteristics.
21446 Possible values are:
21457 Default is same as input.
21459 @item matrixin, min
21460 Set the input colorspace matrix.
21462 Possible value are:
21474 Set the output chroma location.
21476 Possible values are:
21487 @item chromalin, cin
21488 Set the input chroma location.
21490 Possible values are:
21502 Set the nominal peak luminance.
21505 The values of the @option{w} and @option{h} options are expressions
21506 containing the following constants:
21511 The input width and height
21515 These are the same as @var{in_w} and @var{in_h}.
21519 The output (scaled) width and height
21523 These are the same as @var{out_w} and @var{out_h}
21526 The same as @var{iw} / @var{ih}
21529 input sample aspect ratio
21532 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
21536 horizontal and vertical input chroma subsample values. For example for the
21537 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21541 horizontal and vertical output chroma subsample values. For example for the
21542 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
21545 @subsection Commands
21547 This filter supports the following commands:
21551 Set the output video dimension expression.
21552 The command accepts the same syntax of the corresponding option.
21554 If the specified expression is not valid, it is kept at its current
21558 @c man end VIDEO FILTERS
21560 @chapter OpenCL Video Filters
21561 @c man begin OPENCL VIDEO FILTERS
21563 Below is a description of the currently available OpenCL video filters.
21565 To enable compilation of these filters you need to configure FFmpeg with
21566 @code{--enable-opencl}.
21568 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
21571 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
21572 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
21573 given device parameters.
21575 @item -filter_hw_device @var{name}
21576 Pass the hardware device called @var{name} to all filters in any filter graph.
21580 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
21584 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
21586 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
21590 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.
21592 @section avgblur_opencl
21594 Apply average blur filter.
21596 The filter accepts the following options:
21600 Set horizontal radius size.
21601 Range is @code{[1, 1024]} and default value is @code{1}.
21604 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
21607 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
21610 @subsection Example
21614 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.
21616 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
21620 @section boxblur_opencl
21622 Apply a boxblur algorithm to the input video.
21624 It accepts the following parameters:
21628 @item luma_radius, lr
21629 @item luma_power, lp
21630 @item chroma_radius, cr
21631 @item chroma_power, cp
21632 @item alpha_radius, ar
21633 @item alpha_power, ap
21637 A description of the accepted options follows.
21640 @item luma_radius, lr
21641 @item chroma_radius, cr
21642 @item alpha_radius, ar
21643 Set an expression for the box radius in pixels used for blurring the
21644 corresponding input plane.
21646 The radius value must be a non-negative number, and must not be
21647 greater than the value of the expression @code{min(w,h)/2} for the
21648 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
21651 Default value for @option{luma_radius} is "2". If not specified,
21652 @option{chroma_radius} and @option{alpha_radius} default to the
21653 corresponding value set for @option{luma_radius}.
21655 The expressions can contain the following constants:
21659 The input width and height in pixels.
21663 The input chroma image width and height in pixels.
21667 The horizontal and vertical chroma subsample values. For example, for the
21668 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
21671 @item luma_power, lp
21672 @item chroma_power, cp
21673 @item alpha_power, ap
21674 Specify how many times the boxblur filter is applied to the
21675 corresponding plane.
21677 Default value for @option{luma_power} is 2. If not specified,
21678 @option{chroma_power} and @option{alpha_power} default to the
21679 corresponding value set for @option{luma_power}.
21681 A value of 0 will disable the effect.
21684 @subsection Examples
21686 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.
21690 Apply a boxblur filter with the luma, chroma, and alpha radius
21691 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.
21693 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
21694 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
21698 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.
21700 For the luma plane, a 2x2 box radius will be run once.
21702 For the chroma plane, a 4x4 box radius will be run 5 times.
21704 For the alpha plane, a 3x3 box radius will be run 7 times.
21706 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
21710 @section colorkey_opencl
21711 RGB colorspace color keying.
21713 The filter accepts the following options:
21717 The color which will be replaced with transparency.
21720 Similarity percentage with the key color.
21722 0.01 matches only the exact key color, while 1.0 matches everything.
21727 0.0 makes pixels either fully transparent, or not transparent at all.
21729 Higher values result in semi-transparent pixels, with a higher transparency
21730 the more similar the pixels color is to the key color.
21733 @subsection Examples
21737 Make every semi-green pixel in the input transparent with some slight blending:
21739 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
21743 @section convolution_opencl
21745 Apply convolution of 3x3, 5x5, 7x7 matrix.
21747 The filter accepts the following options:
21754 Set matrix for each plane.
21755 Matrix is sequence of 9, 25 or 49 signed numbers.
21756 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
21762 Set multiplier for calculated value for each plane.
21763 If unset or 0, it will be sum of all matrix elements.
21764 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
21770 Set bias for each plane. This value is added to the result of the multiplication.
21771 Useful for making the overall image brighter or darker.
21772 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
21776 @subsection Examples
21782 -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
21788 -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
21792 Apply edge enhance:
21794 -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
21800 -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
21804 Apply laplacian edge detector which includes diagonals:
21806 -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
21812 -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
21816 @section erosion_opencl
21818 Apply erosion effect to the video.
21820 This filter replaces the pixel by the local(3x3) minimum.
21822 It accepts the following options:
21829 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21830 If @code{0}, plane will remain unchanged.
21833 Flag which specifies the pixel to refer to.
21834 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21836 Flags to local 3x3 coordinates region centered on @code{x}:
21845 @subsection Example
21849 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.
21851 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21855 @section deshake_opencl
21856 Feature-point based video stabilization filter.
21858 The filter accepts the following options:
21862 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
21865 Whether or not additional debug info should be displayed, both in the processed output and in the console.
21867 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
21869 Viewing point matches in the output video is only supported for RGB input.
21871 Defaults to @code{0}.
21873 @item adaptive_crop
21874 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
21876 Defaults to @code{1}.
21878 @item refine_features
21879 Whether or not feature points should be refined at a sub-pixel level.
21881 This can be turned off for a slight performance gain at the cost of precision.
21883 Defaults to @code{1}.
21885 @item smooth_strength
21886 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
21888 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
21890 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
21892 Defaults to @code{0.0}.
21894 @item smooth_window_multiplier
21895 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
21897 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
21899 Acceptable values range from @code{0.1} to @code{10.0}.
21901 Larger values increase the amount of motion data available for determining how to smooth the camera path,
21902 potentially improving smoothness, but also increase latency and memory usage.
21904 Defaults to @code{2.0}.
21908 @subsection Examples
21912 Stabilize a video with a fixed, medium smoothing strength:
21914 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
21918 Stabilize a video with debugging (both in console and in rendered video):
21920 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
21924 @section dilation_opencl
21926 Apply dilation effect to the video.
21928 This filter replaces the pixel by the local(3x3) maximum.
21930 It accepts the following options:
21937 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
21938 If @code{0}, plane will remain unchanged.
21941 Flag which specifies the pixel to refer to.
21942 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
21944 Flags to local 3x3 coordinates region centered on @code{x}:
21953 @subsection Example
21957 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.
21959 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
21963 @section nlmeans_opencl
21965 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
21967 @section overlay_opencl
21969 Overlay one video on top of another.
21971 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
21972 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
21974 The filter accepts the following options:
21979 Set the x coordinate of the overlaid video on the main video.
21980 Default value is @code{0}.
21983 Set the y coordinate of the overlaid video on the main video.
21984 Default value is @code{0}.
21988 @subsection Examples
21992 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
21994 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
21997 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
21999 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22004 @section pad_opencl
22006 Add paddings to the input image, and place the original input at the
22007 provided @var{x}, @var{y} coordinates.
22009 It accepts the following options:
22014 Specify an expression for the size of the output image with the
22015 paddings added. If the value for @var{width} or @var{height} is 0, the
22016 corresponding input size is used for the output.
22018 The @var{width} expression can reference the value set by the
22019 @var{height} expression, and vice versa.
22021 The default value of @var{width} and @var{height} is 0.
22025 Specify the offsets to place the input image at within the padded area,
22026 with respect to the top/left border of the output image.
22028 The @var{x} expression can reference the value set by the @var{y}
22029 expression, and vice versa.
22031 The default value of @var{x} and @var{y} is 0.
22033 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22034 so the input image is centered on the padded area.
22037 Specify the color of the padded area. For the syntax of this option,
22038 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22039 manual,ffmpeg-utils}.
22042 Pad to an aspect instead to a resolution.
22045 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22046 options are expressions containing the following constants:
22051 The input video width and height.
22055 These are the same as @var{in_w} and @var{in_h}.
22059 The output width and height (the size of the padded area), as
22060 specified by the @var{width} and @var{height} expressions.
22064 These are the same as @var{out_w} and @var{out_h}.
22068 The x and y offsets as specified by the @var{x} and @var{y}
22069 expressions, or NAN if not yet specified.
22072 same as @var{iw} / @var{ih}
22075 input sample aspect ratio
22078 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22081 @section prewitt_opencl
22083 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22085 The filter accepts the following option:
22089 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22092 Set value which will be multiplied with filtered result.
22093 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22096 Set value which will be added to filtered result.
22097 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22100 @subsection Example
22104 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22106 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22110 @anchor{program_opencl}
22111 @section program_opencl
22113 Filter video using an OpenCL program.
22118 OpenCL program source file.
22121 Kernel name in program.
22124 Number of inputs to the filter. Defaults to 1.
22127 Size of output frames. Defaults to the same as the first input.
22131 The @code{program_opencl} filter also supports the @ref{framesync} options.
22133 The program source file must contain a kernel function with the given name,
22134 which will be run once for each plane of the output. Each run on a plane
22135 gets enqueued as a separate 2D global NDRange with one work-item for each
22136 pixel to be generated. The global ID offset for each work-item is therefore
22137 the coordinates of a pixel in the destination image.
22139 The kernel function needs to take the following arguments:
22142 Destination image, @var{__write_only image2d_t}.
22144 This image will become the output; the kernel should write all of it.
22146 Frame index, @var{unsigned int}.
22148 This is a counter starting from zero and increasing by one for each frame.
22150 Source images, @var{__read_only image2d_t}.
22152 These are the most recent images on each input. The kernel may read from
22153 them to generate the output, but they can't be written to.
22160 Copy the input to the output (output must be the same size as the input).
22162 __kernel void copy(__write_only image2d_t destination,
22163 unsigned int index,
22164 __read_only image2d_t source)
22166 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22168 int2 location = (int2)(get_global_id(0), get_global_id(1));
22170 float4 value = read_imagef(source, sampler, location);
22172 write_imagef(destination, location, value);
22177 Apply a simple transformation, rotating the input by an amount increasing
22178 with the index counter. Pixel values are linearly interpolated by the
22179 sampler, and the output need not have the same dimensions as the input.
