1 @chapter Filtering Introduction
2 @c man begin FILTERING INTRODUCTION
4 Filtering in FFmpeg is enabled through the libavfilter library.
6 In libavfilter, a filter can have multiple inputs and multiple
8 To illustrate the sorts of things that are possible, we consider the
13 input --> split ---------------------> overlay --> output
16 +-----> crop --> vflip -------+
19 This filtergraph splits the input stream in two streams, then sends one
20 stream through the crop filter and the vflip filter, before merging it
21 back with the other stream by overlaying it on top. You can use the
22 following command to achieve this:
25 ffmpeg -i INPUT -vf "split [main][tmp]; [tmp] crop=iw:ih/2:0:0, vflip [flip]; [main][flip] overlay=0:H/2" OUTPUT
28 The result will be that the top half of the video is mirrored
29 onto the bottom half of the output video.
31 Filters in the same linear chain are separated by commas, and distinct
32 linear chains of filters are separated by semicolons. In our example,
33 @var{crop,vflip} are in one linear chain, @var{split} and
34 @var{overlay} are separately in another. The points where the linear
35 chains join are labelled by names enclosed in square brackets. In the
36 example, the split filter generates two outputs that are associated to
37 the labels @var{[main]} and @var{[tmp]}.
39 The stream sent to the second output of @var{split}, labelled as
40 @var{[tmp]}, is processed through the @var{crop} filter, which crops
41 away the lower half part of the video, and then vertically flipped. The
42 @var{overlay} filter takes in input the first unchanged output of the
43 split filter (which was labelled as @var{[main]}), and overlay on its
44 lower half the output generated by the @var{crop,vflip} filterchain.
46 Some filters take in input a list of parameters: they are specified
47 after the filter name and an equal sign, and are separated from each other
50 There exist so-called @var{source filters} that do not have an
51 audio/video input, and @var{sink filters} that will not have audio/video
54 @c man end FILTERING INTRODUCTION
57 @c man begin GRAPH2DOT
59 The @file{graph2dot} program included in the FFmpeg @file{tools}
60 directory can be used to parse a filtergraph description and issue a
61 corresponding textual representation in the dot language.
68 to see how to use @file{graph2dot}.
70 You can then pass the dot description to the @file{dot} program (from
71 the graphviz suite of programs) and obtain a graphical representation
74 For example the sequence of commands:
76 echo @var{GRAPH_DESCRIPTION} | \
77 tools/graph2dot -o graph.tmp && \
78 dot -Tpng graph.tmp -o graph.png && \
82 can be used to create and display an image representing the graph
83 described by the @var{GRAPH_DESCRIPTION} string. Note that this string must be
84 a complete self-contained graph, with its inputs and outputs explicitly defined.
85 For example if your command line is of the form:
87 ffmpeg -i infile -vf scale=640:360 outfile
89 your @var{GRAPH_DESCRIPTION} string will need to be of the form:
91 nullsrc,scale=640:360,nullsink
93 you may also need to set the @var{nullsrc} parameters and add a @var{format}
94 filter in order to simulate a specific input file.
98 @chapter Filtergraph description
99 @c man begin FILTERGRAPH DESCRIPTION
101 A filtergraph is a directed graph of connected filters. It can contain
102 cycles, and there can be multiple links between a pair of
103 filters. Each link has one input pad on one side connecting it to one
104 filter from which it takes its input, and one output pad on the other
105 side connecting it to one filter accepting its output.
107 Each filter in a filtergraph is an instance of a filter class
108 registered in the application, which defines the features and the
109 number of input and output pads of the filter.
111 A filter with no input pads is called a "source", and a filter with no
112 output pads is called a "sink".
114 @anchor{Filtergraph syntax}
115 @section Filtergraph syntax
117 A filtergraph has a textual representation, which is recognized by the
118 @option{-filter}/@option{-vf}/@option{-af} and
119 @option{-filter_complex} options in @command{ffmpeg} and
120 @option{-vf}/@option{-af} in @command{ffplay}, and by the
121 @code{avfilter_graph_parse_ptr()} function defined in
122 @file{libavfilter/avfilter.h}.
124 A filterchain consists of a sequence of connected filters, each one
125 connected to the previous one in the sequence. A filterchain is
126 represented by a list of ","-separated filter descriptions.
128 A filtergraph consists of a sequence of filterchains. A sequence of
129 filterchains is represented by a list of ";"-separated filterchain
132 A filter is represented by a string of the form:
133 [@var{in_link_1}]...[@var{in_link_N}]@var{filter_name}@@@var{id}=@var{arguments}[@var{out_link_1}]...[@var{out_link_M}]
135 @var{filter_name} is the name of the filter class of which the
136 described filter is an instance of, and has to be the name of one of
137 the filter classes registered in the program optionally followed by "@@@var{id}".
138 The name of the filter class is optionally followed by a string
141 @var{arguments} is a string which contains the parameters used to
142 initialize the filter instance. It may have one of two forms:
146 A ':'-separated list of @var{key=value} pairs.
149 A ':'-separated list of @var{value}. In this case, the keys are assumed to be
150 the option names in the order they are declared. E.g. the @code{fade} filter
151 declares three options in this order -- @option{type}, @option{start_frame} and
152 @option{nb_frames}. Then the parameter list @var{in:0:30} means that the value
153 @var{in} is assigned to the option @option{type}, @var{0} to
154 @option{start_frame} and @var{30} to @option{nb_frames}.
157 A ':'-separated list of mixed direct @var{value} and long @var{key=value}
158 pairs. The direct @var{value} must precede the @var{key=value} pairs, and
159 follow the same constraints order of the previous point. The following
160 @var{key=value} pairs can be set in any preferred order.
164 If the option value itself is a list of items (e.g. the @code{format} filter
165 takes a list of pixel formats), the items in the list are usually separated by
168 The list of arguments can be quoted using the character @samp{'} as initial
169 and ending mark, and the character @samp{\} for escaping the characters
170 within the quoted text; otherwise the argument string is considered
171 terminated when the next special character (belonging to the set
172 @samp{[]=;,}) is encountered.
174 The name and arguments of the filter are optionally preceded and
175 followed by a list of link labels.
176 A link label allows one to name a link and associate it to a filter output
177 or input pad. The preceding labels @var{in_link_1}
178 ... @var{in_link_N}, are associated to the filter input pads,
179 the following labels @var{out_link_1} ... @var{out_link_M}, are
180 associated to the output pads.
182 When two link labels with the same name are found in the
183 filtergraph, a link between the corresponding input and output pad is
186 If an output pad is not labelled, it is linked by default to the first
187 unlabelled input pad of the next filter in the filterchain.
188 For example in the filterchain
190 nullsrc, split[L1], [L2]overlay, nullsink
192 the split filter instance has two output pads, and the overlay filter
193 instance two input pads. The first output pad of split is labelled
194 "L1", the first input pad of overlay is labelled "L2", and the second
195 output pad of split is linked to the second input pad of overlay,
196 which are both unlabelled.
198 In a filter description, if the input label of the first filter is not
199 specified, "in" is assumed; if the output label of the last filter is not
200 specified, "out" is assumed.
202 In a complete filterchain all the unlabelled filter input and output
203 pads must be connected. A filtergraph is considered valid if all the
204 filter input and output pads of all the filterchains are connected.
206 Libavfilter will automatically insert @ref{scale} filters where format
207 conversion is required. It is possible to specify swscale flags
208 for those automatically inserted scalers by prepending
209 @code{sws_flags=@var{flags};}
210 to the filtergraph description.
212 Here is a BNF description of the filtergraph syntax:
214 @var{NAME} ::= sequence of alphanumeric characters and '_'
215 @var{FILTER_NAME} ::= @var{NAME}["@@"@var{NAME}]
216 @var{LINKLABEL} ::= "[" @var{NAME} "]"
217 @var{LINKLABELS} ::= @var{LINKLABEL} [@var{LINKLABELS}]
218 @var{FILTER_ARGUMENTS} ::= sequence of chars (possibly quoted)
219 @var{FILTER} ::= [@var{LINKLABELS}] @var{FILTER_NAME} ["=" @var{FILTER_ARGUMENTS}] [@var{LINKLABELS}]
220 @var{FILTERCHAIN} ::= @var{FILTER} [,@var{FILTERCHAIN}]
221 @var{FILTERGRAPH} ::= [sws_flags=@var{flags};] @var{FILTERCHAIN} [;@var{FILTERGRAPH}]
224 @anchor{filtergraph escaping}
225 @section Notes on filtergraph escaping
227 Filtergraph description composition entails several levels of
228 escaping. See @ref{quoting_and_escaping,,the "Quoting and escaping"
229 section in the ffmpeg-utils(1) manual,ffmpeg-utils} for more
230 information about the employed escaping procedure.
232 A first level escaping affects the content of each filter option
233 value, which may contain the special character @code{:} used to
234 separate values, or one of the escaping characters @code{\'}.
236 A second level escaping affects the whole filter description, which
237 may contain the escaping characters @code{\'} or the special
238 characters @code{[],;} used by the filtergraph description.
240 Finally, when you specify a filtergraph on a shell commandline, you
241 need to perform a third level escaping for the shell special
242 characters contained within it.
244 For example, consider the following string to be embedded in
245 the @ref{drawtext} filter description @option{text} value:
247 this is a 'string': may contain one, or more, special characters
250 This string contains the @code{'} special escaping character, and the
251 @code{:} special character, so it needs to be escaped in this way:
253 text=this is a \'string\'\: may contain one, or more, special characters
256 A second level of escaping is required when embedding the filter
257 description in a filtergraph description, in order to escape all the
258 filtergraph special characters. Thus the example above becomes:
260 drawtext=text=this is a \\\'string\\\'\\: may contain one\, or more\, special characters
262 (note that in addition to the @code{\'} escaping special characters,
263 also @code{,} needs to be escaped).
265 Finally an additional level of escaping is needed when writing the
266 filtergraph description in a shell command, which depends on the
267 escaping rules of the adopted shell. For example, assuming that
268 @code{\} is special and needs to be escaped with another @code{\}, the
269 previous string will finally result in:
271 -vf "drawtext=text=this is a \\\\\\'string\\\\\\'\\\\: may contain one\\, or more\\, special characters"
274 @chapter Timeline editing
276 Some filters support a generic @option{enable} option. For the filters
277 supporting timeline editing, this option can be set to an expression which is
278 evaluated before sending a frame to the filter. If the evaluation is non-zero,
279 the filter will be enabled, otherwise the frame will be sent unchanged to the
280 next filter in the filtergraph.
282 The expression accepts the following values:
285 timestamp expressed in seconds, NAN if the input timestamp is unknown
288 sequential number of the input frame, starting from 0
291 the position in the file of the input frame, NAN if unknown
295 width and height of the input frame if video
298 Additionally, these filters support an @option{enable} command that can be used
299 to re-define the expression.
301 Like any other filtering option, the @option{enable} option follows the same
304 For example, to enable a blur filter (@ref{smartblur}) from 10 seconds to 3
305 minutes, and a @ref{curves} filter starting at 3 seconds:
307 smartblur = enable='between(t,10,3*60)',
308 curves = enable='gte(t,3)' : preset=cross_process
311 See @code{ffmpeg -filters} to view which filters have timeline support.
313 @c man end FILTERGRAPH DESCRIPTION
316 @chapter Changing options at runtime with a command
318 Some options can be changed during the operation of the filter using
319 a command. These options are marked 'T' on the output of
320 @command{ffmpeg} @option{-h filter=<name of filter>}.
321 The name of the command is the name of the option and the argument is
325 @chapter Options for filters with several inputs (framesync)
326 @c man begin OPTIONS FOR FILTERS WITH SEVERAL INPUTS
328 Some filters with several inputs support a common set of options.
329 These options can only be set by name, not with the short notation.
333 The action to take when EOF is encountered on the secondary input; it accepts
334 one of the following values:
338 Repeat the last frame (the default).
342 Pass the main input through.
346 If set to 1, force the output to terminate when the shortest input
347 terminates. Default value is 0.
350 If set to 1, force the filter to extend the last frame of secondary streams
351 until the end of the primary stream. A value of 0 disables this behavior.
355 @c man end OPTIONS FOR FILTERS WITH SEVERAL INPUTS
357 @chapter Audio Filters
358 @c man begin AUDIO FILTERS
360 When you configure your FFmpeg build, you can disable any of the
361 existing filters using @code{--disable-filters}.
362 The configure output will show the audio filters included in your
365 Below is a description of the currently available audio filters.
369 A compressor is mainly used to reduce the dynamic range of a signal.
370 Especially modern music is mostly compressed at a high ratio to
371 improve the overall loudness. It's done to get the highest attention
372 of a listener, "fatten" the sound and bring more "power" to the track.
373 If a signal is compressed too much it may sound dull or "dead"
374 afterwards or it may start to "pump" (which could be a powerful effect
375 but can also destroy a track completely).
376 The right compression is the key to reach a professional sound and is
377 the high art of mixing and mastering. Because of its complex settings
378 it may take a long time to get the right feeling for this kind of effect.
380 Compression is done by detecting the volume above a chosen level
381 @code{threshold} and dividing it by the factor set with @code{ratio}.
382 So if you set the threshold to -12dB and your signal reaches -6dB a ratio
383 of 2:1 will result in a signal at -9dB. Because an exact manipulation of
384 the signal would cause distortion of the waveform the reduction can be
385 levelled over the time. This is done by setting "Attack" and "Release".
386 @code{attack} determines how long the signal has to rise above the threshold
387 before any reduction will occur and @code{release} sets the time the signal
388 has to fall below the threshold to reduce the reduction again. Shorter signals
389 than the chosen attack time will be left untouched.
390 The overall reduction of the signal can be made up afterwards with the
391 @code{makeup} setting. So compressing the peaks of a signal about 6dB and
392 raising the makeup to this level results in a signal twice as loud than the
393 source. To gain a softer entry in the compression the @code{knee} flattens the
394 hard edge at the threshold in the range of the chosen decibels.
396 The filter accepts the following options:
400 Set input gain. Default is 1. Range is between 0.015625 and 64.
403 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
404 Default is @code{downward}.
407 If a signal of stream rises above this level it will affect the gain
409 By default it is 0.125. Range is between 0.00097563 and 1.
412 Set a ratio by which the signal is reduced. 1:2 means that if the level
413 rose 4dB above the threshold, it will be only 2dB above after the reduction.
414 Default is 2. Range is between 1 and 20.
417 Amount of milliseconds the signal has to rise above the threshold before gain
418 reduction starts. Default is 20. Range is between 0.01 and 2000.
421 Amount of milliseconds the signal has to fall below the threshold before
422 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
425 Set the amount by how much signal will be amplified after processing.
426 Default is 1. Range is from 1 to 64.
429 Curve the sharp knee around the threshold to enter gain reduction more softly.
430 Default is 2.82843. Range is between 1 and 8.
433 Choose if the @code{average} level between all channels of input stream
434 or the louder(@code{maximum}) channel of input stream affects the
435 reduction. Default is @code{average}.
438 Should the exact signal be taken in case of @code{peak} or an RMS one in case
439 of @code{rms}. Default is @code{rms} which is mostly smoother.
442 How much to use compressed signal in output. Default is 1.
443 Range is between 0 and 1.
448 This filter supports the all above options as @ref{commands}.
451 Simple audio dynamic range compression/expansion filter.
453 The filter accepts the following options:
457 Set contrast. Default is 33. Allowed range is between 0 and 100.
462 Copy the input audio source unchanged to the output. This is mainly useful for
467 Apply cross fade from one input audio stream to another input audio stream.
468 The cross fade is applied for specified duration near the end of first stream.
470 The filter accepts the following options:
474 Specify the number of samples for which the cross fade effect has to last.
475 At the end of the cross fade effect the first input audio will be completely
476 silent. Default is 44100.
479 Specify the duration of the cross fade effect. See
480 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
481 for the accepted syntax.
482 By default the duration is determined by @var{nb_samples}.
483 If set this option is used instead of @var{nb_samples}.
486 Should first stream end overlap with second stream start. Default is enabled.
489 Set curve for cross fade transition for first stream.
492 Set curve for cross fade transition for second stream.
494 For description of available curve types see @ref{afade} filter description.
501 Cross fade from one input to another:
503 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:c1=exp:c2=exp output.flac
507 Cross fade from one input to another but without overlapping:
509 ffmpeg -i first.flac -i second.flac -filter_complex acrossfade=d=10:o=0:c1=exp:c2=exp output.flac
514 Split audio stream into several bands.
516 This filter splits audio stream into two or more frequency ranges.
517 Summing all streams back will give flat output.
519 The filter accepts the following options:
523 Set split frequencies. Those must be positive and increasing.
526 Set filter order for each band split. This controls filter roll-off or steepness
527 of filter transfer function.
528 Available values are:
553 Default is @var{4th}.
556 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
559 Set output gain for each band. Default value is 1 for all bands.
566 Split input audio stream into two bands (low and high) with split frequency of 1500 Hz,
567 each band will be in separate stream:
569 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
573 Same as above, but with higher filter order:
575 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500:order=8th[LOW][HIGH]' -map '[LOW]' low.wav -map '[HIGH]' high.wav
579 Same as above, but also with additional middle band (frequencies between 1500 and 8000):
581 ffmpeg -i in.flac -filter_complex 'acrossover=split=1500 8000:order=8th[LOW][MID][HIGH]' -map '[LOW]' low.wav -map '[MID]' mid.wav -map '[HIGH]' high.wav
587 Reduce audio bit resolution.
589 This filter is bit crusher with enhanced functionality. A bit crusher
590 is used to audibly reduce number of bits an audio signal is sampled
591 with. This doesn't change the bit depth at all, it just produces the
592 effect. Material reduced in bit depth sounds more harsh and "digital".
593 This filter is able to even round to continuous values instead of discrete
595 Additionally it has a D/C offset which results in different crushing of
596 the lower and the upper half of the signal.
597 An Anti-Aliasing setting is able to produce "softer" crushing sounds.
599 Another feature of this filter is the logarithmic mode.
600 This setting switches from linear distances between bits to logarithmic ones.
601 The result is a much more "natural" sounding crusher which doesn't gate low
602 signals for example. The human ear has a logarithmic perception,
603 so this kind of crushing is much more pleasant.
604 Logarithmic crushing is also able to get anti-aliased.
606 The filter accepts the following options:
622 Can be linear: @code{lin} or logarithmic: @code{log}.
631 Set sample reduction.
634 Enable LFO. By default disabled.
645 Delay audio filtering until a given wallclock timestamp. See the @ref{cue}
649 Remove impulsive noise from input audio.
651 Samples detected as impulsive noise are replaced by interpolated samples using
652 autoregressive modelling.
656 Set window size, in milliseconds. Allowed range is from @code{10} to
657 @code{100}. Default value is @code{55} milliseconds.
658 This sets size of window which will be processed at once.
661 Set window overlap, in percentage of window size. Allowed range is from
662 @code{50} to @code{95}. Default value is @code{75} percent.
663 Setting this to a very high value increases impulsive noise removal but makes
664 whole process much slower.
667 Set autoregression order, in percentage of window size. Allowed range is from
668 @code{0} to @code{25}. Default value is @code{2} percent. This option also
669 controls quality of interpolated samples using neighbour good samples.
672 Set threshold value. Allowed range is from @code{1} to @code{100}.
673 Default value is @code{2}.
674 This controls the strength of impulsive noise which is going to be removed.
675 The lower value, the more samples will be detected as impulsive noise.
678 Set burst fusion, in percentage of window size. Allowed range is @code{0} to
679 @code{10}. Default value is @code{2}.
680 If any two samples detected as noise are spaced less than this value then any
681 sample between those two samples will be also detected as noise.
686 It accepts the following values:
689 Select overlap-add method. Even not interpolated samples are slightly
690 changed with this method.
693 Select overlap-save method. Not interpolated samples remain unchanged.
696 Default value is @code{a}.
700 Remove clipped samples from input audio.
702 Samples detected as clipped are replaced by interpolated samples using
703 autoregressive modelling.
707 Set window size, in milliseconds. Allowed range is from @code{10} to @code{100}.
708 Default value is @code{55} milliseconds.
709 This sets size of window which will be processed at once.
712 Set window overlap, in percentage of window size. Allowed range is from @code{50}
713 to @code{95}. Default value is @code{75} percent.
716 Set autoregression order, in percentage of window size. Allowed range is from
717 @code{0} to @code{25}. Default value is @code{8} percent. This option also controls
718 quality of interpolated samples using neighbour good samples.
721 Set threshold value. Allowed range is from @code{1} to @code{100}.
722 Default value is @code{10}. Higher values make clip detection less aggressive.
725 Set size of histogram used to detect clips. Allowed range is from @code{100} to @code{9999}.
726 Default value is @code{1000}. Higher values make clip detection less aggressive.
731 It accepts the following values:
734 Select overlap-add method. Even not interpolated samples are slightly changed
738 Select overlap-save method. Not interpolated samples remain unchanged.
741 Default value is @code{a}.
746 Delay one or more audio channels.
748 Samples in delayed channel are filled with silence.
750 The filter accepts the following option:
754 Set list of delays in milliseconds for each channel separated by '|'.
755 Unused delays will be silently ignored. If number of given delays is
756 smaller than number of channels all remaining channels will not be delayed.
757 If you want to delay exact number of samples, append 'S' to number.
758 If you want instead to delay in seconds, append 's' to number.
761 Use last set delay for all remaining channels. By default is disabled.
762 This option if enabled changes how option @code{delays} is interpreted.
769 Delay first channel by 1.5 seconds, the third channel by 0.5 seconds and leave
770 the second channel (and any other channels that may be present) unchanged.
776 Delay second channel by 500 samples, the third channel by 700 samples and leave
777 the first channel (and any other channels that may be present) unchanged.
783 Delay all channels by same number of samples:
785 adelay=delays=64S:all=1
790 Remedy denormals in audio by adding extremely low-level noise.
792 This filter shall be placed before any filter that can produce denormals.
794 A description of the accepted parameters follows.
798 Set level of added noise in dB. Default is @code{-351}.
799 Allowed range is from -451 to -90.
802 Set type of added noise.
815 Default is @code{dc}.
820 This filter supports the all above options as @ref{commands}.
822 @section aderivative, aintegral
824 Compute derivative/integral of audio stream.
826 Applying both filters one after another produces original audio.
830 Apply echoing to the input audio.
832 Echoes are reflected sound and can occur naturally amongst mountains
833 (and sometimes large buildings) when talking or shouting; digital echo
834 effects emulate this behaviour and are often used to help fill out the
835 sound of a single instrument or vocal. The time difference between the
836 original signal and the reflection is the @code{delay}, and the
837 loudness of the reflected signal is the @code{decay}.
838 Multiple echoes can have different delays and decays.
840 A description of the accepted parameters follows.
844 Set input gain of reflected signal. Default is @code{0.6}.
847 Set output gain of reflected signal. Default is @code{0.3}.
850 Set list of time intervals in milliseconds between original signal and reflections
851 separated by '|'. Allowed range for each @code{delay} is @code{(0 - 90000.0]}.
852 Default is @code{1000}.
855 Set list of loudness of reflected signals separated by '|'.
856 Allowed range for each @code{decay} is @code{(0 - 1.0]}.
857 Default is @code{0.5}.
864 Make it sound as if there are twice as many instruments as are actually playing:
866 aecho=0.8:0.88:60:0.4
870 If delay is very short, then it sounds like a (metallic) robot playing music:
876 A longer delay will sound like an open air concert in the mountains:
878 aecho=0.8:0.9:1000:0.3
882 Same as above but with one more mountain:
884 aecho=0.8:0.9:1000|1800:0.3|0.25
889 Audio emphasis filter creates or restores material directly taken from LPs or
890 emphased CDs with different filter curves. E.g. to store music on vinyl the
891 signal has to be altered by a filter first to even out the disadvantages of
892 this recording medium.
893 Once the material is played back the inverse filter has to be applied to
894 restore the distortion of the frequency response.
896 The filter accepts the following options:
906 Set filter mode. For restoring material use @code{reproduction} mode, otherwise
907 use @code{production} mode. Default is @code{reproduction} mode.
910 Set filter type. Selects medium. Can be one of the following:
922 select Compact Disc (CD).
928 select 50µs (FM-KF).
930 select 75µs (FM-KF).
936 This filter supports the all above options as @ref{commands}.
940 Modify an audio signal according to the specified expressions.
942 This filter accepts one or more expressions (one for each channel),
943 which are evaluated and used to modify a corresponding audio signal.
945 It accepts the following parameters:
949 Set the '|'-separated expressions list for each separate channel. If
950 the number of input channels is greater than the number of
951 expressions, the last specified expression is used for the remaining
954 @item channel_layout, c
955 Set output channel layout. If not specified, the channel layout is
956 specified by the number of expressions. If set to @samp{same}, it will
957 use by default the same input channel layout.
960 Each expression in @var{exprs} can contain the following constants and functions:
964 channel number of the current expression
967 number of the evaluated sample, starting from 0
973 time of the evaluated sample expressed in seconds
976 @item nb_out_channels
977 input and output number of channels
980 the value of input channel with number @var{CH}
983 Note: this filter is slow. For faster processing you should use a
992 aeval=val(ch)/2:c=same
996 Invert phase of the second channel:
1005 Apply fade-in/out effect to input audio.
1007 A description of the accepted parameters follows.
1011 Specify the effect type, can be either @code{in} for fade-in, or
1012 @code{out} for a fade-out effect. Default is @code{in}.
1014 @item start_sample, ss
1015 Specify the number of the start sample for starting to apply the fade
1016 effect. Default is 0.
1018 @item nb_samples, ns
1019 Specify the number of samples for which the fade effect has to last. At
1020 the end of the fade-in effect the output audio will have the same
1021 volume as the input audio, at the end of the fade-out transition
1022 the output audio will be silence. Default is 44100.
1024 @item start_time, st
1025 Specify the start time of the fade effect. Default is 0.
1026 The value must be specified as a time duration; see
1027 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1028 for the accepted syntax.
1029 If set this option is used instead of @var{start_sample}.
1032 Specify the duration of the fade effect. See
1033 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1034 for the accepted syntax.
1035 At the end of the fade-in effect the output audio will have the same
1036 volume as the input audio, at the end of the fade-out transition
1037 the output audio will be silence.
1038 By default the duration is determined by @var{nb_samples}.
1039 If set this option is used instead of @var{nb_samples}.
1042 Set curve for fade transition.
1044 It accepts the following values:
1047 select triangular, linear slope (default)
1049 select quarter of sine wave
1051 select half of sine wave
1053 select exponential sine wave
1057 select inverted parabola
1071 select inverted quarter of sine wave
1073 select inverted half of sine wave
1075 select double-exponential seat
1077 select double-exponential sigmoid
1079 select logistic sigmoid
1081 select sine cardinal function
1083 select inverted sine cardinal function
1089 @subsection Commands
1091 This filter supports the all above options as @ref{commands}.
1093 @subsection Examples
1097 Fade in first 15 seconds of audio:
1099 afade=t=in:ss=0:d=15
1103 Fade out last 25 seconds of a 900 seconds audio:
1105 afade=t=out:st=875:d=25
1110 Denoise audio samples with FFT.
1112 A description of the accepted parameters follows.
1116 Set the noise reduction in dB, allowed range is 0.01 to 97.
1117 Default value is 12 dB.
1120 Set the noise floor in dB, allowed range is -80 to -20.
1121 Default value is -50 dB.
1126 It accepts the following values:
1135 Select shellac noise.
1138 Select custom noise, defined in @code{bn} option.
1140 Default value is white noise.
1144 Set custom band noise for every one of 15 bands.
1145 Bands are separated by ' ' or '|'.
1148 Set the residual floor in dB, allowed range is -80 to -20.
1149 Default value is -38 dB.
1152 Enable noise tracking. By default is disabled.
1153 With this enabled, noise floor is automatically adjusted.
1156 Enable residual tracking. By default is disabled.
1159 Set the output mode.
1161 It accepts the following values:
1164 Pass input unchanged.
1167 Pass noise filtered out.
1172 Default value is @var{o}.
1176 @subsection Commands
1178 This filter supports the following commands:
1180 @item sample_noise, sn
1181 Start or stop measuring noise profile.
1182 Syntax for the command is : "start" or "stop" string.
1183 After measuring noise profile is stopped it will be
1184 automatically applied in filtering.
1186 @item noise_reduction, nr
1187 Change noise reduction. Argument is single float number.
1188 Syntax for the command is : "@var{noise_reduction}"
1190 @item noise_floor, nf
1191 Change noise floor. Argument is single float number.
1192 Syntax for the command is : "@var{noise_floor}"
1194 @item output_mode, om
1195 Change output mode operation.
1196 Syntax for the command is : "i", "o" or "n" string.
1200 Apply arbitrary expressions to samples in frequency domain.
1204 Set frequency domain real expression for each separate channel separated
1205 by '|'. Default is "re".
1206 If the number of input channels is greater than the number of
1207 expressions, the last specified expression is used for the remaining
1211 Set frequency domain imaginary expression for each separate channel
1212 separated by '|'. Default is "im".
1214 Each expression in @var{real} and @var{imag} can contain the following
1215 constants and functions:
1222 current frequency bin number
1225 number of available bins
1228 channel number of the current expression
1237 current real part of frequency bin of current channel
1240 current imaginary part of frequency bin of current channel
1243 Return the value of real part of frequency bin at location (@var{bin},@var{channel})
1246 Return the value of imaginary part of frequency bin at location (@var{bin},@var{channel})
1250 Set window size. Allowed range is from 16 to 131072.
1251 Default is @code{4096}
1254 Set window function. Default is @code{hann}.
1257 Set window overlap. If set to 1, the recommended overlap for selected
1258 window function will be picked. Default is @code{0.75}.
1261 @subsection Examples
1265 Leave almost only low frequencies in audio:
1267 afftfilt="'real=re * (1-clip((b/nb)*b,0,1))':imag='im * (1-clip((b/nb)*b,0,1))'"
1271 Apply robotize effect:
1273 afftfilt="real='hypot(re,im)*sin(0)':imag='hypot(re,im)*cos(0)':win_size=512:overlap=0.75"
1277 Apply whisper effect:
1279 afftfilt="real='hypot(re,im)*cos((random(0)*2-1)*2*3.14)':imag='hypot(re,im)*sin((random(1)*2-1)*2*3.14)':win_size=128:overlap=0.8"
1286 Apply an arbitrary Finite Impulse Response filter.
1288 This filter is designed for applying long FIR filters,
1289 up to 60 seconds long.
1291 It can be used as component for digital crossover filters,
1292 room equalization, cross talk cancellation, wavefield synthesis,
1293 auralization, ambiophonics, ambisonics and spatialization.
1295 This filter uses the streams higher than first one as FIR coefficients.
1296 If the non-first stream holds a single channel, it will be used
1297 for all input channels in the first stream, otherwise
1298 the number of channels in the non-first stream must be same as
1299 the number of channels in the first stream.
1301 It accepts the following parameters:
1305 Set dry gain. This sets input gain.
1308 Set wet gain. This sets final output gain.
1311 Set Impulse Response filter length. Default is 1, which means whole IR is processed.
1314 Enable applying gain measured from power of IR.
1316 Set which approach to use for auto gain measurement.
1320 Do not apply any gain.
1323 select peak gain, very conservative approach. This is default value.
1326 select DC gain, limited application.
1329 select gain to noise approach, this is most popular one.
1333 Set gain to be applied to IR coefficients before filtering.
1334 Allowed range is 0 to 1. This gain is applied after any gain applied with @var{gtype} option.
1337 Set format of IR stream. Can be @code{mono} or @code{input}.
1338 Default is @code{input}.
1341 Set max allowed Impulse Response filter duration in seconds. Default is 30 seconds.
1342 Allowed range is 0.1 to 60 seconds.
1345 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1346 By default it is disabled.
1349 Set for which IR channel to display frequency response. By default is first channel
1350 displayed. This option is used only when @var{response} is enabled.
1353 Set video stream size. This option is used only when @var{response} is enabled.
1356 Set video stream frame rate. This option is used only when @var{response} is enabled.
1359 Set minimal partition size used for convolution. Default is @var{8192}.
1360 Allowed range is from @var{1} to @var{32768}.
1361 Lower values decreases latency at cost of higher CPU usage.
1364 Set maximal partition size used for convolution. Default is @var{8192}.
1365 Allowed range is from @var{8} to @var{32768}.
1366 Lower values may increase CPU usage.
1369 Set number of input impulse responses streams which will be switchable at runtime.
1370 Allowed range is from @var{1} to @var{32}. Default is @var{1}.
1373 Set IR stream which will be used for convolution, starting from @var{0}, should always be
1374 lower than supplied value by @code{nbirs} option. Default is @var{0}.
1375 This option can be changed at runtime via @ref{commands}.
1378 @subsection Examples
1382 Apply reverb to stream using mono IR file as second input, complete command using ffmpeg:
1384 ffmpeg -i input.wav -i middle_tunnel_1way_mono.wav -lavfi afir output.wav
1391 Set output format constraints for the input audio. The framework will
1392 negotiate the most appropriate format to minimize conversions.
1394 It accepts the following parameters:
1397 @item sample_fmts, f
1398 A '|'-separated list of requested sample formats.
1400 @item sample_rates, r
1401 A '|'-separated list of requested sample rates.
1403 @item channel_layouts, cl
1404 A '|'-separated list of requested channel layouts.
1406 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
1407 for the required syntax.
1410 If a parameter is omitted, all values are allowed.
1412 Force the output to either unsigned 8-bit or signed 16-bit stereo
1414 aformat=sample_fmts=u8|s16:channel_layouts=stereo
1418 Apply frequency shift to input audio samples.
1420 The filter accepts the following options:
1424 Specify frequency shift. Allowed range is -INT_MAX to INT_MAX.
1425 Default value is 0.0.
1428 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
1429 Default value is 1.0.
1432 @subsection Commands
1434 This filter supports the all above options as @ref{commands}.
1438 A gate is mainly used to reduce lower parts of a signal. This kind of signal
1439 processing reduces disturbing noise between useful signals.
1441 Gating is done by detecting the volume below a chosen level @var{threshold}
1442 and dividing it by the factor set with @var{ratio}. The bottom of the noise
1443 floor is set via @var{range}. Because an exact manipulation of the signal
1444 would cause distortion of the waveform the reduction can be levelled over
1445 time. This is done by setting @var{attack} and @var{release}.
1447 @var{attack} determines how long the signal has to fall below the threshold
1448 before any reduction will occur and @var{release} sets the time the signal
1449 has to rise above the threshold to reduce the reduction again.
1450 Shorter signals than the chosen attack time will be left untouched.
1454 Set input level before filtering.
1455 Default is 1. Allowed range is from 0.015625 to 64.
1458 Set the mode of operation. Can be @code{upward} or @code{downward}.
1459 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
1460 will be amplified, expanding dynamic range in upward direction.
1461 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
1464 Set the level of gain reduction when the signal is below the threshold.
1465 Default is 0.06125. Allowed range is from 0 to 1.
1466 Setting this to 0 disables reduction and then filter behaves like expander.
1469 If a signal rises above this level the gain reduction is released.
1470 Default is 0.125. Allowed range is from 0 to 1.
1473 Set a ratio by which the signal is reduced.
1474 Default is 2. Allowed range is from 1 to 9000.
1477 Amount of milliseconds the signal has to rise above the threshold before gain
1479 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
1482 Amount of milliseconds the signal has to fall below the threshold before the
1483 reduction is increased again. Default is 250 milliseconds.
1484 Allowed range is from 0.01 to 9000.
1487 Set amount of amplification of signal after processing.
1488 Default is 1. Allowed range is from 1 to 64.
1491 Curve the sharp knee around the threshold to enter gain reduction more softly.
1492 Default is 2.828427125. Allowed range is from 1 to 8.
1495 Choose if exact signal should be taken for detection or an RMS like one.
1496 Default is @code{rms}. Can be @code{peak} or @code{rms}.
1499 Choose if the average level between all channels or the louder channel affects
1501 Default is @code{average}. Can be @code{average} or @code{maximum}.
1504 @subsection Commands
1506 This filter supports the all above options as @ref{commands}.
1510 Apply an arbitrary Infinite Impulse Response filter.
1512 It accepts the following parameters:
1516 Set B/numerator/zeros/reflection coefficients.
1519 Set A/denominator/poles/ladder coefficients.
1531 Set coefficients format.
1535 lattice-ladder function
1537 analog transfer function
1539 digital transfer function
1541 Z-plane zeros/poles, cartesian (default)
1543 Z-plane zeros/poles, polar radians
1545 Z-plane zeros/poles, polar degrees
1551 Set type of processing.
1563 Set filtering precision.
1567 double-precision floating-point (default)
1569 single-precision floating-point
1577 Normalize filter coefficients, by default is enabled.
1578 Enabling it will normalize magnitude response at DC to 0dB.
1581 How much to use filtered signal in output. Default is 1.
1582 Range is between 0 and 1.
1585 Show IR frequency response, magnitude(magenta), phase(green) and group delay(yellow) in additional video stream.
1586 By default it is disabled.
1589 Set for which IR channel to display frequency response. By default is first channel
1590 displayed. This option is used only when @var{response} is enabled.
1593 Set video stream size. This option is used only when @var{response} is enabled.
1596 Coefficients in @code{tf} and @code{sf} format are separated by spaces and are in ascending
1599 Coefficients in @code{zp} format are separated by spaces and order of coefficients
1600 doesn't matter. Coefficients in @code{zp} format are complex numbers with @var{i}
1603 Different coefficients and gains can be provided for every channel, in such case
1604 use '|' to separate coefficients or gains. Last provided coefficients will be
1605 used for all remaining channels.
1607 @subsection Examples
1611 Apply 2 pole elliptic notch at around 5000Hz for 48000 Hz sample rate:
1613 aiir=k=1:z=7.957584807809675810E-1 -2.575128568908332300 3.674839853930788710 -2.57512875289799137 7.957586296317130880E-1:p=1 -2.86950072432325953 3.63022088054647218 -2.28075678147272232 6.361362326477423500E-1:f=tf:r=d
1617 Same as above but in @code{zp} format:
1619 aiir=k=0.79575848078096756:z=0.80918701+0.58773007i 0.80918701-0.58773007i 0.80884700+0.58784055i 0.80884700-0.58784055i:p=0.63892345+0.59951235i 0.63892345-0.59951235i 0.79582691+0.44198673i 0.79582691-0.44198673i:f=zp:r=s
1623 Apply 3-rd order analog normalized Butterworth low-pass filter, using analog transfer function format:
1625 aiir=z=1.3057 0 0 0:p=1.3057 2.3892 2.1860 1:f=sf:r=d
1631 The limiter prevents an input signal from rising over a desired threshold.
1632 This limiter uses lookahead technology to prevent your signal from distorting.
1633 It means that there is a small delay after the signal is processed. Keep in mind
1634 that the delay it produces is the attack time you set.
1636 The filter accepts the following options:
1640 Set input gain. Default is 1.
1643 Set output gain. Default is 1.
1646 Don't let signals above this level pass the limiter. Default is 1.
1649 The limiter will reach its attenuation level in this amount of time in
1650 milliseconds. Default is 5 milliseconds.
1653 Come back from limiting to attenuation 1.0 in this amount of milliseconds.
1654 Default is 50 milliseconds.
1657 When gain reduction is always needed ASC takes care of releasing to an
1658 average reduction level rather than reaching a reduction of 0 in the release
1662 Select how much the release time is affected by ASC, 0 means nearly no changes
1663 in release time while 1 produces higher release times.
1666 Auto level output signal. Default is enabled.
1667 This normalizes audio back to 0dB if enabled.
1670 Depending on picked setting it is recommended to upsample input 2x or 4x times
1671 with @ref{aresample} before applying this filter.
1675 Apply a two-pole all-pass filter with central frequency (in Hz)
1676 @var{frequency}, and filter-width @var{width}.
1677 An all-pass filter changes the audio's frequency to phase relationship
1678 without changing its frequency to amplitude relationship.
1680 The filter accepts the following options:
1684 Set frequency in Hz.
1687 Set method to specify band-width of filter.
1702 Specify the band-width of a filter in width_type units.
1705 How much to use filtered signal in output. Default is 1.
1706 Range is between 0 and 1.
1709 Specify which channels to filter, by default all available are filtered.
1712 Normalize biquad coefficients, by default is disabled.
1713 Enabling it will normalize magnitude response at DC to 0dB.
1716 Set the filter order, can be 1 or 2. Default is 2.
1719 Set transform type of IIR filter.
1728 Set precison of filtering.
1731 Pick automatic sample format depending on surround filters.
1733 Always use signed 16-bit.
1735 Always use signed 32-bit.
1737 Always use float 32-bit.
1739 Always use float 64-bit.
1743 @subsection Commands
1745 This filter supports the following commands:
1748 Change allpass frequency.
1749 Syntax for the command is : "@var{frequency}"
1752 Change allpass width_type.
1753 Syntax for the command is : "@var{width_type}"
1756 Change allpass width.
1757 Syntax for the command is : "@var{width}"
1761 Syntax for the command is : "@var{mix}"
1768 The filter accepts the following options:
1772 Set the number of loops. Setting this value to -1 will result in infinite loops.
1776 Set maximal number of samples. Default is 0.
1779 Set first sample of loop. Default is 0.
1785 Merge two or more audio streams into a single multi-channel stream.
1787 The filter accepts the following options:
1792 Set the number of inputs. Default is 2.
1796 If the channel layouts of the inputs are disjoint, and therefore compatible,
1797 the channel layout of the output will be set accordingly and the channels
1798 will be reordered as necessary. If the channel layouts of the inputs are not
1799 disjoint, the output will have all the channels of the first input then all
1800 the channels of the second input, in that order, and the channel layout of
1801 the output will be the default value corresponding to the total number of
1804 For example, if the first input is in 2.1 (FL+FR+LF) and the second input
1805 is FC+BL+BR, then the output will be in 5.1, with the channels in the
1806 following order: a1, a2, b1, a3, b2, b3 (a1 is the first channel of the
1807 first input, b1 is the first channel of the second input).
1809 On the other hand, if both input are in stereo, the output channels will be
1810 in the default order: a1, a2, b1, b2, and the channel layout will be
1811 arbitrarily set to 4.0, which may or may not be the expected value.
1813 All inputs must have the same sample rate, and format.
1815 If inputs do not have the same duration, the output will stop with the
1818 @subsection Examples
1822 Merge two mono files into a stereo stream:
1824 amovie=left.wav [l] ; amovie=right.mp3 [r] ; [l] [r] amerge
1828 Multiple merges assuming 1 video stream and 6 audio streams in @file{input.mkv}:
1830 ffmpeg -i input.mkv -filter_complex "[0:1][0:2][0:3][0:4][0:5][0:6] amerge=inputs=6" -c:a pcm_s16le output.mkv
1836 Mixes multiple audio inputs into a single output.
1838 Note that this filter only supports float samples (the @var{amerge}
1839 and @var{pan} audio filters support many formats). If the @var{amix}
1840 input has integer samples then @ref{aresample} will be automatically
1841 inserted to perform the conversion to float samples.
1845 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex amix=inputs=3:duration=first:dropout_transition=3 OUTPUT
1847 will mix 3 input audio streams to a single output with the same duration as the
1848 first input and a dropout transition time of 3 seconds.
1850 It accepts the following parameters:
1854 The number of inputs. If unspecified, it defaults to 2.
1857 How to determine the end-of-stream.
1861 The duration of the longest input. (default)
1864 The duration of the shortest input.
1867 The duration of the first input.
1871 @item dropout_transition
1872 The transition time, in seconds, for volume renormalization when an input
1873 stream ends. The default value is 2 seconds.
1876 Specify weight of each input audio stream as sequence.
1877 Each weight is separated by space. By default all inputs have same weight.
1880 @subsection Commands
1882 This filter supports the following commands:
1885 Syntax is same as option with same name.
1890 Multiply first audio stream with second audio stream and store result
1891 in output audio stream. Multiplication is done by multiplying each
1892 sample from first stream with sample at same position from second stream.
1894 With this element-wise multiplication one can create amplitude fades and
1895 amplitude modulations.
1897 @section anequalizer
1899 High-order parametric multiband equalizer for each channel.
1901 It accepts the following parameters:
1905 This option string is in format:
1906 "c@var{chn} f=@var{cf} w=@var{w} g=@var{g} t=@var{f} | ..."
1907 Each equalizer band is separated by '|'.
1911 Set channel number to which equalization will be applied.
1912 If input doesn't have that channel the entry is ignored.
1915 Set central frequency for band.
1916 If input doesn't have that frequency the entry is ignored.
1919 Set band width in Hertz.
1922 Set band gain in dB.
1925 Set filter type for band, optional, can be:
1929 Butterworth, this is default.
1940 With this option activated frequency response of anequalizer is displayed
1944 Set video stream size. Only useful if curves option is activated.
1947 Set max gain that will be displayed. Only useful if curves option is activated.
1948 Setting this to a reasonable value makes it possible to display gain which is derived from
1949 neighbour bands which are too close to each other and thus produce higher gain
1950 when both are activated.
1953 Set frequency scale used to draw frequency response in video output.
1954 Can be linear or logarithmic. Default is logarithmic.
1957 Set color for each channel curve which is going to be displayed in video stream.
1958 This is list of color names separated by space or by '|'.
1959 Unrecognised or missing colors will be replaced by white color.
1962 @subsection Examples
1966 Lower gain by 10 of central frequency 200Hz and width 100 Hz
1967 for first 2 channels using Chebyshev type 1 filter:
1969 anequalizer=c0 f=200 w=100 g=-10 t=1|c1 f=200 w=100 g=-10 t=1
1973 @subsection Commands
1975 This filter supports the following commands:
1978 Alter existing filter parameters.
1979 Syntax for the commands is : "@var{fN}|f=@var{freq}|w=@var{width}|g=@var{gain}"
1981 @var{fN} is existing filter number, starting from 0, if no such filter is available
1983 @var{freq} set new frequency parameter.
1984 @var{width} set new width parameter in Hertz.
1985 @var{gain} set new gain parameter in dB.
1987 Full filter invocation with asendcmd may look like this:
1988 asendcmd=c='4.0 anequalizer change 0|f=200|w=50|g=1',anequalizer=...
1993 Reduce broadband noise in audio samples using Non-Local Means algorithm.
1995 Each sample is adjusted by looking for other samples with similar contexts. This
1996 context similarity is defined by comparing their surrounding patches of size
1997 @option{p}. Patches are searched in an area of @option{r} around the sample.
1999 The filter accepts the following options:
2003 Set denoising strength. Allowed range is from 0.00001 to 10. Default value is 0.00001.
2006 Set patch radius duration. Allowed range is from 1 to 100 milliseconds.
2007 Default value is 2 milliseconds.
2010 Set research radius duration. Allowed range is from 2 to 300 milliseconds.
2011 Default value is 6 milliseconds.
2014 Set the output mode.
2016 It accepts the following values:
2019 Pass input unchanged.
2022 Pass noise filtered out.
2027 Default value is @var{o}.
2031 Set smooth factor. Default value is @var{11}. Allowed range is from @var{1} to @var{15}.
2034 @subsection Commands
2036 This filter supports the all above options as @ref{commands}.
2039 Apply Normalized Least-Mean-Squares algorithm to the first audio stream using the second audio stream.
2041 This adaptive filter is used to mimic a desired filter by finding the filter coefficients that
2042 relate to producing the least mean square of the error signal (difference between the desired,
2043 2nd input audio stream and the actual signal, the 1st input audio stream).
2045 A description of the accepted options follows.
2058 Set the filter leakage.
2061 It accepts the following values:
2070 Pass filtered samples.
2073 Pass difference between desired and filtered samples.
2075 Default value is @var{o}.
2079 @subsection Examples
2083 One of many usages of this filter is noise reduction, input audio is filtered
2084 with same samples that are delayed by fixed amount, one such example for stereo audio is:
2086 asplit[a][b],[a]adelay=32S|32S[a],[b][a]anlms=order=128:leakage=0.0005:mu=.5:out_mode=o
2090 @subsection Commands
2092 This filter supports the same commands as options, excluding option @code{order}.
2096 Pass the audio source unchanged to the output.
2100 Pad the end of an audio stream with silence.
2102 This can be used together with @command{ffmpeg} @option{-shortest} to
2103 extend audio streams to the same length as the video stream.
2105 A description of the accepted options follows.
2109 Set silence packet size. Default value is 4096.
2112 Set the number of samples of silence to add to the end. After the
2113 value is reached, the stream is terminated. This option is mutually
2114 exclusive with @option{whole_len}.
2117 Set the minimum total number of samples in the output audio stream. If
2118 the value is longer than the input audio length, silence is added to
2119 the end, until the value is reached. This option is mutually exclusive
2120 with @option{pad_len}.
2123 Specify the duration of samples of silence to add. See
2124 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2125 for the accepted syntax. Used only if set to non-zero value.
2128 Specify the minimum total duration in the output audio stream. See
2129 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
2130 for the accepted syntax. Used only if set to non-zero value. If the value is longer than
2131 the input audio length, silence is added to the end, until the value is reached.
2132 This option is mutually exclusive with @option{pad_dur}
2135 If neither the @option{pad_len} nor the @option{whole_len} nor @option{pad_dur}
2136 nor @option{whole_dur} option is set, the filter will add silence to the end of
2137 the input stream indefinitely.
2139 @subsection Examples
2143 Add 1024 samples of silence to the end of the input:
2149 Make sure the audio output will contain at least 10000 samples, pad
2150 the input with silence if required:
2152 apad=whole_len=10000
2156 Use @command{ffmpeg} to pad the audio input with silence, so that the
2157 video stream will always result the shortest and will be converted
2158 until the end in the output file when using the @option{shortest}
2161 ffmpeg -i VIDEO -i AUDIO -filter_complex "[1:0]apad" -shortest OUTPUT
2166 Add a phasing effect to the input audio.
2168 A phaser filter creates series of peaks and troughs in the frequency spectrum.
2169 The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
2171 A description of the accepted parameters follows.
2175 Set input gain. Default is 0.4.
2178 Set output gain. Default is 0.74
2181 Set delay in milliseconds. Default is 3.0.
2184 Set decay. Default is 0.4.
2187 Set modulation speed in Hz. Default is 0.5.
2190 Set modulation type. Default is triangular.
2192 It accepts the following values:
2199 @section aphaseshift
2200 Apply phase shift to input audio samples.
2202 The filter accepts the following options:
2206 Specify phase shift. Allowed range is from -1.0 to 1.0.
2207 Default value is 0.0.
2210 Set output gain applied to final output. Allowed range is from 0.0 to 1.0.
2211 Default value is 1.0.
2214 @subsection Commands
2216 This filter supports the all above options as @ref{commands}.
2220 Audio pulsator is something between an autopanner and a tremolo.
2221 But it can produce funny stereo effects as well. Pulsator changes the volume
2222 of the left and right channel based on a LFO (low frequency oscillator) with
2223 different waveforms and shifted phases.
2224 This filter have the ability to define an offset between left and right
2225 channel. An offset of 0 means that both LFO shapes match each other.
2226 The left and right channel are altered equally - a conventional tremolo.
2227 An offset of 50% means that the shape of the right channel is exactly shifted
2228 in phase (or moved backwards about half of the frequency) - pulsator acts as
2229 an autopanner. At 1 both curves match again. Every setting in between moves the
2230 phase shift gapless between all stages and produces some "bypassing" sounds with
2231 sine and triangle waveforms. The more you set the offset near 1 (starting from
2232 the 0.5) the faster the signal passes from the left to the right speaker.
2234 The filter accepts the following options:
2238 Set input gain. By default it is 1. Range is [0.015625 - 64].
2241 Set output gain. By default it is 1. Range is [0.015625 - 64].
2244 Set waveform shape the LFO will use. Can be one of: sine, triangle, square,
2245 sawup or sawdown. Default is sine.
2248 Set modulation. Define how much of original signal is affected by the LFO.
2251 Set left channel offset. Default is 0. Allowed range is [0 - 1].
2254 Set right channel offset. Default is 0.5. Allowed range is [0 - 1].
2257 Set pulse width. Default is 1. Allowed range is [0 - 2].
2260 Set possible timing mode. Can be one of: bpm, ms or hz. Default is hz.
2263 Set bpm. Default is 120. Allowed range is [30 - 300]. Only used if timing
2267 Set ms. Default is 500. Allowed range is [10 - 2000]. Only used if timing
2271 Set frequency in Hz. Default is 2. Allowed range is [0.01 - 100]. Only used
2272 if timing is set to hz.
2278 Resample the input audio to the specified parameters, using the
2279 libswresample library. If none are specified then the filter will
2280 automatically convert between its input and output.
2282 This filter is also able to stretch/squeeze the audio data to make it match
2283 the timestamps or to inject silence / cut out audio to make it match the
2284 timestamps, do a combination of both or do neither.
2286 The filter accepts the syntax
2287 [@var{sample_rate}:]@var{resampler_options}, where @var{sample_rate}
2288 expresses a sample rate and @var{resampler_options} is a list of
2289 @var{key}=@var{value} pairs, separated by ":". See the
2290 @ref{Resampler Options,,"Resampler Options" section in the
2291 ffmpeg-resampler(1) manual,ffmpeg-resampler}
2292 for the complete list of supported options.
2294 @subsection Examples
2298 Resample the input audio to 44100Hz:
2304 Stretch/squeeze samples to the given timestamps, with a maximum of 1000
2305 samples per second compensation:
2307 aresample=async=1000
2313 Reverse an audio clip.
2315 Warning: This filter requires memory to buffer the entire clip, so trimming
2318 @subsection Examples
2322 Take the first 5 seconds of a clip, and reverse it.
2324 atrim=end=5,areverse
2330 Reduce noise from speech using Recurrent Neural Networks.
2332 This filter accepts the following options:
2336 Set train model file to load. This option is always required.
2339 Set how much to mix filtered samples into final output.
2340 Allowed range is from -1 to 1. Default value is 1.
2341 Negative values are special, they set how much to keep filtered noise
2342 in the final filter output. Set this option to -1 to hear actual
2343 noise removed from input signal.
2346 @section asetnsamples
2348 Set the number of samples per each output audio frame.
2350 The last output packet may contain a different number of samples, as
2351 the filter will flush all the remaining samples when the input audio
2354 The filter accepts the following options:
2358 @item nb_out_samples, n
2359 Set the number of frames per each output audio frame. The number is
2360 intended as the number of samples @emph{per each channel}.
2361 Default value is 1024.
2364 If set to 1, the filter will pad the last audio frame with zeroes, so
2365 that the last frame will contain the same number of samples as the
2366 previous ones. Default value is 1.
2369 For example, to set the number of per-frame samples to 1234 and
2370 disable padding for the last frame, use:
2372 asetnsamples=n=1234:p=0
2377 Set the sample rate without altering the PCM data.
2378 This will result in a change of speed and pitch.
2380 The filter accepts the following options:
2383 @item sample_rate, r
2384 Set the output sample rate. Default is 44100 Hz.
2389 Show a line containing various information for each input audio frame.
2390 The input audio is not modified.
2392 The shown line contains a sequence of key/value pairs of the form
2393 @var{key}:@var{value}.
2395 The following values are shown in the output:
2399 The (sequential) number of the input frame, starting from 0.
2402 The presentation timestamp of the input frame, in time base units; the time base
2403 depends on the filter input pad, and is usually 1/@var{sample_rate}.
2406 The presentation timestamp of the input frame in seconds.
2409 position of the frame in the input stream, -1 if this information in
2410 unavailable and/or meaningless (for example in case of synthetic audio)
2419 The sample rate for the audio frame.
2422 The number of samples (per channel) in the frame.
2425 The Adler-32 checksum (printed in hexadecimal) of the audio data. For planar
2426 audio, the data is treated as if all the planes were concatenated.
2428 @item plane_checksums
2429 A list of Adler-32 checksums for each data plane.
2433 Apply audio soft clipping.
2435 Soft clipping is a type of distortion effect where the amplitude of a signal is saturated
2436 along a smooth curve, rather than the abrupt shape of hard-clipping.
2438 This filter accepts the following options:
2442 Set type of soft-clipping.
2444 It accepts the following values:
2458 Set threshold from where to start clipping. Default value is 0dB or 1.
2461 Set gain applied to output. Default value is 0dB or 1.
2464 Set additional parameter which controls sigmoid function.
2467 Set oversampling factor.
2470 @subsection Commands
2472 This filter supports the all above options as @ref{commands}.
2475 Automatic Speech Recognition
2477 This filter uses PocketSphinx for speech recognition. To enable
2478 compilation of this filter, you need to configure FFmpeg with
2479 @code{--enable-pocketsphinx}.
2481 It accepts the following options:
2485 Set sampling rate of input audio. Defaults is @code{16000}.
2486 This need to match speech models, otherwise one will get poor results.
2489 Set dictionary containing acoustic model files.
2492 Set pronunciation dictionary.
2495 Set language model file.
2498 Set language model set.
2501 Set which language model to use.
2504 Set output for log messages.
2507 The filter exports recognized speech as the frame metadata @code{lavfi.asr.text}.
2512 Display time domain statistical information about the audio channels.
2513 Statistics are calculated and displayed for each audio channel and,
2514 where applicable, an overall figure is also given.
2516 It accepts the following option:
2519 Short window length in seconds, used for peak and trough RMS measurement.
2520 Default is @code{0.05} (50 milliseconds). Allowed range is @code{[0.01 - 10]}.
2524 Set metadata injection. All the metadata keys are prefixed with @code{lavfi.astats.X},
2525 where @code{X} is channel number starting from 1 or string @code{Overall}. Default is
2528 Available keys for each channel are:
2574 For example full key look like this @code{lavfi.astats.1.DC_offset} or
2575 this @code{lavfi.astats.Overall.Peak_count}.
2577 For description what each key means read below.
2580 Set number of frame after which stats are going to be recalculated.
2581 Default is disabled.
2583 @item measure_perchannel
2584 Select the entries which need to be measured per channel. The metadata keys can
2585 be used as flags, default is @option{all} which measures everything.
2586 @option{none} disables all per channel measurement.
2588 @item measure_overall
2589 Select the entries which need to be measured overall. The metadata keys can
2590 be used as flags, default is @option{all} which measures everything.
2591 @option{none} disables all overall measurement.
2595 A description of each shown parameter follows:
2599 Mean amplitude displacement from zero.
2602 Minimal sample level.
2605 Maximal sample level.
2607 @item Min difference
2608 Minimal difference between two consecutive samples.
2610 @item Max difference
2611 Maximal difference between two consecutive samples.
2613 @item Mean difference
2614 Mean difference between two consecutive samples.
2615 The average of each difference between two consecutive samples.
2617 @item RMS difference
2618 Root Mean Square difference between two consecutive samples.
2622 Standard peak and RMS level measured in dBFS.
2626 Peak and trough values for RMS level measured over a short window.
2629 Standard ratio of peak to RMS level (note: not in dB).
2632 Flatness (i.e. consecutive samples with the same value) of the signal at its peak levels
2633 (i.e. either @var{Min level} or @var{Max level}).
2636 Number of occasions (not the number of samples) that the signal attained either
2637 @var{Min level} or @var{Max level}.
2639 @item Noise floor dB
2640 Minimum local peak measured in dBFS over a short window.
2642 @item Noise floor count
2643 Number of occasions (not the number of samples) that the signal attained
2647 Overall bit depth of audio. Number of bits used for each sample.
2650 Measured dynamic range of audio in dB.
2652 @item Zero crossings
2653 Number of points where the waveform crosses the zero level axis.
2655 @item Zero crossings rate
2656 Rate of Zero crossings and number of audio samples.
2660 Boost subwoofer frequencies.
2662 The filter accepts the following options:
2666 Set dry gain, how much of original signal is kept. Allowed range is from 0 to 1.
2667 Default value is 0.7.
2670 Set wet gain, how much of filtered signal is kept. Allowed range is from 0 to 1.
2671 Default value is 0.7.
2674 Set delay line decay gain value. Allowed range is from 0 to 1.
2675 Default value is 0.7.
2678 Set delay line feedback gain value. Allowed range is from 0 to 1.
2679 Default value is 0.9.
2682 Set cutoff frequency in Hertz. Allowed range is 50 to 900.
2683 Default value is 100.
2686 Set slope amount for cutoff frequency. Allowed range is 0.0001 to 1.
2687 Default value is 0.5.
2690 Set delay. Allowed range is from 1 to 100.
2691 Default value is 20.
2694 @subsection Commands
2696 This filter supports the all above options as @ref{commands}.
2699 Cut subwoofer frequencies.
2701 This filter allows to set custom, steeper
2702 roll off than highpass filter, and thus is able to more attenuate
2703 frequency content in stop-band.
2705 The filter accepts the following options:
2709 Set cutoff frequency in Hertz. Allowed range is 2 to 200.
2710 Default value is 20.
2713 Set filter order. Available values are from 3 to 20.
2714 Default value is 10.
2717 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2720 @subsection Commands
2722 This filter supports the all above options as @ref{commands}.
2725 Cut super frequencies.
2727 The filter accepts the following options:
2731 Set cutoff frequency in Hertz. Allowed range is 20000 to 192000.
2732 Default value is 20000.
2735 Set filter order. Available values are from 3 to 20.
2736 Default value is 10.
2739 Set input gain level. Allowed range is from 0 to 1. Default value is 1.
2742 @subsection Commands
2744 This filter supports the all above options as @ref{commands}.
2747 Apply high order Butterworth band-pass filter.
2749 The filter accepts the following options:
2753 Set center frequency in Hertz. Allowed range is 2 to 999999.
2754 Default value is 1000.
2757 Set filter order. Available values are from 4 to 20.
2761 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2764 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2767 @subsection Commands
2769 This filter supports the all above options as @ref{commands}.
2772 Apply high order Butterworth band-stop filter.
2774 The filter accepts the following options:
2778 Set center frequency in Hertz. Allowed range is 2 to 999999.
2779 Default value is 1000.
2782 Set filter order. Available values are from 4 to 20.
2786 Set Q-factor. Allowed range is from 0.01 to 100. Default value is 1.
2789 Set input gain level. Allowed range is from 0 to 2. Default value is 1.
2792 @subsection Commands
2794 This filter supports the all above options as @ref{commands}.
2800 The filter accepts exactly one parameter, the audio tempo. If not
2801 specified then the filter will assume nominal 1.0 tempo. Tempo must
2802 be in the [0.5, 100.0] range.
2804 Note that tempo greater than 2 will skip some samples rather than
2805 blend them in. If for any reason this is a concern it is always
2806 possible to daisy-chain several instances of atempo to achieve the
2807 desired product tempo.
2809 @subsection Examples
2813 Slow down audio to 80% tempo:
2819 To speed up audio to 300% tempo:
2825 To speed up audio to 300% tempo by daisy-chaining two atempo instances:
2827 atempo=sqrt(3),atempo=sqrt(3)
2831 @subsection Commands
2833 This filter supports the following commands:
2836 Change filter tempo scale factor.
2837 Syntax for the command is : "@var{tempo}"
2842 Trim the input so that the output contains one continuous subpart of the input.
2844 It accepts the following parameters:
2847 Timestamp (in seconds) of the start of the section to keep. I.e. the audio
2848 sample with the timestamp @var{start} will be the first sample in the output.
2851 Specify time of the first audio sample that will be dropped, i.e. the
2852 audio sample immediately preceding the one with the timestamp @var{end} will be
2853 the last sample in the output.
2856 Same as @var{start}, except this option sets the start timestamp in samples
2860 Same as @var{end}, except this option sets the end timestamp in samples instead
2864 The maximum duration of the output in seconds.
2867 The number of the first sample that should be output.
2870 The number of the first sample that should be dropped.
2873 @option{start}, @option{end}, and @option{duration} are expressed as time
2874 duration specifications; see
2875 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}.
2877 Note that the first two sets of the start/end options and the @option{duration}
2878 option look at the frame timestamp, while the _sample options simply count the
2879 samples that pass through the filter. So start/end_pts and start/end_sample will
2880 give different results when the timestamps are wrong, inexact or do not start at
2881 zero. Also note that this filter does not modify the timestamps. If you wish
2882 to have the output timestamps start at zero, insert the asetpts filter after the
2885 If multiple start or end options are set, this filter tries to be greedy and
2886 keep all samples that match at least one of the specified constraints. To keep
2887 only the part that matches all the constraints at once, chain multiple atrim
2890 The defaults are such that all the input is kept. So it is possible to set e.g.
2891 just the end values to keep everything before the specified time.
2896 Drop everything except the second minute of input:
2898 ffmpeg -i INPUT -af atrim=60:120
2902 Keep only the first 1000 samples:
2904 ffmpeg -i INPUT -af atrim=end_sample=1000
2909 @section axcorrelate
2910 Calculate normalized cross-correlation between two input audio streams.
2912 Resulted samples are always between -1 and 1 inclusive.
2913 If result is 1 it means two input samples are highly correlated in that selected segment.
2914 Result 0 means they are not correlated at all.
2915 If result is -1 it means two input samples are out of phase, which means they cancel each
2918 The filter accepts the following options:
2922 Set size of segment over which cross-correlation is calculated.
2923 Default is 256. Allowed range is from 2 to 131072.
2926 Set algorithm for cross-correlation. Can be @code{slow} or @code{fast}.
2927 Default is @code{slow}. Fast algorithm assumes mean values over any given segment
2928 are always zero and thus need much less calculations to make.
2929 This is generally not true, but is valid for typical audio streams.
2932 @subsection Examples
2936 Calculate correlation between channels in stereo audio stream:
2938 ffmpeg -i stereo.wav -af channelsplit,axcorrelate=size=1024:algo=fast correlation.wav
2944 Apply a two-pole Butterworth band-pass filter with central
2945 frequency @var{frequency}, and (3dB-point) band-width width.
2946 The @var{csg} option selects a constant skirt gain (peak gain = Q)
2947 instead of the default: constant 0dB peak gain.
2948 The filter roll off at 6dB per octave (20dB per decade).
2950 The filter accepts the following options:
2954 Set the filter's central frequency. Default is @code{3000}.
2957 Constant skirt gain if set to 1. Defaults to 0.
2960 Set method to specify band-width of filter.
2975 Specify the band-width of a filter in width_type units.
2978 How much to use filtered signal in output. Default is 1.
2979 Range is between 0 and 1.
2982 Specify which channels to filter, by default all available are filtered.
2985 Normalize biquad coefficients, by default is disabled.
2986 Enabling it will normalize magnitude response at DC to 0dB.
2989 Set transform type of IIR filter.
2998 Set precison of filtering.
3001 Pick automatic sample format depending on surround filters.
3003 Always use signed 16-bit.
3005 Always use signed 32-bit.
3007 Always use float 32-bit.
3009 Always use float 64-bit.
3013 @subsection Commands
3015 This filter supports the following commands:
3018 Change bandpass frequency.
3019 Syntax for the command is : "@var{frequency}"
3022 Change bandpass width_type.
3023 Syntax for the command is : "@var{width_type}"
3026 Change bandpass width.
3027 Syntax for the command is : "@var{width}"
3030 Change bandpass mix.
3031 Syntax for the command is : "@var{mix}"
3036 Apply a two-pole Butterworth band-reject filter with central
3037 frequency @var{frequency}, and (3dB-point) band-width @var{width}.
3038 The filter roll off at 6dB per octave (20dB per decade).
3040 The filter accepts the following options:
3044 Set the filter's central frequency. Default is @code{3000}.
3047 Set method to specify band-width of filter.
3062 Specify the band-width of a filter in width_type units.
3065 How much to use filtered signal in output. Default is 1.
3066 Range is between 0 and 1.
3069 Specify which channels to filter, by default all available are filtered.
3072 Normalize biquad coefficients, by default is disabled.
3073 Enabling it will normalize magnitude response at DC to 0dB.
3076 Set transform type of IIR filter.
3085 Set precison of filtering.
3088 Pick automatic sample format depending on surround filters.
3090 Always use signed 16-bit.
3092 Always use signed 32-bit.
3094 Always use float 32-bit.
3096 Always use float 64-bit.
3100 @subsection Commands
3102 This filter supports the following commands:
3105 Change bandreject frequency.
3106 Syntax for the command is : "@var{frequency}"
3109 Change bandreject width_type.
3110 Syntax for the command is : "@var{width_type}"
3113 Change bandreject width.
3114 Syntax for the command is : "@var{width}"
3117 Change bandreject mix.
3118 Syntax for the command is : "@var{mix}"
3121 @section bass, lowshelf
3123 Boost or cut the bass (lower) frequencies of the audio using a two-pole
3124 shelving filter with a response similar to that of a standard
3125 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
3127 The filter accepts the following options:
3131 Give the gain at 0 Hz. Its useful range is about -20
3132 (for a large cut) to +20 (for a large boost).
3133 Beware of clipping when using a positive gain.
3136 Set the filter's central frequency and so can be used
3137 to extend or reduce the frequency range to be boosted or cut.
3138 The default value is @code{100} Hz.
3141 Set method to specify band-width of filter.
3156 Determine how steep is the filter's shelf transition.
3159 Set number of poles. Default is 2.
3162 How much to use filtered signal in output. Default is 1.
3163 Range is between 0 and 1.
3166 Specify which channels to filter, by default all available are filtered.
3169 Normalize biquad coefficients, by default is disabled.
3170 Enabling it will normalize magnitude response at DC to 0dB.
3173 Set transform type of IIR filter.
3182 Set precison of filtering.
3185 Pick automatic sample format depending on surround filters.
3187 Always use signed 16-bit.
3189 Always use signed 32-bit.
3191 Always use float 32-bit.
3193 Always use float 64-bit.
3197 @subsection Commands
3199 This filter supports the following commands:
3202 Change bass frequency.
3203 Syntax for the command is : "@var{frequency}"
3206 Change bass width_type.
3207 Syntax for the command is : "@var{width_type}"
3211 Syntax for the command is : "@var{width}"
3215 Syntax for the command is : "@var{gain}"
3219 Syntax for the command is : "@var{mix}"
3224 Apply a biquad IIR filter with the given coefficients.
3225 Where @var{b0}, @var{b1}, @var{b2} and @var{a0}, @var{a1}, @var{a2}
3226 are the numerator and denominator coefficients respectively.
3227 and @var{channels}, @var{c} specify which channels to filter, by default all
3228 available are filtered.
3230 @subsection Commands
3232 This filter supports the following commands:
3240 Change biquad parameter.
3241 Syntax for the command is : "@var{value}"
3244 How much to use filtered signal in output. Default is 1.
3245 Range is between 0 and 1.
3248 Specify which channels to filter, by default all available are filtered.
3251 Normalize biquad coefficients, by default is disabled.
3252 Enabling it will normalize magnitude response at DC to 0dB.
3255 Set transform type of IIR filter.
3264 Set precison of filtering.
3267 Pick automatic sample format depending on surround filters.
3269 Always use signed 16-bit.
3271 Always use signed 32-bit.
3273 Always use float 32-bit.
3275 Always use float 64-bit.
3280 Bauer stereo to binaural transformation, which improves headphone listening of
3281 stereo audio records.
3283 To enable compilation of this filter you need to configure FFmpeg with
3284 @code{--enable-libbs2b}.
3286 It accepts the following parameters:
3290 Pre-defined crossfeed level.
3294 Default level (fcut=700, feed=50).
3297 Chu Moy circuit (fcut=700, feed=60).
3300 Jan Meier circuit (fcut=650, feed=95).
3305 Cut frequency (in Hz).
3314 Remap input channels to new locations.
3316 It accepts the following parameters:
3319 Map channels from input to output. The argument is a '|'-separated list of
3320 mappings, each in the @code{@var{in_channel}-@var{out_channel}} or
3321 @var{in_channel} form. @var{in_channel} can be either the name of the input
3322 channel (e.g. FL for front left) or its index in the input channel layout.
3323 @var{out_channel} is the name of the output channel or its index in the output
3324 channel layout. If @var{out_channel} is not given then it is implicitly an
3325 index, starting with zero and increasing by one for each mapping.
3327 @item channel_layout
3328 The channel layout of the output stream.
3331 If no mapping is present, the filter will implicitly map input channels to
3332 output channels, preserving indices.
3334 @subsection Examples
3338 For example, assuming a 5.1+downmix input MOV file,
3340 ffmpeg -i in.mov -filter 'channelmap=map=DL-FL|DR-FR' out.wav
3342 will create an output WAV file tagged as stereo from the downmix channels of
3346 To fix a 5.1 WAV improperly encoded in AAC's native channel order
3348 ffmpeg -i in.wav -filter 'channelmap=1|2|0|5|3|4:5.1' out.wav
3352 @section channelsplit
3354 Split each channel from an input audio stream into a separate output stream.
3356 It accepts the following parameters:
3358 @item channel_layout
3359 The channel layout of the input stream. The default is "stereo".
3361 A channel layout describing the channels to be extracted as separate output streams
3362 or "all" to extract each input channel as a separate stream. The default is "all".
3364 Choosing channels not present in channel layout in the input will result in an error.
3367 @subsection Examples
3371 For example, assuming a stereo input MP3 file,
3373 ffmpeg -i in.mp3 -filter_complex channelsplit out.mkv
3375 will create an output Matroska file with two audio streams, one containing only
3376 the left channel and the other the right channel.
3379 Split a 5.1 WAV file into per-channel files:
3381 ffmpeg -i in.wav -filter_complex
3382 'channelsplit=channel_layout=5.1[FL][FR][FC][LFE][SL][SR]'
3383 -map '[FL]' front_left.wav -map '[FR]' front_right.wav -map '[FC]'
3384 front_center.wav -map '[LFE]' lfe.wav -map '[SL]' side_left.wav -map '[SR]'
3389 Extract only LFE from a 5.1 WAV file:
3391 ffmpeg -i in.wav -filter_complex 'channelsplit=channel_layout=5.1:channels=LFE[LFE]'
3392 -map '[LFE]' lfe.wav
3397 Add a chorus effect to the audio.
3399 Can make a single vocal sound like a chorus, but can also be applied to instrumentation.
3401 Chorus resembles an echo effect with a short delay, but whereas with echo the delay is
3402 constant, with chorus, it is varied using using sinusoidal or triangular modulation.
3403 The modulation depth defines the range the modulated delay is played before or after
3404 the delay. Hence the delayed sound will sound slower or faster, that is the delayed
3405 sound tuned around the original one, like in a chorus where some vocals are slightly
3408 It accepts the following parameters:
3411 Set input gain. Default is 0.4.
3414 Set output gain. Default is 0.4.
3417 Set delays. A typical delay is around 40ms to 60ms.
3429 @subsection Examples
3435 chorus=0.7:0.9:55:0.4:0.25:2
3441 chorus=0.6:0.9:50|60:0.4|0.32:0.25|0.4:2|1.3
3445 Fuller sounding chorus with three delays:
3447 chorus=0.5:0.9:50|60|40:0.4|0.32|0.3:0.25|0.4|0.3:2|2.3|1.3
3452 Compress or expand the audio's dynamic range.
3454 It accepts the following parameters:
3460 A list of times in seconds for each channel over which the instantaneous level
3461 of the input signal is averaged to determine its volume. @var{attacks} refers to
3462 increase of volume and @var{decays} refers to decrease of volume. For most
3463 situations, the attack time (response to the audio getting louder) should be
3464 shorter than the decay time, because the human ear is more sensitive to sudden
3465 loud audio than sudden soft audio. A typical value for attack is 0.3 seconds and
3466 a typical value for decay is 0.8 seconds.
3467 If specified number of attacks & decays is lower than number of channels, the last
3468 set attack/decay will be used for all remaining channels.
3471 A list of points for the transfer function, specified in dB relative to the
3472 maximum possible signal amplitude. Each key points list must be defined using
3473 the following syntax: @code{x0/y0|x1/y1|x2/y2|....} or
3474 @code{x0/y0 x1/y1 x2/y2 ....}
3476 The input values must be in strictly increasing order but the transfer function
3477 does not have to be monotonically rising. The point @code{0/0} is assumed but
3478 may be overridden (by @code{0/out-dBn}). Typical values for the transfer
3479 function are @code{-70/-70|-60/-20|1/0}.
3482 Set the curve radius in dB for all joints. It defaults to 0.01.
3485 Set the additional gain in dB to be applied at all points on the transfer
3486 function. This allows for easy adjustment of the overall gain.
3490 Set an initial volume, in dB, to be assumed for each channel when filtering
3491 starts. This permits the user to supply a nominal level initially, so that, for
3492 example, a very large gain is not applied to initial signal levels before the
3493 companding has begun to operate. A typical value for audio which is initially
3494 quiet is -90 dB. It defaults to 0.
3497 Set a delay, in seconds. The input audio is analyzed immediately, but audio is
3498 delayed before being fed to the volume adjuster. Specifying a delay
3499 approximately equal to the attack/decay times allows the filter to effectively
3500 operate in predictive rather than reactive mode. It defaults to 0.
3504 @subsection Examples
3508 Make music with both quiet and loud passages suitable for listening to in a
3511 compand=.3|.3:1|1:-90/-60|-60/-40|-40/-30|-20/-20:6:0:-90:0.2
3514 Another example for audio with whisper and explosion parts:
3516 compand=0|0:1|1:-90/-900|-70/-70|-30/-9|0/-3:6:0:0:0
3520 A noise gate for when the noise is at a lower level than the signal:
3522 compand=.1|.1:.2|.2:-900/-900|-50.1/-900|-50/-50:.01:0:-90:.1
3526 Here is another noise gate, this time for when the noise is at a higher level
3527 than the signal (making it, in some ways, similar to squelch):
3529 compand=.1|.1:.1|.1:-45.1/-45.1|-45/-900|0/-900:.01:45:-90:.1
3533 2:1 compression starting at -6dB:
3535 compand=points=-80/-80|-6/-6|0/-3.8|20/3.5
3539 2:1 compression starting at -9dB:
3541 compand=points=-80/-80|-9/-9|0/-5.3|20/2.9
3545 2:1 compression starting at -12dB:
3547 compand=points=-80/-80|-12/-12|0/-6.8|20/1.9
3551 2:1 compression starting at -18dB:
3553 compand=points=-80/-80|-18/-18|0/-9.8|20/0.7
3557 3:1 compression starting at -15dB:
3559 compand=points=-80/-80|-15/-15|0/-10.8|20/-5.2
3565 compand=points=-80/-105|-62/-80|-15.4/-15.4|0/-12|20/-7.6
3571 compand=attacks=0:points=-80/-169|-54/-80|-49.5/-64.6|-41.1/-41.1|-25.8/-15|-10.8/-4.5|0/0|20/8.3
3575 Hard limiter at -6dB:
3577 compand=attacks=0:points=-80/-80|-6/-6|20/-6
3581 Hard limiter at -12dB:
3583 compand=attacks=0:points=-80/-80|-12/-12|20/-12
3587 Hard noise gate at -35 dB:
3589 compand=attacks=0:points=-80/-115|-35.1/-80|-35/-35|20/20
3595 compand=attacks=0:points=-80/-80|-12.4/-12.4|-6/-8|0/-6.8|20/-2.8
3599 @section compensationdelay
3601 Compensation Delay Line is a metric based delay to compensate differing
3602 positions of microphones or speakers.
3604 For example, you have recorded guitar with two microphones placed in
3605 different locations. Because the front of sound wave has fixed speed in
3606 normal conditions, the phasing of microphones can vary and depends on
3607 their location and interposition. The best sound mix can be achieved when
3608 these microphones are in phase (synchronized). Note that a distance of
3609 ~30 cm between microphones makes one microphone capture the signal in
3610 antiphase to the other microphone. That makes the final mix sound moody.
3611 This filter helps to solve phasing problems by adding different delays
3612 to each microphone track and make them synchronized.
3614 The best result can be reached when you take one track as base and
3615 synchronize other tracks one by one with it.
3616 Remember that synchronization/delay tolerance depends on sample rate, too.
3617 Higher sample rates will give more tolerance.
3619 The filter accepts the following parameters:
3623 Set millimeters distance. This is compensation distance for fine tuning.
3627 Set cm distance. This is compensation distance for tightening distance setup.
3631 Set meters distance. This is compensation distance for hard distance setup.
3635 Set dry amount. Amount of unprocessed (dry) signal.
3639 Set wet amount. Amount of processed (wet) signal.
3643 Set temperature in degrees Celsius. This is the temperature of the environment.
3648 Apply headphone crossfeed filter.
3650 Crossfeed is the process of blending the left and right channels of stereo
3652 It is mainly used to reduce extreme stereo separation of low frequencies.
3654 The intent is to produce more speaker like sound to the listener.
3656 The filter accepts the following options:
3660 Set strength of crossfeed. Default is 0.2. Allowed range is from 0 to 1.
3661 This sets gain of low shelf filter for side part of stereo image.
3662 Default is -6dB. Max allowed is -30db when strength is set to 1.
3665 Set soundstage wideness. Default is 0.5. Allowed range is from 0 to 1.
3666 This sets cut off frequency of low shelf filter. Default is cut off near
3667 1550 Hz. With range set to 1 cut off frequency is set to 2100 Hz.
3670 Set curve slope of low shelf filter. Default is 0.5.
3671 Allowed range is from 0.01 to 1.
3674 Set input gain. Default is 0.9.
3677 Set output gain. Default is 1.
3680 @subsection Commands
3682 This filter supports the all above options as @ref{commands}.
3684 @section crystalizer
3685 Simple algorithm for audio noise sharpening.
3687 This filter linearly increases differences betweeen each audio sample.
3689 The filter accepts the following options:
3693 Sets the intensity of effect (default: 2.0). Must be in range between -10.0 to 0
3694 (unchanged sound) to 10.0 (maximum effect).
3695 To inverse filtering use negative value.
3698 Enable clipping. By default is enabled.
3701 @subsection Commands
3703 This filter supports the all above options as @ref{commands}.
3706 Apply a DC shift to the audio.
3708 This can be useful to remove a DC offset (caused perhaps by a hardware problem
3709 in the recording chain) from the audio. The effect of a DC offset is reduced
3710 headroom and hence volume. The @ref{astats} filter can be used to determine if
3711 a signal has a DC offset.
3715 Set the DC shift, allowed range is [-1, 1]. It indicates the amount to shift
3719 Optional. It should have a value much less than 1 (e.g. 0.05 or 0.02) and is
3720 used to prevent clipping.
3725 Apply de-essing to the audio samples.
3729 Set intensity for triggering de-essing. Allowed range is from 0 to 1.
3733 Set amount of ducking on treble part of sound. Allowed range is from 0 to 1.
3737 How much of original frequency content to keep when de-essing. Allowed range is from 0 to 1.
3741 Set the output mode.
3743 It accepts the following values:
3746 Pass input unchanged.
3749 Pass ess filtered out.
3754 Default value is @var{o}.
3760 Measure audio dynamic range.
3762 DR values of 14 and higher is found in very dynamic material. DR of 8 to 13
3763 is found in transition material. And anything less that 8 have very poor dynamics
3764 and is very compressed.
3766 The filter accepts the following options:
3770 Set window length in seconds used to split audio into segments of equal length.
3771 Default is 3 seconds.
3775 Dynamic Audio Normalizer.
3777 This filter applies a certain amount of gain to the input audio in order
3778 to bring its peak magnitude to a target level (e.g. 0 dBFS). However, in
3779 contrast to more "simple" normalization algorithms, the Dynamic Audio
3780 Normalizer *dynamically* re-adjusts the gain factor to the input audio.
3781 This allows for applying extra gain to the "quiet" sections of the audio
3782 while avoiding distortions or clipping the "loud" sections. In other words:
3783 The Dynamic Audio Normalizer will "even out" the volume of quiet and loud
3784 sections, in the sense that the volume of each section is brought to the
3785 same target level. Note, however, that the Dynamic Audio Normalizer achieves
3786 this goal *without* applying "dynamic range compressing". It will retain 100%
3787 of the dynamic range *within* each section of the audio file.
3791 Set the frame length in milliseconds. In range from 10 to 8000 milliseconds.
3792 Default is 500 milliseconds.
3793 The Dynamic Audio Normalizer processes the input audio in small chunks,
3794 referred to as frames. This is required, because a peak magnitude has no
3795 meaning for just a single sample value. Instead, we need to determine the
3796 peak magnitude for a contiguous sequence of sample values. While a "standard"
3797 normalizer would simply use the peak magnitude of the complete file, the
3798 Dynamic Audio Normalizer determines the peak magnitude individually for each
3799 frame. The length of a frame is specified in milliseconds. By default, the
3800 Dynamic Audio Normalizer uses a frame length of 500 milliseconds, which has
3801 been found to give good results with most files.
3802 Note that the exact frame length, in number of samples, will be determined
3803 automatically, based on the sampling rate of the individual input audio file.
3806 Set the Gaussian filter window size. In range from 3 to 301, must be odd
3807 number. Default is 31.
3808 Probably the most important parameter of the Dynamic Audio Normalizer is the
3809 @code{window size} of the Gaussian smoothing filter. The filter's window size
3810 is specified in frames, centered around the current frame. For the sake of
3811 simplicity, this must be an odd number. Consequently, the default value of 31
3812 takes into account the current frame, as well as the 15 preceding frames and
3813 the 15 subsequent frames. Using a larger window results in a stronger
3814 smoothing effect and thus in less gain variation, i.e. slower gain
3815 adaptation. Conversely, using a smaller window results in a weaker smoothing
3816 effect and thus in more gain variation, i.e. faster gain adaptation.
3817 In other words, the more you increase this value, the more the Dynamic Audio
3818 Normalizer will behave like a "traditional" normalization filter. On the
3819 contrary, the more you decrease this value, the more the Dynamic Audio
3820 Normalizer will behave like a dynamic range compressor.
3823 Set the target peak value. This specifies the highest permissible magnitude
3824 level for the normalized audio input. This filter will try to approach the
3825 target peak magnitude as closely as possible, but at the same time it also
3826 makes sure that the normalized signal will never exceed the peak magnitude.
3827 A frame's maximum local gain factor is imposed directly by the target peak
3828 magnitude. The default value is 0.95 and thus leaves a headroom of 5%*.
3829 It is not recommended to go above this value.
3832 Set the maximum gain factor. In range from 1.0 to 100.0. Default is 10.0.
3833 The Dynamic Audio Normalizer determines the maximum possible (local) gain
3834 factor for each input frame, i.e. the maximum gain factor that does not
3835 result in clipping or distortion. The maximum gain factor is determined by
3836 the frame's highest magnitude sample. However, the Dynamic Audio Normalizer
3837 additionally bounds the frame's maximum gain factor by a predetermined
3838 (global) maximum gain factor. This is done in order to avoid excessive gain
3839 factors in "silent" or almost silent frames. By default, the maximum gain
3840 factor is 10.0, For most inputs the default value should be sufficient and
3841 it usually is not recommended to increase this value. Though, for input
3842 with an extremely low overall volume level, it may be necessary to allow even
3843 higher gain factors. Note, however, that the Dynamic Audio Normalizer does
3844 not simply apply a "hard" threshold (i.e. cut off values above the threshold).
3845 Instead, a "sigmoid" threshold function will be applied. This way, the
3846 gain factors will smoothly approach the threshold value, but never exceed that
3850 Set the target RMS. In range from 0.0 to 1.0. Default is 0.0 - disabled.
3851 By default, the Dynamic Audio Normalizer performs "peak" normalization.
3852 This means that the maximum local gain factor for each frame is defined
3853 (only) by the frame's highest magnitude sample. This way, the samples can
3854 be amplified as much as possible without exceeding the maximum signal
3855 level, i.e. without clipping. Optionally, however, the Dynamic Audio
3856 Normalizer can also take into account the frame's root mean square,
3857 abbreviated RMS. In electrical engineering, the RMS is commonly used to
3858 determine the power of a time-varying signal. It is therefore considered
3859 that the RMS is a better approximation of the "perceived loudness" than
3860 just looking at the signal's peak magnitude. Consequently, by adjusting all
3861 frames to a constant RMS value, a uniform "perceived loudness" can be
3862 established. If a target RMS value has been specified, a frame's local gain
3863 factor is defined as the factor that would result in exactly that RMS value.
3864 Note, however, that the maximum local gain factor is still restricted by the
3865 frame's highest magnitude sample, in order to prevent clipping.
3868 Enable channels coupling. By default is enabled.
3869 By default, the Dynamic Audio Normalizer will amplify all channels by the same
3870 amount. This means the same gain factor will be applied to all channels, i.e.
3871 the maximum possible gain factor is determined by the "loudest" channel.
3872 However, in some recordings, it may happen that the volume of the different
3873 channels is uneven, e.g. one channel may be "quieter" than the other one(s).
3874 In this case, this option can be used to disable the channel coupling. This way,
3875 the gain factor will be determined independently for each channel, depending
3876 only on the individual channel's highest magnitude sample. This allows for
3877 harmonizing the volume of the different channels.
3880 Enable DC bias correction. By default is disabled.
3881 An audio signal (in the time domain) is a sequence of sample values.
3882 In the Dynamic Audio Normalizer these sample values are represented in the
3883 -1.0 to 1.0 range, regardless of the original input format. Normally, the
3884 audio signal, or "waveform", should be centered around the zero point.
3885 That means if we calculate the mean value of all samples in a file, or in a
3886 single frame, then the result should be 0.0 or at least very close to that
3887 value. If, however, there is a significant deviation of the mean value from
3888 0.0, in either positive or negative direction, this is referred to as a
3889 DC bias or DC offset. Since a DC bias is clearly undesirable, the Dynamic
3890 Audio Normalizer provides optional DC bias correction.
3891 With DC bias correction enabled, the Dynamic Audio Normalizer will determine
3892 the mean value, or "DC correction" offset, of each input frame and subtract
3893 that value from all of the frame's sample values which ensures those samples
3894 are centered around 0.0 again. Also, in order to avoid "gaps" at the frame
3895 boundaries, the DC correction offset values will be interpolated smoothly
3896 between neighbouring frames.
3898 @item altboundary, b
3899 Enable alternative boundary mode. By default is disabled.
3900 The Dynamic Audio Normalizer takes into account a certain neighbourhood
3901 around each frame. This includes the preceding frames as well as the
3902 subsequent frames. However, for the "boundary" frames, located at the very
3903 beginning and at the very end of the audio file, not all neighbouring
3904 frames are available. In particular, for the first few frames in the audio
3905 file, the preceding frames are not known. And, similarly, for the last few
3906 frames in the audio file, the subsequent frames are not known. Thus, the
3907 question arises which gain factors should be assumed for the missing frames
3908 in the "boundary" region. The Dynamic Audio Normalizer implements two modes
3909 to deal with this situation. The default boundary mode assumes a gain factor
3910 of exactly 1.0 for the missing frames, resulting in a smooth "fade in" and
3911 "fade out" at the beginning and at the end of the input, respectively.
3914 Set the compress factor. In range from 0.0 to 30.0. Default is 0.0.
3915 By default, the Dynamic Audio Normalizer does not apply "traditional"
3916 compression. This means that signal peaks will not be pruned and thus the
3917 full dynamic range will be retained within each local neighbourhood. However,
3918 in some cases it may be desirable to combine the Dynamic Audio Normalizer's
3919 normalization algorithm with a more "traditional" compression.
3920 For this purpose, the Dynamic Audio Normalizer provides an optional compression
3921 (thresholding) function. If (and only if) the compression feature is enabled,
3922 all input frames will be processed by a soft knee thresholding function prior
3923 to the actual normalization process. Put simply, the thresholding function is
3924 going to prune all samples whose magnitude exceeds a certain threshold value.
3925 However, the Dynamic Audio Normalizer does not simply apply a fixed threshold
3926 value. Instead, the threshold value will be adjusted for each individual
3928 In general, smaller parameters result in stronger compression, and vice versa.
3929 Values below 3.0 are not recommended, because audible distortion may appear.
3932 Set the target threshold value. This specifies the lowest permissible
3933 magnitude level for the audio input which will be normalized.
3934 If input frame volume is above this value frame will be normalized.
3935 Otherwise frame may not be normalized at all. The default value is set
3936 to 0, which means all input frames will be normalized.
3937 This option is mostly useful if digital noise is not wanted to be amplified.
3940 @subsection Commands
3942 This filter supports the all above options as @ref{commands}.
3946 Make audio easier to listen to on headphones.
3948 This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
3949 so that when listened to on headphones the stereo image is moved from
3950 inside your head (standard for headphones) to outside and in front of
3951 the listener (standard for speakers).
3957 Apply a two-pole peaking equalisation (EQ) filter. With this
3958 filter, the signal-level at and around a selected frequency can
3959 be increased or decreased, whilst (unlike bandpass and bandreject
3960 filters) that at all other frequencies is unchanged.
3962 In order to produce complex equalisation curves, this filter can
3963 be given several times, each with a different central frequency.
3965 The filter accepts the following options:
3969 Set the filter's central frequency in Hz.
3972 Set method to specify band-width of filter.
3987 Specify the band-width of a filter in width_type units.
3990 Set the required gain or attenuation in dB.
3991 Beware of clipping when using a positive gain.
3994 How much to use filtered signal in output. Default is 1.
3995 Range is between 0 and 1.
3998 Specify which channels to filter, by default all available are filtered.
4001 Normalize biquad coefficients, by default is disabled.
4002 Enabling it will normalize magnitude response at DC to 0dB.
4005 Set transform type of IIR filter.
4014 Set precison of filtering.
4017 Pick automatic sample format depending on surround filters.
4019 Always use signed 16-bit.
4021 Always use signed 32-bit.
4023 Always use float 32-bit.
4025 Always use float 64-bit.
4029 @subsection Examples
4032 Attenuate 10 dB at 1000 Hz, with a bandwidth of 200 Hz:
4034 equalizer=f=1000:t=h:width=200:g=-10
4038 Apply 2 dB gain at 1000 Hz with Q 1 and attenuate 5 dB at 100 Hz with Q 2:
4040 equalizer=f=1000:t=q:w=1:g=2,equalizer=f=100:t=q:w=2:g=-5
4044 @subsection Commands
4046 This filter supports the following commands:
4049 Change equalizer frequency.
4050 Syntax for the command is : "@var{frequency}"
4053 Change equalizer width_type.
4054 Syntax for the command is : "@var{width_type}"
4057 Change equalizer width.
4058 Syntax for the command is : "@var{width}"
4061 Change equalizer gain.
4062 Syntax for the command is : "@var{gain}"
4065 Change equalizer mix.
4066 Syntax for the command is : "@var{mix}"
4069 @section extrastereo
4071 Linearly increases the difference between left and right channels which
4072 adds some sort of "live" effect to playback.
4074 The filter accepts the following options:
4078 Sets the difference coefficient (default: 2.5). 0.0 means mono sound
4079 (average of both channels), with 1.0 sound will be unchanged, with
4080 -1.0 left and right channels will be swapped.
4083 Enable clipping. By default is enabled.
4086 @subsection Commands
4088 This filter supports the all above options as @ref{commands}.
4090 @section firequalizer
4091 Apply FIR Equalization using arbitrary frequency response.
4093 The filter accepts the following option:
4097 Set gain curve equation (in dB). The expression can contain variables:
4100 the evaluated frequency
4104 channel number, set to 0 when multichannels evaluation is disabled
4106 channel id, see libavutil/channel_layout.h, set to the first channel id when
4107 multichannels evaluation is disabled
4111 channel_layout, see libavutil/channel_layout.h
4116 @item gain_interpolate(f)
4117 interpolate gain on frequency f based on gain_entry
4118 @item cubic_interpolate(f)
4119 same as gain_interpolate, but smoother
4121 This option is also available as command. Default is @code{gain_interpolate(f)}.
4124 Set gain entry for gain_interpolate function. The expression can
4128 store gain entry at frequency f with value g
4130 This option is also available as command.
4133 Set filter delay in seconds. Higher value means more accurate.
4134 Default is @code{0.01}.
4137 Set filter accuracy in Hz. Lower value means more accurate.
4138 Default is @code{5}.
4141 Set window function. Acceptable values are:
4144 rectangular window, useful when gain curve is already smooth
4146 hann window (default)
4152 3-terms continuous 1st derivative nuttall window
4154 minimum 3-terms discontinuous nuttall window
4156 4-terms continuous 1st derivative nuttall window
4158 minimum 4-terms discontinuous nuttall (blackman-nuttall) window
4160 blackman-harris window
4166 If enabled, use fixed number of audio samples. This improves speed when
4167 filtering with large delay. Default is disabled.
4170 Enable multichannels evaluation on gain. Default is disabled.
4173 Enable zero phase mode by subtracting timestamp to compensate delay.
4174 Default is disabled.
4177 Set scale used by gain. Acceptable values are:
4180 linear frequency, linear gain
4182 linear frequency, logarithmic (in dB) gain (default)
4184 logarithmic (in octave scale where 20 Hz is 0) frequency, linear gain
4186 logarithmic frequency, logarithmic gain
4190 Set file for dumping, suitable for gnuplot.
4193 Set scale for dumpfile. Acceptable values are same with scale option.
4197 Enable 2-channel convolution using complex FFT. This improves speed significantly.
4198 Default is disabled.
4201 Enable minimum phase impulse response. Default is disabled.
4204 @subsection Examples
4209 firequalizer=gain='if(lt(f,1000), 0, -INF)'
4212 lowpass at 1000 Hz with gain_entry:
4214 firequalizer=gain_entry='entry(1000,0); entry(1001, -INF)'
4217 custom equalization:
4219 firequalizer=gain_entry='entry(100,0); entry(400, -4); entry(1000, -6); entry(2000, 0)'
4222 higher delay with zero phase to compensate delay:
4224 firequalizer=delay=0.1:fixed=on:zero_phase=on
4227 lowpass on left channel, highpass on right channel:
4229 firequalizer=gain='if(eq(chid,1), gain_interpolate(f), if(eq(chid,2), gain_interpolate(1e6+f), 0))'
4230 :gain_entry='entry(1000, 0); entry(1001,-INF); entry(1e6+1000,0)':multi=on
4235 Apply a flanging effect to the audio.
4237 The filter accepts the following options:
4241 Set base delay in milliseconds. Range from 0 to 30. Default value is 0.
4244 Set added sweep delay in milliseconds. Range from 0 to 10. Default value is 2.
4247 Set percentage regeneration (delayed signal feedback). Range from -95 to 95.
4251 Set percentage of delayed signal mixed with original. Range from 0 to 100.
4252 Default value is 71.
4255 Set sweeps per second (Hz). Range from 0.1 to 10. Default value is 0.5.
4258 Set swept wave shape, can be @var{triangular} or @var{sinusoidal}.
4259 Default value is @var{sinusoidal}.
4262 Set swept wave percentage-shift for multi channel. Range from 0 to 100.
4263 Default value is 25.
4266 Set delay-line interpolation, @var{linear} or @var{quadratic}.
4267 Default is @var{linear}.
4271 Apply Haas effect to audio.
4273 Note that this makes most sense to apply on mono signals.
4274 With this filter applied to mono signals it give some directionality and
4275 stretches its stereo image.
4277 The filter accepts the following options:
4281 Set input level. By default is @var{1}, or 0dB
4284 Set output level. By default is @var{1}, or 0dB.
4287 Set gain applied to side part of signal. By default is @var{1}.
4290 Set kind of middle source. Can be one of the following:
4300 Pick middle part signal of stereo image.
4303 Pick side part signal of stereo image.
4307 Change middle phase. By default is disabled.
4310 Set left channel delay. By default is @var{2.05} milliseconds.
4313 Set left channel balance. By default is @var{-1}.
4316 Set left channel gain. By default is @var{1}.
4319 Change left phase. By default is disabled.
4322 Set right channel delay. By defaults is @var{2.12} milliseconds.
4325 Set right channel balance. By default is @var{1}.
4328 Set right channel gain. By default is @var{1}.
4331 Change right phase. By default is enabled.
4336 Decodes High Definition Compatible Digital (HDCD) data. A 16-bit PCM stream with
4337 embedded HDCD codes is expanded into a 20-bit PCM stream.
4339 The filter supports the Peak Extend and Low-level Gain Adjustment features
4340 of HDCD, and detects the Transient Filter flag.
4343 ffmpeg -i HDCD16.flac -af hdcd OUT24.flac
4346 When using the filter with wav, note the default encoding for wav is 16-bit,
4347 so the resulting 20-bit stream will be truncated back to 16-bit. Use something
4348 like @command{-acodec pcm_s24le} after the filter to get 24-bit PCM output.
4350 ffmpeg -i HDCD16.wav -af hdcd OUT16.wav
4351 ffmpeg -i HDCD16.wav -af hdcd -c:a pcm_s24le OUT24.wav
4354 The filter accepts the following options:
4357 @item disable_autoconvert
4358 Disable any automatic format conversion or resampling in the filter graph.
4360 @item process_stereo
4361 Process the stereo channels together. If target_gain does not match between
4362 channels, consider it invalid and use the last valid target_gain.
4365 Set the code detect timer period in ms.
4368 Always extend peaks above -3dBFS even if PE isn't signaled.
4371 Replace audio with a solid tone and adjust the amplitude to signal some
4372 specific aspect of the decoding process. The output file can be loaded in
4373 an audio editor alongside the original to aid analysis.
4375 @code{analyze_mode=pe:force_pe=true} can be used to see all samples above the PE level.
4382 Gain adjustment level at each sample
4384 Samples where peak extend occurs
4386 Samples where the code detect timer is active
4388 Samples where the target gain does not match between channels
4394 Apply head-related transfer functions (HRTFs) to create virtual
4395 loudspeakers around the user for binaural listening via headphones.
4396 The HRIRs are provided via additional streams, for each channel
4397 one stereo input stream is needed.
4399 The filter accepts the following options:
4403 Set mapping of input streams for convolution.
4404 The argument is a '|'-separated list of channel names in order as they
4405 are given as additional stream inputs for filter.
4406 This also specify number of input streams. Number of input streams
4407 must be not less than number of channels in first stream plus one.
4410 Set gain applied to audio. Value is in dB. Default is 0.
4413 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
4414 processing audio in time domain which is slow.
4415 @var{freq} is processing audio in frequency domain which is fast.
4416 Default is @var{freq}.
4419 Set custom gain for LFE channels. Value is in dB. Default is 0.
4422 Set size of frame in number of samples which will be processed at once.
4423 Default value is @var{1024}. Allowed range is from 1024 to 96000.
4426 Set format of hrir stream.
4427 Default value is @var{stereo}. Alternative value is @var{multich}.
4428 If value is set to @var{stereo}, number of additional streams should
4429 be greater or equal to number of input channels in first input stream.
4430 Also each additional stream should have stereo number of channels.
4431 If value is set to @var{multich}, number of additional streams should
4432 be exactly one. Also number of input channels of additional stream
4433 should be equal or greater than twice number of channels of first input
4437 @subsection Examples
4441 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4442 each amovie filter use stereo file with IR coefficients as input.
4443 The files give coefficients for each position of virtual loudspeaker:
4446 -filter_complex "amovie=azi_270_ele_0_DFC.wav[sr];amovie=azi_90_ele_0_DFC.wav[sl];amovie=azi_225_ele_0_DFC.wav[br];amovie=azi_135_ele_0_DFC.wav[bl];amovie=azi_0_ele_0_DFC.wav,asplit[fc][lfe];amovie=azi_35_ele_0_DFC.wav[fl];amovie=azi_325_ele_0_DFC.wav[fr];[0:a][fl][fr][fc][lfe][bl][br][sl][sr]headphone=FL|FR|FC|LFE|BL|BR|SL|SR"
4451 Full example using wav files as coefficients with amovie filters for 7.1 downmix,
4452 but now in @var{multich} @var{hrir} format.
4454 ffmpeg -i input.wav -filter_complex "amovie=minp.wav[hrirs];[0:a][hrirs]headphone=map=FL|FR|FC|LFE|BL|BR|SL|SR:hrir=multich"
4461 Apply a high-pass filter with 3dB point frequency.
4462 The filter can be either single-pole, or double-pole (the default).
4463 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4465 The filter accepts the following options:
4469 Set frequency in Hz. Default is 3000.
4472 Set number of poles. Default is 2.
4475 Set method to specify band-width of filter.
4490 Specify the band-width of a filter in width_type units.
4491 Applies only to double-pole filter.
4492 The default is 0.707q and gives a Butterworth response.
4495 How much to use filtered signal in output. Default is 1.
4496 Range is between 0 and 1.
4499 Specify which channels to filter, by default all available are filtered.
4502 Normalize biquad coefficients, by default is disabled.
4503 Enabling it will normalize magnitude response at DC to 0dB.
4506 Set transform type of IIR filter.
4515 Set precison of filtering.
4518 Pick automatic sample format depending on surround filters.
4520 Always use signed 16-bit.
4522 Always use signed 32-bit.
4524 Always use float 32-bit.
4526 Always use float 64-bit.
4530 @subsection Commands
4532 This filter supports the following commands:
4535 Change highpass frequency.
4536 Syntax for the command is : "@var{frequency}"
4539 Change highpass width_type.
4540 Syntax for the command is : "@var{width_type}"
4543 Change highpass width.
4544 Syntax for the command is : "@var{width}"
4547 Change highpass mix.
4548 Syntax for the command is : "@var{mix}"
4553 Join multiple input streams into one multi-channel stream.
4555 It accepts the following parameters:
4559 The number of input streams. It defaults to 2.
4561 @item channel_layout
4562 The desired output channel layout. It defaults to stereo.
4565 Map channels from inputs to output. The argument is a '|'-separated list of
4566 mappings, each in the @code{@var{input_idx}.@var{in_channel}-@var{out_channel}}
4567 form. @var{input_idx} is the 0-based index of the input stream. @var{in_channel}
4568 can be either the name of the input channel (e.g. FL for front left) or its
4569 index in the specified input stream. @var{out_channel} is the name of the output
4573 The filter will attempt to guess the mappings when they are not specified
4574 explicitly. It does so by first trying to find an unused matching input channel
4575 and if that fails it picks the first unused input channel.
4577 Join 3 inputs (with properly set channel layouts):
4579 ffmpeg -i INPUT1 -i INPUT2 -i INPUT3 -filter_complex join=inputs=3 OUTPUT
4582 Build a 5.1 output from 6 single-channel streams:
4584 ffmpeg -i fl -i fr -i fc -i sl -i sr -i lfe -filter_complex
4585 'join=inputs=6:channel_layout=5.1:map=0.0-FL|1.0-FR|2.0-FC|3.0-SL|4.0-SR|5.0-LFE'
4591 Load a LADSPA (Linux Audio Developer's Simple Plugin API) plugin.
4593 To enable compilation of this filter you need to configure FFmpeg with
4594 @code{--enable-ladspa}.
4598 Specifies the name of LADSPA plugin library to load. If the environment
4599 variable @env{LADSPA_PATH} is defined, the LADSPA plugin is searched in
4600 each one of the directories specified by the colon separated list in
4601 @env{LADSPA_PATH}, otherwise in the standard LADSPA paths, which are in
4602 this order: @file{HOME/.ladspa/lib/}, @file{/usr/local/lib/ladspa/},
4603 @file{/usr/lib/ladspa/}.
4606 Specifies the plugin within the library. Some libraries contain only
4607 one plugin, but others contain many of them. If this is not set filter
4608 will list all available plugins within the specified library.
4611 Set the '|' separated list of controls which are zero or more floating point
4612 values that determine the behavior of the loaded plugin (for example delay,
4614 Controls need to be defined using the following syntax:
4615 c0=@var{value0}|c1=@var{value1}|c2=@var{value2}|..., where
4616 @var{valuei} is the value set on the @var{i}-th control.
4617 Alternatively they can be also defined using the following syntax:
4618 @var{value0}|@var{value1}|@var{value2}|..., where
4619 @var{valuei} is the value set on the @var{i}-th control.
4620 If @option{controls} is set to @code{help}, all available controls and
4621 their valid ranges are printed.
4623 @item sample_rate, s
4624 Specify the sample rate, default to 44100. Only used if plugin have
4628 Set the number of samples per channel per each output frame, default
4629 is 1024. Only used if plugin have zero inputs.
4632 Set the minimum duration of the sourced audio. See
4633 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4634 for the accepted syntax.
4635 Note that the resulting duration may be greater than the specified duration,
4636 as the generated audio is always cut at the end of a complete frame.
4637 If not specified, or the expressed duration is negative, the audio is
4638 supposed to be generated forever.
4639 Only used if plugin have zero inputs.
4642 Enable latency compensation, by default is disabled.
4643 Only used if plugin have inputs.
4646 @subsection Examples
4650 List all available plugins within amp (LADSPA example plugin) library:
4656 List all available controls and their valid ranges for @code{vcf_notch}
4657 plugin from @code{VCF} library:
4659 ladspa=f=vcf:p=vcf_notch:c=help
4663 Simulate low quality audio equipment using @code{Computer Music Toolkit} (CMT)
4666 ladspa=file=cmt:plugin=lofi:controls=c0=22|c1=12|c2=12
4670 Add reverberation to the audio using TAP-plugins
4671 (Tom's Audio Processing plugins):
4673 ladspa=file=tap_reverb:tap_reverb
4677 Generate white noise, with 0.2 amplitude:
4679 ladspa=file=cmt:noise_source_white:c=c0=.2
4683 Generate 20 bpm clicks using plugin @code{C* Click - Metronome} from the
4684 @code{C* Audio Plugin Suite} (CAPS) library:
4686 ladspa=file=caps:Click:c=c1=20'
4690 Apply @code{C* Eq10X2 - Stereo 10-band equaliser} effect:
4692 ladspa=caps:Eq10X2:c=c0=-48|c9=-24|c3=12|c4=2
4696 Increase volume by 20dB using fast lookahead limiter from Steve Harris
4697 @code{SWH Plugins} collection:
4699 ladspa=fast_lookahead_limiter_1913:fastLookaheadLimiter:20|0|2
4703 Attenuate low frequencies using Multiband EQ from Steve Harris
4704 @code{SWH Plugins} collection:
4706 ladspa=mbeq_1197:mbeq:-24|-24|-24|0|0|0|0|0|0|0|0|0|0|0|0
4710 Reduce stereo image using @code{Narrower} from the @code{C* Audio Plugin Suite}
4713 ladspa=caps:Narrower
4717 Another white noise, now using @code{C* Audio Plugin Suite} (CAPS) library:
4719 ladspa=caps:White:.2
4723 Some fractal noise, using @code{C* Audio Plugin Suite} (CAPS) library:
4725 ladspa=caps:Fractal:c=c1=1
4729 Dynamic volume normalization using @code{VLevel} plugin:
4731 ladspa=vlevel-ladspa:vlevel_mono
4735 @subsection Commands
4737 This filter supports the following commands:
4740 Modify the @var{N}-th control value.
4742 If the specified value is not valid, it is ignored and prior one is kept.
4747 EBU R128 loudness normalization. Includes both dynamic and linear normalization modes.
4748 Support for both single pass (livestreams, files) and double pass (files) modes.
4749 This algorithm can target IL, LRA, and maximum true peak. In dynamic mode, to accurately
4750 detect true peaks, the audio stream will be upsampled to 192 kHz.
4751 Use the @code{-ar} option or @code{aresample} filter to explicitly set an output sample rate.
4753 The filter accepts the following options:
4757 Set integrated loudness target.
4758 Range is -70.0 - -5.0. Default value is -24.0.
4761 Set loudness range target.
4762 Range is 1.0 - 20.0. Default value is 7.0.
4765 Set maximum true peak.
4766 Range is -9.0 - +0.0. Default value is -2.0.
4768 @item measured_I, measured_i
4769 Measured IL of input file.
4770 Range is -99.0 - +0.0.
4772 @item measured_LRA, measured_lra
4773 Measured LRA of input file.
4774 Range is 0.0 - 99.0.
4776 @item measured_TP, measured_tp
4777 Measured true peak of input file.
4778 Range is -99.0 - +99.0.
4780 @item measured_thresh
4781 Measured threshold of input file.
4782 Range is -99.0 - +0.0.
4785 Set offset gain. Gain is applied before the true-peak limiter.
4786 Range is -99.0 - +99.0. Default is +0.0.
4789 Normalize by linearly scaling the source audio.
4790 @code{measured_I}, @code{measured_LRA}, @code{measured_TP},
4791 and @code{measured_thresh} must all be specified. Target LRA shouldn't
4792 be lower than source LRA and the change in integrated loudness shouldn't
4793 result in a true peak which exceeds the target TP. If any of these
4794 conditions aren't met, normalization mode will revert to @var{dynamic}.
4795 Options are @code{true} or @code{false}. Default is @code{true}.
4798 Treat mono input files as "dual-mono". If a mono file is intended for playback
4799 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
4800 If set to @code{true}, this option will compensate for this effect.
4801 Multi-channel input files are not affected by this option.
4802 Options are true or false. Default is false.
4805 Set print format for stats. Options are summary, json, or none.
4806 Default value is none.
4811 Apply a low-pass filter with 3dB point frequency.
4812 The filter can be either single-pole or double-pole (the default).
4813 The filter roll off at 6dB per pole per octave (20dB per pole per decade).
4815 The filter accepts the following options:
4819 Set frequency in Hz. Default is 500.
4822 Set number of poles. Default is 2.
4825 Set method to specify band-width of filter.
4840 Specify the band-width of a filter in width_type units.
4841 Applies only to double-pole filter.
4842 The default is 0.707q and gives a Butterworth response.
4845 How much to use filtered signal in output. Default is 1.
4846 Range is between 0 and 1.
4849 Specify which channels to filter, by default all available are filtered.
4852 Normalize biquad coefficients, by default is disabled.
4853 Enabling it will normalize magnitude response at DC to 0dB.
4856 Set transform type of IIR filter.
4865 Set precison of filtering.
4868 Pick automatic sample format depending on surround filters.
4870 Always use signed 16-bit.
4872 Always use signed 32-bit.
4874 Always use float 32-bit.
4876 Always use float 64-bit.
4880 @subsection Examples
4883 Lowpass only LFE channel, it LFE is not present it does nothing:
4889 @subsection Commands
4891 This filter supports the following commands:
4894 Change lowpass frequency.
4895 Syntax for the command is : "@var{frequency}"
4898 Change lowpass width_type.
4899 Syntax for the command is : "@var{width_type}"
4902 Change lowpass width.
4903 Syntax for the command is : "@var{width}"
4907 Syntax for the command is : "@var{mix}"
4912 Load a LV2 (LADSPA Version 2) plugin.
4914 To enable compilation of this filter you need to configure FFmpeg with
4915 @code{--enable-lv2}.
4919 Specifies the plugin URI. You may need to escape ':'.
4922 Set the '|' separated list of controls which are zero or more floating point
4923 values that determine the behavior of the loaded plugin (for example delay,
4925 If @option{controls} is set to @code{help}, all available controls and
4926 their valid ranges are printed.
4928 @item sample_rate, s
4929 Specify the sample rate, default to 44100. Only used if plugin have
4933 Set the number of samples per channel per each output frame, default
4934 is 1024. Only used if plugin have zero inputs.
4937 Set the minimum duration of the sourced audio. See
4938 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
4939 for the accepted syntax.
4940 Note that the resulting duration may be greater than the specified duration,
4941 as the generated audio is always cut at the end of a complete frame.
4942 If not specified, or the expressed duration is negative, the audio is
4943 supposed to be generated forever.
4944 Only used if plugin have zero inputs.
4947 @subsection Examples
4951 Apply bass enhancer plugin from Calf:
4953 lv2=p=http\\\\://calf.sourceforge.net/plugins/BassEnhancer:c=amount=2
4957 Apply vinyl plugin from Calf:
4959 lv2=p=http\\\\://calf.sourceforge.net/plugins/Vinyl:c=drone=0.2|aging=0.5
4963 Apply bit crusher plugin from ArtyFX:
4965 lv2=p=http\\\\://www.openavproductions.com/artyfx#bitta:c=crush=0.3
4970 Multiband Compress or expand the audio's dynamic range.
4972 The input audio is divided into bands using 4th order Linkwitz-Riley IIRs.
4973 This is akin to the crossover of a loudspeaker, and results in flat frequency
4974 response when absent compander action.
4976 It accepts the following parameters:
4980 This option syntax is:
4981 attack,decay,[attack,decay..] soft-knee points crossover_frequency [delay [initial_volume [gain]]] | attack,decay ...
4982 For explanation of each item refer to compand filter documentation.
4988 Mix channels with specific gain levels. The filter accepts the output
4989 channel layout followed by a set of channels definitions.
4991 This filter is also designed to efficiently remap the channels of an audio
4994 The filter accepts parameters of the form:
4995 "@var{l}|@var{outdef}|@var{outdef}|..."
4999 output channel layout or number of channels
5002 output channel specification, of the form:
5003 "@var{out_name}=[@var{gain}*]@var{in_name}[(+-)[@var{gain}*]@var{in_name}...]"
5006 output channel to define, either a channel name (FL, FR, etc.) or a channel
5007 number (c0, c1, etc.)
5010 multiplicative coefficient for the channel, 1 leaving the volume unchanged
5013 input channel to use, see out_name for details; it is not possible to mix
5014 named and numbered input channels
5017 If the `=' in a channel specification is replaced by `<', then the gains for
5018 that specification will be renormalized so that the total is 1, thus
5019 avoiding clipping noise.
5021 @subsection Mixing examples
5023 For example, if you want to down-mix from stereo to mono, but with a bigger
5024 factor for the left channel:
5026 pan=1c|c0=0.9*c0+0.1*c1
5029 A customized down-mix to stereo that works automatically for 3-, 4-, 5- and
5030 7-channels surround:
5032 pan=stereo| FL < FL + 0.5*FC + 0.6*BL + 0.6*SL | FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
5035 Note that @command{ffmpeg} integrates a default down-mix (and up-mix) system
5036 that should be preferred (see "-ac" option) unless you have very specific
5039 @subsection Remapping examples
5041 The channel remapping will be effective if, and only if:
5044 @item gain coefficients are zeroes or ones,
5045 @item only one input per channel output,
5048 If all these conditions are satisfied, the filter will notify the user ("Pure
5049 channel mapping detected"), and use an optimized and lossless method to do the
5052 For example, if you have a 5.1 source and want a stereo audio stream by
5053 dropping the extra channels:
5055 pan="stereo| c0=FL | c1=FR"
5058 Given the same source, you can also switch front left and front right channels
5059 and keep the input channel layout:
5061 pan="5.1| c0=c1 | c1=c0 | c2=c2 | c3=c3 | c4=c4 | c5=c5"
5064 If the input is a stereo audio stream, you can mute the front left channel (and
5065 still keep the stereo channel layout) with:
5070 Still with a stereo audio stream input, you can copy the right channel in both
5071 front left and right:
5073 pan="stereo| c0=FR | c1=FR"
5078 ReplayGain scanner filter. This filter takes an audio stream as an input and
5079 outputs it unchanged.
5080 At end of filtering it displays @code{track_gain} and @code{track_peak}.
5084 Convert the audio sample format, sample rate and channel layout. It is
5085 not meant to be used directly.
5088 Apply time-stretching and pitch-shifting with librubberband.
5090 To enable compilation of this filter, you need to configure FFmpeg with
5091 @code{--enable-librubberband}.
5093 The filter accepts the following options:
5097 Set tempo scale factor.
5100 Set pitch scale factor.
5103 Set transients detector.
5104 Possible values are:
5113 Possible values are:
5122 Possible values are:
5129 Set processing window size.
5130 Possible values are:
5139 Possible values are:
5146 Enable formant preservation when shift pitching.
5147 Possible values are:
5155 Possible values are:
5164 Possible values are:
5171 @subsection Commands
5173 This filter supports the following commands:
5176 Change filter tempo scale factor.
5177 Syntax for the command is : "@var{tempo}"
5180 Change filter pitch scale factor.
5181 Syntax for the command is : "@var{pitch}"
5184 @section sidechaincompress
5186 This filter acts like normal compressor but has the ability to compress
5187 detected signal using second input signal.
5188 It needs two input streams and returns one output stream.
5189 First input stream will be processed depending on second stream signal.
5190 The filtered signal then can be filtered with other filters in later stages of
5191 processing. See @ref{pan} and @ref{amerge} filter.
5193 The filter accepts the following options:
5197 Set input gain. Default is 1. Range is between 0.015625 and 64.
5200 Set mode of compressor operation. Can be @code{upward} or @code{downward}.
5201 Default is @code{downward}.
5204 If a signal of second stream raises above this level it will affect the gain
5205 reduction of first stream.
5206 By default is 0.125. Range is between 0.00097563 and 1.
5209 Set a ratio about which the signal is reduced. 1:2 means that if the level
5210 raised 4dB above the threshold, it will be only 2dB above after the reduction.
5211 Default is 2. Range is between 1 and 20.
5214 Amount of milliseconds the signal has to rise above the threshold before gain
5215 reduction starts. Default is 20. Range is between 0.01 and 2000.
5218 Amount of milliseconds the signal has to fall below the threshold before
5219 reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
5222 Set the amount by how much signal will be amplified after processing.
5223 Default is 1. Range is from 1 to 64.
5226 Curve the sharp knee around the threshold to enter gain reduction more softly.
5227 Default is 2.82843. Range is between 1 and 8.
5230 Choose if the @code{average} level between all channels of side-chain stream
5231 or the louder(@code{maximum}) channel of side-chain stream affects the
5232 reduction. Default is @code{average}.
5235 Should the exact signal be taken in case of @code{peak} or an RMS one in case
5236 of @code{rms}. Default is @code{rms} which is mainly smoother.
5239 Set sidechain gain. Default is 1. Range is between 0.015625 and 64.
5242 How much to use compressed signal in output. Default is 1.
5243 Range is between 0 and 1.
5246 @subsection Commands
5248 This filter supports the all above options as @ref{commands}.
5250 @subsection Examples
5254 Full ffmpeg example taking 2 audio inputs, 1st input to be compressed
5255 depending on the signal of 2nd input and later compressed signal to be
5256 merged with 2nd input:
5258 ffmpeg -i main.flac -i sidechain.flac -filter_complex "[1:a]asplit=2[sc][mix];[0:a][sc]sidechaincompress[compr];[compr][mix]amerge"
5262 @section sidechaingate
5264 A sidechain gate acts like a normal (wideband) gate but has the ability to
5265 filter the detected signal before sending it to the gain reduction stage.
5266 Normally a gate uses the full range signal to detect a level above the
5268 For example: If you cut all lower frequencies from your sidechain signal
5269 the gate will decrease the volume of your track only if not enough highs
5270 appear. With this technique you are able to reduce the resonation of a
5271 natural drum or remove "rumbling" of muted strokes from a heavily distorted
5273 It needs two input streams and returns one output stream.
5274 First input stream will be processed depending on second stream signal.
5276 The filter accepts the following options:
5280 Set input level before filtering.
5281 Default is 1. Allowed range is from 0.015625 to 64.
5284 Set the mode of operation. Can be @code{upward} or @code{downward}.
5285 Default is @code{downward}. If set to @code{upward} mode, higher parts of signal
5286 will be amplified, expanding dynamic range in upward direction.
5287 Otherwise, in case of @code{downward} lower parts of signal will be reduced.
5290 Set the level of gain reduction when the signal is below the threshold.
5291 Default is 0.06125. Allowed range is from 0 to 1.
5292 Setting this to 0 disables reduction and then filter behaves like expander.
5295 If a signal rises above this level the gain reduction is released.
5296 Default is 0.125. Allowed range is from 0 to 1.
5299 Set a ratio about which the signal is reduced.
5300 Default is 2. Allowed range is from 1 to 9000.
5303 Amount of milliseconds the signal has to rise above the threshold before gain
5305 Default is 20 milliseconds. Allowed range is from 0.01 to 9000.
5308 Amount of milliseconds the signal has to fall below the threshold before the
5309 reduction is increased again. Default is 250 milliseconds.
5310 Allowed range is from 0.01 to 9000.
5313 Set amount of amplification of signal after processing.
5314 Default is 1. Allowed range is from 1 to 64.
5317 Curve the sharp knee around the threshold to enter gain reduction more softly.
5318 Default is 2.828427125. Allowed range is from 1 to 8.
5321 Choose if exact signal should be taken for detection or an RMS like one.
5322 Default is rms. Can be peak or rms.
5325 Choose if the average level between all channels or the louder channel affects
5327 Default is average. Can be average or maximum.
5330 Set sidechain gain. Default is 1. Range is from 0.015625 to 64.
5333 @subsection Commands
5335 This filter supports the all above options as @ref{commands}.
5337 @section silencedetect
5339 Detect silence in an audio stream.
5341 This filter logs a message when it detects that the input audio volume is less
5342 or equal to a noise tolerance value for a duration greater or equal to the
5343 minimum detected noise duration.
5345 The printed times and duration are expressed in seconds. The
5346 @code{lavfi.silence_start} or @code{lavfi.silence_start.X} metadata key
5347 is set on the first frame whose timestamp equals or exceeds the detection
5348 duration and it contains the timestamp of the first frame of the silence.
5350 The @code{lavfi.silence_duration} or @code{lavfi.silence_duration.X}
5351 and @code{lavfi.silence_end} or @code{lavfi.silence_end.X} metadata
5352 keys are set on the first frame after the silence. If @option{mono} is
5353 enabled, and each channel is evaluated separately, the @code{.X}
5354 suffixed keys are used, and @code{X} corresponds to the channel number.
5356 The filter accepts the following options:
5360 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
5361 specified value) or amplitude ratio. Default is -60dB, or 0.001.
5364 Set silence duration until notification (default is 2 seconds). See
5365 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5366 for the accepted syntax.
5369 Process each channel separately, instead of combined. By default is disabled.
5372 @subsection Examples
5376 Detect 5 seconds of silence with -50dB noise tolerance:
5378 silencedetect=n=-50dB:d=5
5382 Complete example with @command{ffmpeg} to detect silence with 0.0001 noise
5383 tolerance in @file{silence.mp3}:
5385 ffmpeg -i silence.mp3 -af silencedetect=noise=0.0001 -f null -
5389 @section silenceremove
5391 Remove silence from the beginning, middle or end of the audio.
5393 The filter accepts the following options:
5397 This value is used to indicate if audio should be trimmed at beginning of
5398 the audio. A value of zero indicates no silence should be trimmed from the
5399 beginning. When specifying a non-zero value, it trims audio up until it
5400 finds non-silence. Normally, when trimming silence from beginning of audio
5401 the @var{start_periods} will be @code{1} but it can be increased to higher
5402 values to trim all audio up to specific count of non-silence periods.
5403 Default value is @code{0}.
5405 @item start_duration
5406 Specify the amount of time that non-silence must be detected before it stops
5407 trimming audio. By increasing the duration, bursts of noises can be treated
5408 as silence and trimmed off. Default value is @code{0}.
5410 @item start_threshold
5411 This indicates what sample value should be treated as silence. For digital
5412 audio, a value of @code{0} may be fine but for audio recorded from analog,
5413 you may wish to increase the value to account for background noise.
5414 Can be specified in dB (in case "dB" is appended to the specified value)
5415 or amplitude ratio. Default value is @code{0}.
5418 Specify max duration of silence at beginning that will be kept after
5419 trimming. Default is 0, which is equal to trimming all samples detected
5423 Specify mode of detection of silence end in start of multi-channel audio.
5424 Can be @var{any} or @var{all}. Default is @var{any}.
5425 With @var{any}, any sample that is detected as non-silence will cause
5426 stopped trimming of silence.
5427 With @var{all}, only if all channels are detected as non-silence will cause
5428 stopped trimming of silence.
5431 Set the count for trimming silence from the end of audio.
5432 To remove silence from the middle of a file, specify a @var{stop_periods}
5433 that is negative. This value is then treated as a positive value and is
5434 used to indicate the effect should restart processing as specified by
5435 @var{start_periods}, making it suitable for removing periods of silence
5436 in the middle of the audio.
5437 Default value is @code{0}.
5440 Specify a duration of silence that must exist before audio is not copied any
5441 more. By specifying a higher duration, silence that is wanted can be left in
5443 Default value is @code{0}.
5445 @item stop_threshold
5446 This is the same as @option{start_threshold} but for trimming silence from
5448 Can be specified in dB (in case "dB" is appended to the specified value)
5449 or amplitude ratio. Default value is @code{0}.
5452 Specify max duration of silence at end that will be kept after
5453 trimming. Default is 0, which is equal to trimming all samples detected
5457 Specify mode of detection of silence start in end of multi-channel audio.
5458 Can be @var{any} or @var{all}. Default is @var{any}.
5459 With @var{any}, any sample that is detected as non-silence will cause
5460 stopped trimming of silence.
5461 With @var{all}, only if all channels are detected as non-silence will cause
5462 stopped trimming of silence.
5465 Set how is silence detected. Can be @code{rms} or @code{peak}. Second is faster
5466 and works better with digital silence which is exactly 0.
5467 Default value is @code{rms}.
5470 Set duration in number of seconds used to calculate size of window in number
5471 of samples for detecting silence.
5472 Default value is @code{0.02}. Allowed range is from @code{0} to @code{10}.
5475 @subsection Examples
5479 The following example shows how this filter can be used to start a recording
5480 that does not contain the delay at the start which usually occurs between
5481 pressing the record button and the start of the performance:
5483 silenceremove=start_periods=1:start_duration=5:start_threshold=0.02
5487 Trim all silence encountered from beginning to end where there is more than 1
5488 second of silence in audio:
5490 silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-90dB
5494 Trim all digital silence samples, using peak detection, from beginning to end
5495 where there is more than 0 samples of digital silence in audio and digital
5496 silence is detected in all channels at same positions in stream:
5498 silenceremove=window=0:detection=peak:stop_mode=all:start_mode=all:stop_periods=-1:stop_threshold=0
5504 SOFAlizer uses head-related transfer functions (HRTFs) to create virtual
5505 loudspeakers around the user for binaural listening via headphones (audio
5506 formats up to 9 channels supported).
5507 The HRTFs are stored in SOFA files (see @url{http://www.sofacoustics.org/} for a database).
5508 SOFAlizer is developed at the Acoustics Research Institute (ARI) of the
5509 Austrian Academy of Sciences.
5511 To enable compilation of this filter you need to configure FFmpeg with
5512 @code{--enable-libmysofa}.
5514 The filter accepts the following options:
5518 Set the SOFA file used for rendering.
5521 Set gain applied to audio. Value is in dB. Default is 0.
5524 Set rotation of virtual loudspeakers in deg. Default is 0.
5527 Set elevation of virtual speakers in deg. Default is 0.
5530 Set distance in meters between loudspeakers and the listener with near-field
5531 HRTFs. Default is 1.
5534 Set processing type. Can be @var{time} or @var{freq}. @var{time} is
5535 processing audio in time domain which is slow.
5536 @var{freq} is processing audio in frequency domain which is fast.
5537 Default is @var{freq}.
5540 Set custom positions of virtual loudspeakers. Syntax for this option is:
5541 <CH> <AZIM> <ELEV>[|<CH> <AZIM> <ELEV>|...].
5542 Each virtual loudspeaker is described with short channel name following with
5543 azimuth and elevation in degrees.
5544 Each virtual loudspeaker description is separated by '|'.
5545 For example to override front left and front right channel positions use:
5546 'speakers=FL 45 15|FR 345 15'.