22181 __kernel void rotate_image(__write_only image2d_t dst,
22182 unsigned int index,
22183 __read_only image2d_t src)
22185 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22186 CLK_FILTER_LINEAR);
22188 float angle = (float)index / 100.0f;
22190 float2 dst_dim = convert_float2(get_image_dim(dst));
22191 float2 src_dim = convert_float2(get_image_dim(src));
22193 float2 dst_cen = dst_dim / 2.0f;
22194 float2 src_cen = src_dim / 2.0f;
22196 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22198 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22200 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22201 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22203 src_pos = src_pos * src_dim / dst_dim;
22205 float2 src_loc = src_pos + src_cen;
22207 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22208 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22209 write_imagef(dst, dst_loc, 0.5f);
22211 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22216 Blend two inputs together, with the amount of each input used varying
22217 with the index counter.
22219 __kernel void blend_images(__write_only image2d_t dst,
22220 unsigned int index,
22221 __read_only image2d_t src1,
22222 __read_only image2d_t src2)
22224 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22225 CLK_FILTER_LINEAR);
22227 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22229 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22230 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22231 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22233 float4 val1 = read_imagef(src1, sampler, src1_loc);
22234 float4 val2 = read_imagef(src2, sampler, src2_loc);
22236 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22242 @section roberts_opencl
22243 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22245 The filter accepts the following option:
22249 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22252 Set value which will be multiplied with filtered result.
22253 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22256 Set value which will be added to filtered result.
22257 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22260 @subsection Example
22264 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22266 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22270 @section sobel_opencl
22272 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22274 The filter accepts the following option:
22278 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22281 Set value which will be multiplied with filtered result.
22282 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22285 Set value which will be added to filtered result.
22286 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22289 @subsection Example
22293 Apply sobel operator with scale set to 2 and delta set to 10
22295 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22299 @section tonemap_opencl
22301 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22303 It accepts the following parameters:
22307 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22310 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22313 Apply desaturation for highlights that exceed this level of brightness. The
22314 higher the parameter, the more color information will be preserved. This
22315 setting helps prevent unnaturally blown-out colors for super-highlights, by
22316 (smoothly) turning into white instead. This makes images feel more natural,
22317 at the cost of reducing information about out-of-range colors.
22319 The default value is 0.5, and the algorithm here is a little different from
22320 the cpu version tonemap currently. A setting of 0.0 disables this option.
22323 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22324 is used to detect whether the scene has changed or not. If the distance between
22325 the current frame average brightness and the current running average exceeds
22326 a threshold value, we would re-calculate scene average and peak brightness.
22327 The default value is 0.2.
22330 Specify the output pixel format.
22332 Currently supported formats are:
22339 Set the output color range.
22341 Possible values are:
22347 Default is same as input.
22350 Set the output color primaries.
22352 Possible values are:
22358 Default is same as input.
22361 Set the output transfer characteristics.
22363 Possible values are:
22372 Set the output colorspace matrix.
22374 Possible value are:
22380 Default is same as input.
22384 @subsection Example
22388 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22390 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22394 @section unsharp_opencl
22396 Sharpen or blur the input video.
22398 It accepts the following parameters:
22401 @item luma_msize_x, lx
22402 Set the luma matrix horizontal size.
22403 Range is @code{[1, 23]} and default value is @code{5}.
22405 @item luma_msize_y, ly
22406 Set the luma matrix vertical size.
22407 Range is @code{[1, 23]} and default value is @code{5}.
22409 @item luma_amount, la
22410 Set the luma effect strength.
22411 Range is @code{[-10, 10]} and default value is @code{1.0}.
22413 Negative values will blur the input video, while positive values will
22414 sharpen it, a value of zero will disable the effect.
22416 @item chroma_msize_x, cx
22417 Set the chroma matrix horizontal size.
22418 Range is @code{[1, 23]} and default value is @code{5}.
22420 @item chroma_msize_y, cy
22421 Set the chroma matrix vertical size.
22422 Range is @code{[1, 23]} and default value is @code{5}.
22424 @item chroma_amount, ca
22425 Set the chroma effect strength.
22426 Range is @code{[-10, 10]} and default value is @code{0.0}.
22428 Negative values will blur the input video, while positive values will
22429 sharpen it, a value of zero will disable the effect.
22433 All parameters are optional and default to the equivalent of the
22434 string '5:5:1.0:5:5:0.0'.
22436 @subsection Examples
22440 Apply strong luma sharpen effect:
22442 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22446 Apply a strong blur of both luma and chroma parameters:
22448 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22452 @section xfade_opencl
22454 Cross fade two videos with custom transition effect by using OpenCL.
22456 It accepts the following options:
22460 Set one of possible transition effects.
22464 Select custom transition effect, the actual transition description
22465 will be picked from source and kernel options.
22477 Default transition is fade.
22481 OpenCL program source file for custom transition.
22484 Set name of kernel to use for custom transition from program source file.
22487 Set duration of video transition.
22490 Set time of start of transition relative to first video.
22493 The program source file must contain a kernel function with the given name,
22494 which will be run once for each plane of the output. Each run on a plane
22495 gets enqueued as a separate 2D global NDRange with one work-item for each
22496 pixel to be generated. The global ID offset for each work-item is therefore
22497 the coordinates of a pixel in the destination image.
22499 The kernel function needs to take the following arguments:
22502 Destination image, @var{__write_only image2d_t}.
22504 This image will become the output; the kernel should write all of it.
22507 First Source image, @var{__read_only image2d_t}.
22508 Second Source image, @var{__read_only image2d_t}.
22510 These are the most recent images on each input. The kernel may read from
22511 them to generate the output, but they can't be written to.
22514 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22521 Apply dots curtain transition effect:
22523 __kernel void blend_images(__write_only image2d_t dst,
22524 __read_only image2d_t src1,
22525 __read_only image2d_t src2,
22528 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22529 CLK_FILTER_LINEAR);
22530 int2 p = (int2)(get_global_id(0), get_global_id(1));
22531 float2 rp = (float2)(get_global_id(0), get_global_id(1));
22532 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
22535 float2 dots = (float2)(20.0, 20.0);
22536 float2 center = (float2)(0,0);
22539 float4 val1 = read_imagef(src1, sampler, p);
22540 float4 val2 = read_imagef(src2, sampler, p);
22541 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
22543 write_imagef(dst, p, next ? val1 : val2);
22549 @c man end OPENCL VIDEO FILTERS
22551 @chapter VAAPI Video Filters
22552 @c man begin VAAPI VIDEO FILTERS
22554 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
22556 To enable compilation of these filters you need to configure FFmpeg with
22557 @code{--enable-vaapi}.
22559 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}
22561 @section tonemap_vaapi
22563 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
22564 It maps the dynamic range of HDR10 content to the SDR content.
22565 It currently only accepts HDR10 as input.
22567 It accepts the following parameters:
22571 Specify the output pixel format.
22573 Currently supported formats are:
22582 Set the output color primaries.
22584 Default is same as input.
22587 Set the output transfer characteristics.
22592 Set the output colorspace matrix.
22594 Default is same as input.
22598 @subsection Example
22602 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
22604 tonemap_vaapi=format=p010:t=bt2020-10
22608 @c man end VAAPI VIDEO FILTERS
22610 @chapter Video Sources
22611 @c man begin VIDEO SOURCES
22613 Below is a description of the currently available video sources.
22617 Buffer video frames, and make them available to the filter chain.
22619 This source is mainly intended for a programmatic use, in particular
22620 through the interface defined in @file{libavfilter/buffersrc.h}.
22622 It accepts the following parameters:
22627 Specify the size (width and height) of the buffered video frames. For the
22628 syntax of this option, check the
22629 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22632 The input video width.
22635 The input video height.
22638 A string representing the pixel format of the buffered video frames.
22639 It may be a number corresponding to a pixel format, or a pixel format
22643 Specify the timebase assumed by the timestamps of the buffered frames.
22646 Specify the frame rate expected for the video stream.
22648 @item pixel_aspect, sar
22649 The sample (pixel) aspect ratio of the input video.
22652 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
22653 to the filtergraph description to specify swscale flags for automatically
22654 inserted scalers. See @ref{Filtergraph syntax}.
22656 @item hw_frames_ctx
22657 When using a hardware pixel format, this should be a reference to an
22658 AVHWFramesContext describing input frames.
22663 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
22666 will instruct the source to accept video frames with size 320x240 and
22667 with format "yuv410p", assuming 1/24 as the timestamps timebase and
22668 square pixels (1:1 sample aspect ratio).
22669 Since the pixel format with name "yuv410p" corresponds to the number 6
22670 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
22671 this example corresponds to:
22673 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
22676 Alternatively, the options can be specified as a flat string, but this
22677 syntax is deprecated:
22679 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
22683 Create a pattern generated by an elementary cellular automaton.
22685 The initial state of the cellular automaton can be defined through the
22686 @option{filename} and @option{pattern} options. If such options are
22687 not specified an initial state is created randomly.
22689 At each new frame a new row in the video is filled with the result of
22690 the cellular automaton next generation. The behavior when the whole
22691 frame is filled is defined by the @option{scroll} option.
22693 This source accepts the following options:
22697 Read the initial cellular automaton state, i.e. the starting row, from
22698 the specified file.
22699 In the file, each non-whitespace character is considered an alive
22700 cell, a newline will terminate the row, and further characters in the
22701 file will be ignored.
22704 Read the initial cellular automaton state, i.e. the starting row, from
22705 the specified string.
22707 Each non-whitespace character in the string is considered an alive
22708 cell, a newline will terminate the row, and further characters in the
22709 string will be ignored.
22712 Set the video rate, that is the number of frames generated per second.
22715 @item random_fill_ratio, ratio
22716 Set the random fill ratio for the initial cellular automaton row. It
22717 is a floating point number value ranging from 0 to 1, defaults to
22720 This option is ignored when a file or a pattern is specified.
22722 @item random_seed, seed
22723 Set the seed for filling randomly the initial row, must be an integer
22724 included between 0 and UINT32_MAX. If not specified, or if explicitly
22725 set to -1, the filter will try to use a good random seed on a best
22729 Set the cellular automaton rule, it is a number ranging from 0 to 255.
22730 Default value is 110.
22733 Set the size of the output video. For the syntax of this option, check the
22734 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22736 If @option{filename} or @option{pattern} is specified, the size is set
22737 by default to the width of the specified initial state row, and the
22738 height is set to @var{width} * PHI.
22740 If @option{size} is set, it must contain the width of the specified
22741 pattern string, and the specified pattern will be centered in the
22744 If a filename or a pattern string is not specified, the size value
22745 defaults to "320x518" (used for a randomly generated initial state).
22748 If set to 1, scroll the output upward when all the rows in the output
22749 have been already filled. If set to 0, the new generated row will be
22750 written over the top row just after the bottom row is filled.