5547 Descriptions with unrecognised channel names are ignored.
5550 Set custom gain for LFE channels. Value is in dB. Default is 0.
5553 Set custom frame size in number of samples. Default is 1024.
5554 Allowed range is from 1024 to 96000. Only used if option @samp{type}
5555 is set to @var{freq}.
5558 Should all IRs be normalized upon importing SOFA file.
5559 By default is enabled.
5562 Should nearest IRs be interpolated with neighbor IRs if exact position
5563 does not match. By default is disabled.
5566 Minphase all IRs upon loading of SOFA file. By default is disabled.
5569 Set neighbor search angle step. Only used if option @var{interpolate} is enabled.
5572 Set neighbor search radius step. Only used if option @var{interpolate} is enabled.
5575 @subsection Examples
5579 Using ClubFritz6 sofa file:
5581 sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=1
5585 Using ClubFritz12 sofa file and bigger radius with small rotation:
5587 sofalizer=sofa=/path/to/ClubFritz12.sofa:type=freq:radius=2:rotation=5
5591 Similar as above but with custom speaker positions for front left, front right, back left and back right
5592 and also with custom gain:
5594 "sofalizer=sofa=/path/to/ClubFritz6.sofa:type=freq:radius=2:speakers=FL 45|FR 315|BL 135|BR 225:gain=28"
5601 This filter expands or compresses each half-cycle of audio samples
5602 (local set of samples all above or all below zero and between two nearest zero crossings) depending
5603 on threshold value, so audio reaches target peak value under conditions controlled by below options.
5605 The filter accepts the following options:
5609 Set the expansion target peak value. This specifies the highest allowed absolute amplitude
5610 level for the normalized audio input. Default value is 0.95. Allowed range is from 0.0 to 1.0.
5613 Set the maximum expansion factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5614 This option controls maximum local half-cycle of samples expansion. The maximum expansion
5615 would be such that local peak value reaches target peak value but never to surpass it and that
5616 ratio between new and previous peak value does not surpass this option value.
5618 @item compression, c
5619 Set the maximum compression factor. Allowed range is from 1.0 to 50.0. Default value is 2.0.
5620 This option controls maximum local half-cycle of samples compression. This option is used
5621 only if @option{threshold} option is set to value greater than 0.0, then in such cases
5622 when local peak is lower or same as value set by @option{threshold} all samples belonging to
5623 that peak's half-cycle will be compressed by current compression factor.
5626 Set the threshold value. Default value is 0.0. Allowed range is from 0.0 to 1.0.
5627 This option specifies which half-cycles of samples will be compressed and which will be expanded.
5628 Any half-cycle samples with their local peak value below or same as this option value will be
5629 compressed by current compression factor, otherwise, if greater than threshold value they will be
5630 expanded with expansion factor so that it could reach peak target value but never surpass it.
5633 Set the expansion raising amount per each half-cycle of samples. Default value is 0.001.
5634 Allowed range is from 0.0 to 1.0. This controls how fast expansion factor is raised per
5635 each new half-cycle until it reaches @option{expansion} value.
5636 Setting this options too high may lead to distortions.
5639 Set the compression raising amount per each half-cycle of samples. Default value is 0.001.
5640 Allowed range is from 0.0 to 1.0. This controls how fast compression factor is raised per
5641 each new half-cycle until it reaches @option{compression} value.
5644 Specify which channels to filter, by default all available channels are filtered.
5647 Enable inverted filtering, by default is disabled. This inverts interpretation of @option{threshold}
5648 option. When enabled any half-cycle of samples with their local peak value below or same as
5649 @option{threshold} option will be expanded otherwise it will be compressed.
5652 Link channels when calculating gain applied to each filtered channel sample, by default is disabled.
5653 When disabled each filtered channel gain calculation is independent, otherwise when this option
5654 is enabled the minimum of all possible gains for each filtered channel is used.
5657 @subsection Commands
5659 This filter supports the all above options as @ref{commands}.
5661 @section stereotools
5663 This filter has some handy utilities to manage stereo signals, for converting
5664 M/S stereo recordings to L/R signal while having control over the parameters
5665 or spreading the stereo image of master track.
5667 The filter accepts the following options:
5671 Set input level before filtering for both channels. Defaults is 1.
5672 Allowed range is from 0.015625 to 64.
5675 Set output level after filtering for both channels. Defaults is 1.
5676 Allowed range is from 0.015625 to 64.
5679 Set input balance between both channels. Default is 0.
5680 Allowed range is from -1 to 1.
5683 Set output balance between both channels. Default is 0.
5684 Allowed range is from -1 to 1.
5687 Enable softclipping. Results in analog distortion instead of harsh digital 0dB
5688 clipping. Disabled by default.
5691 Mute the left channel. Disabled by default.
5694 Mute the right channel. Disabled by default.
5697 Change the phase of the left channel. Disabled by default.
5700 Change the phase of the right channel. Disabled by default.
5703 Set stereo mode. Available values are:
5707 Left/Right to Left/Right, this is default.
5710 Left/Right to Mid/Side.
5713 Mid/Side to Left/Right.
5716 Left/Right to Left/Left.
5719 Left/Right to Right/Right.
5722 Left/Right to Left + Right.
5725 Left/Right to Right/Left.
5728 Mid/Side to Left/Left.
5731 Mid/Side to Right/Right.
5734 Mid/Side to Right/Left.
5737 Left/Right to Left - Right.
5741 Set level of side signal. Default is 1.
5742 Allowed range is from 0.015625 to 64.
5745 Set balance of side signal. Default is 0.
5746 Allowed range is from -1 to 1.
5749 Set level of the middle signal. Default is 1.
5750 Allowed range is from 0.015625 to 64.
5753 Set middle signal pan. Default is 0. Allowed range is from -1 to 1.
5756 Set stereo base between mono and inversed channels. Default is 0.
5757 Allowed range is from -1 to 1.
5760 Set delay in milliseconds how much to delay left from right channel and
5761 vice versa. Default is 0. Allowed range is from -20 to 20.
5764 Set S/C level. Default is 1. Allowed range is from 1 to 100.
5767 Set the stereo phase in degrees. Default is 0. Allowed range is from 0 to 360.
5769 @item bmode_in, bmode_out
5770 Set balance mode for balance_in/balance_out option.
5772 Can be one of the following:
5776 Classic balance mode. Attenuate one channel at time.
5777 Gain is raised up to 1.
5780 Similar as classic mode above but gain is raised up to 2.
5783 Equal power distribution, from -6dB to +6dB range.
5787 @subsection Commands
5789 This filter supports the all above options as @ref{commands}.
5791 @subsection Examples
5795 Apply karaoke like effect:
5797 stereotools=mlev=0.015625
5801 Convert M/S signal to L/R:
5803 "stereotools=mode=ms>lr"
5807 @section stereowiden
5809 This filter enhance the stereo effect by suppressing signal common to both
5810 channels and by delaying the signal of left into right and vice versa,
5811 thereby widening the stereo effect.
5813 The filter accepts the following options:
5817 Time in milliseconds of the delay of left signal into right and vice versa.
5818 Default is 20 milliseconds.
5821 Amount of gain in delayed signal into right and vice versa. Gives a delay
5822 effect of left signal in right output and vice versa which gives widening
5823 effect. Default is 0.3.
5826 Cross feed of left into right with inverted phase. This helps in suppressing
5827 the mono. If the value is 1 it will cancel all the signal common to both
5828 channels. Default is 0.3.
5831 Set level of input signal of original channel. Default is 0.8.
5834 @subsection Commands
5836 This filter supports the all above options except @code{delay} as @ref{commands}.
5838 @section superequalizer
5839 Apply 18 band equalizer.
5841 The filter accepts the following options:
5848 Set 131Hz band gain.
5850 Set 185Hz band gain.
5852 Set 262Hz band gain.
5854 Set 370Hz band gain.
5856 Set 523Hz band gain.
5858 Set 740Hz band gain.
5860 Set 1047Hz band gain.
5862 Set 1480Hz band gain.
5864 Set 2093Hz band gain.
5866 Set 2960Hz band gain.
5868 Set 4186Hz band gain.
5870 Set 5920Hz band gain.
5872 Set 8372Hz band gain.
5874 Set 11840Hz band gain.
5876 Set 16744Hz band gain.
5878 Set 20000Hz band gain.
5882 Apply audio surround upmix filter.
5884 This filter allows to produce multichannel output from audio stream.
5886 The filter accepts the following options:
5890 Set output channel layout. By default, this is @var{5.1}.
5892 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5893 for the required syntax.
5896 Set input channel layout. By default, this is @var{stereo}.
5898 See @ref{channel layout syntax,,the Channel Layout section in the ffmpeg-utils(1) manual,ffmpeg-utils}
5899 for the required syntax.
5902 Set input volume level. By default, this is @var{1}.
5905 Set output volume level. By default, this is @var{1}.
5908 Enable LFE channel output if output channel layout has it. By default, this is enabled.
5911 Set LFE low cut off frequency. By default, this is @var{128} Hz.
5914 Set LFE high cut off frequency. By default, this is @var{256} Hz.
5917 Set LFE mode, can be @var{add} or @var{sub}. Default is @var{add}.
5918 In @var{add} mode, LFE channel is created from input audio and added to output.
5919 In @var{sub} mode, LFE channel is created from input audio and added to output but
5920 also all non-LFE output channels are subtracted with output LFE channel.
5923 Set angle of stereo surround transform, Allowed range is from @var{0} to @var{360}.
5924 Default is @var{90}.
5927 Set front center input volume. By default, this is @var{1}.
5930 Set front center output volume. By default, this is @var{1}.
5933 Set front left input volume. By default, this is @var{1}.
5936 Set front left output volume. By default, this is @var{1}.
5939 Set front right input volume. By default, this is @var{1}.
5942 Set front right output volume. By default, this is @var{1}.
5945 Set side left input volume. By default, this is @var{1}.
5948 Set side left output volume. By default, this is @var{1}.
5951 Set side right input volume. By default, this is @var{1}.
5954 Set side right output volume. By default, this is @var{1}.
5957 Set back left input volume. By default, this is @var{1}.
5960 Set back left output volume. By default, this is @var{1}.
5963 Set back right input volume. By default, this is @var{1}.
5966 Set back right output volume. By default, this is @var{1}.
5969 Set back center input volume. By default, this is @var{1}.
5972 Set back center output volume. By default, this is @var{1}.
5975 Set LFE input volume. By default, this is @var{1}.
5978 Set LFE output volume. By default, this is @var{1}.
5981 Set spread usage of stereo image across X axis for all channels.
5984 Set spread usage of stereo image across Y axis for all channels.
5986 @item fcx, flx, frx, blx, brx, slx, srx, bcx
5987 Set spread usage of stereo image across X axis for each channel.
5989 @item fcy, fly, fry, bly, bry, sly, sry, bcy
5990 Set spread usage of stereo image across Y axis for each channel.
5993 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
5996 Set window function.
5998 It accepts the following values:
6021 Default is @code{hann}.
6024 Set window overlap. If set to 1, the recommended overlap for selected
6025 window function will be picked. Default is @code{0.5}.
6028 @section treble, highshelf
6030 Boost or cut treble (upper) frequencies of the audio using a two-pole
6031 shelving filter with a response similar to that of a standard
6032 hi-fi's tone-controls. This is also known as shelving equalisation (EQ).
6034 The filter accepts the following options:
6038 Give the gain at whichever is the lower of ~22 kHz and the
6039 Nyquist frequency. Its useful range is about -20 (for a large cut)
6040 to +20 (for a large boost). Beware of clipping when using a positive gain.
6043 Set the filter's central frequency and so can be used
6044 to extend or reduce the frequency range to be boosted or cut.
6045 The default value is @code{3000} Hz.
6048 Set method to specify band-width of filter.
6063 Determine how steep is the filter's shelf transition.
6066 Set number of poles. Default is 2.
6069 How much to use filtered signal in output. Default is 1.
6070 Range is between 0 and 1.
6073 Specify which channels to filter, by default all available are filtered.
6076 Normalize biquad coefficients, by default is disabled.
6077 Enabling it will normalize magnitude response at DC to 0dB.
6080 Set transform type of IIR filter.
6089 Set precison of filtering.
6092 Pick automatic sample format depending on surround filters.
6094 Always use signed 16-bit.
6096 Always use signed 32-bit.
6098 Always use float 32-bit.
6100 Always use float 64-bit.
6104 @subsection Commands
6106 This filter supports the following commands:
6109 Change treble frequency.
6110 Syntax for the command is : "@var{frequency}"
6113 Change treble width_type.
6114 Syntax for the command is : "@var{width_type}"
6117 Change treble width.
6118 Syntax for the command is : "@var{width}"
6122 Syntax for the command is : "@var{gain}"
6126 Syntax for the command is : "@var{mix}"
6131 Sinusoidal amplitude modulation.
6133 The filter accepts the following options:
6137 Modulation frequency in Hertz. Modulation frequencies in the subharmonic range
6138 (20 Hz or lower) will result in a tremolo effect.
6139 This filter may also be used as a ring modulator by specifying
6140 a modulation frequency higher than 20 Hz.
6141 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6144 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6145 Default value is 0.5.
6150 Sinusoidal phase modulation.
6152 The filter accepts the following options:
6156 Modulation frequency in Hertz.
6157 Range is 0.1 - 20000.0. Default value is 5.0 Hz.
6160 Depth of modulation as a percentage. Range is 0.0 - 1.0.
6161 Default value is 0.5.
6166 Adjust the input audio volume.
6168 It accepts the following parameters:
6172 Set audio volume expression.
6174 Output values are clipped to the maximum value.
6176 The output audio volume is given by the relation:
6178 @var{output_volume} = @var{volume} * @var{input_volume}
6181 The default value for @var{volume} is "1.0".
6184 This parameter represents the mathematical precision.
6186 It determines which input sample formats will be allowed, which affects the
6187 precision of the volume scaling.
6191 8-bit fixed-point; this limits input sample format to U8, S16, and S32.
6193 32-bit floating-point; this limits input sample format to FLT. (default)
6195 64-bit floating-point; this limits input sample format to DBL.
6199 Choose the behaviour on encountering ReplayGain side data in input frames.
6203 Remove ReplayGain side data, ignoring its contents (the default).
6206 Ignore ReplayGain side data, but leave it in the frame.
6209 Prefer the track gain, if present.
6212 Prefer the album gain, if present.
6215 @item replaygain_preamp
6216 Pre-amplification gain in dB to apply to the selected replaygain gain.
6218 Default value for @var{replaygain_preamp} is 0.0.
6220 @item replaygain_noclip
6221 Prevent clipping by limiting the gain applied.
6223 Default value for @var{replaygain_noclip} is 1.
6226 Set when the volume expression is evaluated.
6228 It accepts the following values:
6231 only evaluate expression once during the filter initialization, or
6232 when the @samp{volume} command is sent
6235 evaluate expression for each incoming frame
6238 Default value is @samp{once}.
6241 The volume expression can contain the following parameters.
6245 frame number (starting at zero)
6248 @item nb_consumed_samples
6249 number of samples consumed by the filter
6251 number of samples in the current frame
6253 original frame position in the file
6259 PTS at start of stream
6261 time at start of stream
6267 last set volume value
6270 Note that when @option{eval} is set to @samp{once} only the
6271 @var{sample_rate} and @var{tb} variables are available, all other
6272 variables will evaluate to NAN.
6274 @subsection Commands
6276 This filter supports the following commands:
6279 Modify the volume expression.
6280 The command accepts the same syntax of the corresponding option.
6282 If the specified expression is not valid, it is kept at its current
6286 @subsection Examples
6290 Halve the input audio volume:
6294 volume=volume=-6.0206dB
6297 In all the above example the named key for @option{volume} can be
6298 omitted, for example like in:
6304 Increase input audio power by 6 decibels using fixed-point precision:
6306 volume=volume=6dB:precision=fixed
6310 Fade volume after time 10 with an annihilation period of 5 seconds:
6312 volume='if(lt(t,10),1,max(1-(t-10)/5,0))':eval=frame
6316 @section volumedetect
6318 Detect the volume of the input video.
6320 The filter has no parameters. The input is not modified. Statistics about
6321 the volume will be printed in the log when the input stream end is reached.
6323 In particular it will show the mean volume (root mean square), maximum
6324 volume (on a per-sample basis), and the beginning of a histogram of the
6325 registered volume values (from the maximum value to a cumulated 1/1000 of
6328 All volumes are in decibels relative to the maximum PCM value.
6330 @subsection Examples
6332 Here is an excerpt of the output:
6334 [Parsed_volumedetect_0 @ 0xa23120] mean_volume: -27 dB
6335 [Parsed_volumedetect_0 @ 0xa23120] max_volume: -4 dB
6336 [Parsed_volumedetect_0 @ 0xa23120] histogram_4db: 6
6337 [Parsed_volumedetect_0 @ 0xa23120] histogram_5db: 62
6338 [Parsed_volumedetect_0 @ 0xa23120] histogram_6db: 286
6339 [Parsed_volumedetect_0 @ 0xa23120] histogram_7db: 1042
6340 [Parsed_volumedetect_0 @ 0xa23120] histogram_8db: 2551
6341 [Parsed_volumedetect_0 @ 0xa23120] histogram_9db: 4609
6342 [Parsed_volumedetect_0 @ 0xa23120] histogram_10db: 8409
6348 The mean square energy is approximately -27 dB, or 10^-2.7.
6350 The largest sample is at -4 dB, or more precisely between -4 dB and -5 dB.
6352 There are 6 samples at -4 dB, 62 at -5 dB, 286 at -6 dB, etc.
6355 In other words, raising the volume by +4 dB does not cause any clipping,
6356 raising it by +5 dB causes clipping for 6 samples, etc.
6358 @c man end AUDIO FILTERS
6360 @chapter Audio Sources
6361 @c man begin AUDIO SOURCES
6363 Below is a description of the currently available audio sources.
6367 Buffer audio frames, and make them available to the filter chain.
6369 This source is mainly intended for a programmatic use, in particular
6370 through the interface defined in @file{libavfilter/buffersrc.h}.
6372 It accepts the following parameters:
6376 The timebase which will be used for timestamps of submitted frames. It must be
6377 either a floating-point number or in @var{numerator}/@var{denominator} form.
6380 The sample rate of the incoming audio buffers.
6383 The sample format of the incoming audio buffers.
6384 Either a sample format name or its corresponding integer representation from
6385 the enum AVSampleFormat in @file{libavutil/samplefmt.h}
6387 @item channel_layout
6388 The channel layout of the incoming audio buffers.
6389 Either a channel layout name from channel_layout_map in
6390 @file{libavutil/channel_layout.c} or its corresponding integer representation
6391 from the AV_CH_LAYOUT_* macros in @file{libavutil/channel_layout.h}
6394 The number of channels of the incoming audio buffers.
6395 If both @var{channels} and @var{channel_layout} are specified, then they
6400 @subsection Examples
6403 abuffer=sample_rate=44100:sample_fmt=s16p:channel_layout=stereo
6406 will instruct the source to accept planar 16bit signed stereo at 44100Hz.
6407 Since the sample format with name "s16p" corresponds to the number
6408 6 and the "stereo" channel layout corresponds to the value 0x3, this is
6411 abuffer=sample_rate=44100:sample_fmt=6:channel_layout=0x3
6416 Generate an audio signal specified by an expression.
6418 This source accepts in input one or more expressions (one for each
6419 channel), which are evaluated and used to generate a corresponding
6422 This source accepts the following options:
6426 Set the '|'-separated expressions list for each separate channel. In case the
6427 @option{channel_layout} option is not specified, the selected channel layout
6428 depends on the number of provided expressions. Otherwise the last
6429 specified expression is applied to the remaining output channels.
6431 @item channel_layout, c
6432 Set the channel layout. The number of channels in the specified layout
6433 must be equal to the number of specified expressions.
6436 Set the minimum duration of the sourced audio. See
6437 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6438 for the accepted syntax.
6439 Note that the resulting duration may be greater than the specified
6440 duration, as the generated audio is always cut at the end of a
6443 If not specified, or the expressed duration is negative, the audio is
6444 supposed to be generated forever.
6447 Set the number of samples per channel per each output frame,
6450 @item sample_rate, s
6451 Specify the sample rate, default to 44100.
6454 Each expression in @var{exprs} can contain the following constants:
6458 number of the evaluated sample, starting from 0
6461 time of the evaluated sample expressed in seconds, starting from 0
6468 @subsection Examples
6478 Generate a sin signal with frequency of 440 Hz, set sample rate to
6481 aevalsrc="sin(440*2*PI*t):s=8000"
6485 Generate a two channels signal, specify the channel layout (Front
6486 Center + Back Center) explicitly:
6488 aevalsrc="sin(420*2*PI*t)|cos(430*2*PI*t):c=FC|BC"
6492 Generate white noise:
6494 aevalsrc="-2+random(0)"
6498 Generate an amplitude modulated signal:
6500 aevalsrc="sin(10*2*PI*t)*sin(880*2*PI*t)"
6504 Generate 2.5 Hz binaural beats on a 360 Hz carrier:
6506 aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)"
6513 Generate a FIR coefficients using frequency sampling method.
6515 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6517 The filter accepts the following options:
6521 Set number of filter coefficents in output audio stream.
6522 Default value is 1025.
6525 Set frequency points from where magnitude and phase are set.
6526 This must be in non decreasing order, and first element must be 0, while last element
6527 must be 1. Elements are separated by white spaces.
6530 Set magnitude value for every frequency point set by @option{frequency}.
6531 Number of values must be same as number of frequency points.
6532 Values are separated by white spaces.
6535 Set phase value for every frequency point set by @option{frequency}.
6536 Number of values must be same as number of frequency points.
6537 Values are separated by white spaces.
6539 @item sample_rate, r
6540 Set sample rate, default is 44100.
6543 Set number of samples per each frame. Default is 1024.
6546 Set window function. Default is blackman.
6551 The null audio source, return unprocessed audio frames. It is mainly useful
6552 as a template and to be employed in analysis / debugging tools, or as
6553 the source for filters which ignore the input data (for example the sox
6556 This source accepts the following options:
6560 @item channel_layout, cl
6562 Specifies the channel layout, and can be either an integer or a string
6563 representing a channel layout. The default value of @var{channel_layout}
6566 Check the channel_layout_map definition in
6567 @file{libavutil/channel_layout.c} for the mapping between strings and
6568 channel layout values.
6570 @item sample_rate, r
6571 Specifies the sample rate, and defaults to 44100.
6574 Set the number of samples per requested frames.
6577 Set the duration of the sourced audio. See
6578 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
6579 for the accepted syntax.
6581 If not specified, or the expressed duration is negative, the audio is
6582 supposed to be generated forever.
6585 @subsection Examples
6589 Set the sample rate to 48000 Hz and the channel layout to AV_CH_LAYOUT_MONO.
6591 anullsrc=r=48000:cl=4
6595 Do the same operation with a more obvious syntax:
6597 anullsrc=r=48000:cl=mono
6601 All the parameters need to be explicitly defined.
6605 Synthesize a voice utterance using the libflite library.
6607 To enable compilation of this filter you need to configure FFmpeg with
6608 @code{--enable-libflite}.
6610 Note that versions of the flite library prior to 2.0 are not thread-safe.
6612 The filter accepts the following options:
6617 If set to 1, list the names of the available voices and exit
6618 immediately. Default value is 0.
6621 Set the maximum number of samples per frame. Default value is 512.
6624 Set the filename containing the text to speak.
6627 Set the text to speak.
6630 Set the voice to use for the speech synthesis. Default value is
6631 @code{kal}. See also the @var{list_voices} option.
6634 @subsection Examples
6638 Read from file @file{speech.txt}, and synthesize the text using the
6639 standard flite voice:
6641 flite=textfile=speech.txt
6645 Read the specified text selecting the @code{slt} voice:
6647 flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6651 Input text to ffmpeg:
6653 ffmpeg -f lavfi -i flite=text='So fare thee well, poor devil of a Sub-Sub, whose commentator I am':voice=slt
6657 Make @file{ffplay} speak the specified text, using @code{flite} and
6658 the @code{lavfi} device:
6660 ffplay -f lavfi flite=text='No more be grieved for which that thou hast done.'
6664 For more information about libflite, check:
6665 @url{http://www.festvox.org/flite/}
6669 Generate a noise audio signal.
6671 The filter accepts the following options:
6674 @item sample_rate, r
6675 Specify the sample rate. Default value is 48000 Hz.
6678 Specify the amplitude (0.0 - 1.0) of the generated audio stream. Default value
6682 Specify the duration of the generated audio stream. Not specifying this option
6683 results in noise with an infinite length.
6685 @item color, colour, c
6686 Specify the color of noise. Available noise colors are white, pink, brown,
6687 blue, violet and velvet. Default color is white.
6690 Specify a value used to seed the PRNG.
6693 Set the number of samples per each output frame, default is 1024.
6696 @subsection Examples
6701 Generate 60 seconds of pink noise, with a 44.1 kHz sampling rate and an amplitude of 0.5:
6703 anoisesrc=d=60:c=pink:r=44100:a=0.5
6709 Generate odd-tap Hilbert transform FIR coefficients.
6711 The resulting stream can be used with @ref{afir} filter for phase-shifting
6712 the signal by 90 degrees.
6714 This is used in many matrix coding schemes and for analytic signal generation.
6715 The process is often written as a multiplication by i (or j), the imaginary unit.
6717 The filter accepts the following options:
6721 @item sample_rate, s
6722 Set sample rate, default is 44100.
6725 Set length of FIR filter, default is 22051.
6728 Set number of samples per each frame.
6731 Set window function to be used when generating FIR coefficients.
6736 Generate a sinc kaiser-windowed low-pass, high-pass, band-pass, or band-reject FIR coefficients.
6738 The resulting stream can be used with @ref{afir} filter for filtering the audio signal.
6740 The filter accepts the following options:
6743 @item sample_rate, r
6744 Set sample rate, default is 44100.
6747 Set number of samples per each frame. Default is 1024.
6750 Set high-pass frequency. Default is 0.
6753 Set low-pass frequency. Default is 0.
6754 If high-pass frequency is lower than low-pass frequency and low-pass frequency
6755 is higher than 0 then filter will create band-pass filter coefficients,
6756 otherwise band-reject filter coefficients.
6759 Set filter phase response. Default is 50. Allowed range is from 0 to 100.
6762 Set Kaiser window beta.
6765 Set stop-band attenuation. Default is 120dB, allowed range is from 40 to 180 dB.
6768 Enable rounding, by default is disabled.
6771 Set number of taps for high-pass filter.
6774 Set number of taps for low-pass filter.
6779 Generate an audio signal made of a sine wave with amplitude 1/8.
6781 The audio signal is bit-exact.
6783 The filter accepts the following options:
6788 Set the carrier frequency. Default is 440 Hz.
6790 @item beep_factor, b
6791 Enable a periodic beep every second with frequency @var{beep_factor} times
6792 the carrier frequency. Default is 0, meaning the beep is disabled.
6794 @item sample_rate, r
6795 Specify the sample rate, default is 44100.
6798 Specify the duration of the generated audio stream.
6800 @item samples_per_frame
6801 Set the number of samples per output frame.
6803 The expression can contain the following constants:
6807 The (sequential) number of the output audio frame, starting from 0.
6810 The PTS (Presentation TimeStamp) of the output audio frame,
6811 expressed in @var{TB} units.
6814 The PTS of the output audio frame, expressed in seconds.
6817 The timebase of the output audio frames.
6820 Default is @code{1024}.
6823 @subsection Examples
6828 Generate a simple 440 Hz sine wave:
6834 Generate a 220 Hz sine wave with a 880 Hz beep each second, for 5 seconds:
6838 sine=frequency=220:beep_factor=4:duration=5
6842 Generate a 1 kHz sine wave following @code{1602,1601,1602,1601,1602} NTSC
6845 sine=1000:samples_per_frame='st(0,mod(n,5)); 1602-not(not(eq(ld(0),1)+eq(ld(0),3)))'
6849 @c man end AUDIO SOURCES
6851 @chapter Audio Sinks
6852 @c man begin AUDIO SINKS
6854 Below is a description of the currently available audio sinks.
6856 @section abuffersink
6858 Buffer audio frames, and make them available to the end of filter chain.
6860 This sink is mainly intended for programmatic use, in particular
6861 through the interface defined in @file{libavfilter/buffersink.h}
6862 or the options system.
6864 It accepts a pointer to an AVABufferSinkContext structure, which
6865 defines the incoming buffers' formats, to be passed as the opaque
6866 parameter to @code{avfilter_init_filter} for initialization.
6869 Null audio sink; do absolutely nothing with the input audio. It is
6870 mainly useful as a template and for use in analysis / debugging
6873 @c man end AUDIO SINKS
6875 @chapter Video Filters
6876 @c man begin VIDEO FILTERS
6878 When you configure your FFmpeg build, you can disable any of the
6879 existing filters using @code{--disable-filters}.
6880 The configure output will show the video filters included in your
6883 Below is a description of the currently available video filters.
6887 Mark a region of interest in a video frame.
6889 The frame data is passed through unchanged, but metadata is attached
6890 to the frame indicating regions of interest which can affect the
6891 behaviour of later encoding. Multiple regions can be marked by
6892 applying the filter multiple times.
6896 Region distance in pixels from the left edge of the frame.
6898 Region distance in pixels from the top edge of the frame.
6900 Region width in pixels.
6902 Region height in pixels.
6904 The parameters @var{x}, @var{y}, @var{w} and @var{h} are expressions,
6905 and may contain the following variables:
6908 Width of the input frame.
6910 Height of the input frame.
6914 Quantisation offset to apply within the region.
6916 This must be a real value in the range -1 to +1. A value of zero
6917 indicates no quality change. A negative value asks for better quality
6918 (less quantisation), while a positive value asks for worse quality
6919 (greater quantisation).
6921 The range is calibrated so that the extreme values indicate the
6922 largest possible offset - if the rest of the frame is encoded with the
6923 worst possible quality, an offset of -1 indicates that this region
6924 should be encoded with the best possible quality anyway. Intermediate
6925 values are then interpolated in some codec-dependent way.
6927 For example, in 10-bit H.264 the quantisation parameter varies between
6928 -12 and 51. A typical qoffset value of -1/10 therefore indicates that
6929 this region should be encoded with a QP around one-tenth of the full
6930 range better than the rest of the frame. So, if most of the frame
6931 were to be encoded with a QP of around 30, this region would get a QP
6932 of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
6933 An extreme value of -1 would indicate that this region should be
6934 encoded with the best possible quality regardless of the treatment of
6935 the rest of the frame - that is, should be encoded at a QP of -12.
6937 If set to true, remove any existing regions of interest marked on the
6938 frame before adding the new one.
6941 @subsection Examples
6945 Mark the centre quarter of the frame as interesting.
6947 addroi=iw/4:ih/4:iw/2:ih/2:-1/10
6950 Mark the 100-pixel-wide region on the left edge of the frame as very
6951 uninteresting (to be encoded at much lower quality than the rest of
6954 addroi=0:0:100:ih:+1/5
6958 @section alphaextract
6960 Extract the alpha component from the input as a grayscale video. This
6961 is especially useful with the @var{alphamerge} filter.
6965 Add or replace the alpha component of the primary input with the
6966 grayscale value of a second input. This is intended for use with
6967 @var{alphaextract} to allow the transmission or storage of frame
6968 sequences that have alpha in a format that doesn't support an alpha
6971 For example, to reconstruct full frames from a normal YUV-encoded video
6972 and a separate video created with @var{alphaextract}, you might use:
6974 movie=in_alpha.mkv [alpha]; [in][alpha] alphamerge [out]
6979 Amplify differences between current pixel and pixels of adjacent frames in
6980 same pixel location.
6982 This filter accepts the following options:
6986 Set frame radius. Default is 2. Allowed range is from 1 to 63.
6987 For example radius of 3 will instruct filter to calculate average of 7 frames.
6990 Set factor to amplify difference. Default is 2. Allowed range is from 0 to 65535.
6993 Set threshold for difference amplification. Any difference greater or equal to
6994 this value will not alter source pixel. Default is 10.
6995 Allowed range is from 0 to 65535.
6998 Set tolerance for difference amplification. Any difference lower to
6999 this value will not alter source pixel. Default is 0.
7000 Allowed range is from 0 to 65535.
7003 Set lower limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7004 This option controls maximum possible value that will decrease source pixel value.
7007 Set high limit for changing source pixel. Default is 65535. Allowed range is from 0 to 65535.
7008 This option controls maximum possible value that will increase source pixel value.
7011 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
7014 @subsection Commands
7016 This filter supports the following @ref{commands} that corresponds to option of same name:
7028 Same as the @ref{subtitles} filter, except that it doesn't require libavcodec
7029 and libavformat to work. On the other hand, it is limited to ASS (Advanced
7030 Substation Alpha) subtitles files.
7032 This filter accepts the following option in addition to the common options from
7033 the @ref{subtitles} filter:
7037 Set the shaping engine
7039 Available values are:
7042 The default libass shaping engine, which is the best available.
7044 Fast, font-agnostic shaper that can do only substitutions
7046 Slower shaper using OpenType for substitutions and positioning
7049 The default is @code{auto}.
7053 Apply an Adaptive Temporal Averaging Denoiser to the video input.
7055 The filter accepts the following options:
7059 Set threshold A for 1st plane. Default is 0.02.
7060 Valid range is 0 to 0.3.
7063 Set threshold B for 1st plane. Default is 0.04.
7064 Valid range is 0 to 5.
7067 Set threshold A for 2nd plane. Default is 0.02.
7068 Valid range is 0 to 0.3.
7071 Set threshold B for 2nd plane. Default is 0.04.
7072 Valid range is 0 to 5.
7075 Set threshold A for 3rd plane. Default is 0.02.
7076 Valid range is 0 to 0.3.
7079 Set threshold B for 3rd plane. Default is 0.04.
7080 Valid range is 0 to 5.
7082 Threshold A is designed to react on abrupt changes in the input signal and
7083 threshold B is designed to react on continuous changes in the input signal.
7086 Set number of frames filter will use for averaging. Default is 9. Must be odd
7087 number in range [5, 129].
7090 Set what planes of frame filter will use for averaging. Default is all.
7093 Set what variant of algorithm filter will use for averaging. Default is @code{p} parallel.
7094 Alternatively can be set to @code{s} serial.
7096 Parallel can be faster then serial, while other way around is never true.
7097 Parallel will abort early on first change being greater then thresholds, while serial
7098 will continue processing other side of frames if they are equal or below thresholds.
7101 @subsection Commands
7102 This filter supports same @ref{commands} as options except option @code{s}.
7103 The command accepts the same syntax of the corresponding option.
7107 Apply average blur filter.
7109 The filter accepts the following options:
7113 Set horizontal radius size.
7116 Set which planes to filter. By default all planes are filtered.
7119 Set vertical radius size, if zero it will be same as @code{sizeX}.
7120 Default is @code{0}.
7123 @subsection Commands
7124 This filter supports same commands as options.
7125 The command accepts the same syntax of the corresponding option.
7127 If the specified expression is not valid, it is kept at its current
7132 Compute the bounding box for the non-black pixels in the input frame
7135 This filter computes the bounding box containing all the pixels with a
7136 luminance value greater than the minimum allowed value.
7137 The parameters describing the bounding box are printed on the filter
7140 The filter accepts the following option:
7144 Set the minimal luminance value. Default is @code{16}.
7148 Apply bilateral filter, spatial smoothing while preserving edges.
7150 The filter accepts the following options:
7153 Set sigma of gaussian function to calculate spatial weight.
7154 Allowed range is 0 to 512. Default is 0.1.
7157 Set sigma of gaussian function to calculate range weight.
7158 Allowed range is 0 to 1. Default is 0.1.
7161 Set planes to filter. Default is first only.
7164 @subsection Commands
7166 This filter supports the all above options as @ref{commands}.
7168 @section bitplanenoise
7170 Show and measure bit plane noise.
7172 The filter accepts the following options:
7176 Set which plane to analyze. Default is @code{1}.
7179 Filter out noisy pixels from @code{bitplane} set above.
7180 Default is disabled.
7183 @section blackdetect
7185 Detect video intervals that are (almost) completely black. Can be
7186 useful to detect chapter transitions, commercials, or invalid
7189 The filter outputs its detection analysis to both the log as well as
7190 frame metadata. If a black segment of at least the specified minimum
7191 duration is found, a line with the start and end timestamps as well
7192 as duration is printed to the log with level @code{info}. In addition,
7193 a log line with level @code{debug} is printed per frame showing the
7194 black amount detected for that frame.
7196 The filter also attaches metadata to the first frame of a black
7197 segment with key @code{lavfi.black_start} and to the first frame
7198 after the black segment ends with key @code{lavfi.black_end}. The
7199 value is the frame's timestamp. This metadata is added regardless
7200 of the minimum duration specified.
7202 The filter accepts the following options:
7205 @item black_min_duration, d
7206 Set the minimum detected black duration expressed in seconds. It must
7207 be a non-negative floating point number.
7209 Default value is 2.0.
7211 @item picture_black_ratio_th, pic_th
7212 Set the threshold for considering a picture "black".
7213 Express the minimum value for the ratio:
7215 @var{nb_black_pixels} / @var{nb_pixels}
7218 for which a picture is considered black.
7219 Default value is 0.98.
7221 @item pixel_black_th, pix_th
7222 Set the threshold for considering a pixel "black".
7224 The threshold expresses the maximum pixel luminance value for which a
7225 pixel is considered "black". The provided value is scaled according to
7226 the following equation:
7228 @var{absolute_threshold} = @var{luminance_minimum_value} + @var{pixel_black_th} * @var{luminance_range_size}
7231 @var{luminance_range_size} and @var{luminance_minimum_value} depend on
7232 the input video format, the range is [0-255] for YUV full-range
7233 formats and [16-235] for YUV non full-range formats.
7235 Default value is 0.10.
7238 The following example sets the maximum pixel threshold to the minimum
7239 value, and detects only black intervals of 2 or more seconds:
7241 blackdetect=d=2:pix_th=0.00
7246 Detect frames that are (almost) completely black. Can be useful to
7247 detect chapter transitions or commercials. Output lines consist of
7248 the frame number of the detected frame, the percentage of blackness,
7249 the position in the file if known or -1 and the timestamp in seconds.
7251 In order to display the output lines, you need to set the loglevel at
7252 least to the AV_LOG_INFO value.
7254 This filter exports frame metadata @code{lavfi.blackframe.pblack}.
7255 The value represents the percentage of pixels in the picture that
7256 are below the threshold value.
7258 It accepts the following parameters:
7263 The percentage of the pixels that have to be below the threshold; it defaults to
7266 @item threshold, thresh
7267 The threshold below which a pixel value is considered black; it defaults to
7275 Blend two video frames into each other.
7277 The @code{blend} filter takes two input streams and outputs one
7278 stream, the first input is the "top" layer and second input is
7279 "bottom" layer. By default, the output terminates when the longest input terminates.
7281 The @code{tblend} (time blend) filter takes two consecutive frames
7282 from one single stream, and outputs the result obtained by blending
7283 the new frame on top of the old frame.
7285 A description of the accepted options follows.
7293 Set blend mode for specific pixel component or all pixel components in case
7294 of @var{all_mode}. Default value is @code{normal}.
7296 Available values for component modes are:
7338 Set blend opacity for specific pixel component or all pixel components in case
7339 of @var{all_opacity}. Only used in combination with pixel component blend modes.
7346 Set blend expression for specific pixel component or all pixel components in case
7347 of @var{all_expr}. Note that related mode options will be ignored if those are set.
7349 The expressions can use the following variables:
7353 The sequential number of the filtered frame, starting from @code{0}.
7357 the coordinates of the current sample
7361 the width and height of currently filtered plane
7365 Width and height scale for the plane being filtered. It is the
7366 ratio between the dimensions of the current plane to the luma plane,
7367 e.g. for a @code{yuv420p} frame, the values are @code{1,1} for
7368 the luma plane and @code{0.5,0.5} for the chroma planes.
7371 Time of the current frame, expressed in seconds.
7374 Value of pixel component at current location for first video frame (top layer).
7377 Value of pixel component at current location for second video frame (bottom layer).
7381 The @code{blend} filter also supports the @ref{framesync} options.
7383 @subsection Examples
7387 Apply transition from bottom layer to top layer in first 10 seconds:
7389 blend=all_expr='A*(if(gte(T,10),1,T/10))+B*(1-(if(gte(T,10),1,T/10)))'
7393 Apply linear horizontal transition from top layer to bottom layer:
7395 blend=all_expr='A*(X/W)+B*(1-X/W)'
7399 Apply 1x1 checkerboard effect:
7401 blend=all_expr='if(eq(mod(X,2),mod(Y,2)),A,B)'
7405 Apply uncover left effect:
7407 blend=all_expr='if(gte(N*SW+X,W),A,B)'
7411 Apply uncover down effect:
7413 blend=all_expr='if(gte(Y-N*SH,0),A,B)'
7417 Apply uncover up-left effect:
7419 blend=all_expr='if(gte(T*SH*40+Y,H)*gte((T*40*SW+X)*W/H,W),A,B)'
7423 Split diagonally video and shows top and bottom layer on each side:
7425 blend=all_expr='if(gt(X,Y*(W/H)),A,B)'
7429 Display differences between the current and the previous frame:
7431 tblend=all_mode=grainextract
7437 Denoise frames using Block-Matching 3D algorithm.
7439 The filter accepts the following options.
7443 Set denoising strength. Default value is 1.
7444 Allowed range is from 0 to 999.9.
7445 The denoising algorithm is very sensitive to sigma, so adjust it
7446 according to the source.
7449 Set local patch size. This sets dimensions in 2D.
7452 Set sliding step for processing blocks. Default value is 4.
7453 Allowed range is from 1 to 64.
7454 Smaller values allows processing more reference blocks and is slower.
7457 Set maximal number of similar blocks for 3rd dimension. Default value is 1.
7458 When set to 1, no block matching is done. Larger values allows more blocks
7460 Allowed range is from 1 to 256.
7463 Set radius for search block matching. Default is 9.
7464 Allowed range is from 1 to INT32_MAX.
7467 Set step between two search locations for block matching. Default is 1.
7468 Allowed range is from 1 to 64. Smaller is slower.
7471 Set threshold of mean square error for block matching. Valid range is 0 to
7475 Set thresholding parameter for hard thresholding in 3D transformed domain.
7476 Larger values results in stronger hard-thresholding filtering in frequency
7480 Set filtering estimation mode. Can be @code{basic} or @code{final}.
7481 Default is @code{basic}.
7484 If enabled, filter will use 2nd stream for block matching.
7485 Default is disabled for @code{basic} value of @var{estim} option,
7486 and always enabled if value of @var{estim} is @code{final}.
7489 Set planes to filter. Default is all available except alpha.
7492 @subsection Examples
7496 Basic filtering with bm3d:
7498 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic
7502 Same as above, but filtering only luma:
7504 bm3d=sigma=3:block=4:bstep=2:group=1:estim=basic:planes=1
7508 Same as above, but with both estimation modes:
7510 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
7514 Same as above, but prefilter with @ref{nlmeans} filter instead:
7516 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
7522 Apply a boxblur algorithm to the input video.
7524 It accepts the following parameters:
7528 @item luma_radius, lr
7529 @item luma_power, lp
7530 @item chroma_radius, cr
7531 @item chroma_power, cp
7532 @item alpha_radius, ar
7533 @item alpha_power, ap
7537 A description of the accepted options follows.
7540 @item luma_radius, lr
7541 @item chroma_radius, cr
7542 @item alpha_radius, ar
7543 Set an expression for the box radius in pixels used for blurring the
7544 corresponding input plane.
7546 The radius value must be a non-negative number, and must not be
7547 greater than the value of the expression @code{min(w,h)/2} for the
7548 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
7551 Default value for @option{luma_radius} is "2". If not specified,
7552 @option{chroma_radius} and @option{alpha_radius} default to the
7553 corresponding value set for @option{luma_radius}.
7555 The expressions can contain the following constants:
7559 The input width and height in pixels.
7563 The input chroma image width and height in pixels.
7567 The horizontal and vertical chroma subsample values. For example, for the
7568 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
7571 @item luma_power, lp
7572 @item chroma_power, cp
7573 @item alpha_power, ap
7574 Specify how many times the boxblur filter is applied to the
7575 corresponding plane.
7577 Default value for @option{luma_power} is 2. If not specified,
7578 @option{chroma_power} and @option{alpha_power} default to the
7579 corresponding value set for @option{luma_power}.
7581 A value of 0 will disable the effect.
7584 @subsection Examples
7588 Apply a boxblur filter with the luma, chroma, and alpha radii
7591 boxblur=luma_radius=2:luma_power=1
7596 Set the luma radius to 2, and alpha and chroma radius to 0:
7598 boxblur=2:1:cr=0:ar=0
7602 Set the luma and chroma radii to a fraction of the video dimension:
7604 boxblur=luma_radius=min(h\,w)/10:luma_power=1:chroma_radius=min(cw\,ch)/10:chroma_power=1
7610 Deinterlace the input video ("bwdif" stands for "Bob Weaver
7611 Deinterlacing Filter").
7613 Motion adaptive deinterlacing based on yadif with the use of w3fdif and cubic
7614 interpolation algorithms.
7615 It accepts the following parameters:
7619 The interlacing mode to adopt. It accepts one of the following values:
7623 Output one frame for each frame.
7625 Output one frame for each field.
7628 The default value is @code{send_field}.
7631 The picture field parity assumed for the input interlaced video. It accepts one
7632 of the following values:
7636 Assume the top field is first.
7638 Assume the bottom field is first.
7640 Enable automatic detection of field parity.
7643 The default value is @code{auto}.
7644 If the interlacing is unknown or the decoder does not export this information,
7645 top field first will be assumed.
7648 Specify which frames to deinterlace. Accepts one of the following
7653 Deinterlace all frames.
7655 Only deinterlace frames marked as interlaced.
7658 The default value is @code{all}.
7663 Apply Contrast Adaptive Sharpen filter to video stream.
7665 The filter accepts the following options:
7669 Set the sharpening strength. Default value is 0.
7672 Set planes to filter. Default value is to filter all
7673 planes except alpha plane.
7676 @subsection Commands
7677 This filter supports same @ref{commands} as options.
7680 Remove all color information for all colors except for certain one.
7682 The filter accepts the following options:
7686 The color which will not be replaced with neutral chroma.
7689 Similarity percentage with the above color.
7690 0.01 matches only the exact key color, while 1.0 matches everything.
7694 0.0 makes pixels either fully gray, or not gray at all.
7695 Higher values result in more preserved color.
7698 Signals that the color passed is already in YUV instead of RGB.
7700 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7701 This can be used to pass exact YUV values as hexadecimal numbers.
7704 @subsection Commands
7705 This filter supports same @ref{commands} as options.
7706 The command accepts the same syntax of the corresponding option.
7708 If the specified expression is not valid, it is kept at its current
7712 YUV colorspace color/chroma keying.
7714 The filter accepts the following options:
7718 The color which will be replaced with transparency.
7721 Similarity percentage with the key color.
7723 0.01 matches only the exact key color, while 1.0 matches everything.
7728 0.0 makes pixels either fully transparent, or not transparent at all.
7730 Higher values result in semi-transparent pixels, with a higher transparency
7731 the more similar the pixels color is to the key color.
7734 Signals that the color passed is already in YUV instead of RGB.
7736 Literal colors like "green" or "red" don't make sense with this enabled anymore.
7737 This can be used to pass exact YUV values as hexadecimal numbers.
7740 @subsection Commands
7741 This filter supports same @ref{commands} as options.
7742 The command accepts the same syntax of the corresponding option.
7744 If the specified expression is not valid, it is kept at its current
7747 @subsection Examples
7751 Make every green pixel in the input image transparent:
7753 ffmpeg -i input.png -vf chromakey=green out.png
7757 Overlay a greenscreen-video on top of a static black background.
7759 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
7764 Reduce chrominance noise.
7766 The filter accepts the following options:
7770 Set threshold for averaging chrominance values.
7771 Sum of absolute difference of Y, U and V pixel components of current
7772 pixel and neighbour pixels lower than this threshold will be used in
7773 averaging. Luma component is left unchanged and is copied to output.
7774 Default value is 30. Allowed range is from 1 to 200.
7777 Set horizontal radius of rectangle used for averaging.
7778 Allowed range is from 1 to 100. Default value is 5.
7781 Set vertical radius of rectangle used for averaging.
7782 Allowed range is from 1 to 100. Default value is 5.
7785 Set horizontal step when averaging. Default value is 1.
7786 Allowed range is from 1 to 50.
7787 Mostly useful to speed-up filtering.
7790 Set vertical step when averaging. Default value is 1.
7791 Allowed range is from 1 to 50.
7792 Mostly useful to speed-up filtering.
7795 Set Y threshold for averaging chrominance values.
7796 Set finer control for max allowed difference between Y components
7797 of current pixel and neigbour pixels.
7798 Default value is 200. Allowed range is from 1 to 200.
7801 Set U threshold for averaging chrominance values.
7802 Set finer control for max allowed difference between U components
7803 of current pixel and neigbour pixels.
7804 Default value is 200. Allowed range is from 1 to 200.
7807 Set V threshold for averaging chrominance values.
7808 Set finer control for max allowed difference between V components
7809 of current pixel and neigbour pixels.
7810 Default value is 200. Allowed range is from 1 to 200.
7813 @subsection Commands
7814 This filter supports same @ref{commands} as options.
7815 The command accepts the same syntax of the corresponding option.
7817 @section chromashift
7818 Shift chroma pixels horizontally and/or vertically.
7820 The filter accepts the following options:
7823 Set amount to shift chroma-blue horizontally.
7825 Set amount to shift chroma-blue vertically.
7827 Set amount to shift chroma-red horizontally.
7829 Set amount to shift chroma-red vertically.
7831 Set edge mode, can be @var{smear}, default, or @var{warp}.
7834 @subsection Commands
7836 This filter supports the all above options as @ref{commands}.
7840 Display CIE color diagram with pixels overlaid onto it.
7842 The filter accepts the following options:
7857 @item uhdtv, rec2020
7871 Set what gamuts to draw.
7873 See @code{system} option for available values.
7876 Set ciescope size, by default set to 512.
7879 Set intensity used to map input pixel values to CIE diagram.
7882 Set contrast used to draw tongue colors that are out of active color system gamut.
7885 Correct gamma displayed on scope, by default enabled.
7888 Show white point on CIE diagram, by default disabled.
7891 Set input gamma. Used only with XYZ input color space.
7896 Visualize information exported by some codecs.
7898 Some codecs can export information through frames using side-data or other
7899 means. For example, some MPEG based codecs export motion vectors through the
7900 @var{export_mvs} flag in the codec @option{flags2} option.
7902 The filter accepts the following option:
7906 Set motion vectors to visualize.
7908 Available flags for @var{mv} are:
7912 forward predicted MVs of P-frames
7914 forward predicted MVs of B-frames
7916 backward predicted MVs of B-frames
7920 Display quantization parameters using the chroma planes.
7923 Set motion vectors type to visualize. Includes MVs from all frames unless specified by @var{frame_type} option.
7925 Available flags for @var{mv_type} are:
7929 forward predicted MVs
7931 backward predicted MVs
7934 @item frame_type, ft
7935 Set frame type to visualize motion vectors of.
7937 Available flags for @var{frame_type} are:
7941 intra-coded frames (I-frames)
7943 predicted frames (P-frames)
7945 bi-directionally predicted frames (B-frames)
7949 @subsection Examples
7953 Visualize forward predicted MVs of all frames using @command{ffplay}:
7955 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv_type=fp
7959 Visualize multi-directionals MVs of P and B-Frames using @command{ffplay}:
7961 ffplay -flags2 +export_mvs input.mp4 -vf codecview=mv=pf+bf+bb
7965 @section colorbalance
7966 Modify intensity of primary colors (red, green and blue) of input frames.
7968 The filter allows an input frame to be adjusted in the shadows, midtones or highlights
7969 regions for the red-cyan, green-magenta or blue-yellow balance.
7971 A positive adjustment value shifts the balance towards the primary color, a negative
7972 value towards the complementary color.
7974 The filter accepts the following options:
7980 Adjust red, green and blue shadows (darkest pixels).
7985 Adjust red, green and blue midtones (medium pixels).
7990 Adjust red, green and blue highlights (brightest pixels).
7992 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
7995 Preserve lightness when changing color balance. Default is disabled.
7998 @subsection Examples
8002 Add red color cast to shadows:
8008 @subsection Commands
8010 This filter supports the all above options as @ref{commands}.
8012 @section colorchannelmixer
8014 Adjust video input frames by re-mixing color channels.
8016 This filter modifies a color channel by adding the values associated to
8017 the other channels of the same pixels. For example if the value to
8018 modify is red, the output value will be:
8020 @var{red}=@var{red}*@var{rr} + @var{blue}*@var{rb} + @var{green}*@var{rg} + @var{alpha}*@var{ra}
8023 The filter accepts the following options:
8030 Adjust contribution of input red, green, blue and alpha channels for output red channel.
8031 Default is @code{1} for @var{rr}, and @code{0} for @var{rg}, @var{rb} and @var{ra}.
8037 Adjust contribution of input red, green, blue and alpha channels for output green channel.
8038 Default is @code{1} for @var{gg}, and @code{0} for @var{gr}, @var{gb} and @var{ga}.
8044 Adjust contribution of input red, green, blue and alpha channels for output blue channel.
8045 Default is @code{1} for @var{bb}, and @code{0} for @var{br}, @var{bg} and @var{ba}.
8051 Adjust contribution of input red, green, blue and alpha channels for output alpha channel.
8052 Default is @code{1} for @var{aa}, and @code{0} for @var{ar}, @var{ag} and @var{ab}.
8054 Allowed ranges for options are @code{[-2.0, 2.0]}.
8057 @subsection Examples
8061 Convert source to grayscale:
8063 colorchannelmixer=.3:.4:.3:0:.3:.4:.3:0:.3:.4:.3
8066 Simulate sepia tones:
8068 colorchannelmixer=.393:.769:.189:0:.349:.686:.168:0:.272:.534:.131
8072 @subsection Commands
8074 This filter supports the all above options as @ref{commands}.
8077 RGB colorspace color keying.
8079 The filter accepts the following options:
8083 The color which will be replaced with transparency.
8086 Similarity percentage with the key color.
8088 0.01 matches only the exact key color, while 1.0 matches everything.
8093 0.0 makes pixels either fully transparent, or not transparent at all.
8095 Higher values result in semi-transparent pixels, with a higher transparency
8096 the more similar the pixels color is to the key color.
8099 @subsection Examples
8103 Make every green pixel in the input image transparent:
8105 ffmpeg -i input.png -vf colorkey=green out.png
8109 Overlay a greenscreen-video on top of a static background image.
8111 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
8115 @subsection Commands
8116 This filter supports same @ref{commands} as options.
8117 The command accepts the same syntax of the corresponding option.
8119 If the specified expression is not valid, it is kept at its current
8123 Remove all color information for all RGB colors except for certain one.
8125 The filter accepts the following options:
8129 The color which will not be replaced with neutral gray.
8132 Similarity percentage with the above color.
8133 0.01 matches only the exact key color, while 1.0 matches everything.
8136 Blend percentage. 0.0 makes pixels fully gray.
8137 Higher values result in more preserved color.
8140 @subsection Commands
8141 This filter supports same @ref{commands} as options.
8142 The command accepts the same syntax of the corresponding option.
8144 If the specified expression is not valid, it is kept at its current
8147 @section colorlevels
8149 Adjust video input frames using levels.
8151 The filter accepts the following options:
8158 Adjust red, green, blue and alpha input black point.
8159 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{0}.
8165 Adjust red, green, blue and alpha input white point.
8166 Allowed ranges for options are @code{[-1.0, 1.0]}. Defaults are @code{1}.
8168 Input levels are used to lighten highlights (bright tones), darken shadows
8169 (dark tones), change the balance of bright and dark tones.
8175 Adjust red, green, blue and alpha output black point.
8176 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{0}.
8182 Adjust red, green, blue and alpha output white point.
8183 Allowed ranges for options are @code{[0, 1.0]}. Defaults are @code{1}.
8185 Output levels allows manual selection of a constrained output level range.
8188 @subsection Examples
8192 Make video output darker:
8194 colorlevels=rimin=0.058:gimin=0.058:bimin=0.058
8200 colorlevels=rimin=0.039:gimin=0.039:bimin=0.039:rimax=0.96:gimax=0.96:bimax=0.96
8204 Make video output lighter:
8206 colorlevels=rimax=0.902:gimax=0.902:bimax=0.902
8210 Increase brightness:
8212 colorlevels=romin=0.5:gomin=0.5:bomin=0.5
8216 @subsection Commands
8218 This filter supports the all above options as @ref{commands}.
8220 @section colormatrix
8222 Convert color matrix.
8224 The filter accepts the following options:
8229 Specify the source and destination color matrix. Both values must be
8232 The accepted values are:
8260 For example to convert from BT.601 to SMPTE-240M, use the command:
8262 colormatrix=bt601:smpte240m
8267 Convert colorspace, transfer characteristics or color primaries.
8268 Input video needs to have an even size.
8270 The filter accepts the following options:
8275 Specify all color properties at once.
8277 The accepted values are:
8307 Specify output colorspace.
8309 The accepted values are:
8318 BT.470BG or BT.601-6 625
8321 SMPTE-170M or BT.601-6 525
8330 BT.2020 with non-constant luminance
8336 Specify output transfer characteristics.
8338 The accepted values are:
8350 Constant gamma of 2.2
8353 Constant gamma of 2.8
8356 SMPTE-170M, BT.601-6 625 or BT.601-6 525
8374 BT.2020 for 10-bits content
8377 BT.2020 for 12-bits content
8383 Specify output color primaries.
8385 The accepted values are:
8394 BT.470BG or BT.601-6 625
8397 SMPTE-170M or BT.601-6 525
8421 Specify output color range.
8423 The accepted values are:
8426 TV (restricted) range
8429 MPEG (restricted) range
8440 Specify output color format.
8442 The accepted values are:
8445 YUV 4:2:0 planar 8-bits
8448 YUV 4:2:0 planar 10-bits
8451 YUV 4:2:0 planar 12-bits
8454 YUV 4:2:2 planar 8-bits
8457 YUV 4:2:2 planar 10-bits
8460 YUV 4:2:2 planar 12-bits
8463 YUV 4:4:4 planar 8-bits
8466 YUV 4:4:4 planar 10-bits
8469 YUV 4:4:4 planar 12-bits
8474 Do a fast conversion, which skips gamma/primary correction. This will take
8475 significantly less CPU, but will be mathematically incorrect. To get output
8476 compatible with that produced by the colormatrix filter, use fast=1.
8479 Specify dithering mode.
8481 The accepted values are:
8487 Floyd-Steinberg dithering
8491 Whitepoint adaptation mode.
8493 The accepted values are:
8496 Bradford whitepoint adaptation
8499 von Kries whitepoint adaptation
8502 identity whitepoint adaptation (i.e. no whitepoint adaptation)
8506 Override all input properties at once. Same accepted values as @ref{all}.
8509 Override input colorspace. Same accepted values as @ref{space}.
8512 Override input color primaries. Same accepted values as @ref{primaries}.
8515 Override input transfer characteristics. Same accepted values as @ref{trc}.
8518 Override input color range. Same accepted values as @ref{range}.
8522 The filter converts the transfer characteristics, color space and color
8523 primaries to the specified user values. The output value, if not specified,
8524 is set to a default value based on the "all" property. If that property is
8525 also not specified, the filter will log an error. The output color range and
8526 format default to the same value as the input color range and format. The
8527 input transfer characteristics, color space, color primaries and color range
8528 should be set on the input data. If any of these are missing, the filter will
8529 log an error and no conversion will take place.
8531 For example to convert the input to SMPTE-240M, use the command:
8533 colorspace=smpte240m
8536 @section convolution
8538 Apply convolution of 3x3, 5x5, 7x7 or horizontal/vertical up to 49 elements.
8540 The filter accepts the following options:
8547 Set matrix for each plane.
8548 Matrix is sequence of 9, 25 or 49 signed integers in @var{square} mode,
8549 and from 1 to 49 odd number of signed integers in @var{row} mode.
8555 Set multiplier for calculated value for each plane.
8556 If unset or 0, it will be sum of all matrix elements.
8562 Set bias for each plane. This value is added to the result of the multiplication.
8563 Useful for making the overall image brighter or darker. Default is 0.0.
8569 Set matrix mode for each plane. Can be @var{square}, @var{row} or @var{column}.
8570 Default is @var{square}.
8573 @subsection Commands
8575 This filter supports the all above options as @ref{commands}.
8577 @subsection Examples
8583 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"
8589 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"
8595 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"
8601 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"
8605 Apply laplacian edge detector which includes diagonals:
8607 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"
8613 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"
8619 Apply 2D convolution of video stream in frequency domain using second stream
8622 The filter accepts the following options:
8626 Set which planes to process.
8629 Set which impulse video frames will be processed, can be @var{first}
8630 or @var{all}. Default is @var{all}.
8633 The @code{convolve} filter also supports the @ref{framesync} options.
8637 Copy the input video source unchanged to the output. This is mainly useful for
8642 Video filtering on GPU using Apple's CoreImage API on OSX.
8644 Hardware acceleration is based on an OpenGL context. Usually, this means it is
8645 processed by video hardware. However, software-based OpenGL implementations
8646 exist which means there is no guarantee for hardware processing. It depends on
8649 There are many filters and image generators provided by Apple that come with a
8650 large variety of options. The filter has to be referenced by its name along
8653 The coreimage filter accepts the following options:
8656 List all available filters and generators along with all their respective
8657 options as well as possible minimum and maximum values along with the default
8664 Specify all filters by their respective name and options.
8665 Use @var{list_filters} to determine all valid filter names and options.
8666 Numerical options are specified by a float value and are automatically clamped
8667 to their respective value range. Vector and color options have to be specified
8668 by a list of space separated float values. Character escaping has to be done.
8669 A special option name @code{default} is available to use default options for a
8672 It is required to specify either @code{default} or at least one of the filter options.
8673 All omitted options are used with their default values.
8674 The syntax of the filter string is as follows:
8676 filter=<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...][#<NAME>@@<OPTION>=<VALUE>[@@<OPTION>=<VALUE>][@@...]][#...]
8680 Specify a rectangle where the output of the filter chain is copied into the
8681 input image. It is given by a list of space separated float values:
8683 output_rect=x\ y\ width\ height
8685 If not given, the output rectangle equals the dimensions of the input image.
8686 The output rectangle is automatically cropped at the borders of the input
8687 image. Negative values are valid for each component.
8689 output_rect=25\ 25\ 100\ 100
8693 Several filters can be chained for successive processing without GPU-HOST
8694 transfers allowing for fast processing of complex filter chains.
8695 Currently, only filters with zero (generators) or exactly one (filters) input
8696 image and one output image are supported. Also, transition filters are not yet
8699 Some filters generate output images with additional padding depending on the
8700 respective filter kernel. The padding is automatically removed to ensure the
8701 filter output has the same size as the input image.
8703 For image generators, the size of the output image is determined by the
8704 previous output image of the filter chain or the input image of the whole
8705 filterchain, respectively. The generators do not use the pixel information of
8706 this image to generate their output. However, the generated output is
8707 blended onto this image, resulting in partial or complete coverage of the
8710 The @ref{coreimagesrc} video source can be used for generating input images
8711 which are directly fed into the filter chain. By using it, providing input
8712 images by another video source or an input video is not required.
8714 @subsection Examples
8719 List all filters available:
8721 coreimage=list_filters=true
8725 Use the CIBoxBlur filter with default options to blur an image:
8727 coreimage=filter=CIBoxBlur@@default
8731 Use a filter chain with CISepiaTone at default values and CIVignetteEffect with
8732 its center at 100x100 and a radius of 50 pixels:
8734 coreimage=filter=CIBoxBlur@@default#CIVignetteEffect@@inputCenter=100\ 100@@inputRadius=50
8738 Use nullsrc and CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
8739 given as complete and escaped command-line for Apple's standard bash shell:
8741 ffmpeg -f lavfi -i nullsrc=s=100x100,coreimage=filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
8747 Cover a rectangular object
8749 It accepts the following options:
8753 Filepath of the optional cover image, needs to be in yuv420.
8758 It accepts the following values:
8761 cover it by the supplied image
8763 cover it by interpolating the surrounding pixels
8766 Default value is @var{blur}.
8769 @subsection Examples
8773 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
8775 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
8781 Crop the input video to given dimensions.
8783 It accepts the following parameters:
8787 The width of the output video. It defaults to @code{iw}.
8788 This expression is evaluated only once during the filter
8789 configuration, or when the @samp{w} or @samp{out_w} command is sent.
8792 The height of the output video. It defaults to @code{ih}.
8793 This expression is evaluated only once during the filter
8794 configuration, or when the @samp{h} or @samp{out_h} command is sent.
8797 The horizontal position, in the input video, of the left edge of the output
8798 video. It defaults to @code{(in_w-out_w)/2}.
8799 This expression is evaluated per-frame.
8802 The vertical position, in the input video, of the top edge of the output video.
8803 It defaults to @code{(in_h-out_h)/2}.
8804 This expression is evaluated per-frame.
8807 If set to 1 will force the output display aspect ratio
8808 to be the same of the input, by changing the output sample aspect
8809 ratio. It defaults to 0.
8812 Enable exact cropping. If enabled, subsampled videos will be cropped at exact
8813 width/height/x/y as specified and will not be rounded to nearest smaller value.
8817 The @var{out_w}, @var{out_h}, @var{x}, @var{y} parameters are
8818 expressions containing the following constants:
8823 The computed values for @var{x} and @var{y}. They are evaluated for
8828 The input width and height.
8832 These are the same as @var{in_w} and @var{in_h}.
8836 The output (cropped) width and height.
8840 These are the same as @var{out_w} and @var{out_h}.
8843 same as @var{iw} / @var{ih}
8846 input sample aspect ratio
8849 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
8853 horizontal and vertical chroma subsample values. For example for the
8854 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
8857 The number of the input frame, starting from 0.
8860 the position in the file of the input frame, NAN if unknown
8863 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
8867 The expression for @var{out_w} may depend on the value of @var{out_h},
8868 and the expression for @var{out_h} may depend on @var{out_w}, but they
8869 cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
8870 evaluated after @var{out_w} and @var{out_h}.
8872 The @var{x} and @var{y} parameters specify the expressions for the
8873 position of the top-left corner of the output (non-cropped) area. They
8874 are evaluated for each frame. If the evaluated value is not valid, it
8875 is approximated to the nearest valid value.
8877 The expression for @var{x} may depend on @var{y}, and the expression
8878 for @var{y} may depend on @var{x}.
8880 @subsection Examples
8884 Crop area with size 100x100 at position (12,34).
8889 Using named options, the example above becomes:
8891 crop=w=100:h=100:x=12:y=34
8895 Crop the central input area with size 100x100:
8901 Crop the central input area with size 2/3 of the input video:
8903 crop=2/3*in_w:2/3*in_h
8907 Crop the input video central square:
8914 Delimit the rectangle with the top-left corner placed at position
8915 100:100 and the right-bottom corner corresponding to the right-bottom
8916 corner of the input image.
8918 crop=in_w-100:in_h-100:100:100
8922 Crop 10 pixels from the left and right borders, and 20 pixels from
8923 the top and bottom borders
8925 crop=in_w-2*10:in_h-2*20
8929 Keep only the bottom right quarter of the input image:
8931 crop=in_w/2:in_h/2:in_w/2:in_h/2
8935 Crop height for getting Greek harmony:
8937 crop=in_w:1/PHI*in_w
8941 Apply trembling effect:
8943 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)
8947 Apply erratic camera effect depending on timestamp:
8949 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)"
8953 Set x depending on the value of y:
8955 crop=in_w/2:in_h/2:y:10+10*sin(n/10)
8959 @subsection Commands
8961 This filter supports the following commands:
8967 Set width/height of the output video and the horizontal/vertical position
8969 The command accepts the same syntax of the corresponding option.
8971 If the specified expression is not valid, it is kept at its current
8977 Auto-detect the crop size.
8979 It calculates the necessary cropping parameters and prints the
8980 recommended parameters via the logging system. The detected dimensions
8981 correspond to the non-black area of the input video.
8983 It accepts the following parameters:
8988 Set higher black value threshold, which can be optionally specified
8989 from nothing (0) to everything (255 for 8-bit based formats). An intensity
8990 value greater to the set value is considered non-black. It defaults to 24.
8991 You can also specify a value between 0.0 and 1.0 which will be scaled depending
8992 on the bitdepth of the pixel format.
8995 The value which the width/height should be divisible by. It defaults to
8996 16. The offset is automatically adjusted to center the video. Use 2 to
8997 get only even dimensions (needed for 4:2:2 video). 16 is best when
8998 encoding to most video codecs.
9001 Set the number of initial frames for which evaluation is skipped.
9002 Default is 2. Range is 0 to INT_MAX.
9004 @item reset_count, reset
9005 Set the counter that determines after how many frames cropdetect will
9006 reset the previously detected largest video area and start over to
9007 detect the current optimal crop area. Default value is 0.
9009 This can be useful when channel logos distort the video area. 0
9010 indicates 'never reset', and returns the largest area encountered during
9017 Delay video filtering until a given wallclock timestamp. The filter first
9018 passes on @option{preroll} amount of frames, then it buffers at most
9019 @option{buffer} amount of frames and waits for the cue. After reaching the cue
9020 it forwards the buffered frames and also any subsequent frames coming in its
9023 The filter can be used synchronize the output of multiple ffmpeg processes for
9024 realtime output devices like decklink. By putting the delay in the filtering
9025 chain and pre-buffering frames the process can pass on data to output almost
9026 immediately after the target wallclock timestamp is reached.
9028 Perfect frame accuracy cannot be guaranteed, but the result is good enough for
9034 The cue timestamp expressed in a UNIX timestamp in microseconds. Default is 0.
9037 The duration of content to pass on as preroll expressed in seconds. Default is 0.
9040 The maximum duration of content to buffer before waiting for the cue expressed
9041 in seconds. Default is 0.
9048 Apply color adjustments using curves.
9050 This filter is similar to the Adobe Photoshop and GIMP curves tools. Each
9051 component (red, green and blue) has its values defined by @var{N} key points
9052 tied from each other using a smooth curve. The x-axis represents the pixel
9053 values from the input frame, and the y-axis the new pixel values to be set for
9056 By default, a component curve is defined by the two points @var{(0;0)} and
9057 @var{(1;1)}. This creates a straight line where each original pixel value is
9058 "adjusted" to its own value, which means no change to the image.
9060 The filter allows you to redefine these two points and add some more. A new
9061 curve (using a natural cubic spline interpolation) will be define to pass
9062 smoothly through all these new coordinates. The new defined points needs to be
9063 strictly increasing over the x-axis, and their @var{x} and @var{y} values must
9064 be in the @var{[0;1]} interval. If the computed curves happened to go outside
9065 the vector spaces, the values will be clipped accordingly.
9067 The filter accepts the following options:
9071 Select one of the available color presets. This option can be used in addition
9072 to the @option{r}, @option{g}, @option{b} parameters; in this case, the later
9073 options takes priority on the preset values.
9074 Available presets are:
9077 @item color_negative
9080 @item increase_contrast
9082 @item linear_contrast
9083 @item medium_contrast
9085 @item strong_contrast
9088 Default is @code{none}.
9090 Set the master key points. These points will define a second pass mapping. It
9091 is sometimes called a "luminance" or "value" mapping. It can be used with
9092 @option{r}, @option{g}, @option{b} or @option{all} since it acts like a
9093 post-processing LUT.
9095 Set the key points for the red component.
9097 Set the key points for the green component.
9099 Set the key points for the blue component.
9101 Set the key points for all components (not including master).
9102 Can be used in addition to the other key points component
9103 options. In this case, the unset component(s) will fallback on this
9104 @option{all} setting.
9106 Specify a Photoshop curves file (@code{.acv}) to import the settings from.
9108 Save Gnuplot script of the curves in specified file.
9111 To avoid some filtergraph syntax conflicts, each key points list need to be
9112 defined using the following syntax: @code{x0/y0 x1/y1 x2/y2 ...}.
9114 @subsection Examples
9118 Increase slightly the middle level of blue:
9120 curves=blue='0/0 0.5/0.58 1/1'
9126 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'
9128 Here we obtain the following coordinates for each components:
9131 @code{(0;0.11) (0.42;0.51) (1;0.95)}
9133 @code{(0;0) (0.50;0.48) (1;1)}
9135 @code{(0;0.22) (0.49;0.44) (1;0.80)}
9139 The previous example can also be achieved with the associated built-in preset:
9141 curves=preset=vintage
9151 Use a Photoshop preset and redefine the points of the green component:
9153 curves=psfile='MyCurvesPresets/purple.acv':green='0/0 0.45/0.53 1/1'
9157 Check out the curves of the @code{cross_process} profile using @command{ffmpeg}
9158 and @command{gnuplot}:
9160 ffmpeg -f lavfi -i color -vf curves=cross_process:plot=/tmp/curves.plt -frames:v 1 -f null -
9161 gnuplot -p /tmp/curves.plt
9167 Video data analysis filter.
9169 This filter shows hexadecimal pixel values of part of video.
9171 The filter accepts the following options:
9175 Set output video size.
9178 Set x offset from where to pick pixels.
9181 Set y offset from where to pick pixels.
9184 Set scope mode, can be one of the following:
9187 Draw hexadecimal pixel values with white color on black background.
9190 Draw hexadecimal pixel values with input video pixel color on black
9194 Draw hexadecimal pixel values on color background picked from input video,
9195 the text color is picked in such way so its always visible.
9199 Draw rows and columns numbers on left and top of video.
9202 Set background opacity.
9205 Set display number format. Can be @code{hex}, or @code{dec}. Default is @code{hex}.
9209 Apply Directional blur filter.
9211 The filter accepts the following options:
9215 Set angle of directional blur. Default is @code{45}.
9218 Set radius of directional blur. Default is @code{5}.
9221 Set which planes to filter. By default all planes are filtered.
9224 @subsection Commands
9225 This filter supports same @ref{commands} as options.
9226 The command accepts the same syntax of the corresponding option.
9228 If the specified expression is not valid, it is kept at its current
9233 Denoise frames using 2D DCT (frequency domain filtering).
9235 This filter is not designed for real time.
9237 The filter accepts the following options:
9241 Set the noise sigma constant.
9243 This @var{sigma} defines a hard threshold of @code{3 * sigma}; every DCT
9244 coefficient (absolute value) below this threshold with be dropped.
9246 If you need a more advanced filtering, see @option{expr}.
9248 Default is @code{0}.
9251 Set number overlapping pixels for each block. Since the filter can be slow, you
9252 may want to reduce this value, at the cost of a less effective filter and the
9253 risk of various artefacts.
9255 If the overlapping value doesn't permit processing the whole input width or
9256 height, a warning will be displayed and according borders won't be denoised.
9258 Default value is @var{blocksize}-1, which is the best possible setting.
9261 Set the coefficient factor expression.
9263 For each coefficient of a DCT block, this expression will be evaluated as a
9264 multiplier value for the coefficient.
9266 If this is option is set, the @option{sigma} option will be ignored.
9268 The absolute value of the coefficient can be accessed through the @var{c}
9272 Set the @var{blocksize} using the number of bits. @code{1<<@var{n}} defines the
9273 @var{blocksize}, which is the width and height of the processed blocks.
9275 The default value is @var{3} (8x8) and can be raised to @var{4} for a
9276 @var{blocksize} of 16x16. Note that changing this setting has huge consequences
9277 on the speed processing. Also, a larger block size does not necessarily means a
9281 @subsection Examples
9283 Apply a denoise with a @option{sigma} of @code{4.5}:
9288 The same operation can be achieved using the expression system:
9290 dctdnoiz=e='gte(c, 4.5*3)'
9293 Violent denoise using a block size of @code{16x16}:
9300 Remove banding artifacts from input video.
9301 It works by replacing banded pixels with average value of referenced pixels.
9303 The filter accepts the following options:
9310 Set banding detection threshold for each plane. Default is 0.02.
9311 Valid range is 0.00003 to 0.5.
9312 If difference between current pixel and reference pixel is less than threshold,
9313 it will be considered as banded.
9316 Banding detection range in pixels. Default is 16. If positive, random number
9317 in range 0 to set value will be used. If negative, exact absolute value
9319 The range defines square of four pixels around current pixel.
9322 Set direction in radians from which four pixel will be compared. If positive,
9323 random direction from 0 to set direction will be picked. If negative, exact of
9324 absolute value will be picked. For example direction 0, -PI or -2*PI radians
9325 will pick only pixels on same row and -PI/2 will pick only pixels on same
9329 If enabled, current pixel is compared with average value of all four
9330 surrounding pixels. The default is enabled. If disabled current pixel is
9331 compared with all four surrounding pixels. The pixel is considered banded
9332 if only all four differences with surrounding pixels are less than threshold.
9335 If enabled, current pixel is changed if and only if all pixel components are banded,
9336 e.g. banding detection threshold is triggered for all color components.
9337 The default is disabled.
9342 Remove blocking artifacts from input video.
9344 The filter accepts the following options:
9348 Set filter type, can be @var{weak} or @var{strong}. Default is @var{strong}.
9349 This controls what kind of deblocking is applied.
9352 Set size of block, allowed range is from 4 to 512. Default is @var{8}.
9358 Set blocking detection thresholds. Allowed range is 0 to 1.
9359 Defaults are: @var{0.098} for @var{alpha} and @var{0.05} for the rest.
9360 Using higher threshold gives more deblocking strength.
9361 Setting @var{alpha} controls threshold detection at exact edge of block.
9362 Remaining options controls threshold detection near the edge. Each one for
9363 below/above or left/right. Setting any of those to @var{0} disables
9367 Set planes to filter. Default is to filter all available planes.
9370 @subsection Examples
9374 Deblock using weak filter and block size of 4 pixels.
9376 deblock=filter=weak:block=4
9380 Deblock using strong filter, block size of 4 pixels and custom thresholds for
9381 deblocking more edges.
9383 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05
9387 Similar as above, but filter only first plane.
9389 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=1
9393 Similar as above, but filter only second and third plane.
9395 deblock=filter=strong:block=4:alpha=0.12:beta=0.07:gamma=0.06:delta=0.05:planes=6
9402 Drop duplicated frames at regular intervals.
9404 The filter accepts the following options:
9408 Set the number of frames from which one will be dropped. Setting this to
9409 @var{N} means one frame in every batch of @var{N} frames will be dropped.
9410 Default is @code{5}.
9413 Set the threshold for duplicate detection. If the difference metric for a frame
9414 is less than or equal to this value, then it is declared as duplicate. Default
9418 Set scene change threshold. Default is @code{15}.
9422 Set the size of the x and y-axis blocks used during metric calculations.
9423 Larger blocks give better noise suppression, but also give worse detection of
9424 small movements. Must be a power of two. Default is @code{32}.
9427 Mark main input as a pre-processed input and activate clean source input
9428 stream. This allows the input to be pre-processed with various filters to help
9429 the metrics calculation while keeping the frame selection lossless. When set to
9430 @code{1}, the first stream is for the pre-processed input, and the second
9431 stream is the clean source from where the kept frames are chosen. Default is
9435 Set whether or not chroma is considered in the metric calculations. Default is
9441 Apply 2D deconvolution of video stream in frequency domain using second stream
9444 The filter accepts the following options:
9448 Set which planes to process.
9451 Set which impulse video frames will be processed, can be @var{first}
9452 or @var{all}. Default is @var{all}.
9455 Set noise when doing divisions. Default is @var{0.0000001}. Useful when width
9456 and height are not same and not power of 2 or if stream prior to convolving
9460 The @code{deconvolve} filter also supports the @ref{framesync} options.
9464 Reduce cross-luminance (dot-crawl) and cross-color (rainbows) from video.
9466 It accepts the following options:
9470 Set mode of operation. Can be combination of @var{dotcrawl} for cross-luminance reduction and/or
9471 @var{rainbows} for cross-color reduction.
9474 Set spatial luma threshold. Lower values increases reduction of cross-luminance.
9477 Set tolerance for temporal luma. Higher values increases reduction of cross-luminance.
9480 Set tolerance for chroma temporal variation. Higher values increases reduction of cross-color.
9483 Set temporal chroma threshold. Lower values increases reduction of cross-color.
9488 Apply deflate effect to the video.
9490 This filter replaces the pixel by the local(3x3) average by taking into account
9491 only values lower than the pixel.
9493 It accepts the following options:
9500 Limit the maximum change for each plane, default is 65535.
9501 If 0, plane will remain unchanged.
9504 @subsection Commands
9506 This filter supports the all above options as @ref{commands}.
9510 Remove temporal frame luminance variations.
9512 It accepts the following options:
9516 Set moving-average filter size in frames. Default is 5. Allowed range is 2 - 129.
9519 Set averaging mode to smooth temporal luminance variations.
9521 Available values are:
9546 Do not actually modify frame. Useful when one only wants metadata.
9551 Remove judder produced by partially interlaced telecined content.
9553 Judder can be introduced, for instance, by @ref{pullup} filter. If the original
9554 source was partially telecined content then the output of @code{pullup,dejudder}
9555 will have a variable frame rate. May change the recorded frame rate of the
9556 container. Aside from that change, this filter will not affect constant frame
9559 The option available in this filter is:
9563 Specify the length of the window over which the judder repeats.
9565 Accepts any integer greater than 1. Useful values are:
9569 If the original was telecined from 24 to 30 fps (Film to NTSC).
9572 If the original was telecined from 25 to 30 fps (PAL to NTSC).
9575 If a mixture of the two.
9578 The default is @samp{4}.
9583 Suppress a TV station logo by a simple interpolation of the surrounding
9584 pixels. Just set a rectangle covering the logo and watch it disappear
9585 (and sometimes something even uglier appear - your mileage may vary).
9587 It accepts the following parameters:
9592 Specify the top left corner coordinates of the logo. They must be
9597 Specify the width and height of the logo to clear. They must be
9601 Specify the thickness of the fuzzy edge of the rectangle (added to
9602 @var{w} and @var{h}). The default value is 1. This option is
9603 deprecated, setting higher values should no longer be necessary and
9607 When set to 1, a green rectangle is drawn on the screen to simplify
9608 finding the right @var{x}, @var{y}, @var{w}, and @var{h} parameters.
9609 The default value is 0.
9611 The rectangle is drawn on the outermost pixels which will be (partly)
9612 replaced with interpolated values. The values of the next pixels
9613 immediately outside this rectangle in each direction will be used to
9614 compute the interpolated pixel values inside the rectangle.
9618 @subsection Examples
9622 Set a rectangle covering the area with top left corner coordinates 0,0
9623 and size 100x77, and a band of size 10:
9625 delogo=x=0:y=0:w=100:h=77:band=10
9633 Remove the rain in the input image/video by applying the derain methods based on
9634 convolutional neural networks. Supported models:
9638 Recurrent Squeeze-and-Excitation Context Aggregation Net (RESCAN).
9639 See @url{http://openaccess.thecvf.com/content_ECCV_2018/papers/Xia_Li_Recurrent_Squeeze-and-Excitation_Context_ECCV_2018_paper.pdf}.
9642 Training as well as model generation scripts are provided in
9643 the repository at @url{https://github.com/XueweiMeng/derain_filter.git}.
9645 Native model files (.model) can be generated from TensorFlow model
9646 files (.pb) by using tools/python/convert.py
9648 The filter accepts the following options:
9652 Specify which filter to use. This option accepts the following values:
9656 Derain filter. To conduct derain filter, you need to use a derain model.
9659 Dehaze filter. To conduct dehaze filter, you need to use a dehaze model.
9661 Default value is @samp{derain}.
9664 Specify which DNN backend to use for model loading and execution. This option accepts
9665 the following values:
9669 Native implementation of DNN loading and execution.
9672 TensorFlow backend. To enable this backend you
9673 need to install the TensorFlow for C library (see
9674 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9675 @code{--enable-libtensorflow}
9677 Default value is @samp{native}.
9680 Set path to model file specifying network architecture and its parameters.
9681 Note that different backends use different file formats. TensorFlow and native
9682 backend can load files for only its format.
9685 It can also be finished with @ref{dnn_processing} filter.
9689 Attempt to fix small changes in horizontal and/or vertical shift. This
9690 filter helps remove camera shake from hand-holding a camera, bumping a
9691 tripod, moving on a vehicle, etc.
9693 The filter accepts the following options:
9701 Specify a rectangular area where to limit the search for motion
9703 If desired the search for motion vectors can be limited to a
9704 rectangular area of the frame defined by its top left corner, width
9705 and height. These parameters have the same meaning as the drawbox
9706 filter which can be used to visualise the position of the bounding
9709 This is useful when simultaneous movement of subjects within the frame
9710 might be confused for camera motion by the motion vector search.
9712 If any or all of @var{x}, @var{y}, @var{w} and @var{h} are set to -1
9713 then the full frame is used. This allows later options to be set
9714 without specifying the bounding box for the motion vector search.
9716 Default - search the whole frame.
9720 Specify the maximum extent of movement in x and y directions in the
9721 range 0-64 pixels. Default 16.
9724 Specify how to generate pixels to fill blanks at the edge of the
9725 frame. Available values are:
9728 Fill zeroes at blank locations
9730 Original image at blank locations
9732 Extruded edge value at blank locations
9734 Mirrored edge at blank locations
9736 Default value is @samp{mirror}.
9739 Specify the blocksize to use for motion search. Range 4-128 pixels,
9743 Specify the contrast threshold for blocks. Only blocks with more than
9744 the specified contrast (difference between darkest and lightest
9745 pixels) will be considered. Range 1-255, default 125.
9748 Specify the search strategy. Available values are:
9751 Set exhaustive search
9753 Set less exhaustive search.
9755 Default value is @samp{exhaustive}.
9758 If set then a detailed log of the motion search is written to the
9765 Remove unwanted contamination of foreground colors, caused by reflected color of
9766 greenscreen or bluescreen.
9768 This filter accepts the following options:
9772 Set what type of despill to use.
9775 Set how spillmap will be generated.
9778 Set how much to get rid of still remaining spill.
9781 Controls amount of red in spill area.
9784 Controls amount of green in spill area.
9785 Should be -1 for greenscreen.
9788 Controls amount of blue in spill area.
9789 Should be -1 for bluescreen.
9792 Controls brightness of spill area, preserving colors.
9795 Modify alpha from generated spillmap.
9798 @subsection Commands
9800 This filter supports the all above options as @ref{commands}.
9804 Apply an exact inverse of the telecine operation. It requires a predefined
9805 pattern specified using the pattern option which must be the same as that passed
9806 to the telecine filter.
9808 This filter accepts the following options:
9817 The default value is @code{top}.
9821 A string of numbers representing the pulldown pattern you wish to apply.
9822 The default value is @code{23}.
9825 A number representing position of the first frame with respect to the telecine
9826 pattern. This is to be used if the stream is cut. The default value is @code{0}.
9831 Apply dilation effect to the video.
9833 This filter replaces the pixel by the local(3x3) maximum.
9835 It accepts the following options:
9842 Limit the maximum change for each plane, default is 65535.
9843 If 0, plane will remain unchanged.
9846 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
9849 Flags to local 3x3 coordinates maps like this:
9856 @subsection Commands
9858 This filter supports the all above options as @ref{commands}.
9862 Displace pixels as indicated by second and third input stream.
9864 It takes three input streams and outputs one stream, the first input is the
9865 source, and second and third input are displacement maps.
9867 The second input specifies how much to displace pixels along the
9868 x-axis, while the third input specifies how much to displace pixels
9870 If one of displacement map streams terminates, last frame from that
9871 displacement map will be used.
9873 Note that once generated, displacements maps can be reused over and over again.
9875 A description of the accepted options follows.
9879 Set displace behavior for pixels that are out of range.
9881 Available values are:
9884 Missing pixels are replaced by black pixels.
9887 Adjacent pixels will spread out to replace missing pixels.
9890 Out of range pixels are wrapped so they point to pixels of other side.
9893 Out of range pixels will be replaced with mirrored pixels.
9895 Default is @samp{smear}.
9899 @subsection Examples
9903 Add ripple effect to rgb input of video size hd720:
9905 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
9909 Add wave effect to rgb input of video size hd720:
9911 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
9915 @anchor{dnn_processing}
9916 @section dnn_processing
9918 Do image processing with deep neural networks. It works together with another filter
9919 which converts the pixel format of the Frame to what the dnn network requires.
9921 The filter accepts the following options:
9925 Specify which DNN backend to use for model loading and execution. This option accepts
9926 the following values:
9930 Native implementation of DNN loading and execution.
9933 TensorFlow backend. To enable this backend you
9934 need to install the TensorFlow for C library (see
9935 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
9936 @code{--enable-libtensorflow}
9939 OpenVINO backend. To enable this backend you
9940 need to build and install the OpenVINO for C library (see
9941 @url{https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md}) and configure FFmpeg with
9942 @code{--enable-libopenvino} (--extra-cflags=-I... --extra-ldflags=-L... might
9943 be needed if the header files and libraries are not installed into system path)
9947 Default value is @samp{native}.
9950 Set path to model file specifying network architecture and its parameters.
9951 Note that different backends use different file formats. TensorFlow, OpenVINO and native
9952 backend can load files for only its format.
9954 Native model file (.model) can be generated from TensorFlow model file (.pb) by using tools/python/convert.py
9957 Set the input name of the dnn network.
9960 Set the output name of the dnn network.
9963 use DNN async execution if set (default: set),
9964 roll back to sync execution if the backend does not support async.
9968 @subsection Examples
9972 Remove rain in rgb24 frame with can.pb (see @ref{derain} filter):
9974 ./ffmpeg -i rain.jpg -vf format=rgb24,dnn_processing=dnn_backend=tensorflow:model=can.pb:input=x:output=y derain.jpg
9978 Halve the pixel value of the frame with format gray32f:
9980 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
9984 Handle the Y channel with srcnn.pb (see @ref{sr} filter) for frame with yuv420p (planar YUV formats supported):
9986 ./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
9990 Handle the Y channel with espcn.pb (see @ref{sr} filter), which changes frame size, for format yuv420p (planar YUV formats supported):
9992 ./ffmpeg -i 480p.jpg -vf format=yuv420p,dnn_processing=dnn_backend=tensorflow:model=espcn.pb:input=x:output=y -y tmp.espcn.jpg
9999 Draw a colored box on the input image.
10001 It accepts the following parameters:
10006 The expressions which specify the top left corner coordinates of the box. It defaults to 0.
10010 The expressions which specify the width and height of the box; if 0 they are interpreted as
10011 the input width and height. It defaults to 0.
10014 Specify the color of the box to write. For the general syntax of this option,
10015 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10016 value @code{invert} is used, the box edge color is the same as the
10017 video with inverted luma.
10020 The expression which sets the thickness of the box edge.
10021 A value of @code{fill} will create a filled box. Default value is @code{3}.
10023 See below for the list of accepted constants.
10026 Applicable if the input has alpha. With value @code{1}, the pixels of the painted box
10027 will overwrite the video's color and alpha pixels.
10028 Default is @code{0}, which composites the box onto the input, leaving the video's alpha intact.
10031 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10032 following constants:
10036 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10040 horizontal and vertical chroma subsample values. For example for the
10041 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10045 The input width and height.
10048 The input sample aspect ratio.
10052 The x and y offset coordinates where the box is drawn.
10056 The width and height of the drawn box.
10059 The thickness of the drawn box.
10061 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10062 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10066 @subsection Examples
10070 Draw a black box around the edge of the input image:
10076 Draw a box with color red and an opacity of 50%:
10078 drawbox=10:20:200:60:red@@0.5
10081 The previous example can be specified as:
10083 drawbox=x=10:y=20:w=200:h=60:color=red@@0.5
10087 Fill the box with pink color:
10089 drawbox=x=10:y=10:w=100:h=100:color=pink@@0.5:t=fill
10093 Draw a 2-pixel red 2.40:1 mask:
10095 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
10099 @subsection Commands
10100 This filter supports same commands as options.
10101 The command accepts the same syntax of the corresponding option.
10103 If the specified expression is not valid, it is kept at its current
10108 Draw a graph using input video metadata.
10110 It accepts the following parameters:
10114 Set 1st frame metadata key from which metadata values will be used to draw a graph.
10117 Set 1st foreground color expression.
10120 Set 2nd frame metadata key from which metadata values will be used to draw a graph.
10123 Set 2nd foreground color expression.
10126 Set 3rd frame metadata key from which metadata values will be used to draw a graph.
10129 Set 3rd foreground color expression.
10132 Set 4th frame metadata key from which metadata values will be used to draw a graph.
10135 Set 4th foreground color expression.
10138 Set minimal value of metadata value.
10141 Set maximal value of metadata value.
10144 Set graph background color. Default is white.
10149 Available values for mode is:
10156 Default is @code{line}.
10161 Available values for slide is:
10164 Draw new frame when right border is reached.
10167 Replace old columns with new ones.
10170 Scroll from right to left.
10173 Scroll from left to right.
10176 Draw single picture.
10179 Default is @code{frame}.
10182 Set size of graph video. For the syntax of this option, check the
10183 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
10184 The default value is @code{900x256}.
10187 Set the output frame rate. Default value is @code{25}.
10189 The foreground color expressions can use the following variables:
10192 Minimal value of metadata value.
10195 Maximal value of metadata value.
10198 Current metadata key value.
10201 The color is defined as 0xAABBGGRR.
10204 Example using metadata from @ref{signalstats} filter:
10206 signalstats,drawgraph=lavfi.signalstats.YAVG:min=0:max=255
10209 Example using metadata from @ref{ebur128} filter:
10211 ebur128=metadata=1,adrawgraph=lavfi.r128.M:min=-120:max=5
10216 Draw a grid on the input image.
10218 It accepts the following parameters:
10223 The expressions which specify the coordinates of some point of grid intersection (meant to configure offset). Both default to 0.
10227 The expressions which specify the width and height of the grid cell, if 0 they are interpreted as the
10228 input width and height, respectively, minus @code{thickness}, so image gets
10229 framed. Default to 0.
10232 Specify the color of the grid. For the general syntax of this option,
10233 check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}. If the special
10234 value @code{invert} is used, the grid color is the same as the
10235 video with inverted luma.
10238 The expression which sets the thickness of the grid line. Default value is @code{1}.
10240 See below for the list of accepted constants.
10243 Applicable if the input has alpha. With @code{1} the pixels of the painted grid
10244 will overwrite the video's color and alpha pixels.
10245 Default is @code{0}, which composites the grid onto the input, leaving the video's alpha intact.
10248 The parameters for @var{x}, @var{y}, @var{w} and @var{h} and @var{t} are expressions containing the
10249 following constants:
10253 The input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}.
10257 horizontal and vertical chroma subsample values. For example for the
10258 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10262 The input grid cell width and height.
10265 The input sample aspect ratio.
10269 The x and y coordinates of some point of grid intersection (meant to configure offset).
10273 The width and height of the drawn cell.
10276 The thickness of the drawn cell.
10278 These constants allow the @var{x}, @var{y}, @var{w}, @var{h} and @var{t} expressions to refer to
10279 each other, so you may for example specify @code{y=x/dar} or @code{h=w/dar}.
10283 @subsection Examples
10287 Draw a grid with cell 100x100 pixels, thickness 2 pixels, with color red and an opacity of 50%:
10289 drawgrid=width=100:height=100:thickness=2:color=red@@0.5
10293 Draw a white 3x3 grid with an opacity of 50%:
10295 drawgrid=w=iw/3:h=ih/3:t=2:c=white@@0.5
10299 @subsection Commands
10300 This filter supports same commands as options.
10301 The command accepts the same syntax of the corresponding option.
10303 If the specified expression is not valid, it is kept at its current
10309 Draw a text string or text from a specified file on top of a video, using the
10310 libfreetype library.
10312 To enable compilation of this filter, you need to configure FFmpeg with
10313 @code{--enable-libfreetype}.
10314 To enable default font fallback and the @var{font} option you need to
10315 configure FFmpeg with @code{--enable-libfontconfig}.
10316 To enable the @var{text_shaping} option, you need to configure FFmpeg with
10317 @code{--enable-libfribidi}.
10321 It accepts the following parameters:
10326 Used to draw a box around text using the background color.
10327 The value must be either 1 (enable) or 0 (disable).
10328 The default value of @var{box} is 0.
10331 Set the width of the border to be drawn around the box using @var{boxcolor}.
10332 The default value of @var{boxborderw} is 0.
10335 The color to be used for drawing box around text. For the syntax of this
10336 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10338 The default value of @var{boxcolor} is "white".
10341 Set the line spacing in pixels of the border to be drawn around the box using @var{box}.
10342 The default value of @var{line_spacing} is 0.
10345 Set the width of the border to be drawn around the text using @var{bordercolor}.
10346 The default value of @var{borderw} is 0.
10349 Set the color to be used for drawing border around text. For the syntax of this
10350 option, check the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10352 The default value of @var{bordercolor} is "black".
10355 Select how the @var{text} is expanded. Can be either @code{none},
10356 @code{strftime} (deprecated) or
10357 @code{normal} (default). See the @ref{drawtext_expansion, Text expansion} section
10361 Set a start time for the count. Value is in microseconds. Only applied
10362 in the deprecated strftime expansion mode. To emulate in normal expansion
10363 mode use the @code{pts} function, supplying the start time (in seconds)
10364 as the second argument.
10367 If true, check and fix text coords to avoid clipping.
10370 The color to be used for drawing fonts. For the syntax of this option, check
10371 the @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
10373 The default value of @var{fontcolor} is "black".
10375 @item fontcolor_expr
10376 String which is expanded the same way as @var{text} to obtain dynamic
10377 @var{fontcolor} value. By default this option has empty value and is not
10378 processed. When this option is set, it overrides @var{fontcolor} option.
10381 The font family to be used for drawing text. By default Sans.
10384 The font file to be used for drawing text. The path must be included.
10385 This parameter is mandatory if the fontconfig support is disabled.
10388 Draw the text applying alpha blending. The value can
10389 be a number between 0.0 and 1.0.
10390 The expression accepts the same variables @var{x, y} as well.
10391 The default value is 1.
10392 Please see @var{fontcolor_expr}.
10395 The font size to be used for drawing text.
10396 The default value of @var{fontsize} is 16.
10399 If set to 1, attempt to shape the text (for example, reverse the order of
10400 right-to-left text and join Arabic characters) before drawing it.
10401 Otherwise, just draw the text exactly as given.
10402 By default 1 (if supported).
10404 @item ft_load_flags
10405 The flags to be used for loading the fonts.
10407 The flags map the corresponding flags supported by libfreetype, and are
10408 a combination of the following values:
10415 @item vertical_layout
10416 @item force_autohint
10419 @item ignore_global_advance_width
10421 @item ignore_transform
10423 @item linear_design
10427 Default value is "default".
10429 For more information consult the documentation for the FT_LOAD_*
10433 The color to be used for drawing a shadow behind the drawn text. For the
10434 syntax of this option, check the @ref{color syntax,,"Color" section in the
10435 ffmpeg-utils manual,ffmpeg-utils}.
10437 The default value of @var{shadowcolor} is "black".
10441 The x and y offsets for the text shadow position with respect to the
10442 position of the text. They can be either positive or negative
10443 values. The default value for both is "0".
10446 The starting frame number for the n/frame_num variable. The default value
10450 The size in number of spaces to use for rendering the tab.
10451 Default value is 4.
10454 Set the initial timecode representation in "hh:mm:ss[:;.]ff"
10455 format. It can be used with or without text parameter. @var{timecode_rate}
10456 option must be specified.
10458 @item timecode_rate, rate, r
10459 Set the timecode frame rate (timecode only). Value will be rounded to nearest
10460 integer. Minimum value is "1".
10461 Drop-frame timecode is supported for frame rates 30 & 60.
10464 If set to 1, the output of the timecode option will wrap around at 24 hours.
10465 Default is 0 (disabled).
10468 The text string to be drawn. The text must be a sequence of UTF-8
10469 encoded characters.
10470 This parameter is mandatory if no file is specified with the parameter
10474 A text file containing text to be drawn. The text must be a sequence
10475 of UTF-8 encoded characters.
10477 This parameter is mandatory if no text string is specified with the
10478 parameter @var{text}.
10480 If both @var{text} and @var{textfile} are specified, an error is thrown.
10483 If set to 1, the @var{textfile} will be reloaded before each frame.
10484 Be sure to update it atomically, or it may be read partially, or even fail.
10488 The expressions which specify the offsets where text will be drawn
10489 within the video frame. They are relative to the top/left border of the
10492 The default value of @var{x} and @var{y} is "0".
10494 See below for the list of accepted constants and functions.
10497 The parameters for @var{x} and @var{y} are expressions containing the
10498 following constants and functions:
10502 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
10506 horizontal and vertical chroma subsample values. For example for the
10507 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
10510 the height of each text line
10518 @item max_glyph_a, ascent
10519 the maximum distance from the baseline to the highest/upper grid
10520 coordinate used to place a glyph outline point, for all the rendered
10522 It is a positive value, due to the grid's orientation with the Y axis
10525 @item max_glyph_d, descent
10526 the maximum distance from the baseline to the lowest grid coordinate
10527 used to place a glyph outline point, for all the rendered glyphs.
10528 This is a negative value, due to the grid's orientation, with the Y axis
10532 maximum glyph height, that is the maximum height for all the glyphs
10533 contained in the rendered text, it is equivalent to @var{ascent} -
10537 maximum glyph width, that is the maximum width for all the glyphs
10538 contained in the rendered text
10541 the number of input frame, starting from 0
10543 @item rand(min, max)
10544 return a random number included between @var{min} and @var{max}
10547 The input sample aspect ratio.
10550 timestamp expressed in seconds, NAN if the input timestamp is unknown
10553 the height of the rendered text
10556 the width of the rendered text
10560 the x and y offset coordinates where the text is drawn.
10562 These parameters allow the @var{x} and @var{y} expressions to refer
10563 to each other, so you can for example specify @code{y=x/dar}.
10566 A one character description of the current frame's picture type.
10569 The current packet's position in the input file or stream
10570 (in bytes, from the start of the input). A value of -1 indicates
10571 this info is not available.
10574 The current packet's duration, in seconds.
10577 The current packet's size (in bytes).
10580 @anchor{drawtext_expansion}
10581 @subsection Text expansion
10583 If @option{expansion} is set to @code{strftime},
10584 the filter recognizes strftime() sequences in the provided text and
10585 expands them accordingly. Check the documentation of strftime(). This
10586 feature is deprecated.
10588 If @option{expansion} is set to @code{none}, the text is printed verbatim.
10590 If @option{expansion} is set to @code{normal} (which is the default),
10591 the following expansion mechanism is used.
10593 The backslash character @samp{\}, followed by any character, always expands to
10594 the second character.
10596 Sequences of the form @code{%@{...@}} are expanded. The text between the
10597 braces is a function name, possibly followed by arguments separated by ':'.
10598 If the arguments contain special characters or delimiters (':' or '@}'),
10599 they should be escaped.
10601 Note that they probably must also be escaped as the value for the
10602 @option{text} option in the filter argument string and as the filter
10603 argument in the filtergraph description, and possibly also for the shell,
10604 that makes up to four levels of escaping; using a text file avoids these
10607 The following functions are available:
10612 The expression evaluation result.
10614 It must take one argument specifying the expression to be evaluated,
10615 which accepts the same constants and functions as the @var{x} and
10616 @var{y} values. Note that not all constants should be used, for
10617 example the text size is not known when evaluating the expression, so
10618 the constants @var{text_w} and @var{text_h} will have an undefined
10621 @item expr_int_format, eif
10622 Evaluate the expression's value and output as formatted integer.
10624 The first argument is the expression to be evaluated, just as for the @var{expr} function.
10625 The second argument specifies the output format. Allowed values are @samp{x},
10626 @samp{X}, @samp{d} and @samp{u}. They are treated exactly as in the
10627 @code{printf} function.
10628 The third parameter is optional and sets the number of positions taken by the output.
10629 It can be used to add padding with zeros from the left.
10632 The time at which the filter is running, expressed in UTC.
10633 It can accept an argument: a strftime() format string.
10636 The time at which the filter is running, expressed in the local time zone.
10637 It can accept an argument: a strftime() format string.
10640 Frame metadata. Takes one or two arguments.
10642 The first argument is mandatory and specifies the metadata key.
10644 The second argument is optional and specifies a default value, used when the
10645 metadata key is not found or empty.
10647 Available metadata can be identified by inspecting entries
10648 starting with TAG included within each frame section
10649 printed by running @code{ffprobe -show_frames}.
10651 String metadata generated in filters leading to
10652 the drawtext filter are also available.
10655 The frame number, starting from 0.
10658 A one character description of the current picture type.
10661 The timestamp of the current frame.
10662 It can take up to three arguments.
10664 The first argument is the format of the timestamp; it defaults to @code{flt}
10665 for seconds as a decimal number with microsecond accuracy; @code{hms} stands
10666 for a formatted @var{[-]HH:MM:SS.mmm} timestamp with millisecond accuracy.
10667 @code{gmtime} stands for the timestamp of the frame formatted as UTC time;
10668 @code{localtime} stands for the timestamp of the frame formatted as
10669 local time zone time.
10671 The second argument is an offset added to the timestamp.
10673 If the format is set to @code{hms}, a third argument @code{24HH} may be
10674 supplied to present the hour part of the formatted timestamp in 24h format
10677 If the format is set to @code{localtime} or @code{gmtime},
10678 a third argument may be supplied: a strftime() format string.
10679 By default, @var{YYYY-MM-DD HH:MM:SS} format will be used.
10682 @subsection Commands
10684 This filter supports altering parameters via commands:
10687 Alter existing filter parameters.
10689 Syntax for the argument is the same as for filter invocation, e.g.
10692 fontsize=56:fontcolor=green:text='Hello World'
10695 Full filter invocation with sendcmd would look like this:
10698 sendcmd=c='56.0 drawtext reinit fontsize=56\:fontcolor=green\:text=Hello\\ World'
10702 If the entire argument can't be parsed or applied as valid values then the filter will
10703 continue with its existing parameters.
10705 @subsection Examples
10709 Draw "Test Text" with font FreeSerif, using the default values for the
10710 optional parameters.
10713 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text'"
10717 Draw 'Test Text' with font FreeSerif of size 24 at position x=100
10718 and y=50 (counting from the top-left corner of the screen), text is
10719 yellow with a red box around it. Both the text and the box have an
10723 drawtext="fontfile=/usr/share/fonts/truetype/freefont/FreeSerif.ttf: text='Test Text':\
10724 x=100: y=50: fontsize=24: fontcolor=yellow@@0.2: box=1: boxcolor=red@@0.2"
10727 Note that the double quotes are not necessary if spaces are not used
10728 within the parameter list.
10731 Show the text at the center of the video frame:
10733 drawtext="fontsize=30:fontfile=FreeSerif.ttf:text='hello world':x=(w-text_w)/2:y=(h-text_h)/2"
10737 Show the text at a random position, switching to a new position every 30 seconds:
10739 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)"
10743 Show a text line sliding from right to left in the last row of the video
10744 frame. The file @file{LONG_LINE} is assumed to contain a single line
10747 drawtext="fontsize=15:fontfile=FreeSerif.ttf:text=LONG_LINE:y=h-line_h:x=-50*t"
10751 Show the content of file @file{CREDITS} off the bottom of the frame and scroll up.
10753 drawtext="fontsize=20:fontfile=FreeSerif.ttf:textfile=CREDITS:y=h-20*t"
10757 Draw a single green letter "g", at the center of the input video.
10758 The glyph baseline is placed at half screen height.
10760 drawtext="fontsize=60:fontfile=FreeSerif.ttf:fontcolor=green:text=g:x=(w-max_glyph_w)/2:y=h/2-ascent"
10764 Show text for 1 second every 3 seconds:
10766 drawtext="fontfile=FreeSerif.ttf:fontcolor=white:x=100:y=x/dar:enable=lt(mod(t\,3)\,1):text='blink'"
10770 Use fontconfig to set the font. Note that the colons need to be escaped.
10772 drawtext='fontfile=Linux Libertine O-40\:style=Semibold:text=FFmpeg'
10776 Draw "Test Text" with font size dependent on height of the video.
10778 drawtext="text='Test Text': fontsize=h/30: x=(w-text_w)/2: y=(h-text_h*2)"
10782 Print the date of a real-time encoding (see strftime(3)):
10784 drawtext='fontfile=FreeSans.ttf:text=%@{localtime\:%a %b %d %Y@}'
10788 Show text fading in and out (appearing/disappearing):
10791 DS=1.0 # display start
10792 DE=10.0 # display end
10793 FID=1.5 # fade in duration
10794 FOD=5 # fade out duration
10795 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 @}"
10799 Horizontally align multiple separate texts. Note that @option{max_glyph_a}
10800 and the @option{fontsize} value are included in the @option{y} offset.
10802 drawtext=fontfile=FreeSans.ttf:text=DOG:fontsize=24:x=10:y=20+24-max_glyph_a,
10803 drawtext=fontfile=FreeSans.ttf:text=cow:fontsize=24:x=80:y=20+24-max_glyph_a
10807 Plot special @var{lavf.image2dec.source_basename} metadata onto each frame if
10808 such metadata exists. Otherwise, plot the string "NA". Note that image2 demuxer
10809 must have option @option{-export_path_metadata 1} for the special metadata fields
10810 to be available for filters.
10812 drawtext="fontsize=20:fontcolor=white:fontfile=FreeSans.ttf:text='%@{metadata\:lavf.image2dec.source_basename\:NA@}':x=10:y=10"
10817 For more information about libfreetype, check:
10818 @url{http://www.freetype.org/}.
10820 For more information about fontconfig, check:
10821 @url{http://freedesktop.org/software/fontconfig/fontconfig-user.html}.
10823 For more information about libfribidi, check:
10824 @url{http://fribidi.org/}.
10826 @section edgedetect
10828 Detect and draw edges. The filter uses the Canny Edge Detection algorithm.
10830 The filter accepts the following options:
10835 Set low and high threshold values used by the Canny thresholding
10838 The high threshold selects the "strong" edge pixels, which are then
10839 connected through 8-connectivity with the "weak" edge pixels selected
10840 by the low threshold.
10842 @var{low} and @var{high} threshold values must be chosen in the range
10843 [0,1], and @var{low} should be lesser or equal to @var{high}.
10845 Default value for @var{low} is @code{20/255}, and default value for @var{high}
10849 Define the drawing mode.
10853 Draw white/gray wires on black background.
10856 Mix the colors to create a paint/cartoon effect.
10859 Apply Canny edge detector on all selected planes.
10861 Default value is @var{wires}.
10864 Select planes for filtering. By default all available planes are filtered.
10867 @subsection Examples
10871 Standard edge detection with custom values for the hysteresis thresholding:
10873 edgedetect=low=0.1:high=0.4
10877 Painting effect without thresholding:
10879 edgedetect=mode=colormix:high=0
10885 Apply a posterize effect using the ELBG (Enhanced LBG) algorithm.
10887 For each input image, the filter will compute the optimal mapping from
10888 the input to the output given the codebook length, that is the number
10889 of distinct output colors.
10891 This filter accepts the following options.
10894 @item codebook_length, l
10895 Set codebook length. The value must be a positive integer, and
10896 represents the number of distinct output colors. Default value is 256.
10899 Set the maximum number of iterations to apply for computing the optimal
10900 mapping. The higher the value the better the result and the higher the
10901 computation time. Default value is 1.
10904 Set a random seed, must be an integer included between 0 and
10905 UINT32_MAX. If not specified, or if explicitly set to -1, the filter
10906 will try to use a good random seed on a best effort basis.
10909 Set pal8 output pixel format. This option does not work with codebook
10910 length greater than 256.
10915 Measure graylevel entropy in histogram of color channels of video frames.
10917 It accepts the following parameters:
10921 Can be either @var{normal} or @var{diff}. Default is @var{normal}.
10923 @var{diff} mode measures entropy of histogram delta values, absolute differences
10924 between neighbour histogram values.
10928 Set brightness, contrast, saturation and approximate gamma adjustment.
10930 The filter accepts the following options:
10934 Set the contrast expression. The value must be a float value in range
10935 @code{-1000.0} to @code{1000.0}. The default value is "1".
10938 Set the brightness expression. The value must be a float value in
10939 range @code{-1.0} to @code{1.0}. The default value is "0".
10942 Set the saturation expression. The value must be a float in
10943 range @code{0.0} to @code{3.0}. The default value is "1".
10946 Set the gamma expression. The value must be a float in range
10947 @code{0.1} to @code{10.0}. The default value is "1".
10950 Set the gamma expression for red. The value must be a float in
10951 range @code{0.1} to @code{10.0}. The default value is "1".
10954 Set the gamma expression for green. The value must be a float in range
10955 @code{0.1} to @code{10.0}. The default value is "1".
10958 Set the gamma expression for blue. The value must be a float in range
10959 @code{0.1} to @code{10.0}. The default value is "1".
10962 Set the gamma weight expression. It can be used to reduce the effect
10963 of a high gamma value on bright image areas, e.g. keep them from
10964 getting overamplified and just plain white. The value must be a float
10965 in range @code{0.0} to @code{1.0}. A value of @code{0.0} turns the
10966 gamma correction all the way down while @code{1.0} leaves it at its
10967 full strength. Default is "1".
10970 Set when the expressions for brightness, contrast, saturation and
10971 gamma expressions are evaluated.
10973 It accepts the following values:
10976 only evaluate expressions once during the filter initialization or
10977 when a command is processed
10980 evaluate expressions for each incoming frame
10983 Default value is @samp{init}.
10986 The expressions accept the following parameters:
10989 frame count of the input frame starting from 0
10992 byte position of the corresponding packet in the input file, NAN if
10996 frame rate of the input video, NAN if the input frame rate is unknown
10999 timestamp expressed in seconds, NAN if the input timestamp is unknown
11002 @subsection Commands
11003 The filter supports the following commands:
11007 Set the contrast expression.
11010 Set the brightness expression.
11013 Set the saturation expression.
11016 Set the gamma expression.
11019 Set the gamma_r expression.
11022 Set gamma_g expression.
11025 Set gamma_b expression.
11028 Set gamma_weight expression.
11030 The command accepts the same syntax of the corresponding option.
11032 If the specified expression is not valid, it is kept at its current
11039 Apply erosion effect to the video.
11041 This filter replaces the pixel by the local(3x3) minimum.
11043 It accepts the following options:
11050 Limit the maximum change for each plane, default is 65535.
11051 If 0, plane will remain unchanged.
11054 Flag which specifies the pixel to refer to. Default is 255 i.e. all eight
11057 Flags to local 3x3 coordinates maps like this:
11064 @subsection Commands
11066 This filter supports the all above options as @ref{commands}.
11070 Deinterlace the input video ("estdif" stands for "Edge Slope
11071 Tracing Deinterlacing Filter").
11073 Spatial only filter that uses edge slope tracing algorithm
11074 to interpolate missing lines.
11075 It accepts the following parameters:
11079 The interlacing mode to adopt. It accepts one of the following values:
11083 Output one frame for each frame.
11085 Output one frame for each field.
11088 The default value is @code{field}.
11091 The picture field parity assumed for the input interlaced video. It accepts one
11092 of the following values:
11096 Assume the top field is first.
11098 Assume the bottom field is first.
11100 Enable automatic detection of field parity.
11103 The default value is @code{auto}.
11104 If the interlacing is unknown or the decoder does not export this information,
11105 top field first will be assumed.
11108 Specify which frames to deinterlace. Accepts one of the following
11113 Deinterlace all frames.
11115 Only deinterlace frames marked as interlaced.
11118 The default value is @code{all}.
11121 Specify the search radius for edge slope tracing. Default value is 1.
11122 Allowed range is from 1 to 15.
11125 Specify the search radius for best edge matching. Default value is 2.
11126 Allowed range is from 0 to 15.
11129 @subsection Commands
11130 This filter supports same @ref{commands} as options.
11132 @section extractplanes
11134 Extract color channel components from input video stream into
11135 separate grayscale video streams.
11137 The filter accepts the following option:
11141 Set plane(s) to extract.
11143 Available values for planes are:
11154 Choosing planes not available in the input will result in an error.
11155 That means you cannot select @code{r}, @code{g}, @code{b} planes
11156 with @code{y}, @code{u}, @code{v} planes at same time.
11159 @subsection Examples
11163 Extract luma, u and v color channel component from input video frame
11164 into 3 grayscale outputs:
11166 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
11172 Apply a fade-in/out effect to the input video.
11174 It accepts the following parameters:
11178 The effect type can be either "in" for a fade-in, or "out" for a fade-out
11180 Default is @code{in}.
11182 @item start_frame, s
11183 Specify the number of the frame to start applying the fade
11184 effect at. Default is 0.
11187 The number of frames that the fade effect lasts. At the end of the
11188 fade-in effect, the output video will have the same intensity as the input video.
11189 At the end of the fade-out transition, the output video will be filled with the
11190 selected @option{color}.
11194 If set to 1, fade only alpha channel, if one exists on the input.
11195 Default value is 0.
11197 @item start_time, st
11198 Specify the timestamp (in seconds) of the frame to start to apply the fade
11199 effect. If both start_frame and start_time are specified, the fade will start at
11200 whichever comes last. Default is 0.
11203 The number of seconds for which the fade effect has to last. At the end of the
11204 fade-in effect the output video will have the same intensity as the input video,
11205 at the end of the fade-out transition the output video will be filled with the
11206 selected @option{color}.
11207 If both duration and nb_frames are specified, duration is used. Default is 0
11208 (nb_frames is used by default).
11211 Specify the color of the fade. Default is "black".
11214 @subsection Examples
11218 Fade in the first 30 frames of video:
11223 The command above is equivalent to:
11229 Fade out the last 45 frames of a 200-frame video:
11232 fade=type=out:start_frame=155:nb_frames=45
11236 Fade in the first 25 frames and fade out the last 25 frames of a 1000-frame video:
11238 fade=in:0:25, fade=out:975:25
11242 Make the first 5 frames yellow, then fade in from frame 5-24:
11244 fade=in:5:20:color=yellow
11248 Fade in alpha over first 25 frames of video:
11250 fade=in:0:25:alpha=1
11254 Make the first 5.5 seconds black, then fade in for 0.5 seconds:
11256 fade=t=in:st=5.5:d=0.5
11262 Denoise frames using 3D FFT (frequency domain filtering).
11264 The filter accepts the following options:
11268 Set the noise sigma constant. This sets denoising strength.
11269 Default value is 1. Allowed range is from 0 to 30.
11270 Using very high sigma with low overlap may give blocking artifacts.
11273 Set amount of denoising. By default all detected noise is reduced.
11274 Default value is 1. Allowed range is from 0 to 1.
11277 Set size of block, Default is 4, can be 3, 4, 5 or 6.
11278 Actual size of block in pixels is 2 to power of @var{block}, so by default
11279 block size in pixels is 2^4 which is 16.
11282 Set block overlap. Default is 0.5. Allowed range is from 0.2 to 0.8.
11285 Set number of previous frames to use for denoising. By default is set to 0.
11288 Set number of next frames to to use for denoising. By default is set to 0.
11291 Set planes which will be filtered, by default are all available filtered
11296 Apply arbitrary expressions to samples in frequency domain
11300 Adjust the dc value (gain) of the luma plane of the image. The filter
11301 accepts an integer value in range @code{0} to @code{1000}. The default
11302 value is set to @code{0}.
11305 Adjust the dc value (gain) of the 1st chroma plane of the image. The
11306 filter accepts an integer value in range @code{0} to @code{1000}. The
11307 default value is set to @code{0}.
11310 Adjust the dc value (gain) of the 2nd chroma plane of the image. The
11311 filter accepts an integer value in range @code{0} to @code{1000}. The
11312 default value is set to @code{0}.
11315 Set the frequency domain weight expression for the luma plane.
11318 Set the frequency domain weight expression for the 1st chroma plane.
11321 Set the frequency domain weight expression for the 2nd chroma plane.
11324 Set when the expressions are evaluated.
11326 It accepts the following values:
11329 Only evaluate expressions once during the filter initialization.
11332 Evaluate expressions for each incoming frame.
11335 Default value is @samp{init}.
11337 The filter accepts the following variables:
11340 The coordinates of the current sample.
11344 The width and height of the image.
11347 The number of input frame, starting from 0.
11350 @subsection Examples
11356 fftfilt=dc_Y=128:weight_Y='squish(1-(Y+X)/100)'
11362 fftfilt=dc_Y=0:weight_Y='squish((Y+X)/100-1)'
11368 fftfilt=dc_Y=0:weight_Y='1+squish(1-(Y+X)/100)'
11374 fftfilt=dc_Y=0:weight_Y='exp(-4 * ((Y+X)/(W+H)))'
11381 Extract a single field from an interlaced image using stride
11382 arithmetic to avoid wasting CPU time. The output frames are marked as
11385 The filter accepts the following options:
11389 Specify whether to extract the top (if the value is @code{0} or
11390 @code{top}) or the bottom field (if the value is @code{1} or
11396 Create new frames by copying the top and bottom fields from surrounding frames
11397 supplied as numbers by the hint file.
11401 Set file containing hints: absolute/relative frame numbers.
11403 There must be one line for each frame in a clip. Each line must contain two
11404 numbers separated by the comma, optionally followed by @code{-} or @code{+}.
11405 Numbers supplied on each line of file can not be out of [N-1,N+1] where N
11406 is current frame number for @code{absolute} mode or out of [-1, 1] range
11407 for @code{relative} mode. First number tells from which frame to pick up top
11408 field and second number tells from which frame to pick up bottom field.
11410 If optionally followed by @code{+} output frame will be marked as interlaced,
11411 else if followed by @code{-} output frame will be marked as progressive, else
11412 it will be marked same as input frame.
11413 If optionally followed by @code{t} output frame will use only top field, or in
11414 case of @code{b} it will use only bottom field.
11415 If line starts with @code{#} or @code{;} that line is skipped.
11418 Can be item @code{absolute} or @code{relative}. Default is @code{absolute}.
11421 Example of first several lines of @code{hint} file for @code{relative} mode:
11423 0,0 - # first frame
11424 1,0 - # second frame, use third's frame top field and second's frame bottom field
11425 1,0 - # third frame, use fourth's frame top field and third's frame bottom field
11440 @section fieldmatch
11442 Field matching filter for inverse telecine. It is meant to reconstruct the
11443 progressive frames from a telecined stream. The filter does not drop duplicated
11444 frames, so to achieve a complete inverse telecine @code{fieldmatch} needs to be
11445 followed by a decimation filter such as @ref{decimate} in the filtergraph.
11447 The separation of the field matching and the decimation is notably motivated by
11448 the possibility of inserting a de-interlacing filter fallback between the two.
11449 If the source has mixed telecined and real interlaced content,
11450 @code{fieldmatch} will not be able to match fields for the interlaced parts.
11451 But these remaining combed frames will be marked as interlaced, and thus can be
11452 de-interlaced by a later filter such as @ref{yadif} before decimation.
11454 In addition to the various configuration options, @code{fieldmatch} can take an
11455 optional second stream, activated through the @option{ppsrc} option. If
11456 enabled, the frames reconstruction will be based on the fields and frames from
11457 this second stream. This allows the first input to be pre-processed in order to
11458 help the various algorithms of the filter, while keeping the output lossless
11459 (assuming the fields are matched properly). Typically, a field-aware denoiser,
11460 or brightness/contrast adjustments can help.
11462 Note that this filter uses the same algorithms as TIVTC/TFM (AviSynth project)
11463 and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
11464 which @code{fieldmatch} is based on. While the semantic and usage are very
11465 close, some behaviour and options names can differ.
11467 The @ref{decimate} filter currently only works for constant frame rate input.
11468 If your input has mixed telecined (30fps) and progressive content with a lower
11469 framerate like 24fps use the following filterchain to produce the necessary cfr
11470 stream: @code{dejudder,fps=30000/1001,fieldmatch,decimate}.
11472 The filter accepts the following options:
11476 Specify the assumed field order of the input stream. Available values are:
11480 Auto detect parity (use FFmpeg's internal parity value).
11482 Assume bottom field first.
11484 Assume top field first.
11487 Note that it is sometimes recommended not to trust the parity announced by the
11490 Default value is @var{auto}.
11493 Set the matching mode or strategy to use. @option{pc} mode is the safest in the
11494 sense that it won't risk creating jerkiness due to duplicate frames when
11495 possible, but if there are bad edits or blended fields it will end up
11496 outputting combed frames when a good match might actually exist. On the other
11497 hand, @option{pcn_ub} mode is the most risky in terms of creating jerkiness,
11498 but will almost always find a good frame if there is one. The other values are
11499 all somewhere in between @option{pc} and @option{pcn_ub} in terms of risking
11500 jerkiness and creating duplicate frames versus finding good matches in sections
11501 with bad edits, orphaned fields, blended fields, etc.
11503 More details about p/c/n/u/b are available in @ref{p/c/n/u/b meaning} section.
11505 Available values are:
11509 2-way matching (p/c)
11511 2-way matching, and trying 3rd match if still combed (p/c + n)
11513 2-way matching, and trying 3rd match (same order) if still combed (p/c + u)
11515 2-way matching, trying 3rd match if still combed, and trying 4th/5th matches if
11516 still combed (p/c + n + u/b)
11518 3-way matching (p/c/n)
11520 3-way matching, and trying 4th/5th matches if all 3 of the original matches are
11521 detected as combed (p/c/n + u/b)
11524 The parenthesis at the end indicate the matches that would be used for that
11525 mode assuming @option{order}=@var{tff} (and @option{field} on @var{auto} or
11528 In terms of speed @option{pc} mode is by far the fastest and @option{pcn_ub} is
11531 Default value is @var{pc_n}.
11534 Mark the main input stream as a pre-processed input, and enable the secondary
11535 input stream as the clean source to pick the fields from. See the filter
11536 introduction for more details. It is similar to the @option{clip2} feature from
11539 Default value is @code{0} (disabled).
11542 Set the field to match from. It is recommended to set this to the same value as
11543 @option{order} unless you experience matching failures with that setting. In
11544 certain circumstances changing the field that is used to match from can have a
11545 large impact on matching performance. Available values are:
11549 Automatic (same value as @option{order}).
11551 Match from the bottom field.
11553 Match from the top field.
11556 Default value is @var{auto}.
11559 Set whether or not chroma is included during the match comparisons. In most
11560 cases it is recommended to leave this enabled. You should set this to @code{0}
11561 only if your clip has bad chroma problems such as heavy rainbowing or other
11562 artifacts. Setting this to @code{0} could also be used to speed things up at
11563 the cost of some accuracy.
11565 Default value is @code{1}.
11569 These define an exclusion band which excludes the lines between @option{y0} and
11570 @option{y1} from being included in the field matching decision. An exclusion
11571 band can be used to ignore subtitles, a logo, or other things that may
11572 interfere with the matching. @option{y0} sets the starting scan line and
11573 @option{y1} sets the ending line; all lines in between @option{y0} and
11574 @option{y1} (including @option{y0} and @option{y1}) will be ignored. Setting
11575 @option{y0} and @option{y1} to the same value will disable the feature.
11576 @option{y0} and @option{y1} defaults to @code{0}.
11579 Set the scene change detection threshold as a percentage of maximum change on
11580 the luma plane. Good values are in the @code{[8.0, 14.0]} range. Scene change
11581 detection is only relevant in case @option{combmatch}=@var{sc}. The range for
11582 @option{scthresh} is @code{[0.0, 100.0]}.
11584 Default value is @code{12.0}.
11587 When @option{combatch} is not @var{none}, @code{fieldmatch} will take into
11588 account the combed scores of matches when deciding what match to use as the
11589 final match. Available values are:
11593 No final matching based on combed scores.
11595 Combed scores are only used when a scene change is detected.
11597 Use combed scores all the time.
11600 Default is @var{sc}.
11603 Force @code{fieldmatch} to calculate the combed metrics for certain matches and
11604 print them. This setting is known as @option{micout} in TFM/VFM vocabulary.
11605 Available values are:
11609 No forced calculation.
11611 Force p/c/n calculations.
11613 Force p/c/n/u/b calculations.
11616 Default value is @var{none}.
11619 This is the area combing threshold used for combed frame detection. This
11620 essentially controls how "strong" or "visible" combing must be to be detected.
11621 Larger values mean combing must be more visible and smaller values mean combing
11622 can be less visible or strong and still be detected. Valid settings are from
11623 @code{-1} (every pixel will be detected as combed) to @code{255} (no pixel will
11624 be detected as combed). This is basically a pixel difference value. A good
11625 range is @code{[8, 12]}.
11627 Default value is @code{9}.
11630 Sets whether or not chroma is considered in the combed frame decision. Only
11631 disable this if your source has chroma problems (rainbowing, etc.) that are
11632 causing problems for the combed frame detection with chroma enabled. Actually,
11633 using @option{chroma}=@var{0} is usually more reliable, except for the case
11634 where there is chroma only combing in the source.
11636 Default value is @code{0}.
11640 Respectively set the x-axis and y-axis size of the window used during combed
11641 frame detection. This has to do with the size of the area in which
11642 @option{combpel} pixels are required to be detected as combed for a frame to be
11643 declared combed. See the @option{combpel} parameter description for more info.
11644 Possible values are any number that is a power of 2 starting at 4 and going up
11647 Default value is @code{16}.
11650 The number of combed pixels inside any of the @option{blocky} by
11651 @option{blockx} size blocks on the frame for the frame to be detected as
11652 combed. While @option{cthresh} controls how "visible" the combing must be, this
11653 setting controls "how much" combing there must be in any localized area (a
11654 window defined by the @option{blockx} and @option{blocky} settings) on the
11655 frame. Minimum value is @code{0} and maximum is @code{blocky x blockx} (at
11656 which point no frames will ever be detected as combed). This setting is known
11657 as @option{MI} in TFM/VFM vocabulary.
11659 Default value is @code{80}.
11662 @anchor{p/c/n/u/b meaning}
11663 @subsection p/c/n/u/b meaning
11665 @subsubsection p/c/n
11667 We assume the following telecined stream:
11670 Top fields: 1 2 2 3 4
11671 Bottom fields: 1 2 3 4 4
11674 The numbers correspond to the progressive frame the fields relate to. Here, the
11675 first two frames are progressive, the 3rd and 4th are combed, and so on.
11677 When @code{fieldmatch} is configured to run a matching from bottom
11678 (@option{field}=@var{bottom}) this is how this input stream get transformed:
11683 B 1 2 3 4 4 <-- matching reference
11692 As a result of the field matching, we can see that some frames get duplicated.
11693 To perform a complete inverse telecine, you need to rely on a decimation filter
11694 after this operation. See for instance the @ref{decimate} filter.