22753 @item start_full, full
22754 If set to 1, completely fill the output with generated rows before
22755 outputting the first frame.
22756 This is the default behavior, for disabling set the value to 0.
22759 If set to 1, stitch the left and right row edges together.
22760 This is the default behavior, for disabling set the value to 0.
22763 @subsection Examples
22767 Read the initial state from @file{pattern}, and specify an output of
22770 cellauto=f=pattern:s=200x400
22774 Generate a random initial row with a width of 200 cells, with a fill
22777 cellauto=ratio=2/3:s=200x200
22781 Create a pattern generated by rule 18 starting by a single alive cell
22782 centered on an initial row with width 100:
22784 cellauto=p=@@:s=100x400:full=0:rule=18
22788 Specify a more elaborated initial pattern:
22790 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
22795 @anchor{coreimagesrc}
22796 @section coreimagesrc
22797 Video source generated on GPU using Apple's CoreImage API on OSX.
22799 This video source is a specialized version of the @ref{coreimage} video filter.
22800 Use a core image generator at the beginning of the applied filterchain to
22801 generate the content.
22803 The coreimagesrc video source accepts the following options:
22805 @item list_generators
22806 List all available generators along with all their respective options as well as
22807 possible minimum and maximum values along with the default values.
22809 list_generators=true
22813 Specify the size of the sourced video. For the syntax of this option, check the
22814 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
22815 The default value is @code{320x240}.
22818 Specify the frame rate of the sourced video, as the number of frames
22819 generated per second. It has to be a string in the format
22820 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22821 number or a valid video frame rate abbreviation. The default value is
22825 Set the sample aspect ratio of the sourced video.
22828 Set the duration of the sourced video. See
22829 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22830 for the accepted syntax.
22832 If not specified, or the expressed duration is negative, the video is
22833 supposed to be generated forever.
22836 Additionally, all options of the @ref{coreimage} video filter are accepted.
22837 A complete filterchain can be used for further processing of the
22838 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
22839 and examples for details.
22841 @subsection Examples
22846 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
22847 given as complete and escaped command-line for Apple's standard bash shell:
22849 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
22851 This example is equivalent to the QRCode example of @ref{coreimage} without the
22852 need for a nullsrc video source.
22857 Generate several gradients.
22861 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22862 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22865 Set frame rate, expressed as number of frames per second. Default
22868 @item c0, c1, c2, c3, c4, c5, c6, c7
22869 Set 8 colors. Default values for colors is to pick random one.
22871 @item x0, y0, y0, y1
22872 Set gradient line source and destination points. If negative or out of range, random ones
22876 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
22879 Set seed for picking gradient line points.
22882 Set the duration of the sourced video. See
22883 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22884 for the accepted syntax.
22886 If not specified, or the expressed duration is negative, the video is
22887 supposed to be generated forever.
22890 Set speed of gradients rotation.
22894 @section mandelbrot
22896 Generate a Mandelbrot set fractal, and progressively zoom towards the
22897 point specified with @var{start_x} and @var{start_y}.
22899 This source accepts the following options:
22904 Set the terminal pts value. Default value is 400.
22907 Set the terminal scale value.
22908 Must be a floating point value. Default value is 0.3.
22911 Set the inner coloring mode, that is the algorithm used to draw the
22912 Mandelbrot fractal internal region.
22914 It shall assume one of the following values:
22919 Show time until convergence.
22921 Set color based on point closest to the origin of the iterations.
22926 Default value is @var{mincol}.
22929 Set the bailout value. Default value is 10.0.
22932 Set the maximum of iterations performed by the rendering
22933 algorithm. Default value is 7189.
22936 Set outer coloring mode.
22937 It shall assume one of following values:
22939 @item iteration_count
22940 Set iteration count mode.
22941 @item normalized_iteration_count
22942 set normalized iteration count mode.
22944 Default value is @var{normalized_iteration_count}.
22947 Set frame rate, expressed as number of frames per second. Default
22951 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
22952 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
22955 Set the initial scale value. Default value is 3.0.
22958 Set the initial x position. Must be a floating point value between
22959 -100 and 100. Default value is -0.743643887037158704752191506114774.
22962 Set the initial y position. Must be a floating point value between
22963 -100 and 100. Default value is -0.131825904205311970493132056385139.
22968 Generate various test patterns, as generated by the MPlayer test filter.
22970 The size of the generated video is fixed, and is 256x256.
22971 This source is useful in particular for testing encoding features.
22973 This source accepts the following options:
22978 Specify the frame rate of the sourced video, as the number of frames
22979 generated per second. It has to be a string in the format
22980 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
22981 number or a valid video frame rate abbreviation. The default value is
22985 Set the duration of the sourced video. See
22986 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
22987 for the accepted syntax.
22989 If not specified, or the expressed duration is negative, the video is
22990 supposed to be generated forever.
22994 Set the number or the name of the test to perform. Supported tests are:
23008 @item max_frames, m
23009 Set the maximum number of frames generated for each test, default value is 30.
23013 Default value is "all", which will cycle through the list of all tests.
23018 mptestsrc=t=dc_luma
23021 will generate a "dc_luma" test pattern.
23023 @section frei0r_src
23025 Provide a frei0r source.
23027 To enable compilation of this filter you need to install the frei0r
23028 header and configure FFmpeg with @code{--enable-frei0r}.
23030 This source accepts the following parameters:
23035 The size of the video to generate. For the syntax of this option, check the
23036 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23039 The framerate of the generated video. It may be a string of the form
23040 @var{num}/@var{den} or a frame rate abbreviation.
23043 The name to the frei0r source to load. For more information regarding frei0r and
23044 how to set the parameters, read the @ref{frei0r} section in the video filters
23047 @item filter_params
23048 A '|'-separated list of parameters to pass to the frei0r source.
23052 For example, to generate a frei0r partik0l source with size 200x200
23053 and frame rate 10 which is overlaid on the overlay filter main input:
23055 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23060 Generate a life pattern.
23062 This source is based on a generalization of John Conway's life game.
23064 The sourced input represents a life grid, each pixel represents a cell
23065 which can be in one of two possible states, alive or dead. Every cell
23066 interacts with its eight neighbours, which are the cells that are
23067 horizontally, vertically, or diagonally adjacent.
23069 At each interaction the grid evolves according to the adopted rule,
23070 which specifies the number of neighbor alive cells which will make a
23071 cell stay alive or born. The @option{rule} option allows one to specify
23074 This source accepts the following options:
23078 Set the file from which to read the initial grid state. In the file,
23079 each non-whitespace character is considered an alive cell, and newline
23080 is used to delimit the end of each row.
23082 If this option is not specified, the initial grid is generated
23086 Set the video rate, that is the number of frames generated per second.
23089 @item random_fill_ratio, ratio
23090 Set the random fill ratio for the initial random grid. It is a
23091 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23092 It is ignored when a file is specified.
23094 @item random_seed, seed
23095 Set the seed for filling the initial random grid, must be an integer
23096 included between 0 and UINT32_MAX. If not specified, or if explicitly
23097 set to -1, the filter will try to use a good random seed on a best
23103 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23104 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23105 @var{NS} specifies the number of alive neighbor cells which make a
23106 live cell stay alive, and @var{NB} the number of alive neighbor cells
23107 which make a dead cell to become alive (i.e. to "born").
23108 "s" and "b" can be used in place of "S" and "B", respectively.
23110 Alternatively a rule can be specified by an 18-bits integer. The 9
23111 high order bits are used to encode the next cell state if it is alive
23112 for each number of neighbor alive cells, the low order bits specify
23113 the rule for "borning" new cells. Higher order bits encode for an
23114 higher number of neighbor cells.
23115 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23116 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23118 Default value is "S23/B3", which is the original Conway's game of life
23119 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23120 cells, and will born a new cell if there are three alive cells around
23124 Set the size of the output video. For the syntax of this option, check the
23125 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23127 If @option{filename} is specified, the size is set by default to the
23128 same size of the input file. If @option{size} is set, it must contain
23129 the size specified in the input file, and the initial grid defined in
23130 that file is centered in the larger resulting area.
23132 If a filename is not specified, the size value defaults to "320x240"
23133 (used for a randomly generated initial grid).
23136 If set to 1, stitch the left and right grid edges together, and the
23137 top and bottom edges also. Defaults to 1.
23140 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23141 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23142 value from 0 to 255.
23145 Set the color of living (or new born) cells.
23148 Set the color of dead cells. If @option{mold} is set, this is the first color
23149 used to represent a dead cell.
23152 Set mold color, for definitely dead and moldy cells.
23154 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23155 ffmpeg-utils manual,ffmpeg-utils}.
23158 @subsection Examples
23162 Read a grid from @file{pattern}, and center it on a grid of size
23165 life=f=pattern:s=300x300
23169 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23171 life=ratio=2/3:s=200x200
23175 Specify a custom rule for evolving a randomly generated grid:
23181 Full example with slow death effect (mold) using @command{ffplay}:
23183 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23190 @anchor{haldclutsrc}
23193 @anchor{pal100bars}
23194 @anchor{rgbtestsrc}
23196 @anchor{smptehdbars}
23199 @anchor{yuvtestsrc}
23200 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23202 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23204 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23206 The @code{color} source provides an uniformly colored input.
23208 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23209 @ref{haldclut} filter.
23211 The @code{nullsrc} source returns unprocessed video frames. It is
23212 mainly useful to be employed in analysis / debugging tools, or as the
23213 source for filters which ignore the input data.
23215 The @code{pal75bars} source generates a color bars pattern, based on
23216 EBU PAL recommendations with 75% color levels.
23218 The @code{pal100bars} source generates a color bars pattern, based on
23219 EBU PAL recommendations with 100% color levels.
23221 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23222 detecting RGB vs BGR issues. You should see a red, green and blue
23223 stripe from top to bottom.
23225 The @code{smptebars} source generates a color bars pattern, based on
23226 the SMPTE Engineering Guideline EG 1-1990.
23228 The @code{smptehdbars} source generates a color bars pattern, based on
23229 the SMPTE RP 219-2002.
23231 The @code{testsrc} source generates a test video pattern, showing a
23232 color pattern, a scrolling gradient and a timestamp. This is mainly
23233 intended for testing purposes.
23235 The @code{testsrc2} source is similar to testsrc, but supports more
23236 pixel formats instead of just @code{rgb24}. This allows using it as an
23237 input for other tests without requiring a format conversion.
23239 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23240 see a y, cb and cr stripe from top to bottom.
23242 The sources accept the following parameters:
23247 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23248 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23249 pixels to be used as identity matrix for 3D lookup tables. Each component is
23250 coded on a @code{1/(N*N)} scale.
23253 Specify the color of the source, only available in the @code{color}
23254 source. For the syntax of this option, check the
23255 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23258 Specify the size of the sourced video. For the syntax of this option, check the
23259 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23260 The default value is @code{320x240}.