11696 The same operation now matching from top fields (@option{field}=@var{top})
11701 T 1 2 2 3 4 <-- matching reference
11711 In these examples, we can see what @var{p}, @var{c} and @var{n} mean;
11712 basically, they refer to the frame and field of the opposite parity:
11715 @item @var{p} matches the field of the opposite parity in the previous frame
11716 @item @var{c} matches the field of the opposite parity in the current frame
11717 @item @var{n} matches the field of the opposite parity in the next frame
11722 The @var{u} and @var{b} matching are a bit special in the sense that they match
11723 from the opposite parity flag. In the following examples, we assume that we are
11724 currently matching the 2nd frame (Top:2, bottom:2). According to the match, a
11725 'x' is placed above and below each matched fields.
11727 With bottom matching (@option{field}=@var{bottom}):
11732 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11733 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11741 With top matching (@option{field}=@var{top}):
11746 Top 1 2 2 1 2 2 1 2 2 1 2 2 1 2 2
11747 Bottom 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
11755 @subsection Examples
11757 Simple IVTC of a top field first telecined stream:
11759 fieldmatch=order=tff:combmatch=none, decimate
11762 Advanced IVTC, with fallback on @ref{yadif} for still combed frames:
11764 fieldmatch=order=tff:combmatch=full, yadif=deint=interlaced, decimate
11767 @section fieldorder
11769 Transform the field order of the input video.
11771 It accepts the following parameters:
11776 The output field order. Valid values are @var{tff} for top field first or @var{bff}
11777 for bottom field first.
11780 The default value is @samp{tff}.
11782 The transformation is done by shifting the picture content up or down
11783 by one line, and filling the remaining line with appropriate picture content.
11784 This method is consistent with most broadcast field order converters.
11786 If the input video is not flagged as being interlaced, or it is already
11787 flagged as being of the required output field order, then this filter does
11788 not alter the incoming video.
11790 It is very useful when converting to or from PAL DV material,
11791 which is bottom field first.
11795 ffmpeg -i in.vob -vf "fieldorder=bff" out.dv
11798 @section fifo, afifo
11800 Buffer input images and send them when they are requested.
11802 It is mainly useful when auto-inserted by the libavfilter
11805 It does not take parameters.
11807 @section fillborders
11809 Fill borders of the input video, without changing video stream dimensions.
11810 Sometimes video can have garbage at the four edges and you may not want to
11811 crop video input to keep size multiple of some number.
11813 This filter accepts the following options:
11817 Number of pixels to fill from left border.
11820 Number of pixels to fill from right border.
11823 Number of pixels to fill from top border.
11826 Number of pixels to fill from bottom border.
11831 It accepts the following values:
11834 fill pixels using outermost pixels
11837 fill pixels using mirroring (half sample symmetric)
11840 fill pixels with constant value
11843 fill pixels using reflecting (whole sample symmetric)
11846 fill pixels using wrapping
11849 fade pixels to constant value
11852 Default is @var{smear}.
11855 Set color for pixels in fixed or fade mode. Default is @var{black}.
11858 @subsection Commands
11859 This filter supports same @ref{commands} as options.
11860 The command accepts the same syntax of the corresponding option.
11862 If the specified expression is not valid, it is kept at its current
11867 Find a rectangular object
11869 It accepts the following options:
11873 Filepath of the object image, needs to be in gray8.
11876 Detection threshold, default is 0.5.
11879 Number of mipmaps, default is 3.
11881 @item xmin, ymin, xmax, ymax
11882 Specifies the rectangle in which to search.
11885 @subsection Examples
11889 Cover a rectangular object by the supplied image of a given video using @command{ffmpeg}:
11891 ffmpeg -i file.ts -vf find_rect=newref.pgm,cover_rect=cover.jpg:mode=cover new.mkv
11897 Flood area with values of same pixel components with another values.
11899 It accepts the following options:
11902 Set pixel x coordinate.
11905 Set pixel y coordinate.
11908 Set source #0 component value.
11911 Set source #1 component value.
11914 Set source #2 component value.
11917 Set source #3 component value.
11920 Set destination #0 component value.
11923 Set destination #1 component value.
11926 Set destination #2 component value.
11929 Set destination #3 component value.
11935 Convert the input video to one of the specified pixel formats.
11936 Libavfilter will try to pick one that is suitable as input to
11939 It accepts the following parameters:
11943 A '|'-separated list of pixel format names, such as
11944 "pix_fmts=yuv420p|monow|rgb24".
11948 @subsection Examples
11952 Convert the input video to the @var{yuv420p} format
11954 format=pix_fmts=yuv420p
11957 Convert the input video to any of the formats in the list
11959 format=pix_fmts=yuv420p|yuv444p|yuv410p
11966 Convert the video to specified constant frame rate by duplicating or dropping
11967 frames as necessary.
11969 It accepts the following parameters:
11973 The desired output frame rate. The default is @code{25}.
11976 Assume the first PTS should be the given value, in seconds. This allows for
11977 padding/trimming at the start of stream. By default, no assumption is made
11978 about the first frame's expected PTS, so no padding or trimming is done.
11979 For example, this could be set to 0 to pad the beginning with duplicates of
11980 the first frame if a video stream starts after the audio stream or to trim any
11981 frames with a negative PTS.
11984 Timestamp (PTS) rounding method.
11986 Possible values are:
11993 round towards -infinity
11995 round towards +infinity
11999 The default is @code{near}.
12002 Action performed when reading the last frame.
12004 Possible values are:
12007 Use same timestamp rounding method as used for other frames.
12009 Pass through last frame if input duration has not been reached yet.
12011 The default is @code{round}.
12015 Alternatively, the options can be specified as a flat string:
12016 @var{fps}[:@var{start_time}[:@var{round}]].
12018 See also the @ref{setpts} filter.
12020 @subsection Examples
12024 A typical usage in order to set the fps to 25:
12030 Sets the fps to 24, using abbreviation and rounding method to round to nearest:
12032 fps=fps=film:round=near
12038 Pack two different video streams into a stereoscopic video, setting proper
12039 metadata on supported codecs. The two views should have the same size and
12040 framerate and processing will stop when the shorter video ends. Please note
12041 that you may conveniently adjust view properties with the @ref{scale} and
12044 It accepts the following parameters:
12048 The desired packing format. Supported values are:
12053 The views are next to each other (default).
12056 The views are on top of each other.
12059 The views are packed by line.
12062 The views are packed by column.
12065 The views are temporally interleaved.
12074 # Convert left and right views into a frame-sequential video
12075 ffmpeg -i LEFT -i RIGHT -filter_complex framepack=frameseq OUTPUT
12077 # Convert views into a side-by-side video with the same output resolution as the input
12078 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
12083 Change the frame rate by interpolating new video output frames from the source
12086 This filter is not designed to function correctly with interlaced media. If
12087 you wish to change the frame rate of interlaced media then you are required
12088 to deinterlace before this filter and re-interlace after this filter.
12090 A description of the accepted options follows.
12094 Specify the output frames per second. This option can also be specified
12095 as a value alone. The default is @code{50}.
12098 Specify the start of a range where the output frame will be created as a
12099 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12100 the default is @code{15}.
12103 Specify the end of a range where the output frame will be created as a
12104 linear interpolation of two frames. The range is [@code{0}-@code{255}],
12105 the default is @code{240}.
12108 Specify the level at which a scene change is detected as a value between
12109 0 and 100 to indicate a new scene; a low value reflects a low
12110 probability for the current frame to introduce a new scene, while a higher
12111 value means the current frame is more likely to be one.
12112 The default is @code{8.2}.
12115 Specify flags influencing the filter process.
12117 Available value for @var{flags} is:
12120 @item scene_change_detect, scd
12121 Enable scene change detection using the value of the option @var{scene}.
12122 This flag is enabled by default.
12128 Select one frame every N-th frame.
12130 This filter accepts the following option:
12133 Select frame after every @code{step} frames.
12134 Allowed values are positive integers higher than 0. Default value is @code{1}.
12137 @section freezedetect
12139 Detect frozen video.
12141 This filter logs a message and sets frame metadata when it detects that the
12142 input video has no significant change in content during a specified duration.
12143 Video freeze detection calculates the mean average absolute difference of all
12144 the components of video frames and compares it to a noise floor.
12146 The printed times and duration are expressed in seconds. The
12147 @code{lavfi.freezedetect.freeze_start} metadata key is set on the first frame
12148 whose timestamp equals or exceeds the detection duration and it contains the
12149 timestamp of the first frame of the freeze. The
12150 @code{lavfi.freezedetect.freeze_duration} and
12151 @code{lavfi.freezedetect.freeze_end} metadata keys are set on the first frame
12154 The filter accepts the following options:
12158 Set noise tolerance. Can be specified in dB (in case "dB" is appended to the
12159 specified value) or as a difference ratio between 0 and 1. Default is -60dB, or
12163 Set freeze duration until notification (default is 2 seconds).
12166 @section freezeframes
12168 Freeze video frames.
12170 This filter freezes video frames using frame from 2nd input.
12172 The filter accepts the following options:
12176 Set number of first frame from which to start freeze.
12179 Set number of last frame from which to end freeze.
12182 Set number of frame from 2nd input which will be used instead of replaced frames.
12188 Apply a frei0r effect to the input video.
12190 To enable the compilation of this filter, you need to install the frei0r
12191 header and configure FFmpeg with @code{--enable-frei0r}.
12193 It accepts the following parameters:
12198 The name of the frei0r effect to load. If the environment variable
12199 @env{FREI0R_PATH} is defined, the frei0r effect is searched for in each of the
12200 directories specified by the colon-separated list in @env{FREI0R_PATH}.
12201 Otherwise, the standard frei0r paths are searched, in this order:
12202 @file{HOME/.frei0r-1/lib/}, @file{/usr/local/lib/frei0r-1/},
12203 @file{/usr/lib/frei0r-1/}.
12205 @item filter_params
12206 A '|'-separated list of parameters to pass to the frei0r effect.
12210 A frei0r effect parameter can be a boolean (its value is either
12211 "y" or "n"), a double, a color (specified as
12212 @var{R}/@var{G}/@var{B}, where @var{R}, @var{G}, and @var{B} are floating point
12213 numbers between 0.0 and 1.0, inclusive) or a color description as specified in the
12214 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils},
12215 a position (specified as @var{X}/@var{Y}, where
12216 @var{X} and @var{Y} are floating point numbers) and/or a string.
12218 The number and types of parameters depend on the loaded effect. If an
12219 effect parameter is not specified, the default value is set.
12221 @subsection Examples
12225 Apply the distort0r effect, setting the first two double parameters:
12227 frei0r=filter_name=distort0r:filter_params=0.5|0.01
12231 Apply the colordistance effect, taking a color as the first parameter:
12233 frei0r=colordistance:0.2/0.3/0.4
12234 frei0r=colordistance:violet
12235 frei0r=colordistance:0x112233
12239 Apply the perspective effect, specifying the top left and top right image
12242 frei0r=perspective:0.2/0.2|0.8/0.2
12246 For more information, see
12247 @url{http://frei0r.dyne.org}
12249 @subsection Commands
12251 This filter supports the @option{filter_params} option as @ref{commands}.
12255 Apply fast and simple postprocessing. It is a faster version of @ref{spp}.
12257 It splits (I)DCT into horizontal/vertical passes. Unlike the simple post-
12258 processing filter, one of them is performed once per block, not per pixel.
12259 This allows for much higher speed.
12261 The filter accepts the following options:
12265 Set quality. This option defines the number of levels for averaging. It accepts
12266 an integer in the range 4-5. Default value is @code{4}.
12269 Force a constant quantization parameter. It accepts an integer in range 0-63.
12270 If not set, the filter will use the QP from the video stream (if available).
12273 Set filter strength. It accepts an integer in range -15 to 32. Lower values mean
12274 more details but also more artifacts, while higher values make the image smoother
12275 but also blurrier. Default value is @code{0} − PSNR optimal.
12277 @item use_bframe_qp
12278 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
12279 option may cause flicker since the B-Frames have often larger QP. Default is
12280 @code{0} (not enabled).
12286 Apply Gaussian blur filter.
12288 The filter accepts the following options:
12292 Set horizontal sigma, standard deviation of Gaussian blur. Default is @code{0.5}.
12295 Set number of steps for Gaussian approximation. Default is @code{1}.
12298 Set which planes to filter. By default all planes are filtered.
12301 Set vertical sigma, if negative it will be same as @code{sigma}.
12302 Default is @code{-1}.
12305 @subsection Commands
12306 This filter supports same commands as options.
12307 The command accepts the same syntax of the corresponding option.
12309 If the specified expression is not valid, it is kept at its current
12314 Apply generic equation to each pixel.
12316 The filter accepts the following options:
12319 @item lum_expr, lum
12320 Set the luminance expression.
12322 Set the chrominance blue expression.
12324 Set the chrominance red expression.
12325 @item alpha_expr, a
12326 Set the alpha expression.
12328 Set the red expression.
12329 @item green_expr, g
12330 Set the green expression.
12332 Set the blue expression.
12335 The colorspace is selected according to the specified options. If one
12336 of the @option{lum_expr}, @option{cb_expr}, or @option{cr_expr}
12337 options is specified, the filter will automatically select a YCbCr
12338 colorspace. If one of the @option{red_expr}, @option{green_expr}, or
12339 @option{blue_expr} options is specified, it will select an RGB
12342 If one of the chrominance expression is not defined, it falls back on the other
12343 one. If no alpha expression is specified it will evaluate to opaque value.
12344 If none of chrominance expressions are specified, they will evaluate
12345 to the luminance expression.
12347 The expressions can use the following variables and functions:
12351 The sequential number of the filtered frame, starting from @code{0}.
12355 The coordinates of the current sample.
12359 The width and height of the image.
12363 Width and height scale depending on the currently filtered plane. It is the
12364 ratio between the corresponding luma plane number of pixels and the current
12365 plane ones. E.g. for YUV4:2:0 the values are @code{1,1} for the luma plane, and
12366 @code{0.5,0.5} for chroma planes.
12369 Time of the current frame, expressed in seconds.
12372 Return the value of the pixel at location (@var{x},@var{y}) of the current
12376 Return the value of the pixel at location (@var{x},@var{y}) of the luminance
12380 Return the value of the pixel at location (@var{x},@var{y}) of the
12381 blue-difference chroma plane. Return 0 if there is no such plane.
12384 Return the value of the pixel at location (@var{x},@var{y}) of the
12385 red-difference chroma plane. Return 0 if there is no such plane.
12390 Return the value of the pixel at location (@var{x},@var{y}) of the
12391 red/green/blue component. Return 0 if there is no such component.
12394 Return the value of the pixel at location (@var{x},@var{y}) of the alpha
12395 plane. Return 0 if there is no such plane.
12397 @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)
12398 Sum of sample values in the rectangle from (0,0) to (x,y), this allows obtaining
12399 sums of samples within a rectangle. See the functions without the sum postfix.
12401 @item interpolation
12402 Set one of interpolation methods:
12407 Default is bilinear.
12410 For functions, if @var{x} and @var{y} are outside the area, the value will be
12411 automatically clipped to the closer edge.
12413 Please note that this filter can use multiple threads in which case each slice
12414 will have its own expression state. If you want to use only a single expression
12415 state because your expressions depend on previous state then you should limit
12416 the number of filter threads to 1.
12418 @subsection Examples
12422 Flip the image horizontally:
12428 Generate a bidimensional sine wave, with angle @code{PI/3} and a
12429 wavelength of 100 pixels:
12431 geq=128 + 100*sin(2*(PI/100)*(cos(PI/3)*(X-50*T) + sin(PI/3)*Y)):128:128
12435 Generate a fancy enigmatic moving light:
12437 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
12441 Generate a quick emboss effect:
12443 format=gray,geq=lum_expr='(p(X,Y)+(256-p(X-4,Y-4)))/2'
12447 Modify RGB components depending on pixel position:
12449 geq=r='X/W*r(X,Y)':g='(1-X/W)*g(X,Y)':b='(H-Y)/H*b(X,Y)'
12453 Create a radial gradient that is the same size as the input (also see
12454 the @ref{vignette} filter):
12456 geq=lum=255*gauss((X/W-0.5)*3)*gauss((Y/H-0.5)*3)/gauss(0)/gauss(0),format=gray
12462 Fix the banding artifacts that are sometimes introduced into nearly flat
12463 regions by truncation to 8-bit color depth.
12464 Interpolate the gradients that should go where the bands are, and
12467 It is designed for playback only. Do not use it prior to
12468 lossy compression, because compression tends to lose the dither and
12469 bring back the bands.
12471 It accepts the following parameters:
12476 The maximum amount by which the filter will change any one pixel. This is also
12477 the threshold for detecting nearly flat regions. Acceptable values range from
12478 .51 to 64; the default value is 1.2. Out-of-range values will be clipped to the
12482 The neighborhood to fit the gradient to. A larger radius makes for smoother
12483 gradients, but also prevents the filter from modifying the pixels near detailed
12484 regions. Acceptable values are 8-32; the default value is 16. Out-of-range
12485 values will be clipped to the valid range.
12489 Alternatively, the options can be specified as a flat string:
12490 @var{strength}[:@var{radius}]
12492 @subsection Examples
12496 Apply the filter with a @code{3.5} strength and radius of @code{8}:
12502 Specify radius, omitting the strength (which will fall-back to the default
12510 @anchor{graphmonitor}
12511 @section graphmonitor
12512 Show various filtergraph stats.
12514 With this filter one can debug complete filtergraph.
12515 Especially issues with links filling with queued frames.
12517 The filter accepts the following options:
12521 Set video output size. Default is @var{hd720}.
12524 Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}.
12527 Set output mode, can be @var{fulll} or @var{compact}.
12528 In @var{compact} mode only filters with some queued frames have displayed stats.
12531 Set flags which enable which stats are shown in video.
12533 Available values for flags are:
12536 Display number of queued frames in each link.
12538 @item frame_count_in
12539 Display number of frames taken from filter.
12541 @item frame_count_out
12542 Display number of frames given out from filter.
12545 Display current filtered frame pts.
12548 Display current filtered frame time.
12551 Display time base for filter link.
12554 Display used format for filter link.
12557 Display video size or number of audio channels in case of audio used by filter link.
12560 Display video frame rate or sample rate in case of audio used by filter link.
12563 Display link output status.
12567 Set upper limit for video rate of output stream, Default value is @var{25}.
12568 This guarantee that output video frame rate will not be higher than this value.
12572 A color constancy variation filter which estimates scene illumination via grey edge algorithm
12573 and corrects the scene colors accordingly.
12575 See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
12577 The filter accepts the following options:
12581 The order of differentiation to be applied on the scene. Must be chosen in the range
12582 [0,2] and default value is 1.
12585 The Minkowski parameter to be used for calculating the Minkowski distance. Must
12586 be chosen in the range [0,20] and default value is 1. Set to 0 for getting
12587 max value instead of calculating Minkowski distance.
12590 The standard deviation of Gaussian blur to be applied on the scene. Must be
12591 chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
12592 can't be equal to 0 if @var{difford} is greater than 0.
12595 @subsection Examples
12601 greyedge=difford=1:minknorm=5:sigma=2
12607 greyedge=difford=1:minknorm=0:sigma=2
12615 Apply a Hald CLUT to a video stream.
12617 First input is the video stream to process, and second one is the Hald CLUT.
12618 The Hald CLUT input can be a simple picture or a complete video stream.
12620 The filter accepts the following options:
12624 Force termination when the shortest input terminates. Default is @code{0}.
12626 Continue applying the last CLUT after the end of the stream. A value of
12627 @code{0} disable the filter after the last frame of the CLUT is reached.
12628 Default is @code{1}.
12631 @code{haldclut} also has the same interpolation options as @ref{lut3d} (both
12632 filters share the same internals).
12634 This filter also supports the @ref{framesync} options.
12636 More information about the Hald CLUT can be found on Eskil Steenberg's website
12637 (Hald CLUT author) at @url{http://www.quelsolaar.com/technology/clut.html}.
12639 @subsection Workflow examples
12641 @subsubsection Hald CLUT video stream
12643 Generate an identity Hald CLUT stream altered with various effects:
12645 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
12648 Note: make sure you use a lossless codec.
12650 Then use it with @code{haldclut} to apply it on some random stream:
12652 ffmpeg -f lavfi -i mandelbrot -i clut.nut -filter_complex '[0][1] haldclut' -t 20 mandelclut.mkv
12655 The Hald CLUT will be applied to the 10 first seconds (duration of
12656 @file{clut.nut}), then the latest picture of that CLUT stream will be applied
12657 to the remaining frames of the @code{mandelbrot} stream.
12659 @subsubsection Hald CLUT with preview
12661 A Hald CLUT is supposed to be a squared image of @code{Level*Level*Level} by
12662 @code{Level*Level*Level} pixels. For a given Hald CLUT, FFmpeg will select the
12663 biggest possible square starting at the top left of the picture. The remaining
12664 padding pixels (bottom or right) will be ignored. This area can be used to add
12665 a preview of the Hald CLUT.
12667 Typically, the following generated Hald CLUT will be supported by the
12668 @code{haldclut} filter:
12671 ffmpeg -f lavfi -i @ref{haldclutsrc}=8 -vf "
12672 pad=iw+320 [padded_clut];
12673 smptebars=s=320x256, split [a][b];
12674 [padded_clut][a] overlay=W-320:h, curves=color_negative [main];
12675 [main][b] overlay=W-320" -frames:v 1 clut.png
12678 It contains the original and a preview of the effect of the CLUT: SMPTE color
12679 bars are displayed on the right-top, and below the same color bars processed by
12682 Then, the effect of this Hald CLUT can be visualized with:
12684 ffplay input.mkv -vf "movie=clut.png, [in] haldclut"
12689 Flip the input video horizontally.
12691 For example, to horizontally flip the input video with @command{ffmpeg}:
12693 ffmpeg -i in.avi -vf "hflip" out.avi
12697 This filter applies a global color histogram equalization on a
12700 It can be used to correct video that has a compressed range of pixel
12701 intensities. The filter redistributes the pixel intensities to
12702 equalize their distribution across the intensity range. It may be
12703 viewed as an "automatically adjusting contrast filter". This filter is
12704 useful only for correcting degraded or poorly captured source
12707 The filter accepts the following options:
12711 Determine the amount of equalization to be applied. As the strength
12712 is reduced, the distribution of pixel intensities more-and-more
12713 approaches that of the input frame. The value must be a float number
12714 in the range [0,1] and defaults to 0.200.
12717 Set the maximum intensity that can generated and scale the output
12718 values appropriately. The strength should be set as desired and then
12719 the intensity can be limited if needed to avoid washing-out. The value
12720 must be a float number in the range [0,1] and defaults to 0.210.
12723 Set the antibanding level. If enabled the filter will randomly vary
12724 the luminance of output pixels by a small amount to avoid banding of
12725 the histogram. Possible values are @code{none}, @code{weak} or
12726 @code{strong}. It defaults to @code{none}.
12732 Compute and draw a color distribution histogram for the input video.
12734 The computed histogram is a representation of the color component
12735 distribution in an image.
12737 Standard histogram displays the color components distribution in an image.
12738 Displays color graph for each color component. Shows distribution of
12739 the Y, U, V, A or R, G, B components, depending on input format, in the
12740 current frame. Below each graph a color component scale meter is shown.
12742 The filter accepts the following options:
12746 Set height of level. Default value is @code{200}.
12747 Allowed range is [50, 2048].
12750 Set height of color scale. Default value is @code{12}.
12751 Allowed range is [0, 40].
12755 It accepts the following values:
12758 Per color component graphs are placed below each other.
12761 Per color component graphs are placed side by side.
12764 Presents information identical to that in the @code{parade}, except
12765 that the graphs representing color components are superimposed directly
12768 Default is @code{stack}.
12771 Set mode. Can be either @code{linear}, or @code{logarithmic}.
12772 Default is @code{linear}.
12775 Set what color components to display.
12776 Default is @code{7}.
12779 Set foreground opacity. Default is @code{0.7}.
12782 Set background opacity. Default is @code{0.5}.
12785 @subsection Examples
12790 Calculate and draw histogram:
12792 ffplay -i input -vf histogram
12800 This is a high precision/quality 3d denoise filter. It aims to reduce
12801 image noise, producing smooth images and making still images really
12802 still. It should enhance compressibility.
12804 It accepts the following optional parameters:
12808 A non-negative floating point number which specifies spatial luma strength.
12809 It defaults to 4.0.
12811 @item chroma_spatial
12812 A non-negative floating point number which specifies spatial chroma strength.
12813 It defaults to 3.0*@var{luma_spatial}/4.0.
12816 A floating point number which specifies luma temporal strength. It defaults to
12817 6.0*@var{luma_spatial}/4.0.
12820 A floating point number which specifies chroma temporal strength. It defaults to
12821 @var{luma_tmp}*@var{chroma_spatial}/@var{luma_spatial}.
12824 @subsection Commands
12825 This filter supports same @ref{commands} as options.
12826 The command accepts the same syntax of the corresponding option.
12828 If the specified expression is not valid, it is kept at its current
12831 @anchor{hwdownload}
12832 @section hwdownload
12834 Download hardware frames to system memory.
12836 The input must be in hardware frames, and the output a non-hardware format.
12837 Not all formats will be supported on the output - it may be necessary to insert
12838 an additional @option{format} filter immediately following in the graph to get
12839 the output in a supported format.
12843 Map hardware frames to system memory or to another device.
12845 This filter has several different modes of operation; which one is used depends
12846 on the input and output formats:
12849 Hardware frame input, normal frame output
12851 Map the input frames to system memory and pass them to the output. If the
12852 original hardware frame is later required (for example, after overlaying
12853 something else on part of it), the @option{hwmap} filter can be used again
12854 in the next mode to retrieve it.
12856 Normal frame input, hardware frame output
12858 If the input is actually a software-mapped hardware frame, then unmap it -
12859 that is, return the original hardware frame.
12861 Otherwise, a device must be provided. Create new hardware surfaces on that
12862 device for the output, then map them back to the software format at the input
12863 and give those frames to the preceding filter. This will then act like the
12864 @option{hwupload} filter, but may be able to avoid an additional copy when
12865 the input is already in a compatible format.
12867 Hardware frame input and output
12869 A device must be supplied for the output, either directly or with the
12870 @option{derive_device} option. The input and output devices must be of
12871 different types and compatible - the exact meaning of this is
12872 system-dependent, but typically it means that they must refer to the same
12873 underlying hardware context (for example, refer to the same graphics card).
12875 If the input frames were originally created on the output device, then unmap
12876 to retrieve the original frames.
12878 Otherwise, map the frames to the output device - create new hardware frames
12879 on the output corresponding to the frames on the input.
12882 The following additional parameters are accepted:
12886 Set the frame mapping mode. Some combination of:
12889 The mapped frame should be readable.
12891 The mapped frame should be writeable.
12893 The mapping will always overwrite the entire frame.
12895 This may improve performance in some cases, as the original contents of the
12896 frame need not be loaded.
12898 The mapping must not involve any copying.
12900 Indirect mappings to copies of frames are created in some cases where either
12901 direct mapping is not possible or it would have unexpected properties.
12902 Setting this flag ensures that the mapping is direct and will fail if that is
12905 Defaults to @var{read+write} if not specified.
12907 @item derive_device @var{type}
12908 Rather than using the device supplied at initialisation, instead derive a new
12909 device of type @var{type} from the device the input frames exist on.
12912 In a hardware to hardware mapping, map in reverse - create frames in the sink
12913 and map them back to the source. This may be necessary in some cases where
12914 a mapping in one direction is required but only the opposite direction is
12915 supported by the devices being used.
12917 This option is dangerous - it may break the preceding filter in undefined
12918 ways if there are any additional constraints on that filter's output.
12919 Do not use it without fully understanding the implications of its use.
12925 Upload system memory frames to hardware surfaces.
12927 The device to upload to must be supplied when the filter is initialised. If
12928 using ffmpeg, select the appropriate device with the @option{-filter_hw_device}
12929 option or with the @option{derive_device} option. The input and output devices
12930 must be of different types and compatible - the exact meaning of this is
12931 system-dependent, but typically it means that they must refer to the same
12932 underlying hardware context (for example, refer to the same graphics card).
12934 The following additional parameters are accepted:
12937 @item derive_device @var{type}
12938 Rather than using the device supplied at initialisation, instead derive a new
12939 device of type @var{type} from the device the input frames exist on.
12942 @anchor{hwupload_cuda}
12943 @section hwupload_cuda
12945 Upload system memory frames to a CUDA device.
12947 It accepts the following optional parameters:
12951 The number of the CUDA device to use
12956 Apply a high-quality magnification filter designed for pixel art. This filter
12957 was originally created by Maxim Stepin.
12959 It accepts the following option:
12963 Set the scaling dimension: @code{2} for @code{hq2x}, @code{3} for
12964 @code{hq3x} and @code{4} for @code{hq4x}.
12965 Default is @code{3}.
12969 Stack input videos horizontally.
12971 All streams must be of same pixel format and of same height.
12973 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
12974 to create same output.
12976 The filter accepts the following option:
12980 Set number of input streams. Default is 2.
12983 If set to 1, force the output to terminate when the shortest input
12984 terminates. Default value is 0.
12989 Modify the hue and/or the saturation of the input.
12991 It accepts the following parameters:
12995 Specify the hue angle as a number of degrees. It accepts an expression,
12996 and defaults to "0".
12999 Specify the saturation in the [-10,10] range. It accepts an expression and
13003 Specify the hue angle as a number of radians. It accepts an
13004 expression, and defaults to "0".
13007 Specify the brightness in the [-10,10] range. It accepts an expression and
13011 @option{h} and @option{H} are mutually exclusive, and can't be
13012 specified at the same time.
13014 The @option{b}, @option{h}, @option{H} and @option{s} option values are
13015 expressions containing the following constants:
13019 frame count of the input frame starting from 0
13022 presentation timestamp of the input frame expressed in time base units
13025 frame rate of the input video, NAN if the input frame rate is unknown
13028 timestamp expressed in seconds, NAN if the input timestamp is unknown
13031 time base of the input video
13034 @subsection Examples
13038 Set the hue to 90 degrees and the saturation to 1.0:
13044 Same command but expressing the hue in radians:
13050 Rotate hue and make the saturation swing between 0
13051 and 2 over a period of 1 second:
13053 hue="H=2*PI*t: s=sin(2*PI*t)+1"
13057 Apply a 3 seconds saturation fade-in effect starting at 0:
13059 hue="s=min(t/3\,1)"
13062 The general fade-in expression can be written as:
13064 hue="s=min(0\, max((t-START)/DURATION\, 1))"
13068 Apply a 3 seconds saturation fade-out effect starting at 5 seconds:
13070 hue="s=max(0\, min(1\, (8-t)/3))"
13073 The general fade-out expression can be written as:
13075 hue="s=max(0\, min(1\, (START+DURATION-t)/DURATION))"
13080 @subsection Commands
13082 This filter supports the following commands:
13088 Modify the hue and/or the saturation and/or brightness of the input video.
13089 The command accepts the same syntax of the corresponding option.
13091 If the specified expression is not valid, it is kept at its current
13095 @section hysteresis
13097 Grow first stream into second stream by connecting components.
13098 This makes it possible to build more robust edge masks.
13100 This filter accepts the following options:
13104 Set which planes will be processed as bitmap, unprocessed planes will be
13105 copied from first stream.
13106 By default value 0xf, all planes will be processed.
13109 Set threshold which is used in filtering. If pixel component value is higher than
13110 this value filter algorithm for connecting components is activated.
13111 By default value is 0.
13114 The @code{hysteresis} filter also supports the @ref{framesync} options.
13118 Detect video interlacing type.
13120 This filter tries to detect if the input frames are interlaced, progressive,
13121 top or bottom field first. It will also try to detect fields that are
13122 repeated between adjacent frames (a sign of telecine).
13124 Single frame detection considers only immediately adjacent frames when classifying each frame.
13125 Multiple frame detection incorporates the classification history of previous frames.
13127 The filter will log these metadata values:
13130 @item single.current_frame
13131 Detected type of current frame using single-frame detection. One of:
13132 ``tff'' (top field first), ``bff'' (bottom field first),
13133 ``progressive'', or ``undetermined''
13136 Cumulative number of frames detected as top field first using single-frame detection.
13139 Cumulative number of frames detected as top field first using multiple-frame detection.
13142 Cumulative number of frames detected as bottom field first using single-frame detection.
13144 @item multiple.current_frame
13145 Detected type of current frame using multiple-frame detection. One of:
13146 ``tff'' (top field first), ``bff'' (bottom field first),
13147 ``progressive'', or ``undetermined''
13150 Cumulative number of frames detected as bottom field first using multiple-frame detection.
13152 @item single.progressive
13153 Cumulative number of frames detected as progressive using single-frame detection.
13155 @item multiple.progressive
13156 Cumulative number of frames detected as progressive using multiple-frame detection.
13158 @item single.undetermined
13159 Cumulative number of frames that could not be classified using single-frame detection.
13161 @item multiple.undetermined
13162 Cumulative number of frames that could not be classified using multiple-frame detection.
13164 @item repeated.current_frame
13165 Which field in the current frame is repeated from the last. One of ``neither'', ``top'', or ``bottom''.
13167 @item repeated.neither
13168 Cumulative number of frames with no repeated field.
13171 Cumulative number of frames with the top field repeated from the previous frame's top field.
13173 @item repeated.bottom
13174 Cumulative number of frames with the bottom field repeated from the previous frame's bottom field.
13177 The filter accepts the following options:
13181 Set interlacing threshold.
13183 Set progressive threshold.
13185 Threshold for repeated field detection.
13187 Number of frames after which a given frame's contribution to the
13188 statistics is halved (i.e., it contributes only 0.5 to its
13189 classification). The default of 0 means that all frames seen are given
13190 full weight of 1.0 forever.
13191 @item analyze_interlaced_flag
13192 When this is not 0 then idet will use the specified number of frames to determine
13193 if the interlaced flag is accurate, it will not count undetermined frames.
13194 If the flag is found to be accurate it will be used without any further
13195 computations, if it is found to be inaccurate it will be cleared without any
13196 further computations. This allows inserting the idet filter as a low computational
13197 method to clean up the interlaced flag
13202 Deinterleave or interleave fields.
13204 This filter allows one to process interlaced images fields without
13205 deinterlacing them. Deinterleaving splits the input frame into 2
13206 fields (so called half pictures). Odd lines are moved to the top
13207 half of the output image, even lines to the bottom half.
13208 You can process (filter) them independently and then re-interleave them.
13210 The filter accepts the following options:
13214 @item chroma_mode, c
13215 @item alpha_mode, a
13216 Available values for @var{luma_mode}, @var{chroma_mode} and
13217 @var{alpha_mode} are:
13223 @item deinterleave, d
13224 Deinterleave fields, placing one above the other.
13226 @item interleave, i
13227 Interleave fields. Reverse the effect of deinterleaving.
13229 Default value is @code{none}.
13231 @item luma_swap, ls
13232 @item chroma_swap, cs
13233 @item alpha_swap, as
13234 Swap luma/chroma/alpha fields. Exchange even & odd lines. Default value is @code{0}.
13237 @subsection Commands
13239 This filter supports the all above options as @ref{commands}.
13243 Apply inflate effect to the video.
13245 This filter replaces the pixel by the local(3x3) average by taking into account
13246 only values higher than the pixel.
13248 It accepts the following options:
13255 Limit the maximum change for each plane, default is 65535.
13256 If 0, plane will remain unchanged.
13259 @subsection Commands
13261 This filter supports the all above options as @ref{commands}.
13265 Simple interlacing filter from progressive contents. This interleaves upper (or
13266 lower) lines from odd frames with lower (or upper) lines from even frames,
13267 halving the frame rate and preserving image height.
13270 Original Original New Frame
13271 Frame 'j' Frame 'j+1' (tff)
13272 ========== =========== ==================
13273 Line 0 --------------------> Frame 'j' Line 0
13274 Line 1 Line 1 ----> Frame 'j+1' Line 1
13275 Line 2 ---------------------> Frame 'j' Line 2
13276 Line 3 Line 3 ----> Frame 'j+1' Line 3
13278 New Frame + 1 will be generated by Frame 'j+2' and Frame 'j+3' and so on
13281 It accepts the following optional parameters:
13285 This determines whether the interlaced frame is taken from the even
13286 (tff - default) or odd (bff) lines of the progressive frame.
13289 Vertical lowpass filter to avoid twitter interlacing and
13290 reduce moire patterns.
13294 Disable vertical lowpass filter
13297 Enable linear filter (default)
13300 Enable complex filter. This will slightly less reduce twitter and moire
13301 but better retain detail and subjective sharpness impression.
13308 Deinterlace input video by applying Donald Graft's adaptive kernel
13309 deinterling. Work on interlaced parts of a video to produce
13310 progressive frames.
13312 The description of the accepted parameters follows.
13316 Set the threshold which affects the filter's tolerance when
13317 determining if a pixel line must be processed. It must be an integer
13318 in the range [0,255] and defaults to 10. A value of 0 will result in
13319 applying the process on every pixels.
13322 Paint pixels exceeding the threshold value to white if set to 1.
13326 Set the fields order. Swap fields if set to 1, leave fields alone if
13330 Enable additional sharpening if set to 1. Default is 0.
13333 Enable twoway sharpening if set to 1. Default is 0.
13336 @subsection Examples
13340 Apply default values:
13342 kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
13346 Enable additional sharpening:
13352 Paint processed pixels in white:
13360 Slowly update darker pixels.
13362 This filter makes short flashes of light appear longer.
13363 This filter accepts the following options:
13367 Set factor for decaying. Default is .95. Allowed range is from 0 to 1.
13370 Set which planes to filter. Default is all. Allowed range is from 0 to 15.
13373 @section lenscorrection
13375 Correct radial lens distortion
13377 This filter can be used to correct for radial distortion as can result from the use
13378 of wide angle lenses, and thereby re-rectify the image. To find the right parameters
13379 one can use tools available for example as part of opencv or simply trial-and-error.
13380 To use opencv use the calibration sample (under samples/cpp) from the opencv sources
13381 and extract the k1 and k2 coefficients from the resulting matrix.
13383 Note that effectively the same filter is available in the open-source tools Krita and
13384 Digikam from the KDE project.
13386 In contrast to the @ref{vignette} filter, which can also be used to compensate lens errors,
13387 this filter corrects the distortion of the image, whereas @ref{vignette} corrects the
13388 brightness distribution, so you may want to use both filters together in certain
13389 cases, though you will have to take care of ordering, i.e. whether vignetting should
13390 be applied before or after lens correction.
13392 @subsection Options
13394 The filter accepts the following options:
13398 Relative x-coordinate of the focal point of the image, and thereby the center of the
13399 distortion. This value has a range [0,1] and is expressed as fractions of the image
13400 width. Default is 0.5.
13402 Relative y-coordinate of the focal point of the image, and thereby the center of the
13403 distortion. This value has a range [0,1] and is expressed as fractions of the image
13404 height. Default is 0.5.
13406 Coefficient of the quadratic correction term. This value has a range [-1,1]. 0 means
13407 no correction. Default is 0.
13409 Coefficient of the double quadratic correction term. This value has a range [-1,1].
13410 0 means no correction. Default is 0.
13413 The formula that generates the correction is:
13415 @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)
13417 where @var{r_0} is halve of the image diagonal and @var{r_src} and @var{r_tgt} are the
13418 distances from the focal point in the source and target images, respectively.
13422 Apply lens correction via the lensfun library (@url{http://lensfun.sourceforge.net/}).
13424 The @code{lensfun} filter requires the camera make, camera model, and lens model
13425 to apply the lens correction. The filter will load the lensfun database and
13426 query it to find the corresponding camera and lens entries in the database. As
13427 long as these entries can be found with the given options, the filter can
13428 perform corrections on frames. Note that incomplete strings will result in the
13429 filter choosing the best match with the given options, and the filter will
13430 output the chosen camera and lens models (logged with level "info"). You must
13431 provide the make, camera model, and lens model as they are required.
13433 The filter accepts the following options:
13437 The make of the camera (for example, "Canon"). This option is required.
13440 The model of the camera (for example, "Canon EOS 100D"). This option is
13444 The model of the lens (for example, "Canon EF-S 18-55mm f/3.5-5.6 IS STM"). This
13445 option is required.
13448 The type of correction to apply. The following values are valid options:
13452 Enables fixing lens vignetting.
13455 Enables fixing lens geometry. This is the default.
13458 Enables fixing chromatic aberrations.
13461 Enables fixing lens vignetting and lens geometry.
13464 Enables fixing lens vignetting and chromatic aberrations.
13467 Enables fixing both lens geometry and chromatic aberrations.
13470 Enables all possible corrections.
13474 The focal length of the image/video (zoom; expected constant for video). For
13475 example, a 18--55mm lens has focal length range of [18--55], so a value in that
13476 range should be chosen when using that lens. Default 18.
13479 The aperture of the image/video (expected constant for video). Note that
13480 aperture is only used for vignetting correction. Default 3.5.
13482 @item focus_distance
13483 The focus distance of the image/video (expected constant for video). Note that
13484 focus distance is only used for vignetting and only slightly affects the
13485 vignetting correction process. If unknown, leave it at the default value (which
13489 The scale factor which is applied after transformation. After correction the
13490 video is no longer necessarily rectangular. This parameter controls how much of
13491 the resulting image is visible. The value 0 means that a value will be chosen
13492 automatically such that there is little or no unmapped area in the output
13493 image. 1.0 means that no additional scaling is done. Lower values may result
13494 in more of the corrected image being visible, while higher values may avoid
13495 unmapped areas in the output.
13497 @item target_geometry
13498 The target geometry of the output image/video. The following values are valid
13502 @item rectilinear (default)
13505 @item equirectangular
13506 @item fisheye_orthographic
13507 @item fisheye_stereographic
13508 @item fisheye_equisolid
13509 @item fisheye_thoby
13512 Apply the reverse of image correction (instead of correcting distortion, apply
13515 @item interpolation
13516 The type of interpolation used when correcting distortion. The following values
13521 @item linear (default)
13526 @subsection Examples
13530 Apply lens correction with make "Canon", camera model "Canon EOS 100D", and lens
13531 model "Canon EF-S 18-55mm f/3.5-5.6 IS STM" with focal length of "18" and
13535 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
13539 Apply the same as before, but only for the first 5 seconds of video.
13542 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
13549 Obtain the VMAF (Video Multi-Method Assessment Fusion)
13550 score between two input videos.
13552 The obtained VMAF score is printed through the logging system.
13554 It requires Netflix's vmaf library (libvmaf) as a pre-requisite.
13555 After installing the library it can be enabled using:
13556 @code{./configure --enable-libvmaf}.
13557 If no model path is specified it uses the default model: @code{vmaf_v0.6.1.pkl}.
13559 The filter has following options:
13563 Set the model path which is to be used for SVM.
13564 Default value: @code{"/usr/local/share/model/vmaf_v0.6.1.pkl"}
13567 Set the file path to be used to store logs.
13570 Set the format of the log file (csv, json or xml).
13572 @item enable_transform
13573 This option can enable/disable the @code{score_transform} applied to the final predicted VMAF score,
13574 if you have specified score_transform option in the input parameter file passed to @code{run_vmaf_training.py}
13575 Default value: @code{false}
13578 Invokes the phone model which will generate VMAF scores higher than in the
13579 regular model, which is more suitable for laptop, TV, etc. viewing conditions.
13580 Default value: @code{false}
13583 Enables computing psnr along with vmaf.
13584 Default value: @code{false}
13587 Enables computing ssim along with vmaf.
13588 Default value: @code{false}
13591 Enables computing ms_ssim along with vmaf.
13592 Default value: @code{false}
13595 Set the pool method to be used for computing vmaf.
13596 Options are @code{min}, @code{harmonic_mean} or @code{mean} (default).
13599 Set number of threads to be used when computing vmaf.
13600 Default value: @code{0}, which makes use of all available logical processors.
13603 Set interval for frame subsampling used when computing vmaf.
13604 Default value: @code{1}
13606 @item enable_conf_interval
13607 Enables confidence interval.
13608 Default value: @code{false}
13611 This filter also supports the @ref{framesync} options.
13613 @subsection Examples
13616 On the below examples the input file @file{main.mpg} being processed is
13617 compared with the reference file @file{ref.mpg}.
13620 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf -f null -
13624 Example with options:
13626 ffmpeg -i main.mpg -i ref.mpg -lavfi libvmaf="psnr=1:log_fmt=json" -f null -
13630 Example with options and different containers:
13632 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 -
13638 Limits the pixel components values to the specified range [min, max].
13640 The filter accepts the following options:
13644 Lower bound. Defaults to the lowest allowed value for the input.
13647 Upper bound. Defaults to the highest allowed value for the input.
13650 Specify which planes will be processed. Defaults to all available.
13653 @subsection Commands
13655 This filter supports the all above options as @ref{commands}.
13661 The filter accepts the following options:
13665 Set the number of loops. Setting this value to -1 will result in infinite loops.
13669 Set maximal size in number of frames. Default is 0.
13672 Set first frame of loop. Default is 0.
13675 @subsection Examples
13679 Loop single first frame infinitely:
13681 loop=loop=-1:size=1:start=0
13685 Loop single first frame 10 times:
13687 loop=loop=10:size=1:start=0
13691 Loop 10 first frames 5 times:
13693 loop=loop=5:size=10:start=0
13699 Apply a 1D LUT to an input video.
13701 The filter accepts the following options:
13705 Set the 1D LUT file name.
13707 Currently supported formats:
13716 Select interpolation mode.
13718 Available values are:
13722 Use values from the nearest defined point.
13724 Interpolate values using the linear interpolation.
13726 Interpolate values using the cosine interpolation.
13728 Interpolate values using the cubic interpolation.
13730 Interpolate values using the spline interpolation.
13737 Apply a 3D LUT to an input video.
13739 The filter accepts the following options:
13743 Set the 3D LUT file name.
13745 Currently supported formats:
13759 Select interpolation mode.
13761 Available values are:
13765 Use values from the nearest defined point.
13767 Interpolate values using the 8 points defining a cube.
13769 Interpolate values using a tetrahedron.
13775 Turn certain luma values into transparency.
13777 The filter accepts the following options:
13781 Set the luma which will be used as base for transparency.
13782 Default value is @code{0}.
13785 Set the range of luma values to be keyed out.
13786 Default value is @code{0.01}.
13789 Set the range of softness. Default value is @code{0}.
13790 Use this to control gradual transition from zero to full transparency.
13793 @subsection Commands
13794 This filter supports same @ref{commands} as options.
13795 The command accepts the same syntax of the corresponding option.
13797 If the specified expression is not valid, it is kept at its current
13800 @section lut, lutrgb, lutyuv
13802 Compute a look-up table for binding each pixel component input value
13803 to an output value, and apply it to the input video.
13805 @var{lutyuv} applies a lookup table to a YUV input video, @var{lutrgb}
13806 to an RGB input video.
13808 These filters accept the following parameters:
13811 set first pixel component expression
13813 set second pixel component expression
13815 set third pixel component expression
13817 set fourth pixel component expression, corresponds to the alpha component
13820 set red component expression
13822 set green component expression
13824 set blue component expression
13826 alpha component expression
13829 set Y/luminance component expression
13831 set U/Cb component expression
13833 set V/Cr component expression
13836 Each of them specifies the expression to use for computing the lookup table for
13837 the corresponding pixel component values.
13839 The exact component associated to each of the @var{c*} options depends on the
13842 The @var{lut} filter requires either YUV or RGB pixel formats in input,
13843 @var{lutrgb} requires RGB pixel formats in input, and @var{lutyuv} requires YUV.
13845 The expressions can contain the following constants and functions:
13850 The input width and height.
13853 The input value for the pixel component.
13856 The input value, clipped to the @var{minval}-@var{maxval} range.
13859 The maximum value for the pixel component.
13862 The minimum value for the pixel component.
13865 The negated value for the pixel component value, clipped to the
13866 @var{minval}-@var{maxval} range; it corresponds to the expression
13867 "maxval-clipval+minval".
13870 The computed value in @var{val}, clipped to the
13871 @var{minval}-@var{maxval} range.
13873 @item gammaval(gamma)
13874 The computed gamma correction value of the pixel component value,
13875 clipped to the @var{minval}-@var{maxval} range. It corresponds to the
13877 "pow((clipval-minval)/(maxval-minval)\,@var{gamma})*(maxval-minval)+minval"
13881 All expressions default to "val".
13883 @subsection Examples
13887 Negate input video:
13889 lutrgb="r=maxval+minval-val:g=maxval+minval-val:b=maxval+minval-val"
13890 lutyuv="y=maxval+minval-val:u=maxval+minval-val:v=maxval+minval-val"
13893 The above is the same as:
13895 lutrgb="r=negval:g=negval:b=negval"
13896 lutyuv="y=negval:u=negval:v=negval"
13906 Remove chroma components, turning the video into a graytone image:
13908 lutyuv="u=128:v=128"
13912 Apply a luma burning effect:
13918 Remove green and blue components:
13924 Set a constant alpha channel value on input:
13926 format=rgba,lutrgb=a="maxval-minval/2"
13930 Correct luminance gamma by a factor of 0.5:
13932 lutyuv=y=gammaval(0.5)
13936 Discard least significant bits of luma:
13938 lutyuv=y='bitand(val, 128+64+32)'
13942 Technicolor like effect:
13944 lutyuv=u='(val-maxval/2)*2+maxval/2':v='(val-maxval/2)*2+maxval/2'
13948 @section lut2, tlut2
13950 The @code{lut2} filter takes two input streams and outputs one
13953 The @code{tlut2} (time lut2) filter takes two consecutive frames
13954 from one single stream.
13956 This filter accepts the following parameters:
13959 set first pixel component expression
13961 set second pixel component expression
13963 set third pixel component expression
13965 set fourth pixel component expression, corresponds to the alpha component
13968 set output bit depth, only available for @code{lut2} filter. By default is 0,
13969 which means bit depth is automatically picked from first input format.
13972 The @code{lut2} filter also supports the @ref{framesync} options.
13974 Each of them specifies the expression to use for computing the lookup table for
13975 the corresponding pixel component values.
13977 The exact component associated to each of the @var{c*} options depends on the
13980 The expressions can contain the following constants:
13985 The input width and height.
13988 The first input value for the pixel component.
13991 The second input value for the pixel component.
13994 The first input video bit depth.
13997 The second input video bit depth.
14000 All expressions default to "x".
14002 @subsection Examples
14006 Highlight differences between two RGB video streams:
14008 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)'
14012 Highlight differences between two YUV video streams:
14014 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)'
14018 Show max difference between two video streams:
14020 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)))'
14024 @section maskedclamp
14026 Clamp the first input stream with the second input and third input stream.
14028 Returns the value of first stream to be between second input
14029 stream - @code{undershoot} and third input stream + @code{overshoot}.
14031 This filter accepts the following options:
14034 Default value is @code{0}.
14037 Default value is @code{0}.
14040 Set which planes will be processed as bitmap, unprocessed planes will be
14041 copied from first stream.
14042 By default value 0xf, all planes will be processed.
14045 @subsection Commands
14047 This filter supports the all above options as @ref{commands}.
14051 Merge the second and third input stream into output stream using absolute differences
14052 between second input stream and first input stream and absolute difference between
14053 third input stream and first input stream. The picked value will be from second input
14054 stream if second absolute difference is greater than first one or from third input stream
14057 This filter accepts the following options:
14060 Set which planes will be processed as bitmap, unprocessed planes will be
14061 copied from first stream.
14062 By default value 0xf, all planes will be processed.
14065 @subsection Commands
14067 This filter supports the all above options as @ref{commands}.
14069 @section maskedmerge
14071 Merge the first input stream with the second input stream using per pixel
14072 weights in the third input stream.
14074 A value of 0 in the third stream pixel component means that pixel component
14075 from first stream is returned unchanged, while maximum value (eg. 255 for
14076 8-bit videos) means that pixel component from second stream is returned
14077 unchanged. Intermediate values define the amount of merging between both
14078 input stream's pixel components.
14080 This filter accepts the following options:
14083 Set which planes will be processed as bitmap, unprocessed planes will be
14084 copied from first stream.
14085 By default value 0xf, all planes will be processed.
14090 Merge the second and third input stream into output stream using absolute differences
14091 between second input stream and first input stream and absolute difference between
14092 third input stream and first input stream. The picked value will be from second input
14093 stream if second absolute difference is less than first one or from third input stream
14096 This filter accepts the following options:
14099 Set which planes will be processed as bitmap, unprocessed planes will be
14100 copied from first stream.
14101 By default value 0xf, all planes will be processed.
14104 @subsection Commands
14106 This filter supports the all above options as @ref{commands}.
14108 @section maskedthreshold
14109 Pick pixels comparing absolute difference of two video streams with fixed
14112 If absolute difference between pixel component of first and second video
14113 stream is equal or lower than user supplied threshold than pixel component
14114 from first video stream is picked, otherwise pixel component from second
14115 video stream is picked.
14117 This filter accepts the following options:
14120 Set threshold used when picking pixels from absolute difference from two input
14124 Set which planes will be processed as bitmap, unprocessed planes will be
14125 copied from second stream.
14126 By default value 0xf, all planes will be processed.
14129 @subsection Commands
14131 This filter supports the all above options as @ref{commands}.
14134 Create mask from input video.
14136 For example it is useful to create motion masks after @code{tblend} filter.
14138 This filter accepts the following options:
14142 Set low threshold. Any pixel component lower or exact than this value will be set to 0.
14145 Set high threshold. Any pixel component higher than this value will be set to max value
14146 allowed for current pixel format.
14149 Set planes to filter, by default all available planes are filtered.
14152 Fill all frame pixels with this value.
14155 Set max average pixel value for frame. If sum of all pixel components is higher that this
14156 average, output frame will be completely filled with value set by @var{fill} option.
14157 Typically useful for scene changes when used in combination with @code{tblend} filter.
14162 Apply motion-compensation deinterlacing.
14164 It needs one field per frame as input and must thus be used together
14165 with yadif=1/3 or equivalent.
14167 This filter accepts the following options:
14170 Set the deinterlacing mode.
14172 It accepts one of the following values:
14177 use iterative motion estimation
14179 like @samp{slow}, but use multiple reference frames.
14181 Default value is @samp{fast}.
14184 Set the picture field parity assumed for the input video. It must be
14185 one of the following values:
14189 assume top field first
14191 assume bottom field first
14194 Default value is @samp{bff}.
14197 Set per-block quantization parameter (QP) used by the internal
14200 Higher values should result in a smoother motion vector field but less
14201 optimal individual vectors. Default value is 1.
14206 Pick median pixel from certain rectangle defined by radius.
14208 This filter accepts the following options:
14212 Set horizontal radius size. Default value is @code{1}.
14213 Allowed range is integer from 1 to 127.
14216 Set which planes to process. Default is @code{15}, which is all available planes.
14219 Set vertical radius size. Default value is @code{0}.
14220 Allowed range is integer from 0 to 127.
14221 If it is 0, value will be picked from horizontal @code{radius} option.
14224 Set median percentile. Default value is @code{0.5}.
14225 Default value of @code{0.5} will pick always median values, while @code{0} will pick
14226 minimum values, and @code{1} maximum values.
14229 @subsection Commands
14230 This filter supports same @ref{commands} as options.
14231 The command accepts the same syntax of the corresponding option.
14233 If the specified expression is not valid, it is kept at its current
14236 @section mergeplanes
14238 Merge color channel components from several video streams.
14240 The filter accepts up to 4 input streams, and merge selected input
14241 planes to the output video.
14243 This filter accepts the following options:
14246 Set input to output plane mapping. Default is @code{0}.
14248 The mappings is specified as a bitmap. It should be specified as a
14249 hexadecimal number in the form 0xAa[Bb[Cc[Dd]]]. 'Aa' describes the
14250 mapping for the first plane of the output stream. 'A' sets the number of
14251 the input stream to use (from 0 to 3), and 'a' the plane number of the
14252 corresponding input to use (from 0 to 3). The rest of the mappings is
14253 similar, 'Bb' describes the mapping for the output stream second
14254 plane, 'Cc' describes the mapping for the output stream third plane and
14255 'Dd' describes the mapping for the output stream fourth plane.
14258 Set output pixel format. Default is @code{yuva444p}.
14261 @subsection Examples
14265 Merge three gray video streams of same width and height into single video stream:
14267 [a0][a1][a2]mergeplanes=0x001020:yuv444p
14271 Merge 1st yuv444p stream and 2nd gray video stream into yuva444p video stream:
14273 [a0][a1]mergeplanes=0x00010210:yuva444p
14277 Swap Y and A plane in yuva444p stream:
14279 format=yuva444p,mergeplanes=0x03010200:yuva444p
14283 Swap U and V plane in yuv420p stream:
14285 format=yuv420p,mergeplanes=0x000201:yuv420p
14289 Cast a rgb24 clip to yuv444p:
14291 format=rgb24,mergeplanes=0x000102:yuv444p
14297 Estimate and export motion vectors using block matching algorithms.
14298 Motion vectors are stored in frame side data to be used by other filters.
14300 This filter accepts the following options:
14303 Specify the motion estimation method. Accepts one of the following values:
14307 Exhaustive search algorithm.
14309 Three step search algorithm.
14311 Two dimensional logarithmic search algorithm.
14313 New three step search algorithm.
14315 Four step search algorithm.
14317 Diamond search algorithm.
14319 Hexagon-based search algorithm.
14321 Enhanced predictive zonal search algorithm.
14323 Uneven multi-hexagon search algorithm.
14325 Default value is @samp{esa}.
14328 Macroblock size. Default @code{16}.
14331 Search parameter. Default @code{7}.
14334 @section midequalizer
14336 Apply Midway Image Equalization effect using two video streams.
14338 Midway Image Equalization adjusts a pair of images to have the same
14339 histogram, while maintaining their dynamics as much as possible. It's
14340 useful for e.g. matching exposures from a pair of stereo cameras.
14342 This filter has two inputs and one output, which must be of same pixel format, but
14343 may be of different sizes. The output of filter is first input adjusted with
14344 midway histogram of both inputs.
14346 This filter accepts the following option:
14350 Set which planes to process. Default is @code{15}, which is all available planes.
14353 @section minterpolate
14355 Convert the video to specified frame rate using motion interpolation.
14357 This filter accepts the following options:
14360 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}.
14363 Motion interpolation mode. Following values are accepted:
14366 Duplicate previous or next frame for interpolating new ones.
14368 Blend source frames. Interpolated frame is mean of previous and next frames.
14370 Motion compensated interpolation. Following options are effective when this mode is selected:
14374 Motion compensation mode. Following values are accepted:
14377 Overlapped block motion compensation.
14379 Adaptive overlapped block motion compensation. Window weighting coefficients are controlled adaptively according to the reliabilities of the neighboring motion vectors to reduce oversmoothing.
14381 Default mode is @samp{obmc}.
14384 Motion estimation mode. Following values are accepted:
14387 Bidirectional motion estimation. Motion vectors are estimated for each source frame in both forward and backward directions.
14389 Bilateral motion estimation. Motion vectors are estimated directly for interpolated frame.
14391 Default mode is @samp{bilat}.
14394 The algorithm to be used for motion estimation. Following values are accepted:
14397 Exhaustive search algorithm.
14399 Three step search algorithm.
14401 Two dimensional logarithmic search algorithm.
14403 New three step search algorithm.
14405 Four step search algorithm.
14407 Diamond search algorithm.
14409 Hexagon-based search algorithm.
14411 Enhanced predictive zonal search algorithm.
14413 Uneven multi-hexagon search algorithm.
14415 Default algorithm is @samp{epzs}.
14418 Macroblock size. Default @code{16}.
14421 Motion estimation search parameter. Default @code{32}.
14424 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).
14429 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:
14432 Disable scene change detection.
14434 Frame difference. Corresponding pixel values are compared and if it satisfies @var{scd_threshold} scene change is detected.
14436 Default method is @samp{fdiff}.
14438 @item scd_threshold
14439 Scene change detection threshold. Default is @code{10.}.
14444 Mix several video input streams into one video stream.
14446 A description of the accepted options follows.
14450 The number of inputs. If unspecified, it defaults to 2.
14453 Specify weight of each input video stream as sequence.
14454 Each weight is separated by space. If number of weights
14455 is smaller than number of @var{frames} last specified
14456 weight will be used for all remaining unset weights.
14459 Specify scale, if it is set it will be multiplied with sum
14460 of each weight multiplied with pixel values to give final destination
14461 pixel value. By default @var{scale} is auto scaled to sum of weights.
14464 Specify how end of stream is determined.
14467 The duration of the longest input. (default)
14470 The duration of the shortest input.
14473 The duration of the first input.
14477 @section mpdecimate
14479 Drop frames that do not differ greatly from the previous frame in
14480 order to reduce frame rate.
14482 The main use of this filter is for very-low-bitrate encoding
14483 (e.g. streaming over dialup modem), but it could in theory be used for
14484 fixing movies that were inverse-telecined incorrectly.
14486 A description of the accepted options follows.
14490 Set the maximum number of consecutive frames which can be dropped (if
14491 positive), or the minimum interval between dropped frames (if
14492 negative). If the value is 0, the frame is dropped disregarding the
14493 number of previous sequentially dropped frames.
14495 Default value is 0.
14500 Set the dropping threshold values.
14502 Values for @option{hi} and @option{lo} are for 8x8 pixel blocks and
14503 represent actual pixel value differences, so a threshold of 64
14504 corresponds to 1 unit of difference for each pixel, or the same spread
14505 out differently over the block.
14507 A frame is a candidate for dropping if no 8x8 blocks differ by more
14508 than a threshold of @option{hi}, and if no more than @option{frac} blocks (1
14509 meaning the whole image) differ by more than a threshold of @option{lo}.
14511 Default value for @option{hi} is 64*12, default value for @option{lo} is
14512 64*5, and default value for @option{frac} is 0.33.
14518 Negate (invert) the input video.
14520 It accepts the following option:
14525 With value 1, it negates the alpha component, if present. Default value is 0.
14531 Denoise frames using Non-Local Means algorithm.
14533 Each pixel is adjusted by looking for other pixels with similar contexts. This
14534 context similarity is defined by comparing their surrounding patches of size
14535 @option{p}x@option{p}. Patches are searched in an area of @option{r}x@option{r}
14538 Note that the research area defines centers for patches, which means some
14539 patches will be made of pixels outside that research area.
14541 The filter accepts the following options.
14545 Set denoising strength. Default is 1.0. Must be in range [1.0, 30.0].
14548 Set patch size. Default is 7. Must be odd number in range [0, 99].
14551 Same as @option{p} but for chroma planes.
14553 The default value is @var{0} and means automatic.
14556 Set research size. Default is 15. Must be odd number in range [0, 99].
14559 Same as @option{r} but for chroma planes.
14561 The default value is @var{0} and means automatic.
14566 Deinterlace video using neural network edge directed interpolation.
14568 This filter accepts the following options:
14572 Mandatory option, without binary file filter can not work.
14573 Currently file can be found here:
14574 https://github.com/dubhater/vapoursynth-nnedi3/blob/master/src/nnedi3_weights.bin
14577 Set which frames to deinterlace, by default it is @code{all}.
14578 Can be @code{all} or @code{interlaced}.
14581 Set mode of operation.
14583 Can be one of the following:
14587 Use frame flags, both fields.
14589 Use frame flags, single field.
14591 Use top field only.
14593 Use bottom field only.
14595 Use both fields, top first.
14597 Use both fields, bottom first.
14601 Set which planes to process, by default filter process all frames.
14604 Set size of local neighborhood around each pixel, used by the predictor neural
14607 Can be one of the following:
14620 Set the number of neurons in predictor neural network.
14621 Can be one of the following:
14632 Controls the number of different neural network predictions that are blended
14633 together to compute the final output value. Can be @code{fast}, default or
14637 Set which set of weights to use in the predictor.
14638 Can be one of the following:
14642 weights trained to minimize absolute error
14644 weights trained to minimize squared error
14648 Controls whether or not the prescreener neural network is used to decide
14649 which pixels should be processed by the predictor neural network and which
14650 can be handled by simple cubic interpolation.
14651 The prescreener is trained to know whether cubic interpolation will be
14652 sufficient for a pixel or whether it should be predicted by the predictor nn.
14653 The computational complexity of the prescreener nn is much less than that of
14654 the predictor nn. Since most pixels can be handled by cubic interpolation,
14655 using the prescreener generally results in much faster processing.
14656 The prescreener is pretty accurate, so the difference between using it and not
14657 using it is almost always unnoticeable.
14659 Can be one of the following:
14667 Default is @code{new}.
14670 Set various debugging flags.
14675 Force libavfilter not to use any of the specified pixel formats for the
14676 input to the next filter.
14678 It accepts the following parameters:
14682 A '|'-separated list of pixel format names, such as
14683 pix_fmts=yuv420p|monow|rgb24".
14687 @subsection Examples
14691 Force libavfilter to use a format different from @var{yuv420p} for the
14692 input to the vflip filter:
14694 noformat=pix_fmts=yuv420p,vflip
14698 Convert the input video to any of the formats not contained in the list:
14700 noformat=yuv420p|yuv444p|yuv410p
14706 Add noise on video input frame.
14708 The filter accepts the following options:
14716 Set noise seed for specific pixel component or all pixel components in case
14717 of @var{all_seed}. Default value is @code{123457}.
14719 @item all_strength, alls
14720 @item c0_strength, c0s
14721 @item c1_strength, c1s
14722 @item c2_strength, c2s
14723 @item c3_strength, c3s
14724 Set noise strength for specific pixel component or all pixel components in case
14725 @var{all_strength}. Default value is @code{0}. Allowed range is [0, 100].
14727 @item all_flags, allf
14728 @item c0_flags, c0f
14729 @item c1_flags, c1f
14730 @item c2_flags, c2f
14731 @item c3_flags, c3f
14732 Set pixel component flags or set flags for all components if @var{all_flags}.
14733 Available values for component flags are:
14736 averaged temporal noise (smoother)
14738 mix random noise with a (semi)regular pattern
14740 temporal noise (noise pattern changes between frames)
14742 uniform noise (gaussian otherwise)
14746 @subsection Examples
14748 Add temporal and uniform noise to input video:
14750 noise=alls=20:allf=t+u
14755 Normalize RGB video (aka histogram stretching, contrast stretching).
14756 See: https://en.wikipedia.org/wiki/Normalization_(image_processing)
14758 For each channel of each frame, the filter computes the input range and maps
14759 it linearly to the user-specified output range. The output range defaults
14760 to the full dynamic range from pure black to pure white.
14762 Temporal smoothing can be used on the input range to reduce flickering (rapid
14763 changes in brightness) caused when small dark or bright objects enter or leave
14764 the scene. This is similar to the auto-exposure (automatic gain control) on a
14765 video camera, and, like a video camera, it may cause a period of over- or
14766 under-exposure of the video.
14768 The R,G,B channels can be normalized independently, which may cause some
14769 color shifting, or linked together as a single channel, which prevents
14770 color shifting. Linked normalization preserves hue. Independent normalization
14771 does not, so it can be used to remove some color casts. Independent and linked
14772 normalization can be combined in any ratio.
14774 The normalize filter accepts the following options:
14779 Colors which define the output range. The minimum input value is mapped to
14780 the @var{blackpt}. The maximum input value is mapped to the @var{whitept}.
14781 The defaults are black and white respectively. Specifying white for
14782 @var{blackpt} and black for @var{whitept} will give color-inverted,
14783 normalized video. Shades of grey can be used to reduce the dynamic range
14784 (contrast). Specifying saturated colors here can create some interesting
14788 The number of previous frames to use for temporal smoothing. The input range
14789 of each channel is smoothed using a rolling average over the current frame
14790 and the @var{smoothing} previous frames. The default is 0 (no temporal
14794 Controls the ratio of independent (color shifting) channel normalization to
14795 linked (color preserving) normalization. 0.0 is fully linked, 1.0 is fully
14796 independent. Defaults to 1.0 (fully independent).
14799 Overall strength of the filter. 1.0 is full strength. 0.0 is a rather
14800 expensive no-op. Defaults to 1.0 (full strength).
14804 @subsection Commands
14805 This filter supports same @ref{commands} as options, excluding @var{smoothing} option.
14806 The command accepts the same syntax of the corresponding option.
14808 If the specified expression is not valid, it is kept at its current
14811 @subsection Examples
14813 Stretch video contrast to use the full dynamic range, with no temporal
14814 smoothing; may flicker depending on the source content:
14816 normalize=blackpt=black:whitept=white:smoothing=0
14819 As above, but with 50 frames of temporal smoothing; flicker should be
14820 reduced, depending on the source content:
14822 normalize=blackpt=black:whitept=white:smoothing=50
14825 As above, but with hue-preserving linked channel normalization:
14827 normalize=blackpt=black:whitept=white:smoothing=50:independence=0
14830 As above, but with half strength:
14832 normalize=blackpt=black:whitept=white:smoothing=50:independence=0:strength=0.5
14835 Map the darkest input color to red, the brightest input color to cyan:
14837 normalize=blackpt=red:whitept=cyan
14842 Pass the video source unchanged to the output.
14845 Optical Character Recognition
14847 This filter uses Tesseract for optical character recognition. To enable
14848 compilation of this filter, you need to configure FFmpeg with
14849 @code{--enable-libtesseract}.
14851 It accepts the following options:
14855 Set datapath to tesseract data. Default is to use whatever was
14856 set at installation.
14859 Set language, default is "eng".
14862 Set character whitelist.
14865 Set character blacklist.
14868 The filter exports recognized text as the frame metadata @code{lavfi.ocr.text}.
14869 The filter exports confidence of recognized words as the frame metadata @code{lavfi.ocr.confidence}.
14873 Apply a video transform using libopencv.
14875 To enable this filter, install the libopencv library and headers and
14876 configure FFmpeg with @code{--enable-libopencv}.
14878 It accepts the following parameters:
14883 The name of the libopencv filter to apply.
14885 @item filter_params
14886 The parameters to pass to the libopencv filter. If not specified, the default
14887 values are assumed.
14891 Refer to the official libopencv documentation for more precise
14893 @url{http://docs.opencv.org/master/modules/imgproc/doc/filtering.html}
14895 Several libopencv filters are supported; see the following subsections.
14900 Dilate an image by using a specific structuring element.
14901 It corresponds to the libopencv function @code{cvDilate}.
14903 It accepts the parameters: @var{struct_el}|@var{nb_iterations}.
14905 @var{struct_el} represents a structuring element, and has the syntax:
14906 @var{cols}x@var{rows}+@var{anchor_x}x@var{anchor_y}/@var{shape}
14908 @var{cols} and @var{rows} represent the number of columns and rows of
14909 the structuring element, @var{anchor_x} and @var{anchor_y} the anchor
14910 point, and @var{shape} the shape for the structuring element. @var{shape}
14911 must be "rect", "cross", "ellipse", or "custom".
14913 If the value for @var{shape} is "custom", it must be followed by a
14914 string of the form "=@var{filename}". The file with name
14915 @var{filename} is assumed to represent a binary image, with each
14916 printable character corresponding to a bright pixel. When a custom
14917 @var{shape} is used, @var{cols} and @var{rows} are ignored, the number
14918 or columns and rows of the read file are assumed instead.
14920 The default value for @var{struct_el} is "3x3+0x0/rect".
14922 @var{nb_iterations} specifies the number of times the transform is
14923 applied to the image, and defaults to 1.
14927 # Use the default values
14930 # Dilate using a structuring element with a 5x5 cross, iterating two times
14931 ocv=filter_name=dilate:filter_params=5x5+2x2/cross|2
14933 # Read the shape from the file diamond.shape, iterating two times.
14934 # The file diamond.shape may contain a pattern of characters like this
14940 # The specified columns and rows are ignored
14941 # but the anchor point coordinates are not
14942 ocv=dilate:0x0+2x2/custom=diamond.shape|2
14947 Erode an image by using a specific structuring element.
14948 It corresponds to the libopencv function @code{cvErode}.
14950 It accepts the parameters: @var{struct_el}:@var{nb_iterations},
14951 with the same syntax and semantics as the @ref{dilate} filter.
14955 Smooth the input video.
14957 The filter takes the following parameters:
14958 @var{type}|@var{param1}|@var{param2}|@var{param3}|@var{param4}.
14960 @var{type} is the type of smooth filter to apply, and must be one of
14961 the following values: "blur", "blur_no_scale", "median", "gaussian",
14962 or "bilateral". The default value is "gaussian".
14964 The meaning of @var{param1}, @var{param2}, @var{param3}, and @var{param4}
14965 depends on the smooth type. @var{param1} and
14966 @var{param2} accept integer positive values or 0. @var{param3} and
14967 @var{param4} accept floating point values.
14969 The default value for @var{param1} is 3. The default value for the
14970 other parameters is 0.
14972 These parameters correspond to the parameters assigned to the
14973 libopencv function @code{cvSmooth}.
14975 @section oscilloscope
14977 2D Video Oscilloscope.
14979 Useful to measure spatial impulse, step responses, chroma delays, etc.
14981 It accepts the following parameters:
14985 Set scope center x position.
14988 Set scope center y position.
14991 Set scope size, relative to frame diagonal.
14994 Set scope tilt/rotation.
15000 Set trace center x position.
15003 Set trace center y position.
15006 Set trace width, relative to width of frame.
15009 Set trace height, relative to height of frame.
15012 Set which components to trace. By default it traces first three components.
15015 Draw trace grid. By default is enabled.
15018 Draw some statistics. By default is enabled.
15021 Draw scope. By default is enabled.
15024 @subsection Commands
15025 This filter supports same @ref{commands} as options.
15026 The command accepts the same syntax of the corresponding option.
15028 If the specified expression is not valid, it is kept at its current
15031 @subsection Examples
15035 Inspect full first row of video frame.
15037 oscilloscope=x=0.5:y=0:s=1
15041 Inspect full last row of video frame.
15043 oscilloscope=x=0.5:y=1:s=1
15047 Inspect full 5th line of video frame of height 1080.
15049 oscilloscope=x=0.5:y=5/1080:s=1
15053 Inspect full last column of video frame.
15055 oscilloscope=x=1:y=0.5:s=1:t=1
15063 Overlay one video on top of another.
15065 It takes two inputs and has one output. The first input is the "main"
15066 video on which the second input is overlaid.
15068 It accepts the following parameters:
15070 A description of the accepted options follows.
15075 Set the expression for the x and y coordinates of the overlaid video
15076 on the main video. Default value is "0" for both expressions. In case
15077 the expression is invalid, it is set to a huge value (meaning that the
15078 overlay will not be displayed within the output visible area).
15081 See @ref{framesync}.
15084 Set when the expressions for @option{x}, and @option{y} are evaluated.
15086 It accepts the following values:
15089 only evaluate expressions once during the filter initialization or
15090 when a command is processed
15093 evaluate expressions for each incoming frame
15096 Default value is @samp{frame}.
15099 See @ref{framesync}.
15102 Set the format for the output video.
15104 It accepts the following values:
15107 force YUV420 output
15110 force YUV420p10 output
15113 force YUV422 output
15116 force YUV422p10 output
15119 force YUV444 output
15122 force packed RGB output
15125 force planar RGB output
15128 automatically pick format
15131 Default value is @samp{yuv420}.
15134 See @ref{framesync}.
15137 Set format of alpha of the overlaid video, it can be @var{straight} or
15138 @var{premultiplied}. Default is @var{straight}.
15141 The @option{x}, and @option{y} expressions can contain the following
15147 The main input width and height.
15151 The overlay input width and height.
15155 The computed values for @var{x} and @var{y}. They are evaluated for
15160 horizontal and vertical chroma subsample values of the output
15161 format. For example for the pixel format "yuv422p" @var{hsub} is 2 and
15165 the number of input frame, starting from 0
15168 the position in the file of the input frame, NAN if unknown
15171 The timestamp, expressed in seconds. It's NAN if the input timestamp is unknown.
15175 This filter also supports the @ref{framesync} options.
15177 Note that the @var{n}, @var{pos}, @var{t} variables are available only
15178 when evaluation is done @emph{per frame}, and will evaluate to NAN
15179 when @option{eval} is set to @samp{init}.
15181 Be aware that frames are taken from each input video in timestamp
15182 order, hence, if their initial timestamps differ, it is a good idea
15183 to pass the two inputs through a @var{setpts=PTS-STARTPTS} filter to
15184 have them begin in the same zero timestamp, as the example for
15185 the @var{movie} filter does.
15187 You can chain together more overlays but you should test the
15188 efficiency of such approach.
15190 @subsection Commands
15192 This filter supports the following commands:
15196 Modify the x and y of the overlay input.
15197 The command accepts the same syntax of the corresponding option.
15199 If the specified expression is not valid, it is kept at its current
15203 @subsection Examples
15207 Draw the overlay at 10 pixels from the bottom right corner of the main
15210 overlay=main_w-overlay_w-10:main_h-overlay_h-10
15213 Using named options the example above becomes:
15215 overlay=x=main_w-overlay_w-10:y=main_h-overlay_h-10
15219 Insert a transparent PNG logo in the bottom left corner of the input,
15220 using the @command{ffmpeg} tool with the @code{-filter_complex} option:
15222 ffmpeg -i input -i logo -filter_complex 'overlay=10:main_h-overlay_h-10' output
15226 Insert 2 different transparent PNG logos (second logo on bottom
15227 right corner) using the @command{ffmpeg} tool:
15229 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
15233 Add a transparent color layer on top of the main video; @code{WxH}
15234 must specify the size of the main input to the overlay filter:
15236 color=color=red@@.3:size=WxH [over]; [in][over] overlay [out]
15240 Play an original video and a filtered version (here with the deshake
15241 filter) side by side using the @command{ffplay} tool:
15243 ffplay input.avi -vf 'split[a][b]; [a]pad=iw*2:ih[src]; [b]deshake[filt]; [src][filt]overlay=w'
15246 The above command is the same as:
15248 ffplay input.avi -vf 'split[b], pad=iw*2[src], [b]deshake, [src]overlay=w'
15252 Make a sliding overlay appearing from the left to the right top part of the
15253 screen starting since time 2:
15255 overlay=x='if(gte(t,2), -w+(t-2)*20, NAN)':y=0
15259 Compose output by putting two input videos side to side:
15261 ffmpeg -i left.avi -i right.avi -filter_complex "
15262 nullsrc=size=200x100 [background];
15263 [0:v] setpts=PTS-STARTPTS, scale=100x100 [left];
15264 [1:v] setpts=PTS-STARTPTS, scale=100x100 [right];
15265 [background][left] overlay=shortest=1 [background+left];
15266 [background+left][right] overlay=shortest=1:x=100 [left+right]
15271 Mask 10-20 seconds of a video by applying the delogo filter to a section
15273 ffmpeg -i test.avi -codec:v:0 wmv2 -ar 11025 -b:v 9000k
15274 -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]'
15279 Chain several overlays in cascade:
15281 nullsrc=s=200x200 [bg];
15282 testsrc=s=100x100, split=4 [in0][in1][in2][in3];
15283 [in0] lutrgb=r=0, [bg] overlay=0:0 [mid0];
15284 [in1] lutrgb=g=0, [mid0] overlay=100:0 [mid1];
15285 [in2] lutrgb=b=0, [mid1] overlay=0:100 [mid2];
15286 [in3] null, [mid2] overlay=100:100 [out0]
15291 @anchor{overlay_cuda}
15292 @section overlay_cuda
15294 Overlay one video on top of another.
15296 This is the CUDA variant of the @ref{overlay} filter.
15297 It only accepts CUDA frames. The underlying input pixel formats have to match.
15299 It takes two inputs and has one output. The first input is the "main"
15300 video on which the second input is overlaid.
15302 It accepts the following parameters:
15307 Set the x and y coordinates of the overlaid video on the main video.
15308 Default value is "0" for both expressions.
15311 See @ref{framesync}.
15314 See @ref{framesync}.
15317 See @ref{framesync}.
15321 This filter also supports the @ref{framesync} options.
15325 Apply Overcomplete Wavelet denoiser.
15327 The filter accepts the following options:
15333 Larger depth values will denoise lower frequency components more, but
15334 slow down filtering.
15336 Must be an int in the range 8-16, default is @code{8}.
15338 @item luma_strength, ls
15341 Must be a double value in the range 0-1000, default is @code{1.0}.
15343 @item chroma_strength, cs
15344 Set chroma strength.
15346 Must be a double value in the range 0-1000, default is @code{1.0}.
15352 Add paddings to the input image, and place the original input at the
15353 provided @var{x}, @var{y} coordinates.
15355 It accepts the following parameters:
15360 Specify an expression for the size of the output image with the
15361 paddings added. If the value for @var{width} or @var{height} is 0, the
15362 corresponding input size is used for the output.
15364 The @var{width} expression can reference the value set by the
15365 @var{height} expression, and vice versa.
15367 The default value of @var{width} and @var{height} is 0.
15371 Specify the offsets to place the input image at within the padded area,
15372 with respect to the top/left border of the output image.
15374 The @var{x} expression can reference the value set by the @var{y}
15375 expression, and vice versa.
15377 The default value of @var{x} and @var{y} is 0.
15379 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
15380 so the input image is centered on the padded area.
15383 Specify the color of the padded area. For the syntax of this option,
15384 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
15385 manual,ffmpeg-utils}.
15387 The default value of @var{color} is "black".
15390 Specify when to evaluate @var{width}, @var{height}, @var{x} and @var{y} expression.
15392 It accepts the following values:
15396 Only evaluate expressions once during the filter initialization or when
15397 a command is processed.
15400 Evaluate expressions for each incoming frame.
15404 Default value is @samp{init}.
15407 Pad to aspect instead to a resolution.
15411 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
15412 options are expressions containing the following constants:
15417 The input video width and height.
15421 These are the same as @var{in_w} and @var{in_h}.
15425 The output width and height (the size of the padded area), as
15426 specified by the @var{width} and @var{height} expressions.
15430 These are the same as @var{out_w} and @var{out_h}.
15434 The x and y offsets as specified by the @var{x} and @var{y}
15435 expressions, or NAN if not yet specified.
15438 same as @var{iw} / @var{ih}
15441 input sample aspect ratio
15444 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
15448 The horizontal and vertical chroma subsample values. For example for the
15449 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
15452 @subsection Examples
15456 Add paddings with the color "violet" to the input video. The output video
15457 size is 640x480, and the top-left corner of the input video is placed at
15460 pad=640:480:0:40:violet
15463 The example above is equivalent to the following command:
15465 pad=width=640:height=480:x=0:y=40:color=violet
15469 Pad the input to get an output with dimensions increased by 3/2,
15470 and put the input video at the center of the padded area:
15472 pad="3/2*iw:3/2*ih:(ow-iw)/2:(oh-ih)/2"
15476 Pad the input to get a squared output with size equal to the maximum
15477 value between the input width and height, and put the input video at
15478 the center of the padded area:
15480 pad="max(iw\,ih):ow:(ow-iw)/2:(oh-ih)/2"
15484 Pad the input to get a final w/h ratio of 16:9:
15486 pad="ih*16/9:ih:(ow-iw)/2:(oh-ih)/2"
15490 In case of anamorphic video, in order to set the output display aspect
15491 correctly, it is necessary to use @var{sar} in the expression,
15492 according to the relation:
15494 (ih * X / ih) * sar = output_dar
15495 X = output_dar / sar
15498 Thus the previous example needs to be modified to:
15500 pad="ih*16/9/sar:ih:(ow-iw)/2:(oh-ih)/2"
15504 Double the output size and put the input video in the bottom-right
15505 corner of the output padded area:
15507 pad="2*iw:2*ih:ow-iw:oh-ih"
15511 @anchor{palettegen}
15512 @section palettegen
15514 Generate one palette for a whole video stream.
15516 It accepts the following options:
15520 Set the maximum number of colors to quantize in the palette.
15521 Note: the palette will still contain 256 colors; the unused palette entries
15524 @item reserve_transparent
15525 Create a palette of 255 colors maximum and reserve the last one for
15526 transparency. Reserving the transparency color is useful for GIF optimization.
15527 If not set, the maximum of colors in the palette will be 256. You probably want
15528 to disable this option for a standalone image.
15531 @item transparency_color
15532 Set the color that will be used as background for transparency.
15535 Set statistics mode.
15537 It accepts the following values:
15540 Compute full frame histograms.
15542 Compute histograms only for the part that differs from previous frame. This
15543 might be relevant to give more importance to the moving part of your input if
15544 the background is static.
15546 Compute new histogram for each frame.
15549 Default value is @var{full}.
15552 The filter also exports the frame metadata @code{lavfi.color_quant_ratio}
15553 (@code{nb_color_in / nb_color_out}) which you can use to evaluate the degree of
15554 color quantization of the palette. This information is also visible at
15555 @var{info} logging level.
15557 @subsection Examples
15561 Generate a representative palette of a given video using @command{ffmpeg}:
15563 ffmpeg -i input.mkv -vf palettegen palette.png
15567 @section paletteuse
15569 Use a palette to downsample an input video stream.
15571 The filter takes two inputs: one video stream and a palette. The palette must
15572 be a 256 pixels image.
15574 It accepts the following options:
15578 Select dithering mode. Available algorithms are:
15581 Ordered 8x8 bayer dithering (deterministic)
15583 Dithering as defined by Paul Heckbert in 1982 (simple error diffusion).
15584 Note: this dithering is sometimes considered "wrong" and is included as a
15586 @item floyd_steinberg
15587 Floyd and Steingberg dithering (error diffusion)
15589 Frankie Sierra dithering v2 (error diffusion)
15591 Frankie Sierra dithering v2 "Lite" (error diffusion)
15594 Default is @var{sierra2_4a}.
15597 When @var{bayer} dithering is selected, this option defines the scale of the
15598 pattern (how much the crosshatch pattern is visible). A low value means more
15599 visible pattern for less banding, and higher value means less visible pattern
15600 at the cost of more banding.
15602 The option must be an integer value in the range [0,5]. Default is @var{2}.
15605 If set, define the zone to process
15609 Only the changing rectangle will be reprocessed. This is similar to GIF
15610 cropping/offsetting compression mechanism. This option can be useful for speed
15611 if only a part of the image is changing, and has use cases such as limiting the
15612 scope of the error diffusal @option{dither} to the rectangle that bounds the
15613 moving scene (it leads to more deterministic output if the scene doesn't change
15614 much, and as a result less moving noise and better GIF compression).
15617 Default is @var{none}.
15620 Take new palette for each output frame.
15622 @item alpha_threshold
15623 Sets the alpha threshold for transparency. Alpha values above this threshold
15624 will be treated as completely opaque, and values below this threshold will be
15625 treated as completely transparent.
15627 The option must be an integer value in the range [0,255]. Default is @var{128}.
15630 @subsection Examples
15634 Use a palette (generated for example with @ref{palettegen}) to encode a GIF
15635 using @command{ffmpeg}:
15637 ffmpeg -i input.mkv -i palette.png -lavfi paletteuse output.gif
15641 @section perspective
15643 Correct perspective of video not recorded perpendicular to the screen.
15645 A description of the accepted parameters follows.
15656 Set coordinates expression for top left, top right, bottom left and bottom right corners.
15657 Default values are @code{0:0:W:0:0:H:W:H} with which perspective will remain unchanged.
15658 If the @code{sense} option is set to @code{source}, then the specified points will be sent
15659 to the corners of the destination. If the @code{sense} option is set to @code{destination},
15660 then the corners of the source will be sent to the specified coordinates.
15662 The expressions can use the following variables:
15667 the width and height of video frame.
15671 Output frame count.
15674 @item interpolation
15675 Set interpolation for perspective correction.
15677 It accepts the following values:
15683 Default value is @samp{linear}.
15686 Set interpretation of coordinate options.
15688 It accepts the following values:
15692 Send point in the source specified by the given coordinates to
15693 the corners of the destination.
15695 @item 1, destination
15697 Send the corners of the source to the point in the destination specified
15698 by the given coordinates.
15700 Default value is @samp{source}.
15704 Set when the expressions for coordinates @option{x0,y0,...x3,y3} are evaluated.
15706 It accepts the following values:
15709 only evaluate expressions once during the filter initialization or
15710 when a command is processed
15713 evaluate expressions for each incoming frame
15716 Default value is @samp{init}.
15721 Delay interlaced video by one field time so that the field order changes.
15723 The intended use is to fix PAL movies that have been captured with the
15724 opposite field order to the film-to-video transfer.
15726 A description of the accepted parameters follows.
15732 It accepts the following values:
15735 Capture field order top-first, transfer bottom-first.
15736 Filter will delay the bottom field.
15739 Capture field order bottom-first, transfer top-first.
15740 Filter will delay the top field.
15743 Capture and transfer with the same field order. This mode only exists
15744 for the documentation of the other options to refer to, but if you
15745 actually select it, the filter will faithfully do nothing.
15748 Capture field order determined automatically by field flags, transfer
15750 Filter selects among @samp{t} and @samp{b} modes on a frame by frame
15751 basis using field flags. If no field information is available,
15752 then this works just like @samp{u}.
15755 Capture unknown or varying, transfer opposite.
15756 Filter selects among @samp{t} and @samp{b} on a frame by frame basis by
15757 analyzing the images and selecting the alternative that produces best
15758 match between the fields.
15761 Capture top-first, transfer unknown or varying.
15762 Filter selects among @samp{t} and @samp{p} using image analysis.
15765 Capture bottom-first, transfer unknown or varying.
15766 Filter selects among @samp{b} and @samp{p} using image analysis.
15769 Capture determined by field flags, transfer unknown or varying.
15770 Filter selects among @samp{t}, @samp{b} and @samp{p} using field flags and
15771 image analysis. If no field information is available, then this works just
15772 like @samp{U}. This is the default mode.
15775 Both capture and transfer unknown or varying.
15776 Filter selects among @samp{t}, @samp{b} and @samp{p} using image analysis only.
15780 @section photosensitivity
15781 Reduce various flashes in video, so to help users with epilepsy.
15783 It accepts the following options:
15786 Set how many frames to use when filtering. Default is 30.
15789 Set detection threshold factor. Default is 1.
15793 Set how many pixels to skip when sampling frames. Default is 1.
15794 Allowed range is from 1 to 1024.
15797 Leave frames unchanged. Default is disabled.
15800 @section pixdesctest
15802 Pixel format descriptor test filter, mainly useful for internal
15803 testing. The output video should be equal to the input video.
15807 format=monow, pixdesctest
15810 can be used to test the monowhite pixel format descriptor definition.
15814 Display sample values of color channels. Mainly useful for checking color
15815 and levels. Minimum supported resolution is 640x480.
15817 The filters accept the following options:
15821 Set scope X position, relative offset on X axis.
15824 Set scope Y position, relative offset on Y axis.
15833 Set window opacity. This window also holds statistics about pixel area.
15836 Set window X position, relative offset on X axis.
15839 Set window Y position, relative offset on Y axis.
15844 Enable the specified chain of postprocessing subfilters using libpostproc. This
15845 library should be automatically selected with a GPL build (@code{--enable-gpl}).
15846 Subfilters must be separated by '/' and can be disabled by prepending a '-'.
15847 Each subfilter and some options have a short and a long name that can be used
15848 interchangeably, i.e. dr/dering are the same.
15850 The filters accept the following options:
15854 Set postprocessing subfilters string.
15857 All subfilters share common options to determine their scope:
15861 Honor the quality commands for this subfilter.
15864 Do chrominance filtering, too (default).
15867 Do luminance filtering only (no chrominance).
15870 Do chrominance filtering only (no luminance).
15873 These options can be appended after the subfilter name, separated by a '|'.
15875 Available subfilters are:
15878 @item hb/hdeblock[|difference[|flatness]]
15879 Horizontal deblocking filter
15882 Difference factor where higher values mean more deblocking (default: @code{32}).
15884 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15887 @item vb/vdeblock[|difference[|flatness]]
15888 Vertical deblocking filter
15891 Difference factor where higher values mean more deblocking (default: @code{32}).
15893 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15896 @item ha/hadeblock[|difference[|flatness]]
15897 Accurate horizontal deblocking filter
15900 Difference factor where higher values mean more deblocking (default: @code{32}).
15902 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15905 @item va/vadeblock[|difference[|flatness]]
15906 Accurate vertical deblocking filter
15909 Difference factor where higher values mean more deblocking (default: @code{32}).
15911 Flatness threshold where lower values mean more deblocking (default: @code{39}).
15915 The horizontal and vertical deblocking filters share the difference and
15916 flatness values so you cannot set different horizontal and vertical
15920 @item h1/x1hdeblock
15921 Experimental horizontal deblocking filter
15923 @item v1/x1vdeblock
15924 Experimental vertical deblocking filter
15929 @item tn/tmpnoise[|threshold1[|threshold2[|threshold3]]], temporal noise reducer
15932 larger -> stronger filtering
15934 larger -> stronger filtering
15936 larger -> stronger filtering
15939 @item al/autolevels[:f/fullyrange], automatic brightness / contrast correction
15942 Stretch luminance to @code{0-255}.
15945 @item lb/linblenddeint
15946 Linear blend deinterlacing filter that deinterlaces the given block by
15947 filtering all lines with a @code{(1 2 1)} filter.
15949 @item li/linipoldeint
15950 Linear interpolating deinterlacing filter that deinterlaces the given block by
15951 linearly interpolating every second line.
15953 @item ci/cubicipoldeint
15954 Cubic interpolating deinterlacing filter deinterlaces the given block by
15955 cubically interpolating every second line.
15957 @item md/mediandeint
15958 Median deinterlacing filter that deinterlaces the given block by applying a
15959 median filter to every second line.
15961 @item fd/ffmpegdeint
15962 FFmpeg deinterlacing filter that deinterlaces the given block by filtering every
15963 second line with a @code{(-1 4 2 4 -1)} filter.
15966 Vertically applied FIR lowpass deinterlacing filter that deinterlaces the given
15967 block by filtering all lines with a @code{(-1 2 6 2 -1)} filter.
15969 @item fq/forceQuant[|quantizer]
15970 Overrides the quantizer table from the input with the constant quantizer you
15978 Default pp filter combination (@code{hb|a,vb|a,dr|a})
15981 Fast pp filter combination (@code{h1|a,v1|a,dr|a})
15984 High quality pp filter combination (@code{ha|a|128|7,va|a,dr|a})
15987 @subsection Examples
15991 Apply horizontal and vertical deblocking, deringing and automatic
15992 brightness/contrast:
15998 Apply default filters without brightness/contrast correction:
16004 Apply default filters and temporal denoiser:
16006 pp=default/tmpnoise|1|2|3
16010 Apply deblocking on luminance only, and switch vertical deblocking on or off
16011 automatically depending on available CPU time:
16018 Apply Postprocessing filter 7. It is variant of the @ref{spp} filter,
16019 similar to spp = 6 with 7 point DCT, where only the center sample is
16022 The filter accepts the following options:
16026 Force a constant quantization parameter. It accepts an integer in range
16027 0 to 63. If not set, the filter will use the QP from the video stream
16031 Set thresholding mode. Available modes are:
16035 Set hard thresholding.
16037 Set soft thresholding (better de-ringing effect, but likely blurrier).
16039 Set medium thresholding (good results, default).
16043 @section premultiply
16044 Apply alpha premultiply effect to input video stream using first plane
16045 of second stream as alpha.
16047 Both streams must have same dimensions and same pixel format.
16049 The filter accepts the following option:
16053 Set which planes will be processed, unprocessed planes will be copied.
16054 By default value 0xf, all planes will be processed.
16057 Do not require 2nd input for processing, instead use alpha plane from input stream.
16061 Apply prewitt operator to input video stream.
16063 The filter accepts the following option:
16067 Set which planes will be processed, unprocessed planes will be copied.
16068 By default value 0xf, all planes will be processed.
16071 Set value which will be multiplied with filtered result.
16074 Set value which will be added to filtered result.
16077 @subsection Commands
16079 This filter supports the all above options as @ref{commands}.
16081 @section pseudocolor
16083 Alter frame colors in video with pseudocolors.
16085 This filter accepts the following options:
16089 set pixel first component expression
16092 set pixel second component expression
16095 set pixel third component expression
16098 set pixel fourth component expression, corresponds to the alpha component
16101 set component to use as base for altering colors
16104 Each of them specifies the expression to use for computing the lookup table for
16105 the corresponding pixel component values.
16107 The expressions can contain the following constants and functions:
16112 The input width and height.
16115 The input value for the pixel component.
16117 @item ymin, umin, vmin, amin
16118 The minimum allowed component value.
16120 @item ymax, umax, vmax, amax
16121 The maximum allowed component value.
16124 All expressions default to "val".
16126 @subsection Examples
16130 Change too high luma values to gradient:
16132 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'"
16138 Obtain the average, maximum and minimum PSNR (Peak Signal to Noise
16139 Ratio) between two input videos.
16141 This filter takes in input two input videos, the first input is
16142 considered the "main" source and is passed unchanged to the
16143 output. The second input is used as a "reference" video for computing
16146 Both video inputs must have the same resolution and pixel format for
16147 this filter to work correctly. Also it assumes that both inputs
16148 have the same number of frames, which are compared one by one.
16150 The obtained average PSNR is printed through the logging system.
16152 The filter stores the accumulated MSE (mean squared error) of each
16153 frame, and at the end of the processing it is averaged across all frames
16154 equally, and the following formula is applied to obtain the PSNR:
16157 PSNR = 10*log10(MAX^2/MSE)
16160 Where MAX is the average of the maximum values of each component of the
16163 The description of the accepted parameters follows.
16166 @item stats_file, f
16167 If specified the filter will use the named file to save the PSNR of
16168 each individual frame. When filename equals "-" the data is sent to
16171 @item stats_version
16172 Specifies which version of the stats file format to use. Details of
16173 each format are written below.
16174 Default value is 1.
16176 @item stats_add_max
16177 Determines whether the max value is output to the stats log.
16178 Default value is 0.
16179 Requires stats_version >= 2. If this is set and stats_version < 2,
16180 the filter will return an error.
16183 This filter also supports the @ref{framesync} options.
16185 The file printed if @var{stats_file} is selected, contains a sequence of
16186 key/value pairs of the form @var{key}:@var{value} for each compared
16189 If a @var{stats_version} greater than 1 is specified, a header line precedes
16190 the list of per-frame-pair stats, with key value pairs following the frame
16191 format with the following parameters:
16194 @item psnr_log_version
16195 The version of the log file format. Will match @var{stats_version}.
16198 A comma separated list of the per-frame-pair parameters included in
16202 A description of each shown per-frame-pair parameter follows:
16206 sequential number of the input frame, starting from 1
16209 Mean Square Error pixel-by-pixel average difference of the compared
16210 frames, averaged over all the image components.
16212 @item mse_y, mse_u, mse_v, mse_r, mse_g, mse_b, mse_a
16213 Mean Square Error pixel-by-pixel average difference of the compared
16214 frames for the component specified by the suffix.
16216 @item psnr_y, psnr_u, psnr_v, psnr_r, psnr_g, psnr_b, psnr_a
16217 Peak Signal to Noise ratio of the compared frames for the component
16218 specified by the suffix.
16220 @item max_avg, max_y, max_u, max_v
16221 Maximum allowed value for each channel, and average over all
16225 @subsection Examples
16230 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
16231 [main][ref] psnr="stats_file=stats.log" [out]
16234 On this example the input file being processed is compared with the
16235 reference file @file{ref_movie.mpg}. The PSNR of each individual frame
16236 is stored in @file{stats.log}.
16239 Another example with different containers:
16241 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 -
16248 Pulldown reversal (inverse telecine) filter, capable of handling mixed
16249 hard-telecine, 24000/1001 fps progressive, and 30000/1001 fps progressive
16252 The pullup filter is designed to take advantage of future context in making
16253 its decisions. This filter is stateless in the sense that it does not lock
16254 onto a pattern to follow, but it instead looks forward to the following
16255 fields in order to identify matches and rebuild progressive frames.
16257 To produce content with an even framerate, insert the fps filter after
16258 pullup, use @code{fps=24000/1001} if the input frame rate is 29.97fps,
16259 @code{fps=24} for 30fps and the (rare) telecined 25fps input.
16261 The filter accepts the following options:
16268 These options set the amount of "junk" to ignore at the left, right, top, and
16269 bottom of the image, respectively. Left and right are in units of 8 pixels,
16270 while top and bottom are in units of 2 lines.
16271 The default is 8 pixels on each side.
16274 Set the strict breaks. Setting this option to 1 will reduce the chances of
16275 filter generating an occasional mismatched frame, but it may also cause an
16276 excessive number of frames to be dropped during high motion sequences.
16277 Conversely, setting it to -1 will make filter match fields more easily.
16278 This may help processing of video where there is slight blurring between
16279 the fields, but may also cause there to be interlaced frames in the output.
16280 Default value is @code{0}.
16283 Set the metric plane to use. It accepts the following values:
16289 Use chroma blue plane.
16292 Use chroma red plane.
16295 This option may be set to use chroma plane instead of the default luma plane
16296 for doing filter's computations. This may improve accuracy on very clean
16297 source material, but more likely will decrease accuracy, especially if there
16298 is chroma noise (rainbow effect) or any grayscale video.
16299 The main purpose of setting @option{mp} to a chroma plane is to reduce CPU
16300 load and make pullup usable in realtime on slow machines.
16303 For best results (without duplicated frames in the output file) it is
16304 necessary to change the output frame rate. For example, to inverse
16305 telecine NTSC input:
16307 ffmpeg -i input -vf pullup -r 24000/1001 ...
16312 Change video quantization parameters (QP).
16314 The filter accepts the following option:
16318 Set expression for quantization parameter.
16321 The expression is evaluated through the eval API and can contain, among others,
16322 the following constants:
16326 1 if index is not 129, 0 otherwise.
16329 Sequential index starting from -129 to 128.
16332 @subsection Examples
16336 Some equation like:
16344 Flush video frames from internal cache of frames into a random order.
16345 No frame is discarded.
16346 Inspired by @ref{frei0r} nervous filter.
16350 Set size in number of frames of internal cache, in range from @code{2} to
16351 @code{512}. Default is @code{30}.
16354 Set seed for random number generator, must be an integer included between
16355 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
16356 less than @code{0}, the filter will try to use a good random seed on a
16360 @section readeia608
16362 Read closed captioning (EIA-608) information from the top lines of a video frame.
16364 This filter adds frame metadata for @code{lavfi.readeia608.X.cc} and
16365 @code{lavfi.readeia608.X.line}, where @code{X} is the number of the identified line
16366 with EIA-608 data (starting from 0). A description of each metadata value follows:
16369 @item lavfi.readeia608.X.cc
16370 The two bytes stored as EIA-608 data (printed in hexadecimal).
16372 @item lavfi.readeia608.X.line
16373 The number of the line on which the EIA-608 data was identified and read.
16376 This filter accepts the following options:
16380 Set the line to start scanning for EIA-608 data. Default is @code{0}.
16383 Set the line to end scanning for EIA-608 data. Default is @code{29}.
16386 Set the ratio of width reserved for sync code detection.
16387 Default is @code{0.27}. Allowed range is @code{[0.1 - 0.7]}.
16390 Enable checking the parity bit. In the event of a parity error, the filter will output
16391 @code{0x00} for that character. Default is false.
16394 Lowpass lines prior to further processing. Default is enabled.
16397 @subsection Commands
16399 This filter supports the all above options as @ref{commands}.
16401 @subsection Examples
16405 Output a csv with presentation time and the first two lines of identified EIA-608 captioning data.
16407 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
16413 Read vertical interval timecode (VITC) information from the top lines of a
16416 The filter adds frame metadata key @code{lavfi.readvitc.tc_str} with the
16417 timecode value, if a valid timecode has been detected. Further metadata key
16418 @code{lavfi.readvitc.found} is set to 0/1 depending on whether
16419 timecode data has been found or not.
16421 This filter accepts the following options:
16425 Set the maximum number of lines to scan for VITC data. If the value is set to
16426 @code{-1} the full video frame is scanned. Default is @code{45}.
16429 Set the luma threshold for black. Accepts float numbers in the range [0.0,1.0],
16430 default value is @code{0.2}. The value must be equal or less than @code{thr_w}.
16433 Set the luma threshold for white. Accepts float numbers in the range [0.0,1.0],
16434 default value is @code{0.6}. The value must be equal or greater than @code{thr_b}.
16437 @subsection Examples
16441 Detect and draw VITC data onto the video frame; if no valid VITC is detected,
16442 draw @code{--:--:--:--} as a placeholder:
16444 ffmpeg -i input.avi -filter:v 'readvitc,drawtext=fontfile=FreeMono.ttf:text=%@{metadata\\:lavfi.readvitc.tc_str\\:--\\\\\\:--\\\\\\:--\\\\\\:--@}:x=(w-tw)/2:y=400-ascent'
16450 Remap pixels using 2nd: Xmap and 3rd: Ymap input video stream.
16452 Destination pixel at position (X, Y) will be picked from source (x, y) position
16453 where x = Xmap(X, Y) and y = Ymap(X, Y). If mapping values are out of range, zero
16454 value for pixel will be used for destination pixel.
16456 Xmap and Ymap input video streams must be of same dimensions. Output video stream
16457 will have Xmap/Ymap video stream dimensions.
16458 Xmap and Ymap input video streams are 16bit depth, single channel.
16462 Specify pixel format of output from this filter. Can be @code{color} or @code{gray}.
16463 Default is @code{color}.
16466 Specify the color of the unmapped pixels. For the syntax of this option,
16467 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
16468 manual,ffmpeg-utils}. Default color is @code{black}.
16471 @section removegrain
16473 The removegrain filter is a spatial denoiser for progressive video.
16477 Set mode for the first plane.
16480 Set mode for the second plane.
16483 Set mode for the third plane.
16486 Set mode for the fourth plane.
16489 Range of mode is from 0 to 24. Description of each mode follows:
16493 Leave input plane unchanged. Default.
16496 Clips the pixel with the minimum and maximum of the 8 neighbour pixels.
16499 Clips the pixel with the second minimum and maximum of the 8 neighbour pixels.
16502 Clips the pixel with the third minimum and maximum of the 8 neighbour pixels.
16505 Clips the pixel with the fourth minimum and maximum of the 8 neighbour pixels.
16506 This is equivalent to a median filter.
16509 Line-sensitive clipping giving the minimal change.
16512 Line-sensitive clipping, intermediate.
16515 Line-sensitive clipping, intermediate.
16518 Line-sensitive clipping, intermediate.
16521 Line-sensitive clipping on a line where the neighbours pixels are the closest.
16524 Replaces the target pixel with the closest neighbour.
16527 [1 2 1] horizontal and vertical kernel blur.
16533 Bob mode, interpolates top field from the line where the neighbours
16534 pixels are the closest.
16537 Bob mode, interpolates bottom field from the line where the neighbours
16538 pixels are the closest.
16541 Bob mode, interpolates top field. Same as 13 but with a more complicated
16542 interpolation formula.
16545 Bob mode, interpolates bottom field. Same as 14 but with a more complicated
16546 interpolation formula.
16549 Clips the pixel with the minimum and maximum of respectively the maximum and
16550 minimum of each pair of opposite neighbour pixels.
16553 Line-sensitive clipping using opposite neighbours whose greatest distance from
16554 the current pixel is minimal.
16557 Replaces the pixel with the average of its 8 neighbours.
16560 Averages the 9 pixels ([1 1 1] horizontal and vertical blur).
16563 Clips pixels using the averages of opposite neighbour.
16566 Same as mode 21 but simpler and faster.
16569 Small edge and halo removal, but reputed useless.
16575 @section removelogo
16577 Suppress a TV station logo, using an image file to determine which
16578 pixels comprise the logo. It works by filling in the pixels that
16579 comprise the logo with neighboring pixels.
16581 The filter accepts the following options:
16585 Set the filter bitmap file, which can be any image format supported by
16586 libavformat. The width and height of the image file must match those of the
16587 video stream being processed.
16590 Pixels in the provided bitmap image with a value of zero are not
16591 considered part of the logo, non-zero pixels are considered part of
16592 the logo. If you use white (255) for the logo and black (0) for the
16593 rest, you will be safe. For making the filter bitmap, it is
16594 recommended to take a screen capture of a black frame with the logo
16595 visible, and then using a threshold filter followed by the erode
16596 filter once or twice.
16598 If needed, little splotches can be fixed manually. Remember that if
16599 logo pixels are not covered, the filter quality will be much
16600 reduced. Marking too many pixels as part of the logo does not hurt as
16601 much, but it will increase the amount of blurring needed to cover over
16602 the image and will destroy more information than necessary, and extra
16603 pixels will slow things down on a large logo.
16605 @section repeatfields
16607 This filter uses the repeat_field flag from the Video ES headers and hard repeats
16608 fields based on its value.
16612 Reverse a video clip.
16614 Warning: This filter requires memory to buffer the entire clip, so trimming
16617 @subsection Examples
16621 Take the first 5 seconds of a clip, and reverse it.
16628 Shift R/G/B/A pixels horizontally and/or vertically.
16630 The filter accepts the following options:
16633 Set amount to shift red horizontally.
16635 Set amount to shift red vertically.
16637 Set amount to shift green horizontally.
16639 Set amount to shift green vertically.
16641 Set amount to shift blue horizontally.
16643 Set amount to shift blue vertically.
16645 Set amount to shift alpha horizontally.
16647 Set amount to shift alpha vertically.
16649 Set edge mode, can be @var{smear}, default, or @var{warp}.
16652 @subsection Commands
16654 This filter supports the all above options as @ref{commands}.
16657 Apply roberts cross operator to input video stream.
16659 The filter accepts the following option:
16663 Set which planes will be processed, unprocessed planes will be copied.
16664 By default value 0xf, all planes will be processed.
16667 Set value which will be multiplied with filtered result.
16670 Set value which will be added to filtered result.
16673 @subsection Commands
16675 This filter supports the all above options as @ref{commands}.
16679 Rotate video by an arbitrary angle expressed in radians.
16681 The filter accepts the following options:
16683 A description of the optional parameters follows.
16686 Set an expression for the angle by which to rotate the input video
16687 clockwise, expressed as a number of radians. A negative value will
16688 result in a counter-clockwise rotation. By default it is set to "0".
16690 This expression is evaluated for each frame.
16693 Set the output width expression, default value is "iw".
16694 This expression is evaluated just once during configuration.
16697 Set the output height expression, default value is "ih".
16698 This expression is evaluated just once during configuration.
16701 Enable bilinear interpolation if set to 1, a value of 0 disables
16702 it. Default value is 1.
16705 Set the color used to fill the output area not covered by the rotated
16706 image. For the general syntax of this option, check the
16707 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
16708 If the special value "none" is selected then no
16709 background is printed (useful for example if the background is never shown).
16711 Default value is "black".
16714 The expressions for the angle and the output size can contain the
16715 following constants and functions:
16719 sequential number of the input frame, starting from 0. It is always NAN
16720 before the first frame is filtered.
16723 time in seconds of the input frame, it is set to 0 when the filter is
16724 configured. It is always NAN before the first frame is filtered.
16728 horizontal and vertical chroma subsample values. For example for the
16729 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
16733 the input video width and height
16737 the output width and height, that is the size of the padded area as
16738 specified by the @var{width} and @var{height} expressions
16742 the minimal width/height required for completely containing the input
16743 video rotated by @var{a} radians.
16745 These are only available when computing the @option{out_w} and
16746 @option{out_h} expressions.
16749 @subsection Examples
16753 Rotate the input by PI/6 radians clockwise:
16759 Rotate the input by PI/6 radians counter-clockwise:
16765 Rotate the input by 45 degrees clockwise:
16771 Apply a constant rotation with period T, starting from an angle of PI/3:
16773 rotate=PI/3+2*PI*t/T
16777 Make the input video rotation oscillating with a period of T
16778 seconds and an amplitude of A radians:
16780 rotate=A*sin(2*PI/T*t)
16784 Rotate the video, output size is chosen so that the whole rotating
16785 input video is always completely contained in the output:
16787 rotate='2*PI*t:ow=hypot(iw,ih):oh=ow'
16791 Rotate the video, reduce the output size so that no background is ever
16794 rotate=2*PI*t:ow='min(iw,ih)/sqrt(2)':oh=ow:c=none
16798 @subsection Commands
16800 The filter supports the following commands:
16804 Set the angle expression.
16805 The command accepts the same syntax of the corresponding option.
16807 If the specified expression is not valid, it is kept at its current
16813 Apply Shape Adaptive Blur.
16815 The filter accepts the following options:
16818 @item luma_radius, lr
16819 Set luma blur filter strength, must be a value in range 0.1-4.0, default
16820 value is 1.0. A greater value will result in a more blurred image, and
16821 in slower processing.
16823 @item luma_pre_filter_radius, lpfr
16824 Set luma pre-filter radius, must be a value in the 0.1-2.0 range, default
16827 @item luma_strength, ls
16828 Set luma maximum difference between pixels to still be considered, must
16829 be a value in the 0.1-100.0 range, default value is 1.0.
16831 @item chroma_radius, cr
16832 Set chroma blur filter strength, must be a value in range -0.9-4.0. A
16833 greater value will result in a more blurred image, and in slower
16836 @item chroma_pre_filter_radius, cpfr
16837 Set chroma pre-filter radius, must be a value in the -0.9-2.0 range.
16839 @item chroma_strength, cs
16840 Set chroma maximum difference between pixels to still be considered,
16841 must be a value in the -0.9-100.0 range.
16844 Each chroma option value, if not explicitly specified, is set to the
16845 corresponding luma option value.
16850 Scale (resize) the input video, using the libswscale library.
16852 The scale filter forces the output display aspect ratio to be the same
16853 of the input, by changing the output sample aspect ratio.
16855 If the input image format is different from the format requested by
16856 the next filter, the scale filter will convert the input to the
16859 @subsection Options
16860 The filter accepts the following options, or any of the options
16861 supported by the libswscale scaler.
16863 See @ref{scaler_options,,the ffmpeg-scaler manual,ffmpeg-scaler} for
16864 the complete list of scaler options.
16869 Set the output video dimension expression. Default value is the input
16872 If the @var{width} or @var{w} value is 0, the input width is used for
16873 the output. If the @var{height} or @var{h} value is 0, the input height
16874 is used for the output.
16876 If one and only one of the values is -n with n >= 1, the scale filter
16877 will use a value that maintains the aspect ratio of the input image,
16878 calculated from the other specified dimension. After that it will,
16879 however, make sure that the calculated dimension is divisible by n and
16880 adjust the value if necessary.
16882 If both values are -n with n >= 1, the behavior will be identical to
16883 both values being set to 0 as previously detailed.
16885 See below for the list of accepted constants for use in the dimension
16889 Specify when to evaluate @var{width} and @var{height} expression. It accepts the following values:
16893 Only evaluate expressions once during the filter initialization or when a command is processed.
16896 Evaluate expressions for each incoming frame.
16900 Default value is @samp{init}.
16904 Set the interlacing mode. It accepts the following values:
16908 Force interlaced aware scaling.
16911 Do not apply interlaced scaling.
16914 Select interlaced aware scaling depending on whether the source frames
16915 are flagged as interlaced or not.
16918 Default value is @samp{0}.
16921 Set libswscale scaling flags. See
16922 @ref{sws_flags,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16923 complete list of values. If not explicitly specified the filter applies
16927 @item param0, param1
16928 Set libswscale input parameters for scaling algorithms that need them. See
16929 @ref{sws_params,,the ffmpeg-scaler manual,ffmpeg-scaler} for the
16930 complete documentation. If not explicitly specified the filter applies
16936 Set the video size. For the syntax of this option, check the
16937 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
16939 @item in_color_matrix
16940 @item out_color_matrix
16941 Set in/output YCbCr color space type.
16943 This allows the autodetected value to be overridden as well as allows forcing
16944 a specific value used for the output and encoder.
16946 If not specified, the color space type depends on the pixel format.
16952 Choose automatically.
16955 Format conforming to International Telecommunication Union (ITU)
16956 Recommendation BT.709.
16959 Set color space conforming to the United States Federal Communications
16960 Commission (FCC) Code of Federal Regulations (CFR) Title 47 (2003) 73.682 (a).
16965 Set color space conforming to:
16969 ITU Radiocommunication Sector (ITU-R) Recommendation BT.601
16972 ITU-R Rec. BT.470-6 (1998) Systems B, B1, and G
16975 Society of Motion Picture and Television Engineers (SMPTE) ST 170:2004
16980 Set color space conforming to SMPTE ST 240:1999.
16983 Set color space conforming to ITU-R BT.2020 non-constant luminance system.
16988 Set in/output YCbCr sample range.
16990 This allows the autodetected value to be overridden as well as allows forcing
16991 a specific value used for the output and encoder. If not specified, the
16992 range depends on the pixel format. Possible values:
16996 Choose automatically.
16999 Set full range (0-255 in case of 8-bit luma).
17001 @item mpeg/limited/tv
17002 Set "MPEG" range (16-235 in case of 8-bit luma).
17005 @item force_original_aspect_ratio
17006 Enable decreasing or increasing output video width or height if necessary to
17007 keep the original aspect ratio. Possible values:
17011 Scale the video as specified and disable this feature.
17014 The output video dimensions will automatically be decreased if needed.
17017 The output video dimensions will automatically be increased if needed.
17021 One useful instance of this option is that when you know a specific device's
17022 maximum allowed resolution, you can use this to limit the output video to
17023 that, while retaining the aspect ratio. For example, device A allows
17024 1280x720 playback, and your video is 1920x800. Using this option (set it to
17025 decrease) and specifying 1280x720 to the command line makes the output
17028 Please note that this is a different thing than specifying -1 for @option{w}
17029 or @option{h}, you still need to specify the output resolution for this option
17032 @item force_divisible_by
17033 Ensures that both the output dimensions, width and height, are divisible by the
17034 given integer when used together with @option{force_original_aspect_ratio}. This
17035 works similar to using @code{-n} in the @option{w} and @option{h} options.
17037 This option respects the value set for @option{force_original_aspect_ratio},
17038 increasing or decreasing the resolution accordingly. The video's aspect ratio
17039 may be slightly modified.
17041 This option can be handy if you need to have a video fit within or exceed
17042 a defined resolution using @option{force_original_aspect_ratio} but also have
17043 encoder restrictions on width or height divisibility.
17047 The values of the @option{w} and @option{h} options are expressions
17048 containing the following constants:
17053 The input width and height
17057 These are the same as @var{in_w} and @var{in_h}.
17061 The output (scaled) width and height
17065 These are the same as @var{out_w} and @var{out_h}
17068 The same as @var{iw} / @var{ih}
17071 input sample aspect ratio
17074 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
17078 horizontal and vertical input chroma subsample values. For example for the
17079 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17083 horizontal and vertical output chroma subsample values. For example for the
17084 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17087 The (sequential) number of the input frame, starting from 0.
17088 Only available with @code{eval=frame}.
17091 The presentation timestamp of the input frame, expressed as a number of
17092 seconds. Only available with @code{eval=frame}.
17095 The position (byte offset) of the frame in the input stream, or NaN if
17096 this information is unavailable and/or meaningless (for example in case of synthetic video).
17097 Only available with @code{eval=frame}.
17100 @subsection Examples
17104 Scale the input video to a size of 200x100
17109 This is equivalent to:
17120 Specify a size abbreviation for the output size:
17125 which can also be written as:
17131 Scale the input to 2x:
17133 scale=w=2*iw:h=2*ih
17137 The above is the same as:
17139 scale=2*in_w:2*in_h
17143 Scale the input to 2x with forced interlaced scaling:
17145 scale=2*iw:2*ih:interl=1
17149 Scale the input to half size:
17151 scale=w=iw/2:h=ih/2
17155 Increase the width, and set the height to the same size:
17161 Seek Greek harmony:
17168 Increase the height, and set the width to 3/2 of the height:
17170 scale=w=3/2*oh:h=3/5*ih
17174 Increase the size, making the size a multiple of the chroma
17177 scale="trunc(3/2*iw/hsub)*hsub:trunc(3/2*ih/vsub)*vsub"
17181 Increase the width to a maximum of 500 pixels,
17182 keeping the same aspect ratio as the input:
17184 scale=w='min(500\, iw*3/2):h=-1'
17188 Make pixels square by combining scale and setsar:
17190 scale='trunc(ih*dar):ih',setsar=1/1
17194 Make pixels square by combining scale and setsar,
17195 making sure the resulting resolution is even (required by some codecs):
17197 scale='trunc(ih*dar/2)*2:trunc(ih/2)*2',setsar=1/1
17201 @subsection Commands
17203 This filter supports the following commands:
17207 Set the output video dimension expression.
17208 The command accepts the same syntax of the corresponding option.
17210 If the specified expression is not valid, it is kept at its current
17216 Use the NVIDIA Performance Primitives (libnpp) to perform scaling and/or pixel
17217 format conversion on CUDA video frames. Setting the output width and height
17218 works in the same way as for the @var{scale} filter.
17220 The following additional options are accepted:
17223 The pixel format of the output CUDA frames. If set to the string "same" (the
17224 default), the input format will be kept. Note that automatic format negotiation
17225 and conversion is not yet supported for hardware frames
17228 The interpolation algorithm used for resizing. One of the following:
17235 @item cubic2p_bspline
17236 2-parameter cubic (B=1, C=0)
17238 @item cubic2p_catmullrom
17239 2-parameter cubic (B=0, C=1/2)
17241 @item cubic2p_b05c03
17242 2-parameter cubic (B=1/2, C=3/10)
17250 @item force_original_aspect_ratio
17251 Enable decreasing or increasing output video width or height if necessary to
17252 keep the original aspect ratio. Possible values:
17256 Scale the video as specified and disable this feature.
17259 The output video dimensions will automatically be decreased if needed.
17262 The output video dimensions will automatically be increased if needed.
17266 One useful instance of this option is that when you know a specific device's
17267 maximum allowed resolution, you can use this to limit the output video to
17268 that, while retaining the aspect ratio. For example, device A allows
17269 1280x720 playback, and your video is 1920x800. Using this option (set it to
17270 decrease) and specifying 1280x720 to the command line makes the output
17273 Please note that this is a different thing than specifying -1 for @option{w}
17274 or @option{h}, you still need to specify the output resolution for this option
17277 @item force_divisible_by
17278 Ensures that both the output dimensions, width and height, are divisible by the
17279 given integer when used together with @option{force_original_aspect_ratio}. This
17280 works similar to using @code{-n} in the @option{w} and @option{h} options.
17282 This option respects the value set for @option{force_original_aspect_ratio},
17283 increasing or decreasing the resolution accordingly. The video's aspect ratio
17284 may be slightly modified.
17286 This option can be handy if you need to have a video fit within or exceed
17287 a defined resolution using @option{force_original_aspect_ratio} but also have
17288 encoder restrictions on width or height divisibility.
17294 Scale (resize) the input video, based on a reference video.
17296 See the scale filter for available options, scale2ref supports the same but
17297 uses the reference video instead of the main input as basis. scale2ref also
17298 supports the following additional constants for the @option{w} and
17299 @option{h} options:
17304 The main input video's width and height
17307 The same as @var{main_w} / @var{main_h}
17310 The main input video's sample aspect ratio
17312 @item main_dar, mdar
17313 The main input video's display aspect ratio. Calculated from
17314 @code{(main_w / main_h) * main_sar}.
17318 The main input video's horizontal and vertical chroma subsample values.
17319 For example for the pixel format "yuv422p" @var{hsub} is 2 and @var{vsub}
17323 The (sequential) number of the main input frame, starting from 0.
17324 Only available with @code{eval=frame}.
17327 The presentation timestamp of the main input frame, expressed as a number of
17328 seconds. Only available with @code{eval=frame}.
17331 The position (byte offset) of the frame in the main input stream, or NaN if
17332 this information is unavailable and/or meaningless (for example in case of synthetic video).
17333 Only available with @code{eval=frame}.
17336 @subsection Examples
17340 Scale a subtitle stream (b) to match the main video (a) in size before overlaying
17342 'scale2ref[b][a];[a][b]overlay'
17346 Scale a logo to 1/10th the height of a video, while preserving its display aspect ratio.
17348 [logo-in][video-in]scale2ref=w=oh*mdar:h=ih/10[logo-out][video-out]
17352 @subsection Commands
17354 This filter supports the following commands:
17358 Set the output video dimension expression.
17359 The command accepts the same syntax of the corresponding option.
17361 If the specified expression is not valid, it is kept at its current
17366 Scroll input video horizontally and/or vertically by constant speed.
17368 The filter accepts the following options:
17370 @item horizontal, h
17371 Set the horizontal scrolling speed. Default is 0. Allowed range is from -1 to 1.
17372 Negative values changes scrolling direction.
17375 Set the vertical scrolling speed. Default is 0. Allowed range is from -1 to 1.
17376 Negative values changes scrolling direction.
17379 Set the initial horizontal scrolling position. Default is 0. Allowed range is from 0 to 1.
17382 Set the initial vertical scrolling position. Default is 0. Allowed range is from 0 to 1.
17385 @subsection Commands
17387 This filter supports the following @ref{commands}:
17389 @item horizontal, h
17390 Set the horizontal scrolling speed.
17392 Set the vertical scrolling speed.
17398 Detect video scene change.
17400 This filter sets frame metadata with mafd between frame, the scene score, and
17401 forward the frame to the next filter, so they can use these metadata to detect
17402 scene change or others.
17404 In addition, this filter logs a message and sets frame metadata when it detects
17405 a scene change by @option{threshold}.
17407 @code{lavfi.scd.mafd} metadata keys are set with mafd for every frame.
17409 @code{lavfi.scd.score} metadata keys are set with scene change score for every frame
17410 to detect scene change.
17412 @code{lavfi.scd.time} metadata keys are set with current filtered frame time which
17413 detect scene change with @option{threshold}.
17415 The filter accepts the following options:
17419 Set the scene change detection threshold as a percentage of maximum change. Good
17420 values are in the @code{[8.0, 14.0]} range. The range for @option{threshold} is
17423 Default value is @code{10.}.
17426 Set the flag to pass scene change frames to the next filter. Default value is @code{0}
17427 You can enable it if you want to get snapshot of scene change frames only.
17430 @anchor{selectivecolor}
17431 @section selectivecolor
17433 Adjust cyan, magenta, yellow and black (CMYK) to certain ranges of colors (such
17434 as "reds", "yellows", "greens", "cyans", ...). The adjustment range is defined
17435 by the "purity" of the color (that is, how saturated it already is).
17437 This filter is similar to the Adobe Photoshop Selective Color tool.
17439 The filter accepts the following options:
17442 @item correction_method
17443 Select color correction method.
17445 Available values are:
17448 Specified adjustments are applied "as-is" (added/subtracted to original pixel
17451 Specified adjustments are relative to the original component value.
17453 Default is @code{absolute}.
17455 Adjustments for red pixels (pixels where the red component is the maximum)
17457 Adjustments for yellow pixels (pixels where the blue component is the minimum)
17459 Adjustments for green pixels (pixels where the green component is the maximum)
17461 Adjustments for cyan pixels (pixels where the red component is the minimum)
17463 Adjustments for blue pixels (pixels where the blue component is the maximum)
17465 Adjustments for magenta pixels (pixels where the green component is the minimum)
17467 Adjustments for white pixels (pixels where all components are greater than 128)
17469 Adjustments for all pixels except pure black and pure white
17471 Adjustments for black pixels (pixels where all components are lesser than 128)
17473 Specify a Photoshop selective color file (@code{.asv}) to import the settings from.
17476 All the adjustment settings (@option{reds}, @option{yellows}, ...) accept up to
17477 4 space separated floating point adjustment values in the [-1,1] range,
17478 respectively to adjust the amount of cyan, magenta, yellow and black for the
17479 pixels of its range.
17481 @subsection Examples
17485 Increase cyan by 50% and reduce yellow by 33% in every green areas, and
17486 increase magenta by 27% in blue areas:
17488 selectivecolor=greens=.5 0 -.33 0:blues=0 .27
17492 Use a Photoshop selective color preset:
17494 selectivecolor=psfile=MySelectiveColorPresets/Misty.asv
17498 @anchor{separatefields}
17499 @section separatefields
17501 The @code{separatefields} takes a frame-based video input and splits
17502 each frame into its components fields, producing a new half height clip
17503 with twice the frame rate and twice the frame count.
17505 This filter use field-dominance information in frame to decide which
17506 of each pair of fields to place first in the output.
17507 If it gets it wrong use @ref{setfield} filter before @code{separatefields} filter.
17509 @section setdar, setsar
17511 The @code{setdar} filter sets the Display Aspect Ratio for the filter
17514 This is done by changing the specified Sample (aka Pixel) Aspect
17515 Ratio, according to the following equation:
17517 @var{DAR} = @var{HORIZONTAL_RESOLUTION} / @var{VERTICAL_RESOLUTION} * @var{SAR}
17520 Keep in mind that the @code{setdar} filter does not modify the pixel
17521 dimensions of the video frame. Also, the display aspect ratio set by
17522 this filter may be changed by later filters in the filterchain,
17523 e.g. in case of scaling or if another "setdar" or a "setsar" filter is
17526 The @code{setsar} filter sets the Sample (aka Pixel) Aspect Ratio for
17527 the filter output video.
17529 Note that as a consequence of the application of this filter, the
17530 output display aspect ratio will change according to the equation
17533 Keep in mind that the sample aspect ratio set by the @code{setsar}
17534 filter may be changed by later filters in the filterchain, e.g. if
17535 another "setsar" or a "setdar" filter is applied.
17537 It accepts the following parameters:
17540 @item r, ratio, dar (@code{setdar} only), sar (@code{setsar} only)
17541 Set the aspect ratio used by the filter.
17543 The parameter can be a floating point number string, an expression, or
17544 a string of the form @var{num}:@var{den}, where @var{num} and
17545 @var{den} are the numerator and denominator of the aspect ratio. If
17546 the parameter is not specified, it is assumed the value "0".
17547 In case the form "@var{num}:@var{den}" is used, the @code{:} character
17551 Set the maximum integer value to use for expressing numerator and
17552 denominator when reducing the expressed aspect ratio to a rational.
17553 Default value is @code{100}.
17557 The parameter @var{sar} is an expression containing
17558 the following constants:
17562 These are approximated values for the mathematical constants e
17563 (Euler's number), pi (Greek pi), and phi (the golden ratio).
17566 The input width and height.
17569 These are the same as @var{w} / @var{h}.
17572 The input sample aspect ratio.
17575 The input display aspect ratio. It is the same as
17576 (@var{w} / @var{h}) * @var{sar}.
17579 Horizontal and vertical chroma subsample values. For example, for the
17580 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
17583 @subsection Examples
17588 To change the display aspect ratio to 16:9, specify one of the following:
17595 To change the sample aspect ratio to 10:11, specify:
17601 To set a display aspect ratio of 16:9, and specify a maximum integer value of
17602 1000 in the aspect ratio reduction, use the command:
17604 setdar=ratio=16/9:max=1000
17612 Force field for the output video frame.
17614 The @code{setfield} filter marks the interlace type field for the
17615 output frames. It does not change the input frame, but only sets the
17616 corresponding property, which affects how the frame is treated by
17617 following filters (e.g. @code{fieldorder} or @code{yadif}).