23262 This option is not available with the @code{allrgb}, @code{allyuv}, and
23263 @code{haldclutsrc} filters.
23266 Specify the frame rate of the sourced video, as the number of frames
23267 generated per second. It has to be a string in the format
23268 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23269 number or a valid video frame rate abbreviation. The default value is
23273 Set the duration of the sourced video. See
23274 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23275 for the accepted syntax.
23277 If not specified, or the expressed duration is negative, the video is
23278 supposed to be generated forever.
23280 Since the frame rate is used as time base, all frames including the last one
23281 will have their full duration. If the specified duration is not a multiple
23282 of the frame duration, it will be rounded up.
23285 Set the sample aspect ratio of the sourced video.
23288 Specify the alpha (opacity) of the background, only available in the
23289 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23290 255 (fully opaque, the default).
23293 Set the number of decimals to show in the timestamp, only available in the
23294 @code{testsrc} source.
23296 The displayed timestamp value will correspond to the original
23297 timestamp value multiplied by the power of 10 of the specified
23298 value. Default value is 0.
23301 @subsection Examples
23305 Generate a video with a duration of 5.3 seconds, with size
23306 176x144 and a frame rate of 10 frames per second:
23308 testsrc=duration=5.3:size=qcif:rate=10
23312 The following graph description will generate a red source
23313 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23316 color=c=red@@0.2:s=qcif:r=10
23320 If the input content is to be ignored, @code{nullsrc} can be used. The
23321 following command generates noise in the luminance plane by employing
23322 the @code{geq} filter:
23324 nullsrc=s=256x256, geq=random(1)*255:128:128
23328 @subsection Commands
23330 The @code{color} source supports the following commands:
23334 Set the color of the created image. Accepts the same syntax of the
23335 corresponding @option{color} option.
23340 Generate video using an OpenCL program.
23345 OpenCL program source file.
23348 Kernel name in program.
23351 Size of frames to generate. This must be set.
23354 Pixel format to use for the generated frames. This must be set.
23357 Number of frames generated every second. Default value is '25'.
23361 For details of how the program loading works, see the @ref{program_opencl}
23368 Generate a colour ramp by setting pixel values from the position of the pixel
23369 in the output image. (Note that this will work with all pixel formats, but
23370 the generated output will not be the same.)
23372 __kernel void ramp(__write_only image2d_t dst,
23373 unsigned int index)
23375 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23378 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23380 write_imagef(dst, loc, val);
23385 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23387 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23388 unsigned int index)
23390 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23392 float4 value = 0.0f;
23393 int x = loc.x + index;
23394 int y = loc.y + index;
23395 while (x > 0 || y > 0) {
23396 if (x % 3 == 1 && y % 3 == 1) {
23404 write_imagef(dst, loc, value);
23410 @section sierpinski
23412 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23414 This source accepts the following options:
23418 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23419 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23422 Set frame rate, expressed as number of frames per second. Default
23426 Set seed which is used for random panning.
23429 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23432 Set fractal type, can be default @code{carpet} or @code{triangle}.
23435 @c man end VIDEO SOURCES
23437 @chapter Video Sinks
23438 @c man begin VIDEO SINKS
23440 Below is a description of the currently available video sinks.
23442 @section buffersink
23444 Buffer video frames, and make them available to the end of the filter
23447 This sink is mainly intended for programmatic use, in particular
23448 through the interface defined in @file{libavfilter/buffersink.h}
23449 or the options system.
23451 It accepts a pointer to an AVBufferSinkContext structure, which
23452 defines the incoming buffers' formats, to be passed as the opaque
23453 parameter to @code{avfilter_init_filter} for initialization.
23457 Null video sink: do absolutely nothing with the input video. It is
23458 mainly useful as a template and for use in analysis / debugging
23461 @c man end VIDEO SINKS
23463 @chapter Multimedia Filters
23464 @c man begin MULTIMEDIA FILTERS
23466 Below is a description of the currently available multimedia filters.
23470 Convert input audio to a video output, displaying the audio bit scope.
23472 The filter accepts the following options:
23476 Set frame rate, expressed as number of frames per second. Default
23480 Specify the video size for the output. For the syntax of this option, check the
23481 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23482 Default value is @code{1024x256}.
23485 Specify list of colors separated by space or by '|' which will be used to
23486 draw channels. Unrecognized or missing colors will be replaced
23490 @section adrawgraph
23491 Draw a graph using input audio metadata.
23493 See @ref{drawgraph}
23495 @section agraphmonitor
23497 See @ref{graphmonitor}.
23499 @section ahistogram
23501 Convert input audio to a video output, displaying the volume histogram.
23503 The filter accepts the following options:
23507 Specify how histogram is calculated.
23509 It accepts the following values:
23512 Use single histogram for all channels.
23514 Use separate histogram for each channel.
23516 Default is @code{single}.
23519 Set frame rate, expressed as number of frames per second. Default
23523 Specify the video size for the output. For the syntax of this option, check the
23524 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23525 Default value is @code{hd720}.
23530 It accepts the following values:
23541 reverse logarithmic
23543 Default is @code{log}.
23546 Set amplitude scale.
23548 It accepts the following values:
23555 Default is @code{log}.
23558 Set how much frames to accumulate in histogram.
23559 Default is 1. Setting this to -1 accumulates all frames.
23562 Set histogram ratio of window height.
23565 Set sonogram sliding.
23567 It accepts the following values:
23570 replace old rows with new ones.
23572 scroll from top to bottom.
23574 Default is @code{replace}.
23577 @section aphasemeter
23579 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
23580 representing mean phase of current audio frame. A video output can also be produced and is
23581 enabled by default. The audio is passed through as first output.
23583 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
23584 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
23585 and @code{1} means channels are in phase.
23587 The filter accepts the following options, all related to its video output:
23591 Set the output frame rate. Default value is @code{25}.
23594 Set the video size for the output. For the syntax of this option, check the
23595 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23596 Default value is @code{800x400}.
23601 Specify the red, green, blue contrast. Default values are @code{2},
23602 @code{7} and @code{1}.
23603 Allowed range is @code{[0, 255]}.
23606 Set color which will be used for drawing median phase. If color is
23607 @code{none} which is default, no median phase value will be drawn.
23610 Enable video output. Default is enabled.
23613 @subsection phasing detection
23615 The filter also detects out of phase and mono sequences in stereo streams.
23616 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
23618 The filter accepts the following options for this detection:
23622 Enable mono and out of phase detection. Default is disabled.
23625 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
23626 Allowed range is @code{[0, 1]}.
23629 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
23630 Allowed range is @code{[90, 180]}.
23633 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
23636 @subsection Examples
23640 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
23642 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
23646 @section avectorscope
23648 Convert input audio to a video output, representing the audio vector
23651 The filter is used to measure the difference between channels of stereo
23652 audio stream. A monaural signal, consisting of identical left and right
23653 signal, results in straight vertical line. Any stereo separation is visible
23654 as a deviation from this line, creating a Lissajous figure.
23655 If the straight (or deviation from it) but horizontal line appears this
23656 indicates that the left and right channels are out of phase.
23658 The filter accepts the following options:
23662 Set the vectorscope mode.
23664 Available values are:
23667 Lissajous rotated by 45 degrees.
23670 Same as above but not rotated.
23673 Shape resembling half of circle.
23676 Default value is @samp{lissajous}.
23679 Set the video size for the output. For the syntax of this option, check the
23680 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23681 Default value is @code{400x400}.
23684 Set the output frame rate. Default value is @code{25}.
23690 Specify the red, green, blue and alpha contrast. Default values are @code{40},
23691 @code{160}, @code{80} and @code{255}.
23692 Allowed range is @code{[0, 255]}.
23698 Specify the red, green, blue and alpha fade. Default values are @code{15},
23699 @code{10}, @code{5} and @code{5}.
23700 Allowed range is @code{[0, 255]}.
23703 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
23704 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
23707 Set the vectorscope drawing mode.
23709 Available values are:
23712 Draw dot for each sample.
23715 Draw line between previous and current sample.
23718 Default value is @samp{dot}.
23721 Specify amplitude scale of audio samples.
23723 Available values are:
23739 Swap left channel axis with right channel axis.
23749 Mirror only x axis.
23752 Mirror only y axis.
23760 @subsection Examples
23764 Complete example using @command{ffplay}:
23766 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
23767 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
23771 @section bench, abench
23773 Benchmark part of a filtergraph.
23775 The filter accepts the following options:
23779 Start or stop a timer.
23781 Available values are:
23784 Get the current time, set it as frame metadata (using the key
23785 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
23788 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
23789 the input frame metadata to get the time difference. Time difference, average,
23790 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
23791 @code{min}) are then printed. The timestamps are expressed in seconds.
23795 @subsection Examples
23799 Benchmark @ref{selectivecolor} filter:
23801 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
23807 Concatenate audio and video streams, joining them together one after the
23810 The filter works on segments of synchronized video and audio streams. All
23811 segments must have the same number of streams of each type, and that will
23812 also be the number of streams at output.
23814 The filter accepts the following options:
23819 Set the number of segments. Default is 2.
23822 Set the number of output video streams, that is also the number of video
23823 streams in each segment. Default is 1.
23826 Set the number of output audio streams, that is also the number of audio
23827 streams in each segment. Default is 0.
23830 Activate unsafe mode: do not fail if segments have a different format.
23834 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
23835 @var{a} audio outputs.
23837 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
23838 segment, in the same order as the outputs, then the inputs for the second
23841 Related streams do not always have exactly the same duration, for various
23842 reasons including codec frame size or sloppy authoring. For that reason,
23843 related synchronized streams (e.g. a video and its audio track) should be
23844 concatenated at once. The concat filter will use the duration of the longest
23845 stream in each segment (except the last one), and if necessary pad shorter
23846 audio streams with silence.
23848 For this filter to work correctly, all segments must start at timestamp 0.
23850 All corresponding streams must have the same parameters in all segments; the
23851 filtering system will automatically select a common pixel format for video
23852 streams, and a common sample format, sample rate and channel layout for
23853 audio streams, but other settings, such as resolution, must be converted
23854 explicitly by the user.
23856 Different frame rates are acceptable but will result in variable frame rate
23857 at output; be sure to configure the output file to handle it.
23859 @subsection Examples
23863 Concatenate an opening, an episode and an ending, all in bilingual version
23864 (video in stream 0, audio in streams 1 and 2):
23866 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
23867 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
23868 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
23869 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
23873 Concatenate two parts, handling audio and video separately, using the
23874 (a)movie sources, and adjusting the resolution:
23876 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
23877 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
23878 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
23880 Note that a desync will happen at the stitch if the audio and video streams
23881 do not have exactly the same duration in the first file.
23885 @subsection Commands
23887 This filter supports the following commands:
23890 Close the current segment and step to the next one
23896 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
23897 level. By default, it logs a message at a frequency of 10Hz with the
23898 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
23899 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
23901 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
23902 sample format is double-precision floating point. The input stream will be converted to
23903 this specification, if needed. Users may need to insert aformat and/or aresample filters
23904 after this filter to obtain the original parameters.
23906 The filter also has a video output (see the @var{video} option) with a real
23907 time graph to observe the loudness evolution. The graphic contains the logged
23908 message mentioned above, so it is not printed anymore when this option is set,
23909 unless the verbose logging is set. The main graphing area contains the
23910 short-term loudness (3 seconds of analysis), and the gauge on the right is for
23911 the momentary loudness (400 milliseconds), but can optionally be configured
23912 to instead display short-term loudness (see @var{gauge}).
23914 The green area marks a +/- 1LU target range around the target loudness
23915 (-23LUFS by default, unless modified through @var{target}).
23917 More information about the Loudness Recommendation EBU R128 on
23918 @url{http://tech.ebu.ch/loudness}.
23920 The filter accepts the following options:
23925 Activate the video output. The audio stream is passed unchanged whether this
23926 option is set or no. The video stream will be the first output stream if
23927 activated. Default is @code{0}.
23930 Set the video size. This option is for video only. For the syntax of this
23932 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23933 Default and minimum resolution is @code{640x480}.
23936 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
23937 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
23938 other integer value between this range is allowed.
23941 Set metadata injection. If set to @code{1}, the audio input will be segmented
23942 into 100ms output frames, each of them containing various loudness information
23943 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
23945 Default is @code{0}.
23948 Force the frame logging level.
23950 Available values are:
23953 information logging level
23955 verbose logging level
23958 By default, the logging level is set to @var{info}. If the @option{video} or
23959 the @option{metadata} options are set, it switches to @var{verbose}.
23964 Available modes can be cumulated (the option is a @code{flag} type). Possible
23968 Disable any peak mode (default).
23970 Enable sample-peak mode.
23972 Simple peak mode looking for the higher sample value. It logs a message
23973 for sample-peak (identified by @code{SPK}).
23975 Enable true-peak mode.
23977 If enabled, the peak lookup is done on an over-sampled version of the input
23978 stream for better peak accuracy. It logs a message for true-peak.
23979 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
23980 This mode requires a build with @code{libswresample}.
23984 Treat mono input files as "dual mono". If a mono file is intended for playback
23985 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
23986 If set to @code{true}, this option will compensate for this effect.
23987 Multi-channel input files are not affected by this option.
23990 Set a specific pan law to be used for the measurement of dual mono files.
23991 This parameter is optional, and has a default value of -3.01dB.
23994 Set a specific target level (in LUFS) used as relative zero in the visualization.
23995 This parameter is optional and has a default value of -23LUFS as specified
23996 by EBU R128. However, material published online may prefer a level of -16LUFS
23997 (e.g. for use with podcasts or video platforms).
24000 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24001 @code{shortterm}. By default the momentary value will be used, but in certain
24002 scenarios it may be more useful to observe the short term value instead (e.g.
24006 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24007 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24008 video output, not the summary or continuous log output.
24011 @subsection Examples
24015 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24017 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24021 Run an analysis with @command{ffmpeg}:
24023 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24027 @section interleave, ainterleave
24029 Temporally interleave frames from several inputs.
24031 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24033 These filters read frames from several inputs and send the oldest
24034 queued frame to the output.
24036 Input streams must have well defined, monotonically increasing frame
24039 In order to submit one frame to output, these filters need to enqueue
24040 at least one frame for each input, so they cannot work in case one
24041 input is not yet terminated and will not receive incoming frames.
24043 For example consider the case when one input is a @code{select} filter
24044 which always drops input frames. The @code{interleave} filter will keep
24045 reading from that input, but it will never be able to send new frames
24046 to output until the input sends an end-of-stream signal.
24048 Also, depending on inputs synchronization, the filters will drop
24049 frames in case one input receives more frames than the other ones, and
24050 the queue is already filled.
24052 These filters accept the following options:
24056 Set the number of different inputs, it is 2 by default.
24059 How to determine the end-of-stream.
24063 The duration of the longest input. (default)
24066 The duration of the shortest input.
24069 The duration of the first input.
24074 @subsection Examples
24078 Interleave frames belonging to different streams using @command{ffmpeg}:
24080 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24084 Add flickering blur effect:
24086 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24090 @section metadata, ametadata
24092 Manipulate frame metadata.
24094 This filter accepts the following options:
24098 Set mode of operation of the filter.
24100 Can be one of the following:
24104 If both @code{value} and @code{key} is set, select frames
24105 which have such metadata. If only @code{key} is set, select
24106 every frame that has such key in metadata.
24109 Add new metadata @code{key} and @code{value}. If key is already available
24113 Modify value of already present key.
24116 If @code{value} is set, delete only keys that have such value.
24117 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24121 Print key and its value if metadata was found. If @code{key} is not set print all
24122 metadata values available in frame.
24126 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24129 Set metadata value which will be used. This option is mandatory for
24130 @code{modify} and @code{add} mode.
24133 Which function to use when comparing metadata value and @code{value}.
24135 Can be one of following:
24139 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24142 Values are interpreted as strings, returns true if metadata value starts with
24143 the @code{value} option string.
24146 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24149 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24152 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24155 Values are interpreted as floats, returns true if expression from option @code{expr}
24159 Values are interpreted as strings, returns true if metadata value ends with
24160 the @code{value} option string.
24164 Set expression which is used when @code{function} is set to @code{expr}.
24165 The expression is evaluated through the eval API and can contain the following
24170 Float representation of @code{value} from metadata key.
24173 Float representation of @code{value} as supplied by user in @code{value} option.
24177 If specified in @code{print} mode, output is written to the named file. Instead of
24178 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24179 for standard output. If @code{file} option is not set, output is written to the log
24180 with AV_LOG_INFO loglevel.
24183 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24187 @subsection Examples
24191 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24194 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24197 Print silencedetect output to file @file{metadata.txt}.
24199 silencedetect,ametadata=mode=print:file=metadata.txt
24202 Direct all metadata to a pipe with file descriptor 4.
24204 metadata=mode=print:file='pipe\:4'
24208 @section perms, aperms
24210 Set read/write permissions for the output frames.
24212 These filters are mainly aimed at developers to test direct path in the
24213 following filter in the filtergraph.
24215 The filters accept the following options:
24219 Select the permissions mode.
24221 It accepts the following values:
24224 Do nothing. This is the default.
24226 Set all the output frames read-only.
24228 Set all the output frames directly writable.
24230 Make the frame read-only if writable, and writable if read-only.
24232 Set each output frame read-only or writable randomly.
24236 Set the seed for the @var{random} mode, must be an integer included between
24237 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24238 @code{-1}, the filter will try to use a good random seed on a best effort
24242 Note: in case of auto-inserted filter between the permission filter and the
24243 following one, the permission might not be received as expected in that
24244 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24245 perms/aperms filter can avoid this problem.
24247 @section realtime, arealtime
24249 Slow down filtering to match real time approximately.
24251 These filters will pause the filtering for a variable amount of time to
24252 match the output rate with the input timestamps.
24253 They are similar to the @option{re} option to @code{ffmpeg}.
24255 They accept the following options:
24259 Time limit for the pauses. Any pause longer than that will be considered
24260 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24262 Speed factor for processing. The value must be a float larger than zero.
24263 Values larger than 1.0 will result in faster than realtime processing,
24264 smaller will slow processing down. The @var{limit} is automatically adapted
24265 accordingly. Default is 1.0.
24267 A processing speed faster than what is possible without these filters cannot
24272 @section select, aselect
24274 Select frames to pass in output.
24276 This filter accepts the following options:
24281 Set expression, which is evaluated for each input frame.
24283 If the expression is evaluated to zero, the frame is discarded.
24285 If the evaluation result is negative or NaN, the frame is sent to the
24286 first output; otherwise it is sent to the output with index
24287 @code{ceil(val)-1}, assuming that the input index starts from 0.
24289 For example a value of @code{1.2} corresponds to the output with index
24290 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24293 Set the number of outputs. The output to which to send the selected
24294 frame is based on the result of the evaluation. Default value is 1.
24297 The expression can contain the following constants:
24301 The (sequential) number of the filtered frame, starting from 0.
24304 The (sequential) number of the selected frame, starting from 0.
24306 @item prev_selected_n
24307 The sequential number of the last selected frame. It's NAN if undefined.
24310 The timebase of the input timestamps.
24313 The PTS (Presentation TimeStamp) of the filtered video frame,
24314 expressed in @var{TB} units. It's NAN if undefined.
24317 The PTS of the filtered video frame,
24318 expressed in seconds. It's NAN if undefined.
24321 The PTS of the previously filtered video frame. It's NAN if undefined.
24323 @item prev_selected_pts
24324 The PTS of the last previously filtered video frame. It's NAN if undefined.
24326 @item prev_selected_t
24327 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24330 The PTS of the first video frame in the video. It's NAN if undefined.
24333 The time of the first video frame in the video. It's NAN if undefined.
24335 @item pict_type @emph{(video only)}
24336 The type of the filtered frame. It can assume one of the following
24348 @item interlace_type @emph{(video only)}
24349 The frame interlace type. It can assume one of the following values:
24352 The frame is progressive (not interlaced).
24354 The frame is top-field-first.
24356 The frame is bottom-field-first.
24359 @item consumed_sample_n @emph{(audio only)}
24360 the number of selected samples before the current frame
24362 @item samples_n @emph{(audio only)}
24363 the number of samples in the current frame
24365 @item sample_rate @emph{(audio only)}
24366 the input sample rate
24369 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24372 the position in the file of the filtered frame, -1 if the information
24373 is not available (e.g. for synthetic video)
24375 @item scene @emph{(video only)}
24376 value between 0 and 1 to indicate a new scene; a low value reflects a low
24377 probability for the current frame to introduce a new scene, while a higher
24378 value means the current frame is more likely to be one (see the example below)
24380 @item concatdec_select
24381 The concat demuxer can select only part of a concat input file by setting an
24382 inpoint and an outpoint, but the output packets may not be entirely contained
24383 in the selected interval. By using this variable, it is possible to skip frames
24384 generated by the concat demuxer which are not exactly contained in the selected
24387 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24388 and the @var{lavf.concat.duration} packet metadata values which are also
24389 present in the decoded frames.
24391 The @var{concatdec_select} variable is -1 if the frame pts is at least
24392 start_time and either the duration metadata is missing or the frame pts is less
24393 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24396 That basically means that an input frame is selected if its pts is within the
24397 interval set by the concat demuxer.