17619 The filter accepts the following options:
17624 Available values are:
17628 Keep the same field property.
17631 Mark the frame as bottom-field-first.
17634 Mark the frame as top-field-first.
17637 Mark the frame as progressive.
17644 Force frame parameter for the output video frame.
17646 The @code{setparams} filter marks interlace and color range for the
17647 output frames. It does not change the input frame, but only sets the
17648 corresponding property, which affects how the frame is treated by
17653 Available values are:
17657 Keep the same field property (default).
17660 Mark the frame as bottom-field-first.
17663 Mark the frame as top-field-first.
17666 Mark the frame as progressive.
17670 Available values are:
17674 Keep the same color range property (default).
17676 @item unspecified, unknown
17677 Mark the frame as unspecified color range.
17679 @item limited, tv, mpeg
17680 Mark the frame as limited range.
17682 @item full, pc, jpeg
17683 Mark the frame as full range.
17686 @item color_primaries
17687 Set the color primaries.
17688 Available values are:
17692 Keep the same color primaries property (default).
17709 Set the color transfer.
17710 Available values are:
17714 Keep the same color trc property (default).
17736 Set the colorspace.
17737 Available values are:
17741 Keep the same colorspace property (default).
17754 @item chroma-derived-nc
17755 @item chroma-derived-c
17762 Show a line containing various information for each input video frame.
17763 The input video is not modified.
17765 This filter supports the following options:
17769 Calculate checksums of each plane. By default enabled.
17772 The shown line contains a sequence of key/value pairs of the form
17773 @var{key}:@var{value}.
17775 The following values are shown in the output:
17779 The (sequential) number of the input frame, starting from 0.
17782 The Presentation TimeStamp of the input frame, expressed as a number of
17783 time base units. The time base unit depends on the filter input pad.
17786 The Presentation TimeStamp of the input frame, expressed as a number of
17790 The position of the frame in the input stream, or -1 if this information is
17791 unavailable and/or meaningless (for example in case of synthetic video).
17794 The pixel format name.
17797 The sample aspect ratio of the input frame, expressed in the form
17798 @var{num}/@var{den}.
17801 The size of the input frame. For the syntax of this option, check the
17802 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
17805 The type of interlaced mode ("P" for "progressive", "T" for top field first, "B"
17806 for bottom field first).
17809 This is 1 if the frame is a key frame, 0 otherwise.
17812 The picture type of the input frame ("I" for an I-frame, "P" for a
17813 P-frame, "B" for a B-frame, or "?" for an unknown type).
17814 Also refer to the documentation of the @code{AVPictureType} enum and of
17815 the @code{av_get_picture_type_char} function defined in
17816 @file{libavutil/avutil.h}.
17819 The Adler-32 checksum (printed in hexadecimal) of all the planes of the input frame.
17821 @item plane_checksum
17822 The Adler-32 checksum (printed in hexadecimal) of each plane of the input frame,
17823 expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3}]".
17826 The mean value of pixels in each plane of the input frame, expressed in the form
17827 "[@var{mean0} @var{mean1} @var{mean2} @var{mean3}]".
17830 The standard deviation of pixel values in each plane of the input frame, expressed
17831 in the form "[@var{stdev0} @var{stdev1} @var{stdev2} @var{stdev3}]".
17835 @section showpalette
17837 Displays the 256 colors palette of each frame. This filter is only relevant for
17838 @var{pal8} pixel format frames.
17840 It accepts the following option:
17844 Set the size of the box used to represent one palette color entry. Default is
17845 @code{30} (for a @code{30x30} pixel box).
17848 @section shuffleframes
17850 Reorder and/or duplicate and/or drop video frames.
17852 It accepts the following parameters:
17856 Set the destination indexes of input frames.
17857 This is space or '|' separated list of indexes that maps input frames to output
17858 frames. Number of indexes also sets maximal value that each index may have.
17859 '-1' index have special meaning and that is to drop frame.
17862 The first frame has the index 0. The default is to keep the input unchanged.
17864 @subsection Examples
17868 Swap second and third frame of every three frames of the input:
17870 ffmpeg -i INPUT -vf "shuffleframes=0 2 1" OUTPUT
17874 Swap 10th and 1st frame of every ten frames of the input:
17876 ffmpeg -i INPUT -vf "shuffleframes=9 1 2 3 4 5 6 7 8 0" OUTPUT
17880 @section shufflepixels
17882 Reorder pixels in video frames.
17884 This filter accepts the following options:
17888 Set shuffle direction. Can be forward or inverse direction.
17889 Default direction is forward.
17892 Set shuffle mode. Can be horizontal, vertical or block mode.
17896 Set shuffle block_size. In case of horizontal shuffle mode only width
17897 part of size is used, and in case of vertical shuffle mode only height
17898 part of size is used.
17901 Set random seed used with shuffling pixels. Mainly useful to set to be able
17902 to reverse filtering process to get original input.
17903 For example, to reverse forward shuffle you need to use same parameters
17904 and exact same seed and to set direction to inverse.
17907 @section shuffleplanes
17909 Reorder and/or duplicate video planes.
17911 It accepts the following parameters:
17916 The index of the input plane to be used as the first output plane.
17919 The index of the input plane to be used as the second output plane.
17922 The index of the input plane to be used as the third output plane.
17925 The index of the input plane to be used as the fourth output plane.
17929 The first plane has the index 0. The default is to keep the input unchanged.
17931 @subsection Examples
17935 Swap the second and third planes of the input:
17937 ffmpeg -i INPUT -vf shuffleplanes=0:2:1:3 OUTPUT
17941 @anchor{signalstats}
17942 @section signalstats
17943 Evaluate various visual metrics that assist in determining issues associated
17944 with the digitization of analog video media.
17946 By default the filter will log these metadata values:
17950 Display the minimal Y value contained within the input frame. Expressed in
17954 Display the Y value at the 10% percentile within the input frame. Expressed in
17958 Display the average Y value within the input frame. Expressed in range of
17962 Display the Y value at the 90% percentile within the input frame. Expressed in
17966 Display the maximum Y value contained within the input frame. Expressed in
17970 Display the minimal U value contained within the input frame. Expressed in
17974 Display the U value at the 10% percentile within the input frame. Expressed in
17978 Display the average U value within the input frame. Expressed in range of
17982 Display the U value at the 90% percentile within the input frame. Expressed in
17986 Display the maximum U value contained within the input frame. Expressed in
17990 Display the minimal V value contained within the input frame. Expressed in
17994 Display the V value at the 10% percentile within the input frame. Expressed in
17998 Display the average V value within the input frame. Expressed in range of
18002 Display the V value at the 90% percentile within the input frame. Expressed in
18006 Display the maximum V value contained within the input frame. Expressed in
18010 Display the minimal saturation value contained within the input frame.
18011 Expressed in range of [0-~181.02].
18014 Display the saturation value at the 10% percentile within the input frame.
18015 Expressed in range of [0-~181.02].
18018 Display the average saturation value within the input frame. Expressed in range
18022 Display the saturation value at the 90% percentile within the input frame.
18023 Expressed in range of [0-~181.02].
18026 Display the maximum saturation value contained within the input frame.
18027 Expressed in range of [0-~181.02].
18030 Display the median value for hue within the input frame. Expressed in range of
18034 Display the average value for hue within the input frame. Expressed in range of
18038 Display the average of sample value difference between all values of the Y
18039 plane in the current frame and corresponding values of the previous input frame.
18040 Expressed in range of [0-255].
18043 Display the average of sample value difference between all values of the U
18044 plane in the current frame and corresponding values of the previous input frame.
18045 Expressed in range of [0-255].
18048 Display the average of sample value difference between all values of the V
18049 plane in the current frame and corresponding values of the previous input frame.
18050 Expressed in range of [0-255].
18053 Display bit depth of Y plane in current frame.
18054 Expressed in range of [0-16].
18057 Display bit depth of U plane in current frame.
18058 Expressed in range of [0-16].
18061 Display bit depth of V plane in current frame.
18062 Expressed in range of [0-16].
18065 The filter accepts the following options:
18071 @option{stat} specify an additional form of image analysis.
18072 @option{out} output video with the specified type of pixel highlighted.
18074 Both options accept the following values:
18078 Identify @var{temporal outliers} pixels. A @var{temporal outlier} is a pixel
18079 unlike the neighboring pixels of the same field. Examples of temporal outliers
18080 include the results of video dropouts, head clogs, or tape tracking issues.
18083 Identify @var{vertical line repetition}. Vertical line repetition includes
18084 similar rows of pixels within a frame. In born-digital video vertical line
18085 repetition is common, but this pattern is uncommon in video digitized from an
18086 analog source. When it occurs in video that results from the digitization of an
18087 analog source it can indicate concealment from a dropout compensator.
18090 Identify pixels that fall outside of legal broadcast range.
18094 Set the highlight color for the @option{out} option. The default color is
18098 @subsection Examples
18102 Output data of various video metrics:
18104 ffprobe -f lavfi movie=example.mov,signalstats="stat=tout+vrep+brng" -show_frames
18108 Output specific data about the minimum and maximum values of the Y plane per frame:
18110 ffprobe -f lavfi movie=example.mov,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YMIN
18114 Playback video while highlighting pixels that are outside of broadcast range in red.
18116 ffplay example.mov -vf signalstats="out=brng:color=red"
18120 Playback video with signalstats metadata drawn over the frame.
18122 ffplay example.mov -vf signalstats=stat=brng+vrep+tout,drawtext=fontfile=FreeSerif.ttf:textfile=signalstat_drawtext.txt
18125 The contents of signalstat_drawtext.txt used in the command are:
18128 Y (%@{metadata:lavfi.signalstats.YMIN@}-%@{metadata:lavfi.signalstats.YMAX@})
18129 U (%@{metadata:lavfi.signalstats.UMIN@}-%@{metadata:lavfi.signalstats.UMAX@})
18130 V (%@{metadata:lavfi.signalstats.VMIN@}-%@{metadata:lavfi.signalstats.VMAX@})
18131 saturation maximum: %@{metadata:lavfi.signalstats.SATMAX@}
18139 Calculates the MPEG-7 Video Signature. The filter can handle more than one
18140 input. In this case the matching between the inputs can be calculated additionally.
18141 The filter always passes through the first input. The signature of each stream can
18142 be written into a file.
18144 It accepts the following options:
18148 Enable or disable the matching process.
18150 Available values are:
18154 Disable the calculation of a matching (default).
18156 Calculate the matching for the whole video and output whether the whole video
18157 matches or only parts.
18159 Calculate only until a matching is found or the video ends. Should be faster in
18164 Set the number of inputs. The option value must be a non negative integer.
18165 Default value is 1.
18168 Set the path to which the output is written. If there is more than one input,
18169 the path must be a prototype, i.e. must contain %d or %0nd (where n is a positive
18170 integer), that will be replaced with the input number. If no filename is
18171 specified, no output will be written. This is the default.
18174 Choose the output format.
18176 Available values are:
18180 Use the specified binary representation (default).
18182 Use the specified xml representation.
18186 Set threshold to detect one word as similar. The option value must be an integer
18187 greater than zero. The default value is 9000.
18190 Set threshold to detect all words as similar. The option value must be an integer
18191 greater than zero. The default value is 60000.
18194 Set threshold to detect frames as similar. The option value must be an integer
18195 greater than zero. The default value is 116.
18198 Set the minimum length of a sequence in frames to recognize it as matching
18199 sequence. The option value must be a non negative integer value.
18200 The default value is 0.
18203 Set the minimum relation, that matching frames to all frames must have.
18204 The option value must be a double value between 0 and 1. The default value is 0.5.
18207 @subsection Examples
18211 To calculate the signature of an input video and store it in signature.bin:
18213 ffmpeg -i input.mkv -vf signature=filename=signature.bin -map 0:v -f null -
18217 To detect whether two videos match and store the signatures in XML format in
18218 signature0.xml and signature1.xml:
18220 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 -
18228 Blur the input video without impacting the outlines.
18230 It accepts the following options:
18233 @item luma_radius, lr
18234 Set the luma radius. The option value must be a float number in
18235 the range [0.1,5.0] that specifies the variance of the gaussian filter
18236 used to blur the image (slower if larger). Default value is 1.0.
18238 @item luma_strength, ls
18239 Set the luma strength. The option value must be a float number
18240 in the range [-1.0,1.0] that configures the blurring. A value included
18241 in [0.0,1.0] will blur the image whereas a value included in
18242 [-1.0,0.0] will sharpen the image. Default value is 1.0.
18244 @item luma_threshold, lt
18245 Set the luma threshold used as a coefficient to determine
18246 whether a pixel should be blurred or not. The option value must be an
18247 integer in the range [-30,30]. A value of 0 will filter all the image,
18248 a value included in [0,30] will filter flat areas and a value included
18249 in [-30,0] will filter edges. Default value is 0.
18251 @item chroma_radius, cr
18252 Set the chroma radius. The option value must be a float number in
18253 the range [0.1,5.0] that specifies the variance of the gaussian filter
18254 used to blur the image (slower if larger). Default value is @option{luma_radius}.
18256 @item chroma_strength, cs
18257 Set the chroma strength. The option value must be a float number
18258 in the range [-1.0,1.0] that configures the blurring. A value included
18259 in [0.0,1.0] will blur the image whereas a value included in
18260 [-1.0,0.0] will sharpen the image. Default value is @option{luma_strength}.
18262 @item chroma_threshold, ct
18263 Set the chroma threshold used as a coefficient to determine
18264 whether a pixel should be blurred or not. The option value must be an
18265 integer in the range [-30,30]. A value of 0 will filter all the image,
18266 a value included in [0,30] will filter flat areas and a value included
18267 in [-30,0] will filter edges. Default value is @option{luma_threshold}.
18270 If a chroma option is not explicitly set, the corresponding luma value
18274 Apply sobel operator to input video stream.
18276 The filter accepts the following option:
18280 Set which planes will be processed, unprocessed planes will be copied.
18281 By default value 0xf, all planes will be processed.
18284 Set value which will be multiplied with filtered result.
18287 Set value which will be added to filtered result.
18290 @subsection Commands
18292 This filter supports the all above options as @ref{commands}.
18297 Apply a simple postprocessing filter that compresses and decompresses the image
18298 at several (or - in the case of @option{quality} level @code{6} - all) shifts
18299 and average the results.
18301 The filter accepts the following options:
18305 Set quality. This option defines the number of levels for averaging. It accepts
18306 an integer in the range 0-6. If set to @code{0}, the filter will have no
18307 effect. A value of @code{6} means the higher quality. For each increment of
18308 that value the speed drops by a factor of approximately 2. Default value is
18312 Force a constant quantization parameter. If not set, the filter will use the QP
18313 from the video stream (if available).
18316 Set thresholding mode. Available modes are:
18320 Set hard thresholding (default).
18322 Set soft thresholding (better de-ringing effect, but likely blurrier).
18325 @item use_bframe_qp
18326 Enable the use of the QP from the B-Frames if set to @code{1}. Using this
18327 option may cause flicker since the B-Frames have often larger QP. Default is
18328 @code{0} (not enabled).
18331 @subsection Commands
18333 This filter supports the following commands:
18335 @item quality, level
18336 Set quality level. The value @code{max} can be used to set the maximum level,
18337 currently @code{6}.
18343 Scale the input by applying one of the super-resolution methods based on
18344 convolutional neural networks. Supported models:
18348 Super-Resolution Convolutional Neural Network model (SRCNN).
18349 See @url{https://arxiv.org/abs/1501.00092}.
18352 Efficient Sub-Pixel Convolutional Neural Network model (ESPCN).
18353 See @url{https://arxiv.org/abs/1609.05158}.
18356 Training scripts as well as scripts for model file (.pb) saving can be found at
18357 @url{https://github.com/XueweiMeng/sr/tree/sr_dnn_native}. Original repository
18358 is at @url{https://github.com/HighVoltageRocknRoll/sr.git}.
18360 Native model files (.model) can be generated from TensorFlow model
18361 files (.pb) by using tools/python/convert.py
18363 The filter accepts the following options:
18367 Specify which DNN backend to use for model loading and execution. This option accepts
18368 the following values:
18372 Native implementation of DNN loading and execution.
18375 TensorFlow backend. To enable this backend you
18376 need to install the TensorFlow for C library (see
18377 @url{https://www.tensorflow.org/install/install_c}) and configure FFmpeg with
18378 @code{--enable-libtensorflow}
18381 Default value is @samp{native}.
18384 Set path to model file specifying network architecture and its parameters.
18385 Note that different backends use different file formats. TensorFlow backend
18386 can load files for both formats, while native backend can load files for only
18390 Set scale factor for SRCNN model. Allowed values are @code{2}, @code{3} and @code{4}.
18391 Default value is @code{2}. Scale factor is necessary for SRCNN model, because it accepts
18392 input upscaled using bicubic upscaling with proper scale factor.
18395 This feature can also be finished with @ref{dnn_processing} filter.
18399 Obtain the SSIM (Structural SImilarity Metric) between two input videos.
18401 This filter takes in input two input videos, the first input is
18402 considered the "main" source and is passed unchanged to the
18403 output. The second input is used as a "reference" video for computing
18406 Both video inputs must have the same resolution and pixel format for
18407 this filter to work correctly. Also it assumes that both inputs
18408 have the same number of frames, which are compared one by one.
18410 The filter stores the calculated SSIM of each frame.
18412 The description of the accepted parameters follows.
18415 @item stats_file, f
18416 If specified the filter will use the named file to save the SSIM of
18417 each individual frame. When filename equals "-" the data is sent to
18421 The file printed if @var{stats_file} is selected, contains a sequence of
18422 key/value pairs of the form @var{key}:@var{value} for each compared
18425 A description of each shown parameter follows:
18429 sequential number of the input frame, starting from 1
18431 @item Y, U, V, R, G, B
18432 SSIM of the compared frames for the component specified by the suffix.
18435 SSIM of the compared frames for the whole frame.
18438 Same as above but in dB representation.
18441 This filter also supports the @ref{framesync} options.
18443 @subsection Examples
18448 movie=ref_movie.mpg, setpts=PTS-STARTPTS [main];
18449 [main][ref] ssim="stats_file=stats.log" [out]
18452 On this example the input file being processed is compared with the
18453 reference file @file{ref_movie.mpg}. The SSIM of each individual frame
18454 is stored in @file{stats.log}.
18457 Another example with both psnr and ssim at same time:
18459 ffmpeg -i main.mpg -i ref.mpg -lavfi "ssim;[0:v][1:v]psnr" -f null -
18463 Another example with different containers:
18465 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 -
18471 Convert between different stereoscopic image formats.
18473 The filters accept the following options:
18477 Set stereoscopic image format of input.
18479 Available values for input image formats are:
18482 side by side parallel (left eye left, right eye right)
18485 side by side crosseye (right eye left, left eye right)
18488 side by side parallel with half width resolution
18489 (left eye left, right eye right)
18492 side by side crosseye with half width resolution
18493 (right eye left, left eye right)
18497 above-below (left eye above, right eye below)
18501 above-below (right eye above, left eye below)
18505 above-below with half height resolution
18506 (left eye above, right eye below)
18510 above-below with half height resolution
18511 (right eye above, left eye below)
18514 alternating frames (left eye first, right eye second)
18517 alternating frames (right eye first, left eye second)
18520 interleaved rows (left eye has top row, right eye starts on next row)
18523 interleaved rows (right eye has top row, left eye starts on next row)
18526 interleaved columns, left eye first
18529 interleaved columns, right eye first
18531 Default value is @samp{sbsl}.
18535 Set stereoscopic image format of output.
18539 side by side parallel (left eye left, right eye right)
18542 side by side crosseye (right eye left, left eye right)
18545 side by side parallel with half width resolution
18546 (left eye left, right eye right)
18549 side by side crosseye with half width resolution
18550 (right eye left, left eye right)
18554 above-below (left eye above, right eye below)
18558 above-below (right eye above, left eye below)
18562 above-below with half height resolution
18563 (left eye above, right eye below)
18567 above-below with half height resolution
18568 (right eye above, left eye below)
18571 alternating frames (left eye first, right eye second)
18574 alternating frames (right eye first, left eye second)
18577 interleaved rows (left eye has top row, right eye starts on next row)
18580 interleaved rows (right eye has top row, left eye starts on next row)
18583 anaglyph red/blue gray
18584 (red filter on left eye, blue filter on right eye)
18587 anaglyph red/green gray
18588 (red filter on left eye, green filter on right eye)
18591 anaglyph red/cyan gray
18592 (red filter on left eye, cyan filter on right eye)
18595 anaglyph red/cyan half colored
18596 (red filter on left eye, cyan filter on right eye)
18599 anaglyph red/cyan color
18600 (red filter on left eye, cyan filter on right eye)
18603 anaglyph red/cyan color optimized with the least squares projection of dubois
18604 (red filter on left eye, cyan filter on right eye)
18607 anaglyph green/magenta gray
18608 (green filter on left eye, magenta filter on right eye)
18611 anaglyph green/magenta half colored
18612 (green filter on left eye, magenta filter on right eye)
18615 anaglyph green/magenta colored
18616 (green filter on left eye, magenta filter on right eye)
18619 anaglyph green/magenta color optimized with the least squares projection of dubois
18620 (green filter on left eye, magenta filter on right eye)
18623 anaglyph yellow/blue gray
18624 (yellow filter on left eye, blue filter on right eye)
18627 anaglyph yellow/blue half colored
18628 (yellow filter on left eye, blue filter on right eye)
18631 anaglyph yellow/blue colored
18632 (yellow filter on left eye, blue filter on right eye)
18635 anaglyph yellow/blue color optimized with the least squares projection of dubois
18636 (yellow filter on left eye, blue filter on right eye)
18639 mono output (left eye only)
18642 mono output (right eye only)
18645 checkerboard, left eye first
18648 checkerboard, right eye first
18651 interleaved columns, left eye first
18654 interleaved columns, right eye first
18660 Default value is @samp{arcd}.
18663 @subsection Examples
18667 Convert input video from side by side parallel to anaglyph yellow/blue dubois:
18673 Convert input video from above below (left eye above, right eye below) to side by side crosseye.
18679 @section streamselect, astreamselect
18680 Select video or audio streams.
18682 The filter accepts the following options:
18686 Set number of inputs. Default is 2.
18689 Set input indexes to remap to outputs.
18692 @subsection Commands
18694 The @code{streamselect} and @code{astreamselect} filter supports the following
18699 Set input indexes to remap to outputs.
18702 @subsection Examples
18706 Select first 5 seconds 1st stream and rest of time 2nd stream:
18708 sendcmd='5.0 streamselect map 1',streamselect=inputs=2:map=0
18712 Same as above, but for audio:
18714 asendcmd='5.0 astreamselect map 1',astreamselect=inputs=2:map=0
18721 Draw subtitles on top of input video using the libass library.
18723 To enable compilation of this filter you need to configure FFmpeg with
18724 @code{--enable-libass}. This filter also requires a build with libavcodec and
18725 libavformat to convert the passed subtitles file to ASS (Advanced Substation
18726 Alpha) subtitles format.
18728 The filter accepts the following options:
18732 Set the filename of the subtitle file to read. It must be specified.
18734 @item original_size
18735 Specify the size of the original video, the video for which the ASS file
18736 was composed. For the syntax of this option, check the
18737 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
18738 Due to a misdesign in ASS aspect ratio arithmetic, this is necessary to
18739 correctly scale the fonts if the aspect ratio has been changed.
18742 Set a directory path containing fonts that can be used by the filter.
18743 These fonts will be used in addition to whatever the font provider uses.
18746 Process alpha channel, by default alpha channel is untouched.
18749 Set subtitles input character encoding. @code{subtitles} filter only. Only
18750 useful if not UTF-8.
18752 @item stream_index, si
18753 Set subtitles stream index. @code{subtitles} filter only.
18756 Override default style or script info parameters of the subtitles. It accepts a
18757 string containing ASS style format @code{KEY=VALUE} couples separated by ",".
18760 If the first key is not specified, it is assumed that the first value
18761 specifies the @option{filename}.
18763 For example, to render the file @file{sub.srt} on top of the input
18764 video, use the command:
18769 which is equivalent to:
18771 subtitles=filename=sub.srt
18774 To render the default subtitles stream from file @file{video.mkv}, use:
18776 subtitles=video.mkv
18779 To render the second subtitles stream from that file, use:
18781 subtitles=video.mkv:si=1
18784 To make the subtitles stream from @file{sub.srt} appear in 80% transparent blue
18785 @code{DejaVu Serif}, use:
18787 subtitles=sub.srt:force_style='Fontname=DejaVu Serif,PrimaryColour=&HCCFF0000'
18790 @section super2xsai
18792 Scale the input by 2x and smooth using the Super2xSaI (Scale and
18793 Interpolate) pixel art scaling algorithm.
18795 Useful for enlarging pixel art images without reducing sharpness.
18799 Swap two rectangular objects in video.
18801 This filter accepts the following options:
18811 Set 1st rect x coordinate.
18814 Set 1st rect y coordinate.
18817 Set 2nd rect x coordinate.
18820 Set 2nd rect y coordinate.
18822 All expressions are evaluated once for each frame.
18825 The all options are expressions containing the following constants:
18830 The input width and height.
18833 same as @var{w} / @var{h}
18836 input sample aspect ratio
18839 input display aspect ratio, it is the same as (@var{w} / @var{h}) * @var{sar}
18842 The number of the input frame, starting from 0.
18845 The timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
18848 the position in the file of the input frame, NAN if unknown
18855 Blend successive video frames.
18861 Apply telecine process to the video.
18863 This filter accepts the following options:
18872 The default value is @code{top}.
18876 A string of numbers representing the pulldown pattern you wish to apply.
18877 The default value is @code{23}.
18881 Some typical patterns:
18886 24p: 2332 (preferred)
18893 24p: 222222222223 ("Euro pulldown")
18898 @section thistogram
18900 Compute and draw a color distribution histogram for the input video across time.
18902 Unlike @ref{histogram} video filter which only shows histogram of single input frame
18903 at certain time, this filter shows also past histograms of number of frames defined
18904 by @code{width} option.
18906 The computed histogram is a representation of the color component
18907 distribution in an image.
18909 The filter accepts the following options:
18913 Set width of single color component output. Default value is @code{0}.
18914 Value of @code{0} means width will be picked from input video.
18915 This also set number of passed histograms to keep.
18916 Allowed range is [0, 8192].
18918 @item display_mode, d
18920 It accepts the following values:
18923 Per color component graphs are placed below each other.
18926 Per color component graphs are placed side by side.
18929 Presents information identical to that in the @code{parade}, except
18930 that the graphs representing color components are superimposed directly
18933 Default is @code{stack}.
18935 @item levels_mode, m
18936 Set mode. Can be either @code{linear}, or @code{logarithmic}.
18937 Default is @code{linear}.
18939 @item components, c
18940 Set what color components to display.
18941 Default is @code{7}.
18944 Set background opacity. Default is @code{0.9}.
18947 Show envelope. Default is disabled.
18950 Set envelope color. Default is @code{gold}.
18955 Available values for slide is:
18958 Draw new frame when right border is reached.
18961 Replace old columns with new ones.
18964 Scroll from right to left.
18967 Scroll from left to right.
18970 Draw single picture.
18973 Default is @code{replace}.
18978 Apply threshold effect to video stream.
18980 This filter needs four video streams to perform thresholding.
18981 First stream is stream we are filtering.
18982 Second stream is holding threshold values, third stream is holding min values,
18983 and last, fourth stream is holding max values.
18985 The filter accepts the following option:
18989 Set which planes will be processed, unprocessed planes will be copied.
18990 By default value 0xf, all planes will be processed.
18993 For example if first stream pixel's component value is less then threshold value
18994 of pixel component from 2nd threshold stream, third stream value will picked,
18995 otherwise fourth stream pixel component value will be picked.
18997 Using color source filter one can perform various types of thresholding:
18999 @subsection Examples
19003 Binary threshold, using gray color as threshold:
19005 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=black -f lavfi -i color=white -lavfi threshold output.avi
19009 Inverted binary threshold, using gray color as threshold:
19011 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -f lavfi -i color=black -lavfi threshold output.avi
19015 Truncate binary threshold, using gray color as threshold:
19017 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=gray -lavfi threshold output.avi
19021 Threshold to zero, using gray color as threshold:
19023 ffmpeg -i 320x240.avi -f lavfi -i color=gray -f lavfi -i color=white -i 320x240.avi -lavfi threshold output.avi
19027 Inverted threshold to zero, using gray color as threshold:
19029 ffmpeg -i 320x240.avi -f lavfi -i color=gray -i 320x240.avi -f lavfi -i color=white -lavfi threshold output.avi
19034 Select the most representative frame in a given sequence of consecutive frames.
19036 The filter accepts the following options:
19040 Set the frames batch size to analyze; in a set of @var{n} frames, the filter
19041 will pick one of them, and then handle the next batch of @var{n} frames until
19042 the end. Default is @code{100}.
19045 Since the filter keeps track of the whole frames sequence, a bigger @var{n}
19046 value will result in a higher memory usage, so a high value is not recommended.
19048 @subsection Examples
19052 Extract one picture each 50 frames:
19058 Complete example of a thumbnail creation with @command{ffmpeg}:
19060 ffmpeg -i in.avi -vf thumbnail,scale=300:200 -frames:v 1 out.png
19067 Tile several successive frames together.
19069 The @ref{untile} filter can do the reverse.
19071 The filter accepts the following options:
19076 Set the grid size (i.e. the number of lines and columns). For the syntax of
19077 this option, check the
19078 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19081 Set the maximum number of frames to render in the given area. It must be less
19082 than or equal to @var{w}x@var{h}. The default value is @code{0}, meaning all
19083 the area will be used.
19086 Set the outer border margin in pixels.
19089 Set the inner border thickness (i.e. the number of pixels between frames). For
19090 more advanced padding options (such as having different values for the edges),
19091 refer to the pad video filter.
19094 Specify the color of the unused area. For the syntax of this option, check the
19095 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
19096 The default value of @var{color} is "black".
19099 Set the number of frames to overlap when tiling several successive frames together.
19100 The value must be between @code{0} and @var{nb_frames - 1}.
19103 Set the number of frames to initially be empty before displaying first output frame.
19104 This controls how soon will one get first output frame.
19105 The value must be between @code{0} and @var{nb_frames - 1}.
19108 @subsection Examples
19112 Produce 8x8 PNG tiles of all keyframes (@option{-skip_frame nokey}) in a movie:
19114 ffmpeg -skip_frame nokey -i file.avi -vf 'scale=128:72,tile=8x8' -an -vsync 0 keyframes%03d.png
19116 The @option{-vsync 0} is necessary to prevent @command{ffmpeg} from
19117 duplicating each output frame to accommodate the originally detected frame
19121 Display @code{5} pictures in an area of @code{3x2} frames,
19122 with @code{7} pixels between them, and @code{2} pixels of initial margin, using
19123 mixed flat and named options:
19125 tile=3x2:nb_frames=5:padding=7:margin=2
19129 @section tinterlace
19131 Perform various types of temporal field interlacing.
19133 Frames are counted starting from 1, so the first input frame is
19136 The filter accepts the following options:
19141 Specify the mode of the interlacing. This option can also be specified
19142 as a value alone. See below for a list of values for this option.
19144 Available values are:
19148 Move odd frames into the upper field, even into the lower field,
19149 generating a double height frame at half frame rate.
19153 Frame 1 Frame 2 Frame 3 Frame 4
19155 11111 22222 33333 44444
19156 11111 22222 33333 44444
19157 11111 22222 33333 44444
19158 11111 22222 33333 44444
19172 Only output odd frames, even frames are dropped, generating a frame with
19173 unchanged height at half frame rate.
19178 Frame 1 Frame 2 Frame 3 Frame 4
19180 11111 22222 33333 44444
19181 11111 22222 33333 44444
19182 11111 22222 33333 44444
19183 11111 22222 33333 44444
19193 Only output even frames, odd frames are dropped, generating a frame with
19194 unchanged height at half frame rate.
19199 Frame 1 Frame 2 Frame 3 Frame 4
19201 11111 22222 33333 44444
19202 11111 22222 33333 44444
19203 11111 22222 33333 44444
19204 11111 22222 33333 44444
19214 Expand each frame to full height, but pad alternate lines with black,
19215 generating a frame with double height at the same input frame rate.
19220 Frame 1 Frame 2 Frame 3 Frame 4
19222 11111 22222 33333 44444
19223 11111 22222 33333 44444
19224 11111 22222 33333 44444
19225 11111 22222 33333 44444
19228 11111 ..... 33333 .....
19229 ..... 22222 ..... 44444
19230 11111 ..... 33333 .....
19231 ..... 22222 ..... 44444
19232 11111 ..... 33333 .....
19233 ..... 22222 ..... 44444
19234 11111 ..... 33333 .....
19235 ..... 22222 ..... 44444
19239 @item interleave_top, 4
19240 Interleave the upper field from odd frames with the lower field from
19241 even frames, generating a frame with unchanged height at half frame rate.
19246 Frame 1 Frame 2 Frame 3 Frame 4
19248 11111<- 22222 33333<- 44444
19249 11111 22222<- 33333 44444<-
19250 11111<- 22222 33333<- 44444
19251 11111 22222<- 33333 44444<-
19261 @item interleave_bottom, 5
19262 Interleave the lower field from odd frames with the upper field from
19263 even frames, generating a frame with unchanged height at half frame rate.
19268 Frame 1 Frame 2 Frame 3 Frame 4
19270 11111 22222<- 33333 44444<-
19271 11111<- 22222 33333<- 44444
19272 11111 22222<- 33333 44444<-
19273 11111<- 22222 33333<- 44444
19283 @item interlacex2, 6
19284 Double frame rate with unchanged height. Frames are inserted each
19285 containing the second temporal field from the previous input frame and
19286 the first temporal field from the next input frame. This mode relies on
19287 the top_field_first flag. Useful for interlaced video displays with no
19288 field synchronisation.
19293 Frame 1 Frame 2 Frame 3 Frame 4
19295 11111 22222 33333 44444
19296 11111 22222 33333 44444
19297 11111 22222 33333 44444
19298 11111 22222 33333 44444
19301 11111 22222 22222 33333 33333 44444 44444
19302 11111 11111 22222 22222 33333 33333 44444
19303 11111 22222 22222 33333 33333 44444 44444
19304 11111 11111 22222 22222 33333 33333 44444
19309 Move odd frames into the upper field, even into the lower field,
19310 generating a double height frame at same frame rate.
19315 Frame 1 Frame 2 Frame 3 Frame 4
19317 11111 22222 33333 44444
19318 11111 22222 33333 44444
19319 11111 22222 33333 44444
19320 11111 22222 33333 44444
19323 11111 33333 33333 55555
19324 22222 22222 44444 44444
19325 11111 33333 33333 55555
19326 22222 22222 44444 44444
19327 11111 33333 33333 55555
19328 22222 22222 44444 44444
19329 11111 33333 33333 55555
19330 22222 22222 44444 44444
19335 Numeric values are deprecated but are accepted for backward
19336 compatibility reasons.
19338 Default mode is @code{merge}.
19341 Specify flags influencing the filter process.
19343 Available value for @var{flags} is:
19346 @item low_pass_filter, vlpf
19347 Enable linear vertical low-pass filtering in the filter.
19348 Vertical low-pass filtering is required when creating an interlaced
19349 destination from a progressive source which contains high-frequency
19350 vertical detail. Filtering will reduce interlace 'twitter' and Moire
19353 @item complex_filter, cvlpf
19354 Enable complex vertical low-pass filtering.
19355 This will slightly less reduce interlace 'twitter' and Moire
19356 patterning but better retain detail and subjective sharpness impression.
19359 Bypass already interlaced frames, only adjust the frame rate.
19362 Vertical low-pass filtering and bypassing already interlaced frames can only be
19363 enabled for @option{mode} @var{interleave_top} and @var{interleave_bottom}.
19368 Pick median pixels from several successive input video frames.
19370 The filter accepts the following options:
19374 Set radius of median filter.
19375 Default is 1. Allowed range is from 1 to 127.
19378 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
19381 Set median percentile. Default value is @code{0.5}.
19382 Default value of @code{0.5} will pick always median values, while @code{0} will pick
19383 minimum values, and @code{1} maximum values.
19386 @subsection Commands
19388 This filter supports all above options as @ref{commands}, excluding option @code{radius}.
19390 @section tmidequalizer
19392 Apply Temporal Midway Video Equalization effect.
19394 Midway Video Equalization adjusts a sequence of video frames to have the same
19395 histograms, while maintaining their dynamics as much as possible. It's
19396 useful for e.g. matching exposures from a video frames sequence.
19398 This filter accepts the following option:
19402 Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127.
19405 Set filtering sigma. Default is @code{0.5}. This controls strength of filtering.
19406 Setting this option to 0 effectively does nothing.
19409 Set which planes to process. Default is @code{15}, which is all available planes.
19414 Mix successive video frames.
19416 A description of the accepted options follows.
19420 The number of successive frames to mix. If unspecified, it defaults to 3.
19423 Specify weight of each input video frame.
19424 Each weight is separated by space. If number of weights is smaller than
19425 number of @var{frames} last specified weight will be used for all remaining
19429 Specify scale, if it is set it will be multiplied with sum
19430 of each weight multiplied with pixel values to give final destination
19431 pixel value. By default @var{scale} is auto scaled to sum of weights.
19434 @subsection Examples
19438 Average 7 successive frames:
19440 tmix=frames=7:weights="1 1 1 1 1 1 1"
19444 Apply simple temporal convolution:
19446 tmix=frames=3:weights="-1 3 -1"
19450 Similar as above but only showing temporal differences:
19452 tmix=frames=3:weights="-1 2 -1":scale=1
19458 Tone map colors from different dynamic ranges.
19460 This filter expects data in single precision floating point, as it needs to
19461 operate on (and can output) out-of-range values. Another filter, such as
19462 @ref{zscale}, is needed to convert the resulting frame to a usable format.
19464 The tonemapping algorithms implemented only work on linear light, so input
19465 data should be linearized beforehand (and possibly correctly tagged).
19468 ffmpeg -i INPUT -vf zscale=transfer=linear,tonemap=clip,zscale=transfer=bt709,format=yuv420p OUTPUT
19471 @subsection Options
19472 The filter accepts the following options.
19476 Set the tone map algorithm to use.
19478 Possible values are:
19481 Do not apply any tone map, only desaturate overbright pixels.
19484 Hard-clip any out-of-range values. Use it for perfect color accuracy for
19485 in-range values, while distorting out-of-range values.
19488 Stretch the entire reference gamut to a linear multiple of the display.
19491 Fit a logarithmic transfer between the tone curves.
19494 Preserve overall image brightness with a simple curve, using nonlinear
19495 contrast, which results in flattening details and degrading color accuracy.
19498 Preserve both dark and bright details better than @var{reinhard}, at the cost
19499 of slightly darkening everything. Use it when detail preservation is more
19500 important than color and brightness accuracy.
19503 Smoothly map out-of-range values, while retaining contrast and colors for
19504 in-range material as much as possible. Use it when color accuracy is more
19505 important than detail preservation.
19511 Tune the tone mapping algorithm.
19513 This affects the following algorithms:
19519 Specifies the scale factor to use while stretching.
19523 Specifies the exponent of the function.
19527 Specify an extra linear coefficient to multiply into the signal before clipping.
19531 Specify the local contrast coefficient at the display peak.
19532 Default to 0.5, which means that in-gamut values will be about half as bright
19539 Specify the transition point from linear to mobius transform. Every value
19540 below this point is guaranteed to be mapped 1:1. The higher the value, the
19541 more accurate the result will be, at the cost of losing bright details.
19542 Default to 0.3, which due to the steep initial slope still preserves in-range
19543 colors fairly accurately.
19547 Apply desaturation for highlights that exceed this level of brightness. The
19548 higher the parameter, the more color information will be preserved. This
19549 setting helps prevent unnaturally blown-out colors for super-highlights, by
19550 (smoothly) turning into white instead. This makes images feel more natural,
19551 at the cost of reducing information about out-of-range colors.
19553 The default of 2.0 is somewhat conservative and will mostly just apply to
19554 skies or directly sunlit surfaces. A setting of 0.0 disables this option.
19556 This option works only if the input frame has a supported color tag.
19559 Override signal/nominal/reference peak with this value. Useful when the
19560 embedded peak information in display metadata is not reliable or when tone
19561 mapping from a lower range to a higher range.
19566 Temporarily pad video frames.
19568 The filter accepts the following options:
19572 Specify number of delay frames before input video stream. Default is 0.
19575 Specify number of padding frames after input video stream.
19576 Set to -1 to pad indefinitely. Default is 0.
19579 Set kind of frames added to beginning of stream.
19580 Can be either @var{add} or @var{clone}.
19581 With @var{add} frames of solid-color are added.
19582 With @var{clone} frames are clones of first frame.
19583 Default is @var{add}.
19586 Set kind of frames added to end of stream.
19587 Can be either @var{add} or @var{clone}.
19588 With @var{add} frames of solid-color are added.
19589 With @var{clone} frames are clones of last frame.
19590 Default is @var{add}.
19592 @item start_duration, stop_duration
19593 Specify the duration of the start/stop delay. See
19594 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19595 for the accepted syntax.
19596 These options override @var{start} and @var{stop}. Default is 0.
19599 Specify the color of the padded area. For the syntax of this option,
19600 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
19601 manual,ffmpeg-utils}.
19603 The default value of @var{color} is "black".
19609 Transpose rows with columns in the input video and optionally flip it.
19611 It accepts the following parameters:
19616 Specify the transposition direction.
19618 Can assume the following values:
19620 @item 0, 4, cclock_flip
19621 Rotate by 90 degrees counterclockwise and vertically flip (default), that is:
19629 Rotate by 90 degrees clockwise, that is:
19637 Rotate by 90 degrees counterclockwise, that is:
19644 @item 3, 7, clock_flip
19645 Rotate by 90 degrees clockwise and vertically flip, that is:
19653 For values between 4-7, the transposition is only done if the input
19654 video geometry is portrait and not landscape. These values are
19655 deprecated, the @code{passthrough} option should be used instead.
19657 Numerical values are deprecated, and should be dropped in favor of
19658 symbolic constants.
19661 Do not apply the transposition if the input geometry matches the one
19662 specified by the specified value. It accepts the following values:
19665 Always apply transposition.
19667 Preserve portrait geometry (when @var{height} >= @var{width}).
19669 Preserve landscape geometry (when @var{width} >= @var{height}).
19672 Default value is @code{none}.
19675 For example to rotate by 90 degrees clockwise and preserve portrait
19678 transpose=dir=1:passthrough=portrait
19681 The command above can also be specified as:
19683 transpose=1:portrait
19686 @section transpose_npp
19688 Transpose rows with columns in the input video and optionally flip it.
19689 For more in depth examples see the @ref{transpose} video filter, which shares mostly the same options.
19691 It accepts the following parameters:
19696 Specify the transposition direction.
19698 Can assume the following values:
19701 Rotate by 90 degrees counterclockwise and vertically flip. (default)
19704 Rotate by 90 degrees clockwise.
19707 Rotate by 90 degrees counterclockwise.
19710 Rotate by 90 degrees clockwise and vertically flip.
19714 Do not apply the transposition if the input geometry matches the one
19715 specified by the specified value. It accepts the following values:
19718 Always apply transposition. (default)
19720 Preserve portrait geometry (when @var{height} >= @var{width}).
19722 Preserve landscape geometry (when @var{width} >= @var{height}).
19728 Trim the input so that the output contains one continuous subpart of the input.
19730 It accepts the following parameters:
19733 Specify the time of the start of the kept section, i.e. the frame with the
19734 timestamp @var{start} will be the first frame in the output.
19737 Specify the time of the first frame that will be dropped, i.e. the frame
19738 immediately preceding the one with the timestamp @var{end} will be the last
19739 frame in the output.
19742 This is the same as @var{start}, except this option sets the start timestamp
19743 in timebase units instead of seconds.
19746 This is the same as @var{end}, except this option sets the end timestamp
19747 in timebase units instead of seconds.
19750 The maximum duration of the output in seconds.
19753 The number of the first frame that should be passed to the output.
19756 The number of the first frame that should be dropped.
19759 @option{start}, @option{end}, and @option{duration} are expressed as time
19760 duration specifications; see
19761 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
19762 for the accepted syntax.
19764 Note that the first two sets of the start/end options and the @option{duration}
19765 option look at the frame timestamp, while the _frame variants simply count the
19766 frames that pass through the filter. Also note that this filter does not modify
19767 the timestamps. If you wish for the output timestamps to start at zero, insert a
19768 setpts filter after the trim filter.
19770 If multiple start or end options are set, this filter tries to be greedy and
19771 keep all the frames that match at least one of the specified constraints. To keep
19772 only the part that matches all the constraints at once, chain multiple trim
19775 The defaults are such that all the input is kept. So it is possible to set e.g.
19776 just the end values to keep everything before the specified time.
19781 Drop everything except the second minute of input:
19783 ffmpeg -i INPUT -vf trim=60:120
19787 Keep only the first second:
19789 ffmpeg -i INPUT -vf trim=duration=1
19794 @section unpremultiply
19795 Apply alpha unpremultiply effect to input video stream using first plane
19796 of second stream as alpha.
19798 Both streams must have same dimensions and same pixel format.
19800 The filter accepts the following option:
19804 Set which planes will be processed, unprocessed planes will be copied.
19805 By default value 0xf, all planes will be processed.
19807 If the format has 1 or 2 components, then luma is bit 0.
19808 If the format has 3 or 4 components:
19809 for RGB formats bit 0 is green, bit 1 is blue and bit 2 is red;
19810 for YUV formats bit 0 is luma, bit 1 is chroma-U and bit 2 is chroma-V.
19811 If present, the alpha channel is always the last bit.
19814 Do not require 2nd input for processing, instead use alpha plane from input stream.
19820 Sharpen or blur the input video.
19822 It accepts the following parameters:
19825 @item luma_msize_x, lx
19826 Set the luma matrix horizontal size. It must be an odd integer between
19827 3 and 23. The default value is 5.
19829 @item luma_msize_y, ly
19830 Set the luma matrix vertical size. It must be an odd integer between 3
19831 and 23. The default value is 5.
19833 @item luma_amount, la
19834 Set the luma effect strength. It must be a floating point number, reasonable
19835 values lay between -1.5 and 1.5.
19837 Negative values will blur the input video, while positive values will
19838 sharpen it, a value of zero will disable the effect.
19840 Default value is 1.0.
19842 @item chroma_msize_x, cx
19843 Set the chroma matrix horizontal size. It must be an odd integer
19844 between 3 and 23. The default value is 5.
19846 @item chroma_msize_y, cy
19847 Set the chroma matrix vertical size. It must be an odd integer
19848 between 3 and 23. The default value is 5.
19850 @item chroma_amount, ca
19851 Set the chroma effect strength. It must be a floating point number, reasonable
19852 values lay between -1.5 and 1.5.
19854 Negative values will blur the input video, while positive values will
19855 sharpen it, a value of zero will disable the effect.
19857 Default value is 0.0.
19861 All parameters are optional and default to the equivalent of the
19862 string '5:5:1.0:5:5:0.0'.
19864 @subsection Examples
19868 Apply strong luma sharpen effect:
19870 unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5
19874 Apply a strong blur of both luma and chroma parameters:
19876 unsharp=7:7:-2:7:7:-2
19883 Decompose a video made of tiled images into the individual images.
19885 The frame rate of the output video is the frame rate of the input video
19886 multiplied by the number of tiles.
19888 This filter does the reverse of @ref{tile}.
19890 The filter accepts the following options:
19895 Set the grid size (i.e. the number of lines and columns). For the syntax of
19896 this option, check the
19897 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
19900 @subsection Examples
19904 Produce a 1-second video from a still image file made of 25 frames stacked
19905 vertically, like an analogic film reel:
19907 ffmpeg -r 1 -i image.jpg -vf untile=1x25 movie.mkv
19913 Apply ultra slow/simple postprocessing filter that compresses and decompresses
19914 the image at several (or - in the case of @option{quality} level @code{8} - all)
19915 shifts and average the results.
19917 The way this differs from the behavior of spp is that uspp actually encodes &
19918 decodes each case with libavcodec Snow, whereas spp uses a simplified intra only 8x8
19919 DCT similar to MJPEG.
19921 The filter accepts the following options:
19925 Set quality. This option defines the number of levels for averaging. It accepts
19926 an integer in the range 0-8. If set to @code{0}, the filter will have no
19927 effect. A value of @code{8} means the higher quality. For each increment of
19928 that value the speed drops by a factor of approximately 2. Default value is
19932 Force a constant quantization parameter. If not set, the filter will use the QP
19933 from the video stream (if available).
19938 Convert 360 videos between various formats.
19940 The filter accepts the following options:
19946 Set format of the input/output video.
19954 Equirectangular projection.
19959 Cubemap with 3x2/6x1/1x6 layout.
19961 Format specific options:
19966 Set padding proportion for the input/output cubemap. Values in decimals.
19973 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)
19976 Default value is @b{@samp{0}}.
19977 Maximum value is @b{@samp{0.1}}.
19981 Set fixed padding for the input/output cubemap. Values in pixels.
19983 Default value is @b{@samp{0}}. If greater than zero it overrides other padding options.
19987 Set order of faces for the input/output cubemap. Choose one direction for each position.
19989 Designation of directions:
20005 Default value is @b{@samp{rludfb}}.
20009 Set rotation of faces for the input/output cubemap. Choose one angle for each position.
20011 Designation of angles:
20014 0 degrees clockwise
20016 90 degrees clockwise
20018 180 degrees clockwise
20020 270 degrees clockwise
20023 Default value is @b{@samp{000000}}.
20027 Equi-Angular Cubemap.
20034 Format specific options:
20039 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20041 If diagonal field of view is set it overrides horizontal and vertical field of view.
20046 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20048 If diagonal field of view is set it overrides horizontal and vertical field of view.
20054 Format specific options:
20059 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20061 If diagonal field of view is set it overrides horizontal and vertical field of view.
20066 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20068 If diagonal field of view is set it overrides horizontal and vertical field of view.
20074 Facebook's 360 formats.
20077 Stereographic format.
20079 Format specific options:
20084 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20086 If diagonal field of view is set it overrides horizontal and vertical field of view.
20091 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20093 If diagonal field of view is set it overrides horizontal and vertical field of view.
20100 Ball format, gives significant distortion toward the back.
20103 Hammer-Aitoff map projection format.
20106 Sinusoidal map projection format.
20109 Fisheye projection.
20111 Format specific options:
20116 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20118 If diagonal field of view is set it overrides horizontal and vertical field of view.
20123 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20125 If diagonal field of view is set it overrides horizontal and vertical field of view.
20129 Pannini projection.
20131 Format specific options:
20134 Set output pannini parameter.
20137 Set input pannini parameter.
20141 Cylindrical projection.
20143 Format specific options:
20148 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20150 If diagonal field of view is set it overrides horizontal and vertical field of view.
20155 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20157 If diagonal field of view is set it overrides horizontal and vertical field of view.
20161 Perspective projection. @i{(output only)}
20163 Format specific options:
20166 Set perspective parameter.
20170 Tetrahedron projection.
20173 Truncated square pyramid projection.
20177 Half equirectangular projection.
20182 Format specific options:
20187 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20189 If diagonal field of view is set it overrides horizontal and vertical field of view.
20194 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20196 If diagonal field of view is set it overrides horizontal and vertical field of view.
20200 Orthographic format.
20202 Format specific options:
20207 Set output horizontal/vertical/diagonal field of view. Values in degrees.
20209 If diagonal field of view is set it overrides horizontal and vertical field of view.
20214 Set input horizontal/vertical/diagonal field of view. Values in degrees.
20216 If diagonal field of view is set it overrides horizontal and vertical field of view.
20220 Octahedron projection.
20224 Set interpolation method.@*
20225 @i{Note: more complex interpolation methods require much more memory to run.}
20235 Bilinear interpolation.
20237 Lagrange9 interpolation.
20240 Bicubic interpolation.
20243 Lanczos interpolation.
20246 Spline16 interpolation.
20249 Gaussian interpolation.
20251 Mitchell interpolation.
20254 Default value is @b{@samp{line}}.
20258 Set the output video resolution.
20260 Default resolution depends on formats.
20264 Set the input/output stereo format.
20275 Default value is @b{@samp{2d}} for input and output format.
20280 Set rotation for the output video. Values in degrees.
20283 Set rotation order for the output video. Choose one item for each position.
20294 Default value is @b{@samp{ypr}}.
20299 Flip the output video horizontally(swaps left-right)/vertically(swaps up-down)/in-depth(swaps back-forward). Boolean values.
20303 Set if input video is flipped horizontally/vertically. Boolean values.
20306 Set if input video is transposed. Boolean value, by default disabled.
20309 Set if output video needs to be transposed. Boolean value, by default disabled.
20312 Build mask in alpha plane for all unmapped pixels by marking them fully transparent. Boolean value, by default disabled.
20315 @subsection Examples
20319 Convert equirectangular video to cubemap with 3x2 layout and 1% padding using bicubic interpolation:
20321 ffmpeg -i input.mkv -vf v360=e:c3x2:cubic:out_pad=0.01 output.mkv
20324 Extract back view of Equi-Angular Cubemap:
20326 ffmpeg -i input.mkv -vf v360=eac:flat:yaw=180 output.mkv
20329 Convert transposed and horizontally flipped Equi-Angular Cubemap in side-by-side stereo format to equirectangular top-bottom stereo format:
20331 v360=eac:equirect:in_stereo=sbs:in_trans=1:ih_flip=1:out_stereo=tb
20335 @subsection Commands
20337 This filter supports subset of above options as @ref{commands}.
20339 @section vaguedenoiser
20341 Apply a wavelet based denoiser.
20343 It transforms each frame from the video input into the wavelet domain,
20344 using Cohen-Daubechies-Feauveau 9/7. Then it applies some filtering to
20345 the obtained coefficients. It does an inverse wavelet transform after.
20346 Due to wavelet properties, it should give a nice smoothed result, and
20347 reduced noise, without blurring picture features.
20349 This filter accepts the following options:
20353 The filtering strength. The higher, the more filtered the video will be.
20354 Hard thresholding can use a higher threshold than soft thresholding
20355 before the video looks overfiltered. Default value is 2.
20358 The filtering method the filter will use.
20360 It accepts the following values:
20363 All values under the threshold will be zeroed.
20366 All values under the threshold will be zeroed. All values above will be
20367 reduced by the threshold.
20370 Scales or nullifies coefficients - intermediary between (more) soft and
20371 (less) hard thresholding.
20374 Default is garrote.
20377 Number of times, the wavelet will decompose the picture. Picture can't
20378 be decomposed beyond a particular point (typically, 8 for a 640x480
20379 frame - as 2^9 = 512 > 480). Valid values are integers between 1 and 32. Default value is 6.
20382 Partial of full denoising (limited coefficients shrinking), from 0 to 100. Default value is 85.
20385 A list of the planes to process. By default all planes are processed.
20388 The threshold type the filter will use.
20390 It accepts the following values:
20393 Threshold used is same for all decompositions.
20396 Threshold used depends also on each decomposition coefficients.
20399 Default is universal.
20402 @section vectorscope
20404 Display 2 color component values in the two dimensional graph (which is called
20407 This filter accepts the following options:
20411 Set vectorscope mode.
20413 It accepts the following values:
20417 Gray values are displayed on graph, higher brightness means more pixels have
20418 same component color value on location in graph. This is the default mode.
20421 Gray values are displayed on graph. Surrounding pixels values which are not
20422 present in video frame are drawn in gradient of 2 color components which are
20423 set by option @code{x} and @code{y}. The 3rd color component is static.
20426 Actual color components values present in video frame are displayed on graph.
20429 Similar as color2 but higher frequency of same values @code{x} and @code{y}
20430 on graph increases value of another color component, which is luminance by
20431 default values of @code{x} and @code{y}.
20434 Actual colors present in video frame are displayed on graph. If two different
20435 colors map to same position on graph then color with higher value of component
20436 not present in graph is picked.
20439 Gray values are displayed on graph. Similar to @code{color} but with 3rd color
20440 component picked from radial gradient.
20444 Set which color component will be represented on X-axis. Default is @code{1}.
20447 Set which color component will be represented on Y-axis. Default is @code{2}.
20450 Set intensity, used by modes: gray, color, color3 and color5 for increasing brightness
20451 of color component which represents frequency of (X, Y) location in graph.
20456 No envelope, this is default.
20459 Instant envelope, even darkest single pixel will be clearly highlighted.
20462 Hold maximum and minimum values presented in graph over time. This way you
20463 can still spot out of range values without constantly looking at vectorscope.
20466 Peak and instant envelope combined together.
20470 Set what kind of graticule to draw.
20479 Set graticule opacity.
20482 Set graticule flags.
20486 Draw graticule for white point.
20489 Draw graticule for black point.
20492 Draw color points short names.
20496 Set background opacity.
20498 @item lthreshold, l
20499 Set low threshold for color component not represented on X or Y axis.
20500 Values lower than this value will be ignored. Default is 0.
20501 Note this value is multiplied with actual max possible value one pixel component
20502 can have. So for 8-bit input and low threshold value of 0.1 actual threshold
20505 @item hthreshold, h
20506 Set high threshold for color component not represented on X or Y axis.
20507 Values higher than this value will be ignored. Default is 1.
20508 Note this value is multiplied with actual max possible value one pixel component
20509 can have. So for 8-bit input and high threshold value of 0.9 actual threshold
20510 is 0.9 * 255 = 230.
20512 @item colorspace, c
20513 Set what kind of colorspace to use when drawing graticule.
20523 Set color tint for gray/tint vectorscope mode. By default both options are zero.
20524 This means no tint, and output will remain gray.
20527 @anchor{vidstabdetect}
20528 @section vidstabdetect
20530 Analyze video stabilization/deshaking. Perform pass 1 of 2, see
20531 @ref{vidstabtransform} for pass 2.
20533 This filter generates a file with relative translation and rotation
20534 transform information about subsequent frames, which is then used by
20535 the @ref{vidstabtransform} filter.
20537 To enable compilation of this filter you need to configure FFmpeg with
20538 @code{--enable-libvidstab}.
20540 This filter accepts the following options:
20544 Set the path to the file used to write the transforms information.
20545 Default value is @file{transforms.trf}.
20548 Set how shaky the video is and how quick the camera is. It accepts an
20549 integer in the range 1-10, a value of 1 means little shakiness, a
20550 value of 10 means strong shakiness. Default value is 5.
20553 Set the accuracy of the detection process. It must be a value in the
20554 range 1-15. A value of 1 means low accuracy, a value of 15 means high
20555 accuracy. Default value is 15.
20558 Set stepsize of the search process. The region around minimum is
20559 scanned with 1 pixel resolution. Default value is 6.
20562 Set minimum contrast. Below this value a local measurement field is
20563 discarded. Must be a floating point value in the range 0-1. Default
20567 Set reference frame number for tripod mode.
20569 If enabled, the motion of the frames is compared to a reference frame
20570 in the filtered stream, identified by the specified number. The idea
20571 is to compensate all movements in a more-or-less static scene and keep
20572 the camera view absolutely still.
20574 If set to 0, it is disabled. The frames are counted starting from 1.
20577 Show fields and transforms in the resulting frames. It accepts an
20578 integer in the range 0-2. Default value is 0, which disables any
20582 @subsection Examples
20586 Use default values:
20592 Analyze strongly shaky movie and put the results in file
20593 @file{mytransforms.trf}:
20595 vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf"
20599 Visualize the result of internal transformations in the resulting
20602 vidstabdetect=show=1
20606 Analyze a video with medium shakiness using @command{ffmpeg}:
20608 ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi
20612 @anchor{vidstabtransform}
20613 @section vidstabtransform
20615 Video stabilization/deshaking: pass 2 of 2,
20616 see @ref{vidstabdetect} for pass 1.
20618 Read a file with transform information for each frame and
20619 apply/compensate them. Together with the @ref{vidstabdetect}
20620 filter this can be used to deshake videos. See also
20621 @url{http://public.hronopik.de/vid.stab}. It is important to also use
20622 the @ref{unsharp} filter, see below.
20624 To enable compilation of this filter you need to configure FFmpeg with
20625 @code{--enable-libvidstab}.
20627 @subsection Options
20631 Set path to the file used to read the transforms. Default value is
20632 @file{transforms.trf}.
20635 Set the number of frames (value*2 + 1) used for lowpass filtering the
20636 camera movements. Default value is 10.
20638 For example a number of 10 means that 21 frames are used (10 in the
20639 past and 10 in the future) to smoothen the motion in the video. A
20640 larger value leads to a smoother video, but limits the acceleration of
20641 the camera (pan/tilt movements). 0 is a special case where a static
20642 camera is simulated.
20645 Set the camera path optimization algorithm.
20647 Accepted values are:
20650 gaussian kernel low-pass filter on camera motion (default)
20652 averaging on transformations
20656 Set maximal number of pixels to translate frames. Default value is -1,
20660 Set maximal angle in radians (degree*PI/180) to rotate frames. Default
20661 value is -1, meaning no limit.
20664 Specify how to deal with borders that may be visible due to movement
20667 Available values are:
20670 keep image information from previous frame (default)
20672 fill the border black
20676 Invert transforms if set to 1. Default value is 0.
20679 Consider transforms as relative to previous frame if set to 1,
20680 absolute if set to 0. Default value is 0.
20683 Set percentage to zoom. A positive value will result in a zoom-in
20684 effect, a negative value in a zoom-out effect. Default value is 0 (no
20688 Set optimal zooming to avoid borders.
20690 Accepted values are:
20695 optimal static zoom value is determined (only very strong movements
20696 will lead to visible borders) (default)
20698 optimal adaptive zoom value is determined (no borders will be
20699 visible), see @option{zoomspeed}
20702 Note that the value given at zoom is added to the one calculated here.
20705 Set percent to zoom maximally each frame (enabled when
20706 @option{optzoom} is set to 2). Range is from 0 to 5, default value is
20710 Specify type of interpolation.
20712 Available values are:
20717 linear only horizontal
20719 linear in both directions (default)
20721 cubic in both directions (slow)
20725 Enable virtual tripod mode if set to 1, which is equivalent to
20726 @code{relative=0:smoothing=0}. Default value is 0.
20728 Use also @code{tripod} option of @ref{vidstabdetect}.
20731 Increase log verbosity if set to 1. Also the detected global motions
20732 are written to the temporary file @file{global_motions.trf}. Default
20736 @subsection Examples
20740 Use @command{ffmpeg} for a typical stabilization with default values:
20742 ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
20745 Note the use of the @ref{unsharp} filter which is always recommended.
20748 Zoom in a bit more and load transform data from a given file:
20750 vidstabtransform=zoom=5:input="mytransforms.trf"
20754 Smoothen the video even more:
20756 vidstabtransform=smoothing=30
20762 Flip the input video vertically.
20764 For example, to vertically flip a video with @command{ffmpeg}:
20766 ffmpeg -i in.avi -vf "vflip" out.avi
20771 Detect variable frame rate video.
20773 This filter tries to detect if the input is variable or constant frame rate.
20775 At end it will output number of frames detected as having variable delta pts,
20776 and ones with constant delta pts.
20777 If there was frames with variable delta, than it will also show min, max and
20778 average delta encountered.
20782 Boost or alter saturation.
20784 The filter accepts the following options:
20787 Set strength of boost if positive value or strength of alter if negative value.
20788 Default is 0. Allowed range is from -2 to 2.
20791 Set the red balance. Default is 1. Allowed range is from -10 to 10.
20794 Set the green balance. Default is 1. Allowed range is from -10 to 10.
20797 Set the blue balance. Default is 1. Allowed range is from -10 to 10.
20800 Set the red luma coefficient.
20803 Set the green luma coefficient.
20806 Set the blue luma coefficient.
20809 If @code{intensity} is negative and this is set to 1, colors will change,
20810 otherwise colors will be less saturated, more towards gray.
20813 @subsection Commands
20815 This filter supports the all above options as @ref{commands}.
20820 Make or reverse a natural vignetting effect.
20822 The filter accepts the following options:
20826 Set lens angle expression as a number of radians.
20828 The value is clipped in the @code{[0,PI/2]} range.
20830 Default value: @code{"PI/5"}
20834 Set center coordinates expressions. Respectively @code{"w/2"} and @code{"h/2"}
20838 Set forward/backward mode.
20840 Available modes are:
20843 The larger the distance from the central point, the darker the image becomes.
20846 The larger the distance from the central point, the brighter the image becomes.
20847 This can be used to reverse a vignette effect, though there is no automatic
20848 detection to extract the lens @option{angle} and other settings (yet). It can
20849 also be used to create a burning effect.
20852 Default value is @samp{forward}.
20855 Set evaluation mode for the expressions (@option{angle}, @option{x0}, @option{y0}).
20857 It accepts the following values:
20860 Evaluate expressions only once during the filter initialization.
20863 Evaluate expressions for each incoming frame. This is way slower than the
20864 @samp{init} mode since it requires all the scalers to be re-computed, but it
20865 allows advanced dynamic expressions.
20868 Default value is @samp{init}.
20871 Set dithering to reduce the circular banding effects. Default is @code{1}
20875 Set vignette aspect. This setting allows one to adjust the shape of the vignette.
20876 Setting this value to the SAR of the input will make a rectangular vignetting
20877 following the dimensions of the video.
20879 Default is @code{1/1}.
20882 @subsection Expressions
20884 The @option{alpha}, @option{x0} and @option{y0} expressions can contain the
20885 following parameters.