24401 The default value of the select expression is "1".
24403 @subsection Examples
24407 Select all frames in input:
24412 The example above is the same as:
24424 Select only I-frames:
24426 select='eq(pict_type\,I)'
24430 Select one frame every 100:
24432 select='not(mod(n\,100))'
24436 Select only frames contained in the 10-20 time interval:
24438 select=between(t\,10\,20)
24442 Select only I-frames contained in the 10-20 time interval:
24444 select=between(t\,10\,20)*eq(pict_type\,I)
24448 Select frames with a minimum distance of 10 seconds:
24450 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24454 Use aselect to select only audio frames with samples number > 100:
24456 aselect='gt(samples_n\,100)'
24460 Create a mosaic of the first scenes:
24462 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24465 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24469 Send even and odd frames to separate outputs, and compose them:
24471 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24475 Select useful frames from an ffconcat file which is using inpoints and
24476 outpoints but where the source files are not intra frame only.
24478 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24482 @section sendcmd, asendcmd
24484 Send commands to filters in the filtergraph.
24486 These filters read commands to be sent to other filters in the
24489 @code{sendcmd} must be inserted between two video filters,
24490 @code{asendcmd} must be inserted between two audio filters, but apart
24491 from that they act the same way.
24493 The specification of commands can be provided in the filter arguments
24494 with the @var{commands} option, or in a file specified by the
24495 @var{filename} option.
24497 These filters accept the following options:
24500 Set the commands to be read and sent to the other filters.
24502 Set the filename of the commands to be read and sent to the other
24506 @subsection Commands syntax
24508 A commands description consists of a sequence of interval
24509 specifications, comprising a list of commands to be executed when a
24510 particular event related to that interval occurs. The occurring event
24511 is typically the current frame time entering or leaving a given time
24514 An interval is specified by the following syntax:
24516 @var{START}[-@var{END}] @var{COMMANDS};
24519 The time interval is specified by the @var{START} and @var{END} times.
24520 @var{END} is optional and defaults to the maximum time.
24522 The current frame time is considered within the specified interval if
24523 it is included in the interval [@var{START}, @var{END}), that is when
24524 the time is greater or equal to @var{START} and is lesser than
24527 @var{COMMANDS} consists of a sequence of one or more command
24528 specifications, separated by ",", relating to that interval. The
24529 syntax of a command specification is given by:
24531 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
24534 @var{FLAGS} is optional and specifies the type of events relating to
24535 the time interval which enable sending the specified command, and must
24536 be a non-null sequence of identifier flags separated by "+" or "|" and
24537 enclosed between "[" and "]".
24539 The following flags are recognized:
24542 The command is sent when the current frame timestamp enters the
24543 specified interval. In other words, the command is sent when the
24544 previous frame timestamp was not in the given interval, and the
24548 The command is sent when the current frame timestamp leaves the
24549 specified interval. In other words, the command is sent when the
24550 previous frame timestamp was in the given interval, and the
24554 The command @var{ARG} is interpreted as expression and result of
24555 expression is passed as @var{ARG}.
24557 The expression is evaluated through the eval API and can contain the following
24562 Original position in the file of the frame, or undefined if undefined
24563 for the current frame.
24566 The presentation timestamp in input.
24569 The count of the input frame for video or audio, starting from 0.
24572 The time in seconds of the current frame.
24575 The start time in seconds of the current command interval.
24578 The end time in seconds of the current command interval.
24581 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
24586 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
24589 @var{TARGET} specifies the target of the command, usually the name of
24590 the filter class or a specific filter instance name.
24592 @var{COMMAND} specifies the name of the command for the target filter.
24594 @var{ARG} is optional and specifies the optional list of argument for
24595 the given @var{COMMAND}.
24597 Between one interval specification and another, whitespaces, or
24598 sequences of characters starting with @code{#} until the end of line,
24599 are ignored and can be used to annotate comments.
24601 A simplified BNF description of the commands specification syntax
24604 @var{COMMAND_FLAG} ::= "enter" | "leave"
24605 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
24606 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
24607 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
24608 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
24609 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
24612 @subsection Examples
24616 Specify audio tempo change at second 4:
24618 asendcmd=c='4.0 atempo tempo 1.5',atempo
24622 Target a specific filter instance:
24624 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
24628 Specify a list of drawtext and hue commands in a file.
24630 # show text in the interval 5-10
24631 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
24632 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
24634 # desaturate the image in the interval 15-20
24635 15.0-20.0 [enter] hue s 0,
24636 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
24638 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
24640 # apply an exponential saturation fade-out effect, starting from time 25
24641 25 [enter] hue s exp(25-t)
24644 A filtergraph allowing to read and process the above command list
24645 stored in a file @file{test.cmd}, can be specified with:
24647 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
24652 @section setpts, asetpts
24654 Change the PTS (presentation timestamp) of the input frames.
24656 @code{setpts} works on video frames, @code{asetpts} on audio frames.
24658 This filter accepts the following options:
24663 The expression which is evaluated for each frame to construct its timestamp.
24667 The expression is evaluated through the eval API and can contain the following
24671 @item FRAME_RATE, FR
24672 frame rate, only defined for constant frame-rate video
24675 The presentation timestamp in input
24678 The count of the input frame for video or the number of consumed samples,
24679 not including the current frame for audio, starting from 0.
24681 @item NB_CONSUMED_SAMPLES
24682 The number of consumed samples, not including the current frame (only
24685 @item NB_SAMPLES, S
24686 The number of samples in the current frame (only audio)
24688 @item SAMPLE_RATE, SR
24689 The audio sample rate.
24692 The PTS of the first frame.
24695 the time in seconds of the first frame
24698 State whether the current frame is interlaced.
24701 the time in seconds of the current frame
24704 original position in the file of the frame, or undefined if undefined
24705 for the current frame
24708 The previous input PTS.
24711 previous input time in seconds
24714 The previous output PTS.
24717 previous output time in seconds
24720 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
24724 The wallclock (RTC) time at the start of the movie in microseconds.
24727 The timebase of the input timestamps.
24731 @subsection Examples
24735 Start counting PTS from zero
24737 setpts=PTS-STARTPTS
24741 Apply fast motion effect:
24747 Apply slow motion effect:
24753 Set fixed rate of 25 frames per second:
24759 Set fixed rate 25 fps with some jitter:
24761 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
24765 Apply an offset of 10 seconds to the input PTS:
24771 Generate timestamps from a "live source" and rebase onto the current timebase:
24773 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
24777 Generate timestamps by counting samples:
24786 Force color range for the output video frame.
24788 The @code{setrange} filter marks the color range property for the
24789 output frames. It does not change the input frame, but only sets the
24790 corresponding property, which affects how the frame is treated by
24793 The filter accepts the following options:
24798 Available values are:
24802 Keep the same color range property.
24804 @item unspecified, unknown
24805 Set the color range as unspecified.
24807 @item limited, tv, mpeg
24808 Set the color range as limited.
24810 @item full, pc, jpeg
24811 Set the color range as full.
24815 @section settb, asettb
24817 Set the timebase to use for the output frames timestamps.
24818 It is mainly useful for testing timebase configuration.
24820 It accepts the following parameters:
24825 The expression which is evaluated into the output timebase.
24829 The value for @option{tb} is an arithmetic expression representing a
24830 rational. The expression can contain the constants "AVTB" (the default
24831 timebase), "intb" (the input timebase) and "sr" (the sample rate,
24832 audio only). Default value is "intb".
24834 @subsection Examples
24838 Set the timebase to 1/25:
24844 Set the timebase to 1/10:
24850 Set the timebase to 1001/1000:
24856 Set the timebase to 2*intb:
24862 Set the default timebase value:
24869 Convert input audio to a video output representing frequency spectrum
24870 logarithmically using Brown-Puckette constant Q transform algorithm with
24871 direct frequency domain coefficient calculation (but the transform itself
24872 is not really constant Q, instead the Q factor is actually variable/clamped),
24873 with musical tone scale, from E0 to D#10.
24875 The filter accepts the following options:
24879 Specify the video size for the output. It must be even. For the syntax of this option,
24880 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24881 Default value is @code{1920x1080}.
24884 Set the output frame rate. Default value is @code{25}.
24887 Set the bargraph height. It must be even. Default value is @code{-1} which
24888 computes the bargraph height automatically.
24891 Set the axis height. It must be even. Default value is @code{-1} which computes
24892 the axis height automatically.
24895 Set the sonogram height. It must be even. Default value is @code{-1} which
24896 computes the sonogram height automatically.
24899 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
24900 instead. Default value is @code{1}.
24902 @item sono_v, volume
24903 Specify the sonogram volume expression. It can contain variables:
24906 the @var{bar_v} evaluated expression
24907 @item frequency, freq, f
24908 the frequency where it is evaluated
24909 @item timeclamp, tc
24910 the value of @var{timeclamp} option
24914 @item a_weighting(f)
24915 A-weighting of equal loudness
24916 @item b_weighting(f)
24917 B-weighting of equal loudness
24918 @item c_weighting(f)
24919 C-weighting of equal loudness.
24921 Default value is @code{16}.
24923 @item bar_v, volume2
24924 Specify the bargraph volume expression. It can contain variables:
24927 the @var{sono_v} evaluated expression
24928 @item frequency, freq, f
24929 the frequency where it is evaluated
24930 @item timeclamp, tc
24931 the value of @var{timeclamp} option
24935 @item a_weighting(f)
24936 A-weighting of equal loudness
24937 @item b_weighting(f)
24938 B-weighting of equal loudness
24939 @item c_weighting(f)
24940 C-weighting of equal loudness.
24942 Default value is @code{sono_v}.
24944 @item sono_g, gamma
24945 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
24946 higher gamma makes the spectrum having more range. Default value is @code{3}.
24947 Acceptable range is @code{[1, 7]}.
24949 @item bar_g, gamma2
24950 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
24954 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
24955 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
24957 @item timeclamp, tc
24958 Specify the transform timeclamp. At low frequency, there is trade-off between
24959 accuracy in time domain and frequency domain. If timeclamp is lower,
24960 event in time domain is represented more accurately (such as fast bass drum),
24961 otherwise event in frequency domain is represented more accurately
24962 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
24965 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
24966 limits future samples by applying asymmetric windowing in time domain, useful
24967 when low latency is required. Accepted range is @code{[0, 1]}.
24970 Specify the transform base frequency. Default value is @code{20.01523126408007475},
24971 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
24974 Specify the transform end frequency. Default value is @code{20495.59681441799654},
24975 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
24978 This option is deprecated and ignored.
24981 Specify the transform length in time domain. Use this option to control accuracy
24982 trade-off between time domain and frequency domain at every frequency sample.
24983 It can contain variables:
24985 @item frequency, freq, f
24986 the frequency where it is evaluated
24987 @item timeclamp, tc
24988 the value of @var{timeclamp} option.
24990 Default value is @code{384*tc/(384+tc*f)}.
24993 Specify the transform count for every video frame. Default value is @code{6}.
24994 Acceptable range is @code{[1, 30]}.
24997 Specify the transform count for every single pixel. Default value is @code{0},
24998 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25001 Specify font file for use with freetype to draw the axis. If not specified,
25002 use embedded font. Note that drawing with font file or embedded font is not
25003 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25007 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25008 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25012 Specify font color expression. This is arithmetic expression that should return
25013 integer value 0xRRGGBB. It can contain variables:
25015 @item frequency, freq, f
25016 the frequency where it is evaluated
25017 @item timeclamp, tc
25018 the value of @var{timeclamp} option
25023 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25024 @item r(x), g(x), b(x)
25025 red, green, and blue value of intensity x.
25027 Default value is @code{st(0, (midi(f)-59.5)/12);
25028 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25029 r(1-ld(1)) + b(ld(1))}.
25032 Specify image file to draw the axis. This option override @var{fontfile} and
25033 @var{fontcolor} option.
25036 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25037 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25038 Default value is @code{1}.
25041 Set colorspace. The accepted values are:
25044 Unspecified (default)
25053 BT.470BG or BT.601-6 625
25056 SMPTE-170M or BT.601-6 525
25062 BT.2020 with non-constant luminance
25067 Set spectrogram color scheme. This is list of floating point values with format
25068 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25069 The default is @code{1|0.5|0|0|0.5|1}.
25073 @subsection Examples
25077 Playing audio while showing the spectrum:
25079 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25083 Same as above, but with frame rate 30 fps:
25085 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25089 Playing at 1280x720:
25091 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25095 Disable sonogram display:
25101 A1 and its harmonics: A1, A2, (near)E3, A3:
25103 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),
25104 asplit[a][out1]; [a] showcqt [out0]'
25108 Same as above, but with more accuracy in frequency domain:
25110 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),
25111 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25117 bar_v=10:sono_v=bar_v*a_weighting(f)
25121 Custom gamma, now spectrum is linear to the amplitude.
25127 Custom tlength equation:
25129 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)))'
25133 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25135 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25139 Custom font using fontconfig:
25141 font='Courier New,Monospace,mono|bold'
25145 Custom frequency range with custom axis using image file:
25147 axisfile=myaxis.png:basefreq=40:endfreq=10000
25153 Convert input audio to video output representing the audio power spectrum.
25154 Audio amplitude is on Y-axis while frequency is on X-axis.
25156 The filter accepts the following options:
25160 Specify size of video. For the syntax of this option, check the
25161 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25162 Default is @code{1024x512}.
25166 This set how each frequency bin will be represented.
25168 It accepts the following values:
25174 Default is @code{bar}.
25177 Set amplitude scale.
25179 It accepts the following values:
25193 Default is @code{log}.
25196 Set frequency scale.
25198 It accepts the following values:
25207 Reverse logarithmic scale.
25209 Default is @code{lin}.
25212 Set window size. Allowed range is from 16 to 65536.
25214 Default is @code{2048}
25217 Set windowing function.
25219 It accepts the following values:
25242 Default is @code{hanning}.
25245 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25246 which means optimal overlap for selected window function will be picked.
25249 Set time averaging. Setting this to 0 will display current maximal peaks.
25250 Default is @code{1}, which means time averaging is disabled.
25253 Specify list of colors separated by space or by '|' which will be used to
25254 draw channel frequencies. Unrecognized or missing colors will be replaced
25258 Set channel display mode.
25260 It accepts the following values:
25265 Default is @code{combined}.
25268 Set minimum amplitude used in @code{log} amplitude scaler.
25271 Set data display mode.
25273 It accepts the following values:
25279 Default is @code{magnitude}.
25282 @section showspatial
25284 Convert stereo input audio to a video output, representing the spatial relationship
25285 between two channels.
25287 The filter accepts the following options:
25291 Specify the video size for the output. For the syntax of this option, check the
25292 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25293 Default value is @code{512x512}.
25296 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25299 Set window function.
25301 It accepts the following values:
25326 Default value is @code{hann}.
25329 Set ratio of overlap window. Default value is @code{0.5}.
25330 When value is @code{1} overlap is set to recommended size for specific
25331 window function currently used.
25334 @anchor{showspectrum}
25335 @section showspectrum
25337 Convert input audio to a video output, representing the audio frequency
25340 The filter accepts the following options:
25344 Specify the video size for the output. For the syntax of this option, check the
25345 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25346 Default value is @code{640x512}.
25349 Specify how the spectrum should slide along the window.
25351 It accepts the following values:
25354 the samples start again on the left when they reach the right
25356 the samples scroll from right to left
25358 frames are only produced when the samples reach the right
25360 the samples scroll from left to right
25363 Default value is @code{replace}.
25366 Specify display mode.
25368 It accepts the following values:
25371 all channels are displayed in the same row
25373 all channels are displayed in separate rows
25376 Default value is @samp{combined}.
25379 Specify display color mode.
25381 It accepts the following values:
25384 each channel is displayed in a separate color
25386 each channel is displayed using the same color scheme
25388 each channel is displayed using the rainbow color scheme
25390 each channel is displayed using the moreland color scheme
25392 each channel is displayed using the nebulae color scheme
25394 each channel is displayed using the fire color scheme
25396 each channel is displayed using the fiery color scheme
25398 each channel is displayed using the fruit color scheme
25400 each channel is displayed using the cool color scheme
25402 each channel is displayed using the magma color scheme
25404 each channel is displayed using the green color scheme
25406 each channel is displayed using the viridis color scheme
25408 each channel is displayed using the plasma color scheme
25410 each channel is displayed using the cividis color scheme
25412 each channel is displayed using the terrain color scheme
25415 Default value is @samp{channel}.
25418 Specify scale used for calculating intensity color values.
25420 It accepts the following values:
25425 square root, default
25436 Default value is @samp{sqrt}.
25439 Specify frequency scale.
25441 It accepts the following values:
25449 Default value is @samp{lin}.
25452 Set saturation modifier for displayed colors. Negative values provide
25453 alternative color scheme. @code{0} is no saturation at all.
25454 Saturation must be in [-10.0, 10.0] range.
25455 Default value is @code{1}.
25458 Set window function.
25460 It accepts the following values:
25485 Default value is @code{hann}.
25488 Set orientation of time vs frequency axis. Can be @code{vertical} or
25489 @code{horizontal}. Default is @code{vertical}.
25492 Set ratio of overlap window. Default value is @code{0}.
25493 When value is @code{1} overlap is set to recommended size for specific
25494 window function currently used.
25497 Set scale gain for calculating intensity color values.
25498 Default value is @code{1}.
25501 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25504 Set color rotation, must be in [-1.0, 1.0] range.
25505 Default value is @code{0}.
25508 Set start frequency from which to display spectrogram. Default is @code{0}.
25511 Set stop frequency to which to display spectrogram. Default is @code{0}.
25514 Set upper frame rate limit. Default is @code{auto}, unlimited.
25517 Draw time and frequency axes and legends. Default is disabled.
25520 The usage is very similar to the showwaves filter; see the examples in that
25523 @subsection Examples
25527 Large window with logarithmic color scaling:
25529 showspectrum=s=1280x480:scale=log
25533 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
25535 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
25536 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
25540 @section showspectrumpic
25542 Convert input audio to a single video frame, representing the audio frequency
25545 The filter accepts the following options:
25549 Specify the video size for the output. For the syntax of this option, check the
25550 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25551 Default value is @code{4096x2048}.
25554 Specify display mode.
25556 It accepts the following values:
25559 all channels are displayed in the same row
25561 all channels are displayed in separate rows
25563 Default value is @samp{combined}.
25566 Specify display color mode.
25568 It accepts the following values:
25571 each channel is displayed in a separate color
25573 each channel is displayed using the same color scheme
25575 each channel is displayed using the rainbow color scheme
25577 each channel is displayed using the moreland color scheme
25579 each channel is displayed using the nebulae color scheme
25581 each channel is displayed using the fire color scheme
25583 each channel is displayed using the fiery color scheme
25585 each channel is displayed using the fruit color scheme
25587 each channel is displayed using the cool color scheme
25589 each channel is displayed using the magma color scheme
25591 each channel is displayed using the green color scheme
25593 each channel is displayed using the viridis color scheme
25595 each channel is displayed using the plasma color scheme
25597 each channel is displayed using the cividis color scheme
25599 each channel is displayed using the terrain color scheme
25601 Default value is @samp{intensity}.
25604 Specify scale used for calculating intensity color values.
25606 It accepts the following values:
25611 square root, default
25621 Default value is @samp{log}.
25624 Specify frequency scale.
25626 It accepts the following values:
25634 Default value is @samp{lin}.
25637 Set saturation modifier for displayed colors. Negative values provide
25638 alternative color scheme. @code{0} is no saturation at all.
25639 Saturation must be in [-10.0, 10.0] range.
25640 Default value is @code{1}.
25643 Set window function.
25645 It accepts the following values:
25669 Default value is @code{hann}.
25672 Set orientation of time vs frequency axis. Can be @code{vertical} or
25673 @code{horizontal}. Default is @code{vertical}.
25676 Set scale gain for calculating intensity color values.
25677 Default value is @code{1}.
25680 Draw time and frequency axes and legends. Default is enabled.
25683 Set color rotation, must be in [-1.0, 1.0] range.
25684 Default value is @code{0}.
25687 Set start frequency from which to display spectrogram. Default is @code{0}.
25690 Set stop frequency to which to display spectrogram. Default is @code{0}.
25693 @subsection Examples
25697 Extract an audio spectrogram of a whole audio track
25698 in a 1024x1024 picture using @command{ffmpeg}:
25700 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
25704 @section showvolume
25706 Convert input audio volume to a video output.
25708 The filter accepts the following options:
25715 Set border width, allowed range is [0, 5]. Default is 1.
25718 Set channel width, allowed range is [80, 8192]. Default is 400.
25721 Set channel height, allowed range is [1, 900]. Default is 20.
25724 Set fade, allowed range is [0, 1]. Default is 0.95.
25727 Set volume color expression.
25729 The expression can use the following variables:
25733 Current max volume of channel in dB.
25739 Current channel number, starting from 0.
25743 If set, displays channel names. Default is enabled.
25746 If set, displays volume values. Default is enabled.
25749 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
25750 default is @code{h}.
25753 Set step size, allowed range is [0, 5]. Default is 0, which means
25757 Set background opacity, allowed range is [0, 1]. Default is 0.
25760 Set metering mode, can be peak: @code{p} or rms: @code{r},
25761 default is @code{p}.
25764 Set display scale, can be linear: @code{lin} or log: @code{log},
25765 default is @code{lin}.
25769 If set to > 0., display a line for the max level
25770 in the previous seconds.
25771 default is disabled: @code{0.}
25774 The color of the max line. Use when @code{dm} option is set to > 0.
25775 default is: @code{orange}
25780 Convert input audio to a video output, representing the samples waves.
25782 The filter accepts the following options:
25786 Specify the video size for the output. For the syntax of this option, check the
25787 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25788 Default value is @code{600x240}.
25793 Available values are:
25796 Draw a point for each sample.
25799 Draw a vertical line for each sample.
25802 Draw a point for each sample and a line between them.
25805 Draw a centered vertical line for each sample.
25808 Default value is @code{point}.
25811 Set the number of samples which are printed on the same column. A
25812 larger value will decrease the frame rate. Must be a positive
25813 integer. This option can be set only if the value for @var{rate}
25814 is not explicitly specified.
25817 Set the (approximate) output frame rate. This is done by setting the
25818 option @var{n}. Default value is "25".
25820 @item split_channels
25821 Set if channels should be drawn separately or overlap. Default value is 0.
25824 Set colors separated by '|' which are going to be used for drawing of each channel.
25827 Set amplitude scale.
25829 Available values are:
25847 Set the draw mode. This is mostly useful to set for high @var{n}.
25849 Available values are:
25852 Scale pixel values for each drawn sample.
25855 Draw every sample directly.
25858 Default value is @code{scale}.
25861 @subsection Examples
25865 Output the input file audio and the corresponding video representation
25868 amovie=a.mp3,asplit[out0],showwaves[out1]
25872 Create a synthetic signal and show it with showwaves, forcing a
25873 frame rate of 30 frames per second:
25875 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
25879 @section showwavespic
25881 Convert input audio to a single video frame, representing the samples waves.
25883 The filter accepts the following options:
25887 Specify the video size for the output. For the syntax of this option, check the
25888 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25889 Default value is @code{600x240}.
25891 @item split_channels
25892 Set if channels should be drawn separately or overlap. Default value is 0.
25895 Set colors separated by '|' which are going to be used for drawing of each channel.
25898 Set amplitude scale.
25900 Available values are:
25920 Available values are:
25923 Scale pixel values for each drawn sample.
25926 Draw every sample directly.
25929 Default value is @code{scale}.
25932 Set the filter mode.
25934 Available values are:
25937 Use average samples values for each drawn sample.
25940 Use peak samples values for each drawn sample.
25943 Default value is @code{average}.
25946 @subsection Examples
25950 Extract a channel split representation of the wave form of a whole audio track
25951 in a 1024x800 picture using @command{ffmpeg}:
25953 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
25957 @section sidedata, asidedata
25959 Delete frame side data, or select frames based on it.
25961 This filter accepts the following options:
25965 Set mode of operation of the filter.
25967 Can be one of the following:
25971 Select every frame with side data of @code{type}.
25974 Delete side data of @code{type}. If @code{type} is not set, delete all side
25980 Set side data type used with all modes. Must be set for @code{select} mode. For
25981 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
25982 in @file{libavutil/frame.h}. For example, to choose
25983 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
25987 @section spectrumsynth
25989 Synthesize audio from 2 input video spectrums, first input stream represents
25990 magnitude across time and second represents phase across time.
25991 The filter will transform from frequency domain as displayed in videos back
25992 to time domain as presented in audio output.
25994 This filter is primarily created for reversing processed @ref{showspectrum}
25995 filter outputs, but can synthesize sound from other spectrograms too.
25996 But in such case results are going to be poor if the phase data is not
25997 available, because in such cases phase data need to be recreated, usually
25998 it's just recreated from random noise.
25999 For best results use gray only output (@code{channel} color mode in
26000 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26001 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26002 @code{data} option. Inputs videos should generally use @code{fullframe}
26003 slide mode as that saves resources needed for decoding video.
26005 The filter accepts the following options:
26009 Specify sample rate of output audio, the sample rate of audio from which
26010 spectrum was generated may differ.
26013 Set number of channels represented in input video spectrums.
26016 Set scale which was used when generating magnitude input spectrum.
26017 Can be @code{lin} or @code{log}. Default is @code{log}.
26020 Set slide which was used when generating inputs spectrums.
26021 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26022 Default is @code{fullframe}.
26025 Set window function used for resynthesis.
26028 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26029 which means optimal overlap for selected window function will be picked.
26032 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26033 Default is @code{vertical}.
26036 @subsection Examples
26040 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26041 then resynthesize videos back to audio with spectrumsynth:
26043 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
26044 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
26045 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26049 @section split, asplit
26051 Split input into several identical outputs.
26053 @code{asplit} works with audio input, @code{split} with video.
26055 The filter accepts a single parameter which specifies the number of outputs. If
26056 unspecified, it defaults to 2.
26058 @subsection Examples
26062 Create two separate outputs from the same input:
26064 [in] split [out0][out1]
26068 To create 3 or more outputs, you need to specify the number of
26071 [in] asplit=3 [out0][out1][out2]
26075 Create two separate outputs from the same input, one cropped and
26078 [in] split [splitout1][splitout2];
26079 [splitout1] crop=100:100:0:0 [cropout];
26080 [splitout2] pad=200:200:100:100 [padout];
26084 Create 5 copies of the input audio with @command{ffmpeg}:
26086 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26092 Receive commands sent through a libzmq client, and forward them to
26093 filters in the filtergraph.
26095 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26096 must be inserted between two video filters, @code{azmq} between two
26097 audio filters. Both are capable to send messages to any filter type.
26099 To enable these filters you need to install the libzmq library and
26100 headers and configure FFmpeg with @code{--enable-libzmq}.
26102 For more information about libzmq see:
26103 @url{http://www.zeromq.org/}
26105 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26106 receives messages sent through a network interface defined by the
26107 @option{bind_address} (or the abbreviation "@option{b}") option.
26108 Default value of this option is @file{tcp://localhost:5555}. You may
26109 want to alter this value to your needs, but do not forget to escape any
26110 ':' signs (see @ref{filtergraph escaping}).
26112 The received message must be in the form:
26114 @var{TARGET} @var{COMMAND} [@var{ARG}]
26117 @var{TARGET} specifies the target of the command, usually the name of
26118 the filter class or a specific filter instance name. The default
26119 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26120 but you can override this by using the @samp{filter_name@@id} syntax
26121 (see @ref{Filtergraph syntax}).
26123 @var{COMMAND} specifies the name of the command for the target filter.
26125 @var{ARG} is optional and specifies the optional argument list for the
26126 given @var{COMMAND}.
26128 Upon reception, the message is processed and the corresponding command
26129 is injected into the filtergraph. Depending on the result, the filter
26130 will send a reply to the client, adopting the format:
26132 @var{ERROR_CODE} @var{ERROR_REASON}
26136 @var{MESSAGE} is optional.
26138 @subsection Examples
26140 Look at @file{tools/zmqsend} for an example of a zmq client which can
26141 be used to send commands processed by these filters.
26143 Consider the following filtergraph generated by @command{ffplay}.
26144 In this example the last overlay filter has an instance name. All other
26145 filters will have default instance names.
26148 ffplay -dumpgraph 1 -f lavfi "
26149 color=s=100x100:c=red [l];
26150 color=s=100x100:c=blue [r];
26151 nullsrc=s=200x100, zmq [bg];
26152 [bg][l] overlay [bg+l];
26153 [bg+l][r] overlay@@my=x=100 "
26156 To change the color of the left side of the video, the following
26157 command can be used:
26159 echo Parsed_color_0 c yellow | tools/zmqsend
26162 To change the right side:
26164 echo Parsed_color_1 c pink | tools/zmqsend
26167 To change the position of the right side:
26169 echo overlay@@my x 150 | tools/zmqsend
26173 @c man end MULTIMEDIA FILTERS
26175 @chapter Multimedia Sources
26176 @c man begin MULTIMEDIA SOURCES
26178 Below is a description of the currently available multimedia sources.
26182 This is the same as @ref{movie} source, except it selects an audio
26188 Read audio and/or video stream(s) from a movie container.
26190 It accepts the following parameters:
26194 The name of the resource to read (not necessarily a file; it can also be a
26195 device or a stream accessed through some protocol).
26197 @item format_name, f
26198 Specifies the format assumed for the movie to read, and can be either
26199 the name of a container or an input device. If not specified, the
26200 format is guessed from @var{movie_name} or by probing.
26202 @item seek_point, sp
26203 Specifies the seek point in seconds. The frames will be output
26204 starting from this seek point. The parameter is evaluated with
26205 @code{av_strtod}, so the numerical value may be suffixed by an IS
26206 postfix. The default value is "0".
26209 Specifies the streams to read. Several streams can be specified,
26210 separated by "+". The source will then have as many outputs, in the
26211 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26212 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26213 respectively the default (best suited) video and audio stream. Default
26214 is "dv", or "da" if the filter is called as "amovie".
26216 @item stream_index, si
26217 Specifies the index of the video stream to read. If the value is -1,
26218 the most suitable video stream will be automatically selected. The default
26219 value is "-1". Deprecated. If the filter is called "amovie", it will select
26220 audio instead of video.
26223 Specifies how many times to read the stream in sequence.
26224 If the value is 0, the stream will be looped infinitely.
26225 Default value is "1".
26227 Note that when the movie is looped the source timestamps are not
26228 changed, so it will generate non monotonically increasing timestamps.
26230 @item discontinuity
26231 Specifies the time difference between frames above which the point is
26232 considered a timestamp discontinuity which is removed by adjusting the later
26236 It allows overlaying a second video on top of the main input of
26237 a filtergraph, as shown in this graph:
26239 input -----------> deltapts0 --> overlay --> output
26242 movie --> scale--> deltapts1 -------+
26244 @subsection Examples
26248 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26249 on top of the input labelled "in":
26251 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26252 [in] setpts=PTS-STARTPTS [main];
26253 [main][over] overlay=16:16 [out]
26257 Read from a video4linux2 device, and overlay it on top of the input
26260 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26261 [in] setpts=PTS-STARTPTS [main];
26262 [main][over] overlay=16:16 [out]
26266 Read the first video stream and the audio stream with id 0x81 from
26267 dvd.vob; the video is connected to the pad named "video" and the audio is
26268 connected to the pad named "audio":
26270 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26274 @subsection Commands
26276 Both movie and amovie support the following commands:
26279 Perform seek using "av_seek_frame".
26280 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26283 @var{stream_index}: If stream_index is -1, a default
26284 stream is selected, and @var{timestamp} is automatically converted
26285 from AV_TIME_BASE units to the stream specific time_base.
26287 @var{timestamp}: Timestamp in AVStream.time_base units
26288 or, if no stream is specified, in AV_TIME_BASE units.
26290 @var{flags}: Flags which select direction and seeking mode.
26294 Get movie duration in AV_TIME_BASE units.
26298 @c man end MULTIMEDIA SOURCES