20890 input width and height
20893 the number of input frame, starting from 0
20896 the PTS (Presentation TimeStamp) time of the filtered video frame, expressed in
20897 @var{TB} units, NAN if undefined
20900 frame rate of the input video, NAN if the input frame rate is unknown
20903 the PTS (Presentation TimeStamp) of the filtered video frame,
20904 expressed in seconds, NAN if undefined
20907 time base of the input video
20911 @subsection Examples
20915 Apply simple strong vignetting effect:
20921 Make a flickering vignetting:
20923 vignette='PI/4+random(1)*PI/50':eval=frame
20928 @section vmafmotion
20930 Obtain the average VMAF motion score of a video.
20931 It is one of the component metrics of VMAF.
20933 The obtained average motion score is printed through the logging system.
20935 The filter accepts the following options:
20939 If specified, the filter will use the named file to save the motion score of
20940 each frame with respect to the previous frame.
20941 When filename equals "-" the data is sent to standard output.
20946 ffmpeg -i ref.mpg -vf vmafmotion -f null -
20950 Stack input videos vertically.
20952 All streams must be of same pixel format and of same width.
20954 Note that this filter is faster than using @ref{overlay} and @ref{pad} filter
20955 to create same output.
20957 The filter accepts the following options:
20961 Set number of input streams. Default is 2.
20964 If set to 1, force the output to terminate when the shortest input
20965 terminates. Default value is 0.
20970 Deinterlace the input video ("w3fdif" stands for "Weston 3 Field
20971 Deinterlacing Filter").
20973 Based on the process described by Martin Weston for BBC R&D, and
20974 implemented based on the de-interlace algorithm written by Jim
20975 Easterbrook for BBC R&D, the Weston 3 field deinterlacing filter
20976 uses filter coefficients calculated by BBC R&D.
20978 This filter uses field-dominance information in frame to decide which
20979 of each pair of fields to place first in the output.
20980 If it gets it wrong use @ref{setfield} filter before @code{w3fdif} filter.
20982 There are two sets of filter coefficients, so called "simple"
20983 and "complex". Which set of filter coefficients is used can
20984 be set by passing an optional parameter:
20988 Set the interlacing filter coefficients. Accepts one of the following values:
20992 Simple filter coefficient set.
20994 More-complex filter coefficient set.
20996 Default value is @samp{complex}.
20999 The interlacing mode to adopt. It accepts one of the following values:
21003 Output one frame for each frame.
21005 Output one frame for each field.
21008 The default value is @code{field}.
21011 The picture field parity assumed for the input interlaced video. It accepts one
21012 of the following values:
21016 Assume the top field is first.
21018 Assume the bottom field is first.
21020 Enable automatic detection of field parity.
21023 The default value is @code{auto}.
21024 If the interlacing is unknown or the decoder does not export this information,
21025 top field first will be assumed.
21028 Specify which frames to deinterlace. Accepts one of the following values:
21032 Deinterlace all frames,
21034 Only deinterlace frames marked as interlaced.
21037 Default value is @samp{all}.
21040 @subsection Commands
21041 This filter supports same @ref{commands} as options.
21044 Video waveform monitor.
21046 The waveform monitor plots color component intensity. By default luminance
21047 only. Each column of the waveform corresponds to a column of pixels in the
21050 It accepts the following options:
21054 Can be either @code{row}, or @code{column}. Default is @code{column}.
21055 In row mode, the graph on the left side represents color component value 0 and
21056 the right side represents value = 255. In column mode, the top side represents
21057 color component value = 0 and bottom side represents value = 255.
21060 Set intensity. Smaller values are useful to find out how many values of the same
21061 luminance are distributed across input rows/columns.
21062 Default value is @code{0.04}. Allowed range is [0, 1].
21065 Set mirroring mode. @code{0} means unmirrored, @code{1} means mirrored.
21066 In mirrored mode, higher values will be represented on the left
21067 side for @code{row} mode and at the top for @code{column} mode. Default is
21068 @code{1} (mirrored).
21072 It accepts the following values:
21075 Presents information identical to that in the @code{parade}, except
21076 that the graphs representing color components are superimposed directly
21079 This display mode makes it easier to spot relative differences or similarities
21080 in overlapping areas of the color components that are supposed to be identical,
21081 such as neutral whites, grays, or blacks.
21084 Display separate graph for the color components side by side in
21085 @code{row} mode or one below the other in @code{column} mode.
21088 Display separate graph for the color components side by side in
21089 @code{column} mode or one below the other in @code{row} mode.
21091 Using this display mode makes it easy to spot color casts in the highlights
21092 and shadows of an image, by comparing the contours of the top and the bottom
21093 graphs of each waveform. Since whites, grays, and blacks are characterized
21094 by exactly equal amounts of red, green, and blue, neutral areas of the picture
21095 should display three waveforms of roughly equal width/height. If not, the
21096 correction is easy to perform by making level adjustments the three waveforms.
21098 Default is @code{stack}.
21100 @item components, c
21101 Set which color components to display. Default is 1, which means only luminance
21102 or red color component if input is in RGB colorspace. If is set for example to
21103 7 it will display all 3 (if) available color components.
21108 No envelope, this is default.
21111 Instant envelope, minimum and maximum values presented in graph will be easily
21112 visible even with small @code{step} value.
21115 Hold minimum and maximum values presented in graph across time. This way you
21116 can still spot out of range values without constantly looking at waveforms.
21119 Peak and instant envelope combined together.
21125 No filtering, this is default.
21128 Luma and chroma combined together.
21131 Similar as above, but shows difference between blue and red chroma.
21134 Similar as above, but use different colors.
21137 Similar as above, but again with different colors.
21140 Displays only chroma.
21143 Displays actual color value on waveform.
21146 Similar as above, but with luma showing frequency of chroma values.
21150 Set which graticule to display.
21154 Do not display graticule.
21157 Display green graticule showing legal broadcast ranges.
21160 Display orange graticule showing legal broadcast ranges.
21163 Display invert graticule showing legal broadcast ranges.
21167 Set graticule opacity.
21170 Set graticule flags.
21174 Draw numbers above lines. By default enabled.
21177 Draw dots instead of lines.
21181 Set scale used for displaying graticule.
21188 Default is digital.
21191 Set background opacity.
21195 Set tint for output.
21196 Only used with lowpass filter and when display is not overlay and input
21197 pixel formats are not RGB.
21200 @section weave, doubleweave
21202 The @code{weave} takes a field-based video input and join
21203 each two sequential fields into single frame, producing a new double
21204 height clip with half the frame rate and half the frame count.
21206 The @code{doubleweave} works same as @code{weave} but without
21207 halving frame rate and frame count.
21209 It accepts the following option:
21213 Set first field. Available values are:
21217 Set the frame as top-field-first.
21220 Set the frame as bottom-field-first.
21224 @subsection Examples
21228 Interlace video using @ref{select} and @ref{separatefields} filter:
21230 separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
21235 Apply the xBR high-quality magnification filter which is designed for pixel
21236 art. It follows a set of edge-detection rules, see
21237 @url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
21239 It accepts the following option:
21243 Set the scaling dimension: @code{2} for @code{2xBR}, @code{3} for
21244 @code{3xBR} and @code{4} for @code{4xBR}.
21245 Default is @code{3}.
21250 Apply cross fade from one input video stream to another input video stream.
21251 The cross fade is applied for specified duration.
21253 The filter accepts the following options:
21257 Set one of available transition effects:
21305 Default transition effect is fade.
21308 Set cross fade duration in seconds.
21309 Default duration is 1 second.
21312 Set cross fade start relative to first input stream in seconds.
21313 Default offset is 0.
21316 Set expression for custom transition effect.
21318 The expressions can use the following variables and functions:
21323 The coordinates of the current sample.
21327 The width and height of the image.
21330 Progress of transition effect.
21333 Currently processed plane.
21336 Return value of first input at current location and plane.
21339 Return value of second input at current location and plane.
21345 Return the value of the pixel at location (@var{x},@var{y}) of the
21346 first/second/third/fourth component of first input.
21352 Return the value of the pixel at location (@var{x},@var{y}) of the
21353 first/second/third/fourth component of second input.
21357 @subsection Examples
21361 Cross fade from one input video to another input video, with fade transition and duration of transition
21362 of 2 seconds starting at offset of 5 seconds:
21364 ffmpeg -i first.mp4 -i second.mp4 -filter_complex xfade=transition=fade:duration=2:offset=5 output.mp4
21369 Pick median pixels from several input videos.
21371 The filter accepts the following options:
21375 Set number of inputs.
21376 Default is 3. Allowed range is from 3 to 255.
21377 If number of inputs is even number, than result will be mean value between two median values.
21380 Set which planes to filter. Default value is @code{15}, by which all planes are processed.
21383 Set median percentile. Default value is @code{0.5}.
21384 Default value of @code{0.5} will pick always median values, while @code{0} will pick
21385 minimum values, and @code{1} maximum values.
21388 @subsection Commands
21390 This filter supports all above options as @ref{commands}, excluding option @code{inputs}.
21393 Stack video inputs into custom layout.
21395 All streams must be of same pixel format.
21397 The filter accepts the following options:
21401 Set number of input streams. Default is 2.
21404 Specify layout of inputs.
21405 This option requires the desired layout configuration to be explicitly set by the user.
21406 This sets position of each video input in output. Each input
21407 is separated by '|'.
21408 The first number represents the column, and the second number represents the row.
21409 Numbers start at 0 and are separated by '_'. Optionally one can use wX and hX,
21410 where X is video input from which to take width or height.
21411 Multiple values can be used when separated by '+'. In such
21412 case values are summed together.
21414 Note that if inputs are of different sizes gaps may appear, as not all of
21415 the output video frame will be filled. Similarly, videos can overlap each
21416 other if their position doesn't leave enough space for the full frame of
21419 For 2 inputs, a default layout of @code{0_0|w0_0} is set. In all other cases,
21420 a layout must be set by the user.
21423 If set to 1, force the output to terminate when the shortest input
21424 terminates. Default value is 0.
21427 If set to valid color, all unused pixels will be filled with that color.
21428 By default fill is set to none, so it is disabled.
21431 @subsection Examples
21435 Display 4 inputs into 2x2 grid.
21439 input1(0, 0) | input3(w0, 0)
21440 input2(0, h0) | input4(w0, h0)
21444 xstack=inputs=4:layout=0_0|0_h0|w0_0|w0_h0
21447 Note that if inputs are of different sizes, gaps or overlaps may occur.
21450 Display 4 inputs into 1x4 grid.
21457 input4(0, h0+h1+h2)
21461 xstack=inputs=4:layout=0_0|0_h0|0_h0+h1|0_h0+h1+h2
21464 Note that if inputs are of different widths, unused space will appear.
21467 Display 9 inputs into 3x3 grid.
21471 input1(0, 0) | input4(w0, 0) | input7(w0+w3, 0)
21472 input2(0, h0) | input5(w0, h0) | input8(w0+w3, h0)
21473 input3(0, h0+h1) | input6(w0, h0+h1) | input9(w0+w3, h0+h1)
21477 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
21480 Note that if inputs are of different sizes, gaps or overlaps may occur.
21483 Display 16 inputs into 4x4 grid.
21487 input1(0, 0) | input5(w0, 0) | input9 (w0+w4, 0) | input13(w0+w4+w8, 0)
21488 input2(0, h0) | input6(w0, h0) | input10(w0+w4, h0) | input14(w0+w4+w8, h0)
21489 input3(0, h0+h1) | input7(w0, h0+h1) | input11(w0+w4, h0+h1) | input15(w0+w4+w8, h0+h1)
21490 input4(0, h0+h1+h2)| input8(w0, h0+h1+h2)| input12(w0+w4, h0+h1+h2)| input16(w0+w4+w8, h0+h1+h2)
21494 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|
21495 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
21498 Note that if inputs are of different sizes, gaps or overlaps may occur.
21505 Deinterlace the input video ("yadif" means "yet another deinterlacing
21508 It accepts the following parameters:
21514 The interlacing mode to adopt. It accepts one of the following values:
21517 @item 0, send_frame
21518 Output one frame for each frame.
21519 @item 1, send_field
21520 Output one frame for each field.
21521 @item 2, send_frame_nospatial
21522 Like @code{send_frame}, but it skips the spatial interlacing check.
21523 @item 3, send_field_nospatial
21524 Like @code{send_field}, but it skips the spatial interlacing check.
21527 The default value is @code{send_frame}.
21530 The picture field parity assumed for the input interlaced video. It accepts one
21531 of the following values:
21535 Assume the top field is first.
21537 Assume the bottom field is first.
21539 Enable automatic detection of field parity.
21542 The default value is @code{auto}.
21543 If the interlacing is unknown or the decoder does not export this information,
21544 top field first will be assumed.
21547 Specify which frames to deinterlace. Accepts one of the following
21552 Deinterlace all frames.
21553 @item 1, interlaced
21554 Only deinterlace frames marked as interlaced.
21557 The default value is @code{all}.
21560 @section yadif_cuda
21562 Deinterlace the input video using the @ref{yadif} algorithm, but implemented
21563 in CUDA so that it can work as part of a GPU accelerated pipeline with nvdec
21566 It accepts the following parameters:
21572 The interlacing mode to adopt. It accepts one of the following values:
21575 @item 0, send_frame
21576 Output one frame for each frame.
21577 @item 1, send_field
21578 Output one frame for each field.
21579 @item 2, send_frame_nospatial
21580 Like @code{send_frame}, but it skips the spatial interlacing check.
21581 @item 3, send_field_nospatial
21582 Like @code{send_field}, but it skips the spatial interlacing check.
21585 The default value is @code{send_frame}.
21588 The picture field parity assumed for the input interlaced video. It accepts one
21589 of the following values:
21593 Assume the top field is first.
21595 Assume the bottom field is first.
21597 Enable automatic detection of field parity.
21600 The default value is @code{auto}.
21601 If the interlacing is unknown or the decoder does not export this information,
21602 top field first will be assumed.
21605 Specify which frames to deinterlace. Accepts one of the following
21610 Deinterlace all frames.
21611 @item 1, interlaced
21612 Only deinterlace frames marked as interlaced.
21615 The default value is @code{all}.
21620 Apply blur filter while preserving edges ("yaepblur" means "yet another edge preserving blur filter").
21621 The algorithm is described in
21622 "J. S. Lee, Digital image enhancement and noise filtering by use of local statistics, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-2, 1980."
21624 It accepts the following parameters:
21628 Set the window radius. Default value is 3.
21631 Set which planes to filter. Default is only the first plane.
21634 Set blur strength. Default value is 128.
21637 @subsection Commands
21638 This filter supports same @ref{commands} as options.
21642 Apply Zoom & Pan effect.
21644 This filter accepts the following options:
21648 Set the zoom expression. Range is 1-10. Default is 1.
21652 Set the x and y expression. Default is 0.
21655 Set the duration expression in number of frames.
21656 This sets for how many number of frames effect will last for
21657 single input image.
21660 Set the output image size, default is 'hd720'.
21663 Set the output frame rate, default is '25'.
21666 Each expression can contain the following constants:
21685 Output frame count.
21688 The input timestamp expressed in seconds. It's NAN if the input timestamp is unknown.
21690 @item out_time, time, ot
21691 The output timestamp expressed in seconds.
21695 Last calculated 'x' and 'y' position from 'x' and 'y' expression
21696 for current input frame.
21700 'x' and 'y' of last output frame of previous input frame or 0 when there was
21701 not yet such frame (first input frame).
21704 Last calculated zoom from 'z' expression for current input frame.
21707 Last calculated zoom of last output frame of previous input frame.
21710 Number of output frames for current input frame. Calculated from 'd' expression
21711 for each input frame.
21714 number of output frames created for previous input frame
21717 Rational number: input width / input height
21720 sample aspect ratio
21723 display aspect ratio
21727 @subsection Examples
21731 Zoom in up to 1.5x and pan at same time to some spot near center of picture:
21733 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
21737 Zoom in up to 1.5x and pan always at center of picture:
21739 zoompan=z='min(zoom+0.0015,1.5)':d=700:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21743 Same as above but without pausing:
21745 zoompan=z='min(max(zoom,pzoom)+0.0015,1.5)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21749 Zoom in 2x into center of picture only for the first second of the input video:
21751 zoompan=z='if(between(in_time,0,1),2,1)':d=1:x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'
21758 Scale (resize) the input video, using the z.lib library:
21759 @url{https://github.com/sekrit-twc/zimg}. To enable compilation of this
21760 filter, you need to configure FFmpeg with @code{--enable-libzimg}.
21762 The zscale filter forces the output display aspect ratio to be the same
21763 as the input, by changing the output sample aspect ratio.
21765 If the input image format is different from the format requested by
21766 the next filter, the zscale filter will convert the input to the
21769 @subsection Options
21770 The filter accepts the following options.
21775 Set the output video dimension expression. Default value is the input
21778 If the @var{width} or @var{w} value is 0, the input width is used for
21779 the output. If the @var{height} or @var{h} value is 0, the input height
21780 is used for the output.
21782 If one and only one of the values is -n with n >= 1, the zscale filter
21783 will use a value that maintains the aspect ratio of the input image,
21784 calculated from the other specified dimension. After that it will,
21785 however, make sure that the calculated dimension is divisible by n and
21786 adjust the value if necessary.
21788 If both values are -n with n >= 1, the behavior will be identical to
21789 both values being set to 0 as previously detailed.
21791 See below for the list of accepted constants for use in the dimension
21795 Set the video size. For the syntax of this option, check the
21796 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
21799 Set the dither type.
21801 Possible values are:
21806 @item error_diffusion
21812 Set the resize filter type.
21814 Possible values are:
21824 Default is bilinear.
21827 Set the color range.
21829 Possible values are:
21836 Default is same as input.
21839 Set the color primaries.
21841 Possible values are:
21851 Default is same as input.
21854 Set the transfer characteristics.
21856 Possible values are:
21870 Default is same as input.
21873 Set the colorspace matrix.
21875 Possible value are:
21886 Default is same as input.
21889 Set the input color range.
21891 Possible values are:
21898 Default is same as input.
21900 @item primariesin, pin
21901 Set the input color primaries.
21903 Possible values are:
21913 Default is same as input.
21915 @item transferin, tin
21916 Set the input transfer characteristics.
21918 Possible values are:
21929 Default is same as input.
21931 @item matrixin, min
21932 Set the input colorspace matrix.
21934 Possible value are:
21946 Set the output chroma location.
21948 Possible values are:
21959 @item chromalin, cin
21960 Set the input chroma location.
21962 Possible values are:
21974 Set the nominal peak luminance.
21977 The values of the @option{w} and @option{h} options are expressions
21978 containing the following constants:
21983 The input width and height
21987 These are the same as @var{in_w} and @var{in_h}.
21991 The output (scaled) width and height
21995 These are the same as @var{out_w} and @var{out_h}
21998 The same as @var{iw} / @var{ih}
22001 input sample aspect ratio
22004 The input display aspect ratio. Calculated from @code{(iw / ih) * sar}.
22008 horizontal and vertical input chroma subsample values. For example for the
22009 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22013 horizontal and vertical output chroma subsample values. For example for the
22014 pixel format "yuv422p" @var{hsub} is 2 and @var{vsub} is 1.
22017 @subsection Commands
22019 This filter supports the following commands:
22023 Set the output video dimension expression.
22024 The command accepts the same syntax of the corresponding option.
22026 If the specified expression is not valid, it is kept at its current
22030 @c man end VIDEO FILTERS
22032 @chapter OpenCL Video Filters
22033 @c man begin OPENCL VIDEO FILTERS
22035 Below is a description of the currently available OpenCL video filters.
22037 To enable compilation of these filters you need to configure FFmpeg with
22038 @code{--enable-opencl}.
22040 Running OpenCL filters requires you to initialize a hardware device and to pass that device to all filters in any filter graph.
22043 @item -init_hw_device opencl[=@var{name}][:@var{device}[,@var{key=value}...]]
22044 Initialise a new hardware device of type @var{opencl} called @var{name}, using the
22045 given device parameters.
22047 @item -filter_hw_device @var{name}
22048 Pass the hardware device called @var{name} to all filters in any filter graph.
22052 For more detailed information see @url{https://www.ffmpeg.org/ffmpeg.html#Advanced-Video-options}
22056 Example of choosing the first device on the second platform and running avgblur_opencl filter with default parameters on it.
22058 -init_hw_device opencl=gpu:1.0 -filter_hw_device gpu -i INPUT -vf "hwupload, avgblur_opencl, hwdownload" OUTPUT
22062 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.
22064 @section avgblur_opencl
22066 Apply average blur filter.
22068 The filter accepts the following options:
22072 Set horizontal radius size.
22073 Range is @code{[1, 1024]} and default value is @code{1}.
22076 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22079 Set vertical radius size. Range is @code{[1, 1024]} and default value is @code{0}. If zero, @code{sizeX} value will be used.
22082 @subsection Example
22086 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.
22088 -i INPUT -vf "hwupload, avgblur_opencl=3, hwdownload" OUTPUT
22092 @section boxblur_opencl
22094 Apply a boxblur algorithm to the input video.
22096 It accepts the following parameters:
22100 @item luma_radius, lr
22101 @item luma_power, lp
22102 @item chroma_radius, cr
22103 @item chroma_power, cp
22104 @item alpha_radius, ar
22105 @item alpha_power, ap
22109 A description of the accepted options follows.
22112 @item luma_radius, lr
22113 @item chroma_radius, cr
22114 @item alpha_radius, ar
22115 Set an expression for the box radius in pixels used for blurring the
22116 corresponding input plane.
22118 The radius value must be a non-negative number, and must not be
22119 greater than the value of the expression @code{min(w,h)/2} for the
22120 luma and alpha planes, and of @code{min(cw,ch)/2} for the chroma
22123 Default value for @option{luma_radius} is "2". If not specified,
22124 @option{chroma_radius} and @option{alpha_radius} default to the
22125 corresponding value set for @option{luma_radius}.
22127 The expressions can contain the following constants:
22131 The input width and height in pixels.
22135 The input chroma image width and height in pixels.
22139 The horizontal and vertical chroma subsample values. For example, for the
22140 pixel format "yuv422p", @var{hsub} is 2 and @var{vsub} is 1.
22143 @item luma_power, lp
22144 @item chroma_power, cp
22145 @item alpha_power, ap
22146 Specify how many times the boxblur filter is applied to the
22147 corresponding plane.
22149 Default value for @option{luma_power} is 2. If not specified,
22150 @option{chroma_power} and @option{alpha_power} default to the
22151 corresponding value set for @option{luma_power}.
22153 A value of 0 will disable the effect.
22156 @subsection Examples
22158 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.
22162 Apply a boxblur filter with the luma, chroma, and alpha radius
22163 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.
22165 -i INPUT -vf "hwupload, boxblur_opencl=luma_radius=2:luma_power=3, hwdownload" OUTPUT
22166 -i INPUT -vf "hwupload, boxblur_opencl=2:3, hwdownload" OUTPUT
22170 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.
22172 For the luma plane, a 2x2 box radius will be run once.
22174 For the chroma plane, a 4x4 box radius will be run 5 times.
22176 For the alpha plane, a 3x3 box radius will be run 7 times.
22178 -i INPUT -vf "hwupload, boxblur_opencl=2:1:4:5:3:7, hwdownload" OUTPUT
22182 @section colorkey_opencl
22183 RGB colorspace color keying.
22185 The filter accepts the following options:
22189 The color which will be replaced with transparency.
22192 Similarity percentage with the key color.
22194 0.01 matches only the exact key color, while 1.0 matches everything.
22199 0.0 makes pixels either fully transparent, or not transparent at all.
22201 Higher values result in semi-transparent pixels, with a higher transparency
22202 the more similar the pixels color is to the key color.
22205 @subsection Examples
22209 Make every semi-green pixel in the input transparent with some slight blending:
22211 -i INPUT -vf "hwupload, colorkey_opencl=green:0.3:0.1, hwdownload" OUTPUT
22215 @section convolution_opencl
22217 Apply convolution of 3x3, 5x5, 7x7 matrix.
22219 The filter accepts the following options:
22226 Set matrix for each plane.
22227 Matrix is sequence of 9, 25 or 49 signed numbers.
22228 Default value for each plane is @code{0 0 0 0 1 0 0 0 0}.
22234 Set multiplier for calculated value for each plane.
22235 If unset or 0, it will be sum of all matrix elements.
22236 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{1.0}.
22242 Set bias for each plane. This value is added to the result of the multiplication.
22243 Useful for making the overall image brighter or darker.
22244 The option value must be a float number greater or equal to @code{0.0}. Default value is @code{0.0}.
22248 @subsection Examples
22254 -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
22260 -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
22264 Apply edge enhance:
22266 -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
22272 -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
22276 Apply laplacian edge detector which includes diagonals:
22278 -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
22284 -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
22288 @section erosion_opencl
22290 Apply erosion effect to the video.
22292 This filter replaces the pixel by the local(3x3) minimum.
22294 It accepts the following options:
22301 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22302 If @code{0}, plane will remain unchanged.
22305 Flag which specifies the pixel to refer to.
22306 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22308 Flags to local 3x3 coordinates region centered on @code{x}:
22317 @subsection Example
22321 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.
22323 -i INPUT -vf "hwupload, erosion_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22327 @section deshake_opencl
22328 Feature-point based video stabilization filter.
22330 The filter accepts the following options:
22334 Simulates a tripod by preventing any camera movement whatsoever from the original frame. Defaults to @code{0}.
22337 Whether or not additional debug info should be displayed, both in the processed output and in the console.
22339 Note that in order to see console debug output you will also need to pass @code{-v verbose} to ffmpeg.
22341 Viewing point matches in the output video is only supported for RGB input.
22343 Defaults to @code{0}.
22345 @item adaptive_crop
22346 Whether or not to do a tiny bit of cropping at the borders to cut down on the amount of mirrored pixels.
22348 Defaults to @code{1}.
22350 @item refine_features
22351 Whether or not feature points should be refined at a sub-pixel level.
22353 This can be turned off for a slight performance gain at the cost of precision.
22355 Defaults to @code{1}.
22357 @item smooth_strength
22358 The strength of the smoothing applied to the camera path from @code{0.0} to @code{1.0}.
22360 @code{1.0} is the maximum smoothing strength while values less than that result in less smoothing.
22362 @code{0.0} causes the filter to adaptively choose a smoothing strength on a per-frame basis.
22364 Defaults to @code{0.0}.
22366 @item smooth_window_multiplier
22367 Controls the size of the smoothing window (the number of frames buffered to determine motion information from).
22369 The size of the smoothing window is determined by multiplying the framerate of the video by this number.
22371 Acceptable values range from @code{0.1} to @code{10.0}.
22373 Larger values increase the amount of motion data available for determining how to smooth the camera path,
22374 potentially improving smoothness, but also increase latency and memory usage.
22376 Defaults to @code{2.0}.
22380 @subsection Examples
22384 Stabilize a video with a fixed, medium smoothing strength:
22386 -i INPUT -vf "hwupload, deshake_opencl=smooth_strength=0.5, hwdownload" OUTPUT
22390 Stabilize a video with debugging (both in console and in rendered video):
22392 -i INPUT -filter_complex "[0:v]format=rgba, hwupload, deshake_opencl=debug=1, hwdownload, format=rgba, format=yuv420p" -v verbose OUTPUT
22396 @section dilation_opencl
22398 Apply dilation effect to the video.
22400 This filter replaces the pixel by the local(3x3) maximum.
22402 It accepts the following options:
22409 Limit the maximum change for each plane. Range is @code{[0, 65535]} and default value is @code{65535}.
22410 If @code{0}, plane will remain unchanged.
22413 Flag which specifies the pixel to refer to.
22414 Range is @code{[0, 255]} and default value is @code{255}, i.e. all eight pixels are used.
22416 Flags to local 3x3 coordinates region centered on @code{x}:
22425 @subsection Example
22429 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.
22431 -i INPUT -vf "hwupload, dilation_opencl=30:40:50:coordinates=231, hwdownload" OUTPUT
22435 @section nlmeans_opencl
22437 Non-local Means denoise filter through OpenCL, this filter accepts same options as @ref{nlmeans}.
22439 @section overlay_opencl
22441 Overlay one video on top of another.
22443 It takes two inputs and has one output. The first input is the "main" video on which the second input is overlaid.
22444 This filter requires same memory layout for all the inputs. So, format conversion may be needed.
22446 The filter accepts the following options:
22451 Set the x coordinate of the overlaid video on the main video.
22452 Default value is @code{0}.
22455 Set the y coordinate of the overlaid video on the main video.
22456 Default value is @code{0}.
22460 @subsection Examples
22464 Overlay an image LOGO at the top-left corner of the INPUT video. Both inputs are yuv420p format.
22466 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuv420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22469 The inputs have same memory layout for color channels , the overlay has additional alpha plane, like INPUT is yuv420p, and the LOGO is yuva420p.
22471 -i INPUT -i LOGO -filter_complex "[0:v]hwupload[a], [1:v]format=yuva420p, hwupload[b], [a][b]overlay_opencl, hwdownload" OUTPUT
22476 @section pad_opencl
22478 Add paddings to the input image, and place the original input at the
22479 provided @var{x}, @var{y} coordinates.
22481 It accepts the following options:
22486 Specify an expression for the size of the output image with the
22487 paddings added. If the value for @var{width} or @var{height} is 0, the
22488 corresponding input size is used for the output.
22490 The @var{width} expression can reference the value set by the
22491 @var{height} expression, and vice versa.
22493 The default value of @var{width} and @var{height} is 0.
22497 Specify the offsets to place the input image at within the padded area,
22498 with respect to the top/left border of the output image.
22500 The @var{x} expression can reference the value set by the @var{y}
22501 expression, and vice versa.
22503 The default value of @var{x} and @var{y} is 0.
22505 If @var{x} or @var{y} evaluate to a negative number, they'll be changed
22506 so the input image is centered on the padded area.
22509 Specify the color of the padded area. For the syntax of this option,
22510 check the @ref{color syntax,,"Color" section in the ffmpeg-utils
22511 manual,ffmpeg-utils}.
22514 Pad to an aspect instead to a resolution.
22517 The value for the @var{width}, @var{height}, @var{x}, and @var{y}
22518 options are expressions containing the following constants:
22523 The input video width and height.
22527 These are the same as @var{in_w} and @var{in_h}.
22531 The output width and height (the size of the padded area), as
22532 specified by the @var{width} and @var{height} expressions.
22536 These are the same as @var{out_w} and @var{out_h}.
22540 The x and y offsets as specified by the @var{x} and @var{y}
22541 expressions, or NAN if not yet specified.
22544 same as @var{iw} / @var{ih}
22547 input sample aspect ratio
22550 input display aspect ratio, it is the same as (@var{iw} / @var{ih}) * @var{sar}
22553 @section prewitt_opencl
22555 Apply the Prewitt operator (@url{https://en.wikipedia.org/wiki/Prewitt_operator}) to input video stream.
22557 The filter accepts the following option:
22561 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22564 Set value which will be multiplied with filtered result.
22565 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22568 Set value which will be added to filtered result.
22569 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22572 @subsection Example
22576 Apply the Prewitt operator with scale set to 2 and delta set to 10.
22578 -i INPUT -vf "hwupload, prewitt_opencl=scale=2:delta=10, hwdownload" OUTPUT
22582 @anchor{program_opencl}
22583 @section program_opencl
22585 Filter video using an OpenCL program.
22590 OpenCL program source file.
22593 Kernel name in program.
22596 Number of inputs to the filter. Defaults to 1.
22599 Size of output frames. Defaults to the same as the first input.
22603 The @code{program_opencl} filter also supports the @ref{framesync} options.
22605 The program source file must contain a kernel function with the given name,
22606 which will be run once for each plane of the output. Each run on a plane
22607 gets enqueued as a separate 2D global NDRange with one work-item for each
22608 pixel to be generated. The global ID offset for each work-item is therefore
22609 the coordinates of a pixel in the destination image.
22611 The kernel function needs to take the following arguments:
22614 Destination image, @var{__write_only image2d_t}.
22616 This image will become the output; the kernel should write all of it.
22618 Frame index, @var{unsigned int}.
22620 This is a counter starting from zero and increasing by one for each frame.
22622 Source images, @var{__read_only image2d_t}.
22624 These are the most recent images on each input. The kernel may read from
22625 them to generate the output, but they can't be written to.
22632 Copy the input to the output (output must be the same size as the input).
22634 __kernel void copy(__write_only image2d_t destination,
22635 unsigned int index,
22636 __read_only image2d_t source)
22638 const sampler_t sampler = CLK_NORMALIZED_COORDS_FALSE;
22640 int2 location = (int2)(get_global_id(0), get_global_id(1));
22642 float4 value = read_imagef(source, sampler, location);
22644 write_imagef(destination, location, value);
22649 Apply a simple transformation, rotating the input by an amount increasing
22650 with the index counter. Pixel values are linearly interpolated by the
22651 sampler, and the output need not have the same dimensions as the input.
22653 __kernel void rotate_image(__write_only image2d_t dst,
22654 unsigned int index,
22655 __read_only image2d_t src)
22657 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22658 CLK_FILTER_LINEAR);
22660 float angle = (float)index / 100.0f;
22662 float2 dst_dim = convert_float2(get_image_dim(dst));
22663 float2 src_dim = convert_float2(get_image_dim(src));
22665 float2 dst_cen = dst_dim / 2.0f;
22666 float2 src_cen = src_dim / 2.0f;
22668 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22670 float2 dst_pos = convert_float2(dst_loc) - dst_cen;
22672 cos(angle) * dst_pos.x - sin(angle) * dst_pos.y,
22673 sin(angle) * dst_pos.x + cos(angle) * dst_pos.y
22675 src_pos = src_pos * src_dim / dst_dim;
22677 float2 src_loc = src_pos + src_cen;
22679 if (src_loc.x < 0.0f || src_loc.y < 0.0f ||
22680 src_loc.x > src_dim.x || src_loc.y > src_dim.y)
22681 write_imagef(dst, dst_loc, 0.5f);
22683 write_imagef(dst, dst_loc, read_imagef(src, sampler, src_loc));
22688 Blend two inputs together, with the amount of each input used varying
22689 with the index counter.
22691 __kernel void blend_images(__write_only image2d_t dst,
22692 unsigned int index,
22693 __read_only image2d_t src1,
22694 __read_only image2d_t src2)
22696 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
22697 CLK_FILTER_LINEAR);
22699 float blend = (cos((float)index / 50.0f) + 1.0f) / 2.0f;
22701 int2 dst_loc = (int2)(get_global_id(0), get_global_id(1));
22702 int2 src1_loc = dst_loc * get_image_dim(src1) / get_image_dim(dst);
22703 int2 src2_loc = dst_loc * get_image_dim(src2) / get_image_dim(dst);
22705 float4 val1 = read_imagef(src1, sampler, src1_loc);
22706 float4 val2 = read_imagef(src2, sampler, src2_loc);
22708 write_imagef(dst, dst_loc, val1 * blend + val2 * (1.0f - blend));
22714 @section roberts_opencl
22715 Apply the Roberts cross operator (@url{https://en.wikipedia.org/wiki/Roberts_cross}) to input video stream.
22717 The filter accepts the following option:
22721 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22724 Set value which will be multiplied with filtered result.
22725 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22728 Set value which will be added to filtered result.
22729 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22732 @subsection Example
22736 Apply the Roberts cross operator with scale set to 2 and delta set to 10
22738 -i INPUT -vf "hwupload, roberts_opencl=scale=2:delta=10, hwdownload" OUTPUT
22742 @section sobel_opencl
22744 Apply the Sobel operator (@url{https://en.wikipedia.org/wiki/Sobel_operator}) to input video stream.
22746 The filter accepts the following option:
22750 Set which planes to filter. Default value is @code{0xf}, by which all planes are processed.
22753 Set value which will be multiplied with filtered result.
22754 Range is @code{[0.0, 65535]} and default value is @code{1.0}.
22757 Set value which will be added to filtered result.
22758 Range is @code{[-65535, 65535]} and default value is @code{0.0}.
22761 @subsection Example
22765 Apply sobel operator with scale set to 2 and delta set to 10
22767 -i INPUT -vf "hwupload, sobel_opencl=scale=2:delta=10, hwdownload" OUTPUT
22771 @section tonemap_opencl
22773 Perform HDR(PQ/HLG) to SDR conversion with tone-mapping.
22775 It accepts the following parameters:
22779 Specify the tone-mapping operator to be used. Same as tonemap option in @ref{tonemap}.
22782 Tune the tone mapping algorithm. same as param option in @ref{tonemap}.
22785 Apply desaturation for highlights that exceed this level of brightness. The
22786 higher the parameter, the more color information will be preserved. This
22787 setting helps prevent unnaturally blown-out colors for super-highlights, by
22788 (smoothly) turning into white instead. This makes images feel more natural,
22789 at the cost of reducing information about out-of-range colors.
22791 The default value is 0.5, and the algorithm here is a little different from
22792 the cpu version tonemap currently. A setting of 0.0 disables this option.
22795 The tonemapping algorithm parameters is fine-tuned per each scene. And a threshold
22796 is used to detect whether the scene has changed or not. If the distance between
22797 the current frame average brightness and the current running average exceeds
22798 a threshold value, we would re-calculate scene average and peak brightness.
22799 The default value is 0.2.
22802 Specify the output pixel format.
22804 Currently supported formats are:
22811 Set the output color range.
22813 Possible values are:
22819 Default is same as input.
22822 Set the output color primaries.
22824 Possible values are:
22830 Default is same as input.
22833 Set the output transfer characteristics.
22835 Possible values are:
22844 Set the output colorspace matrix.
22846 Possible value are:
22852 Default is same as input.
22856 @subsection Example
22860 Convert HDR(PQ/HLG) video to bt2020-transfer-characteristic p010 format using linear operator.
22862 -i INPUT -vf "format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010,hwdownload,format=p010" OUTPUT
22866 @section unsharp_opencl
22868 Sharpen or blur the input video.
22870 It accepts the following parameters:
22873 @item luma_msize_x, lx
22874 Set the luma matrix horizontal size.
22875 Range is @code{[1, 23]} and default value is @code{5}.
22877 @item luma_msize_y, ly
22878 Set the luma matrix vertical size.
22879 Range is @code{[1, 23]} and default value is @code{5}.
22881 @item luma_amount, la
22882 Set the luma effect strength.
22883 Range is @code{[-10, 10]} and default value is @code{1.0}.
22885 Negative values will blur the input video, while positive values will
22886 sharpen it, a value of zero will disable the effect.
22888 @item chroma_msize_x, cx
22889 Set the chroma matrix horizontal size.
22890 Range is @code{[1, 23]} and default value is @code{5}.
22892 @item chroma_msize_y, cy
22893 Set the chroma matrix vertical size.
22894 Range is @code{[1, 23]} and default value is @code{5}.
22896 @item chroma_amount, ca
22897 Set the chroma effect strength.
22898 Range is @code{[-10, 10]} and default value is @code{0.0}.
22900 Negative values will blur the input video, while positive values will
22901 sharpen it, a value of zero will disable the effect.
22905 All parameters are optional and default to the equivalent of the
22906 string '5:5:1.0:5:5:0.0'.
22908 @subsection Examples
22912 Apply strong luma sharpen effect:
22914 -i INPUT -vf "hwupload, unsharp_opencl=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5, hwdownload" OUTPUT
22918 Apply a strong blur of both luma and chroma parameters:
22920 -i INPUT -vf "hwupload, unsharp_opencl=7:7:-2:7:7:-2, hwdownload" OUTPUT
22924 @section xfade_opencl
22926 Cross fade two videos with custom transition effect by using OpenCL.
22928 It accepts the following options:
22932 Set one of possible transition effects.
22936 Select custom transition effect, the actual transition description
22937 will be picked from source and kernel options.
22949 Default transition is fade.
22953 OpenCL program source file for custom transition.
22956 Set name of kernel to use for custom transition from program source file.
22959 Set duration of video transition.
22962 Set time of start of transition relative to first video.
22965 The program source file must contain a kernel function with the given name,
22966 which will be run once for each plane of the output. Each run on a plane
22967 gets enqueued as a separate 2D global NDRange with one work-item for each
22968 pixel to be generated. The global ID offset for each work-item is therefore
22969 the coordinates of a pixel in the destination image.
22971 The kernel function needs to take the following arguments:
22974 Destination image, @var{__write_only image2d_t}.
22976 This image will become the output; the kernel should write all of it.
22979 First Source image, @var{__read_only image2d_t}.
22980 Second Source image, @var{__read_only image2d_t}.
22982 These are the most recent images on each input. The kernel may read from
22983 them to generate the output, but they can't be written to.
22986 Transition progress, @var{float}. This value is always between 0 and 1 inclusive.
22993 Apply dots curtain transition effect:
22995 __kernel void blend_images(__write_only image2d_t dst,
22996 __read_only image2d_t src1,
22997 __read_only image2d_t src2,
23000 const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
23001 CLK_FILTER_LINEAR);
23002 int2 p = (int2)(get_global_id(0), get_global_id(1));
23003 float2 rp = (float2)(get_global_id(0), get_global_id(1));
23004 float2 dim = (float2)(get_image_dim(src1).x, get_image_dim(src1).y);
23007 float2 dots = (float2)(20.0, 20.0);
23008 float2 center = (float2)(0,0);
23011 float4 val1 = read_imagef(src1, sampler, p);
23012 float4 val2 = read_imagef(src2, sampler, p);
23013 bool next = distance(fract(rp * dots, &unused), (float2)(0.5, 0.5)) < (progress / distance(rp, center));
23015 write_imagef(dst, p, next ? val1 : val2);
23021 @c man end OPENCL VIDEO FILTERS
23023 @chapter VAAPI Video Filters
23024 @c man begin VAAPI VIDEO FILTERS
23026 VAAPI Video filters are usually used with VAAPI decoder and VAAPI encoder. Below is a description of VAAPI video filters.
23028 To enable compilation of these filters you need to configure FFmpeg with
23029 @code{--enable-vaapi}.
23031 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}
23033 @section tonemap_vaapi
23035 Perform HDR(High Dynamic Range) to SDR(Standard Dynamic Range) conversion with tone-mapping.
23036 It maps the dynamic range of HDR10 content to the SDR content.
23037 It currently only accepts HDR10 as input.
23039 It accepts the following parameters:
23043 Specify the output pixel format.
23045 Currently supported formats are:
23054 Set the output color primaries.
23056 Default is same as input.
23059 Set the output transfer characteristics.
23064 Set the output colorspace matrix.
23066 Default is same as input.
23070 @subsection Example
23074 Convert HDR(HDR10) video to bt2020-transfer-characteristic p010 format
23076 tonemap_vaapi=format=p010:t=bt2020-10
23080 @c man end VAAPI VIDEO FILTERS
23082 @chapter Video Sources
23083 @c man begin VIDEO SOURCES
23085 Below is a description of the currently available video sources.
23089 Buffer video frames, and make them available to the filter chain.
23091 This source is mainly intended for a programmatic use, in particular
23092 through the interface defined in @file{libavfilter/buffersrc.h}.
23094 It accepts the following parameters:
23099 Specify the size (width and height) of the buffered video frames. For the
23100 syntax of this option, check the
23101 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23104 The input video width.
23107 The input video height.
23110 A string representing the pixel format of the buffered video frames.
23111 It may be a number corresponding to a pixel format, or a pixel format
23115 Specify the timebase assumed by the timestamps of the buffered frames.
23118 Specify the frame rate expected for the video stream.
23120 @item pixel_aspect, sar
23121 The sample (pixel) aspect ratio of the input video.
23124 This option is deprecated and ignored. Prepend @code{sws_flags=@var{flags};}
23125 to the filtergraph description to specify swscale flags for automatically
23126 inserted scalers. See @ref{Filtergraph syntax}.
23128 @item hw_frames_ctx
23129 When using a hardware pixel format, this should be a reference to an
23130 AVHWFramesContext describing input frames.
23135 buffer=width=320:height=240:pix_fmt=yuv410p:time_base=1/24:sar=1
23138 will instruct the source to accept video frames with size 320x240 and
23139 with format "yuv410p", assuming 1/24 as the timestamps timebase and
23140 square pixels (1:1 sample aspect ratio).
23141 Since the pixel format with name "yuv410p" corresponds to the number 6
23142 (check the enum AVPixelFormat definition in @file{libavutil/pixfmt.h}),
23143 this example corresponds to:
23145 buffer=size=320x240:pixfmt=6:time_base=1/24:pixel_aspect=1/1
23148 Alternatively, the options can be specified as a flat string, but this
23149 syntax is deprecated:
23151 @var{width}:@var{height}:@var{pix_fmt}:@var{time_base.num}:@var{time_base.den}:@var{pixel_aspect.num}:@var{pixel_aspect.den}
23155 Create a pattern generated by an elementary cellular automaton.
23157 The initial state of the cellular automaton can be defined through the
23158 @option{filename} and @option{pattern} options. If such options are
23159 not specified an initial state is created randomly.
23161 At each new frame a new row in the video is filled with the result of
23162 the cellular automaton next generation. The behavior when the whole
23163 frame is filled is defined by the @option{scroll} option.
23165 This source accepts the following options:
23169 Read the initial cellular automaton state, i.e. the starting row, from
23170 the specified file.
23171 In the file, each non-whitespace character is considered an alive
23172 cell, a newline will terminate the row, and further characters in the
23173 file will be ignored.
23176 Read the initial cellular automaton state, i.e. the starting row, from
23177 the specified string.
23179 Each non-whitespace character in the string is considered an alive
23180 cell, a newline will terminate the row, and further characters in the
23181 string will be ignored.
23184 Set the video rate, that is the number of frames generated per second.
23187 @item random_fill_ratio, ratio
23188 Set the random fill ratio for the initial cellular automaton row. It
23189 is a floating point number value ranging from 0 to 1, defaults to
23192 This option is ignored when a file or a pattern is specified.
23194 @item random_seed, seed
23195 Set the seed for filling randomly the initial row, must be an integer
23196 included between 0 and UINT32_MAX. If not specified, or if explicitly
23197 set to -1, the filter will try to use a good random seed on a best
23201 Set the cellular automaton rule, it is a number ranging from 0 to 255.
23202 Default value is 110.
23205 Set the size of the output video. For the syntax of this option, check the
23206 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23208 If @option{filename} or @option{pattern} is specified, the size is set
23209 by default to the width of the specified initial state row, and the
23210 height is set to @var{width} * PHI.
23212 If @option{size} is set, it must contain the width of the specified
23213 pattern string, and the specified pattern will be centered in the
23216 If a filename or a pattern string is not specified, the size value
23217 defaults to "320x518" (used for a randomly generated initial state).
23220 If set to 1, scroll the output upward when all the rows in the output
23221 have been already filled. If set to 0, the new generated row will be
23222 written over the top row just after the bottom row is filled.
23225 @item start_full, full
23226 If set to 1, completely fill the output with generated rows before
23227 outputting the first frame.
23228 This is the default behavior, for disabling set the value to 0.
23231 If set to 1, stitch the left and right row edges together.
23232 This is the default behavior, for disabling set the value to 0.
23235 @subsection Examples
23239 Read the initial state from @file{pattern}, and specify an output of
23242 cellauto=f=pattern:s=200x400
23246 Generate a random initial row with a width of 200 cells, with a fill
23249 cellauto=ratio=2/3:s=200x200
23253 Create a pattern generated by rule 18 starting by a single alive cell
23254 centered on an initial row with width 100:
23256 cellauto=p=@@:s=100x400:full=0:rule=18
23260 Specify a more elaborated initial pattern:
23262 cellauto=p='@@@@ @@ @@@@':s=100x400:full=0:rule=18
23267 @anchor{coreimagesrc}
23268 @section coreimagesrc
23269 Video source generated on GPU using Apple's CoreImage API on OSX.
23271 This video source is a specialized version of the @ref{coreimage} video filter.
23272 Use a core image generator at the beginning of the applied filterchain to
23273 generate the content.
23275 The coreimagesrc video source accepts the following options:
23277 @item list_generators
23278 List all available generators along with all their respective options as well as
23279 possible minimum and maximum values along with the default values.
23281 list_generators=true
23285 Specify the size of the sourced video. For the syntax of this option, check the
23286 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23287 The default value is @code{320x240}.
23290 Specify the frame rate of the sourced video, as the number of frames
23291 generated per second. It has to be a string in the format
23292 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23293 number or a valid video frame rate abbreviation. The default value is
23297 Set the sample aspect ratio of the sourced video.
23300 Set the duration of the sourced video. See
23301 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23302 for the accepted syntax.
23304 If not specified, or the expressed duration is negative, the video is
23305 supposed to be generated forever.
23308 Additionally, all options of the @ref{coreimage} video filter are accepted.
23309 A complete filterchain can be used for further processing of the
23310 generated input without CPU-HOST transfer. See @ref{coreimage} documentation
23311 and examples for details.
23313 @subsection Examples
23318 Use CIQRCodeGenerator to create a QR code for the FFmpeg homepage,
23319 given as complete and escaped command-line for Apple's standard bash shell:
23321 ffmpeg -f lavfi -i coreimagesrc=s=100x100:filter=CIQRCodeGenerator@@inputMessage=https\\\\\://FFmpeg.org/@@inputCorrectionLevel=H -frames:v 1 QRCode.png
23323 This example is equivalent to the QRCode example of @ref{coreimage} without the
23324 need for a nullsrc video source.
23329 Generate several gradients.
23333 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23334 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23337 Set frame rate, expressed as number of frames per second. Default
23340 @item c0, c1, c2, c3, c4, c5, c6, c7
23341 Set 8 colors. Default values for colors is to pick random one.
23343 @item x0, y0, y0, y1
23344 Set gradient line source and destination points. If negative or out of range, random ones
23348 Set number of colors to use at once. Allowed range is from 2 to 8. Default value is 2.
23351 Set seed for picking gradient line points.
23354 Set the duration of the sourced video. See
23355 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23356 for the accepted syntax.
23358 If not specified, or the expressed duration is negative, the video is
23359 supposed to be generated forever.
23362 Set speed of gradients rotation.
23366 @section mandelbrot
23368 Generate a Mandelbrot set fractal, and progressively zoom towards the
23369 point specified with @var{start_x} and @var{start_y}.
23371 This source accepts the following options:
23376 Set the terminal pts value. Default value is 400.
23379 Set the terminal scale value.
23380 Must be a floating point value. Default value is 0.3.
23383 Set the inner coloring mode, that is the algorithm used to draw the
23384 Mandelbrot fractal internal region.
23386 It shall assume one of the following values:
23391 Show time until convergence.
23393 Set color based on point closest to the origin of the iterations.
23398 Default value is @var{mincol}.
23401 Set the bailout value. Default value is 10.0.
23404 Set the maximum of iterations performed by the rendering
23405 algorithm. Default value is 7189.
23408 Set outer coloring mode.
23409 It shall assume one of following values:
23411 @item iteration_count
23412 Set iteration count mode.
23413 @item normalized_iteration_count
23414 set normalized iteration count mode.
23416 Default value is @var{normalized_iteration_count}.
23419 Set frame rate, expressed as number of frames per second. Default
23423 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23424 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23427 Set the initial scale value. Default value is 3.0.
23430 Set the initial x position. Must be a floating point value between
23431 -100 and 100. Default value is -0.743643887037158704752191506114774.
23434 Set the initial y position. Must be a floating point value between
23435 -100 and 100. Default value is -0.131825904205311970493132056385139.
23440 Generate various test patterns, as generated by the MPlayer test filter.
23442 The size of the generated video is fixed, and is 256x256.
23443 This source is useful in particular for testing encoding features.
23445 This source accepts the following options:
23450 Specify the frame rate of the sourced video, as the number of frames
23451 generated per second. It has to be a string in the format
23452 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23453 number or a valid video frame rate abbreviation. The default value is
23457 Set the duration of the sourced video. See
23458 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23459 for the accepted syntax.
23461 If not specified, or the expressed duration is negative, the video is
23462 supposed to be generated forever.
23466 Set the number or the name of the test to perform. Supported tests are:
23480 @item max_frames, m
23481 Set the maximum number of frames generated for each test, default value is 30.
23485 Default value is "all", which will cycle through the list of all tests.
23490 mptestsrc=t=dc_luma
23493 will generate a "dc_luma" test pattern.
23495 @section frei0r_src
23497 Provide a frei0r source.
23499 To enable compilation of this filter you need to install the frei0r
23500 header and configure FFmpeg with @code{--enable-frei0r}.
23502 This source accepts the following parameters:
23507 The size of the video to generate. For the syntax of this option, check the
23508 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23511 The framerate of the generated video. It may be a string of the form
23512 @var{num}/@var{den} or a frame rate abbreviation.
23515 The name to the frei0r source to load. For more information regarding frei0r and
23516 how to set the parameters, read the @ref{frei0r} section in the video filters
23519 @item filter_params
23520 A '|'-separated list of parameters to pass to the frei0r source.
23524 For example, to generate a frei0r partik0l source with size 200x200
23525 and frame rate 10 which is overlaid on the overlay filter main input:
23527 frei0r_src=size=200x200:framerate=10:filter_name=partik0l:filter_params=1234 [overlay]; [in][overlay] overlay
23532 Generate a life pattern.
23534 This source is based on a generalization of John Conway's life game.
23536 The sourced input represents a life grid, each pixel represents a cell
23537 which can be in one of two possible states, alive or dead. Every cell
23538 interacts with its eight neighbours, which are the cells that are
23539 horizontally, vertically, or diagonally adjacent.
23541 At each interaction the grid evolves according to the adopted rule,
23542 which specifies the number of neighbor alive cells which will make a
23543 cell stay alive or born. The @option{rule} option allows one to specify
23546 This source accepts the following options:
23550 Set the file from which to read the initial grid state. In the file,
23551 each non-whitespace character is considered an alive cell, and newline
23552 is used to delimit the end of each row.
23554 If this option is not specified, the initial grid is generated
23558 Set the video rate, that is the number of frames generated per second.
23561 @item random_fill_ratio, ratio
23562 Set the random fill ratio for the initial random grid. It is a
23563 floating point number value ranging from 0 to 1, defaults to 1/PHI.
23564 It is ignored when a file is specified.
23566 @item random_seed, seed
23567 Set the seed for filling the initial random grid, must be an integer
23568 included between 0 and UINT32_MAX. If not specified, or if explicitly
23569 set to -1, the filter will try to use a good random seed on a best
23575 A rule can be specified with a code of the kind "S@var{NS}/B@var{NB}",
23576 where @var{NS} and @var{NB} are sequences of numbers in the range 0-8,
23577 @var{NS} specifies the number of alive neighbor cells which make a
23578 live cell stay alive, and @var{NB} the number of alive neighbor cells
23579 which make a dead cell to become alive (i.e. to "born").
23580 "s" and "b" can be used in place of "S" and "B", respectively.
23582 Alternatively a rule can be specified by an 18-bits integer. The 9
23583 high order bits are used to encode the next cell state if it is alive
23584 for each number of neighbor alive cells, the low order bits specify
23585 the rule for "borning" new cells. Higher order bits encode for an
23586 higher number of neighbor cells.
23587 For example the number 6153 = @code{(12<<9)+9} specifies a stay alive
23588 rule of 12 and a born rule of 9, which corresponds to "S23/B03".
23590 Default value is "S23/B3", which is the original Conway's game of life
23591 rule, and will keep a cell alive if it has 2 or 3 neighbor alive
23592 cells, and will born a new cell if there are three alive cells around
23596 Set the size of the output video. For the syntax of this option, check the
23597 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23599 If @option{filename} is specified, the size is set by default to the
23600 same size of the input file. If @option{size} is set, it must contain
23601 the size specified in the input file, and the initial grid defined in
23602 that file is centered in the larger resulting area.
23604 If a filename is not specified, the size value defaults to "320x240"
23605 (used for a randomly generated initial grid).
23608 If set to 1, stitch the left and right grid edges together, and the
23609 top and bottom edges also. Defaults to 1.
23612 Set cell mold speed. If set, a dead cell will go from @option{death_color} to
23613 @option{mold_color} with a step of @option{mold}. @option{mold} can have a
23614 value from 0 to 255.
23617 Set the color of living (or new born) cells.
23620 Set the color of dead cells. If @option{mold} is set, this is the first color
23621 used to represent a dead cell.
23624 Set mold color, for definitely dead and moldy cells.
23626 For the syntax of these 3 color options, check the @ref{color syntax,,"Color" section in the
23627 ffmpeg-utils manual,ffmpeg-utils}.
23630 @subsection Examples
23634 Read a grid from @file{pattern}, and center it on a grid of size
23637 life=f=pattern:s=300x300
23641 Generate a random grid of size 200x200, with a fill ratio of 2/3:
23643 life=ratio=2/3:s=200x200
23647 Specify a custom rule for evolving a randomly generated grid:
23653 Full example with slow death effect (mold) using @command{ffplay}:
23655 ffplay -f lavfi life=s=300x200:mold=10:r=60:ratio=0.1:death_color=#C83232:life_color=#00ff00,scale=1200:800:flags=16
23662 @anchor{haldclutsrc}
23665 @anchor{pal100bars}
23666 @anchor{rgbtestsrc}
23668 @anchor{smptehdbars}
23671 @anchor{yuvtestsrc}
23672 @section allrgb, allyuv, color, haldclutsrc, nullsrc, pal75bars, pal100bars, rgbtestsrc, smptebars, smptehdbars, testsrc, testsrc2, yuvtestsrc
23674 The @code{allrgb} source returns frames of size 4096x4096 of all rgb colors.
23676 The @code{allyuv} source returns frames of size 4096x4096 of all yuv colors.
23678 The @code{color} source provides an uniformly colored input.
23680 The @code{haldclutsrc} source provides an identity Hald CLUT. See also
23681 @ref{haldclut} filter.
23683 The @code{nullsrc} source returns unprocessed video frames. It is
23684 mainly useful to be employed in analysis / debugging tools, or as the
23685 source for filters which ignore the input data.
23687 The @code{pal75bars} source generates a color bars pattern, based on
23688 EBU PAL recommendations with 75% color levels.
23690 The @code{pal100bars} source generates a color bars pattern, based on
23691 EBU PAL recommendations with 100% color levels.
23693 The @code{rgbtestsrc} source generates an RGB test pattern useful for
23694 detecting RGB vs BGR issues. You should see a red, green and blue
23695 stripe from top to bottom.
23697 The @code{smptebars} source generates a color bars pattern, based on
23698 the SMPTE Engineering Guideline EG 1-1990.
23700 The @code{smptehdbars} source generates a color bars pattern, based on
23701 the SMPTE RP 219-2002.
23703 The @code{testsrc} source generates a test video pattern, showing a
23704 color pattern, a scrolling gradient and a timestamp. This is mainly
23705 intended for testing purposes.
23707 The @code{testsrc2} source is similar to testsrc, but supports more
23708 pixel formats instead of just @code{rgb24}. This allows using it as an
23709 input for other tests without requiring a format conversion.
23711 The @code{yuvtestsrc} source generates an YUV test pattern. You should
23712 see a y, cb and cr stripe from top to bottom.
23714 The sources accept the following parameters:
23719 Specify the level of the Hald CLUT, only available in the @code{haldclutsrc}
23720 source. A level of @code{N} generates a picture of @code{N*N*N} by @code{N*N*N}
23721 pixels to be used as identity matrix for 3D lookup tables. Each component is
23722 coded on a @code{1/(N*N)} scale.
23725 Specify the color of the source, only available in the @code{color}
23726 source. For the syntax of this option, check the
23727 @ref{color syntax,,"Color" section in the ffmpeg-utils manual,ffmpeg-utils}.
23730 Specify the size of the sourced video. For the syntax of this option, check the
23731 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23732 The default value is @code{320x240}.
23734 This option is not available with the @code{allrgb}, @code{allyuv}, and
23735 @code{haldclutsrc} filters.
23738 Specify the frame rate of the sourced video, as the number of frames
23739 generated per second. It has to be a string in the format
23740 @var{frame_rate_num}/@var{frame_rate_den}, an integer number, a floating point
23741 number or a valid video frame rate abbreviation. The default value is
23745 Set the duration of the sourced video. See
23746 @ref{time duration syntax,,the Time duration section in the ffmpeg-utils(1) manual,ffmpeg-utils}
23747 for the accepted syntax.
23749 If not specified, or the expressed duration is negative, the video is
23750 supposed to be generated forever.
23752 Since the frame rate is used as time base, all frames including the last one
23753 will have their full duration. If the specified duration is not a multiple
23754 of the frame duration, it will be rounded up.
23757 Set the sample aspect ratio of the sourced video.
23760 Specify the alpha (opacity) of the background, only available in the
23761 @code{testsrc2} source. The value must be between 0 (fully transparent) and
23762 255 (fully opaque, the default).
23765 Set the number of decimals to show in the timestamp, only available in the
23766 @code{testsrc} source.
23768 The displayed timestamp value will correspond to the original
23769 timestamp value multiplied by the power of 10 of the specified
23770 value. Default value is 0.
23773 @subsection Examples
23777 Generate a video with a duration of 5.3 seconds, with size
23778 176x144 and a frame rate of 10 frames per second:
23780 testsrc=duration=5.3:size=qcif:rate=10
23784 The following graph description will generate a red source
23785 with an opacity of 0.2, with size "qcif" and a frame rate of 10
23788 color=c=red@@0.2:s=qcif:r=10
23792 If the input content is to be ignored, @code{nullsrc} can be used. The
23793 following command generates noise in the luminance plane by employing
23794 the @code{geq} filter:
23796 nullsrc=s=256x256, geq=random(1)*255:128:128
23800 @subsection Commands
23802 The @code{color} source supports the following commands:
23806 Set the color of the created image. Accepts the same syntax of the
23807 corresponding @option{color} option.
23812 Generate video using an OpenCL program.
23817 OpenCL program source file.
23820 Kernel name in program.
23823 Size of frames to generate. This must be set.
23826 Pixel format to use for the generated frames. This must be set.
23829 Number of frames generated every second. Default value is '25'.
23833 For details of how the program loading works, see the @ref{program_opencl}
23840 Generate a colour ramp by setting pixel values from the position of the pixel
23841 in the output image. (Note that this will work with all pixel formats, but
23842 the generated output will not be the same.)
23844 __kernel void ramp(__write_only image2d_t dst,
23845 unsigned int index)
23847 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23850 val.xy = val.zw = convert_float2(loc) / convert_float2(get_image_dim(dst));
23852 write_imagef(dst, loc, val);
23857 Generate a Sierpinski carpet pattern, panning by a single pixel each frame.
23859 __kernel void sierpinski_carpet(__write_only image2d_t dst,
23860 unsigned int index)
23862 int2 loc = (int2)(get_global_id(0), get_global_id(1));
23864 float4 value = 0.0f;
23865 int x = loc.x + index;
23866 int y = loc.y + index;
23867 while (x > 0 || y > 0) {
23868 if (x % 3 == 1 && y % 3 == 1) {
23876 write_imagef(dst, loc, value);
23882 @section sierpinski
23884 Generate a Sierpinski carpet/triangle fractal, and randomly pan around.
23886 This source accepts the following options:
23890 Set frame size. For the syntax of this option, check the @ref{video size syntax,,"Video
23891 size" section in the ffmpeg-utils manual,ffmpeg-utils}. Default value is "640x480".
23894 Set frame rate, expressed as number of frames per second. Default
23898 Set seed which is used for random panning.
23901 Set max jump for single pan destination. Allowed range is from 1 to 10000.
23904 Set fractal type, can be default @code{carpet} or @code{triangle}.
23907 @c man end VIDEO SOURCES
23909 @chapter Video Sinks
23910 @c man begin VIDEO SINKS
23912 Below is a description of the currently available video sinks.
23914 @section buffersink
23916 Buffer video frames, and make them available to the end of the filter
23919 This sink is mainly intended for programmatic use, in particular
23920 through the interface defined in @file{libavfilter/buffersink.h}
23921 or the options system.
23923 It accepts a pointer to an AVBufferSinkContext structure, which
23924 defines the incoming buffers' formats, to be passed as the opaque
23925 parameter to @code{avfilter_init_filter} for initialization.
23929 Null video sink: do absolutely nothing with the input video. It is
23930 mainly useful as a template and for use in analysis / debugging
23933 @c man end VIDEO SINKS
23935 @chapter Multimedia Filters
23936 @c man begin MULTIMEDIA FILTERS
23938 Below is a description of the currently available multimedia filters.
23942 Convert input audio to a video output, displaying the audio bit scope.
23944 The filter accepts the following options:
23948 Set frame rate, expressed as number of frames per second. Default
23952 Specify the video size for the output. For the syntax of this option, check the
23953 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23954 Default value is @code{1024x256}.
23957 Specify list of colors separated by space or by '|' which will be used to
23958 draw channels. Unrecognized or missing colors will be replaced
23962 @section adrawgraph
23963 Draw a graph using input audio metadata.
23965 See @ref{drawgraph}
23967 @section agraphmonitor
23969 See @ref{graphmonitor}.
23971 @section ahistogram
23973 Convert input audio to a video output, displaying the volume histogram.
23975 The filter accepts the following options:
23979 Specify how histogram is calculated.
23981 It accepts the following values:
23984 Use single histogram for all channels.
23986 Use separate histogram for each channel.
23988 Default is @code{single}.
23991 Set frame rate, expressed as number of frames per second. Default
23995 Specify the video size for the output. For the syntax of this option, check the
23996 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
23997 Default value is @code{hd720}.
24002 It accepts the following values:
24013 reverse logarithmic
24015 Default is @code{log}.
24018 Set amplitude scale.
24020 It accepts the following values:
24027 Default is @code{log}.
24030 Set how much frames to accumulate in histogram.
24031 Default is 1. Setting this to -1 accumulates all frames.
24034 Set histogram ratio of window height.
24037 Set sonogram sliding.
24039 It accepts the following values:
24042 replace old rows with new ones.
24044 scroll from top to bottom.
24046 Default is @code{replace}.
24049 @section aphasemeter
24051 Measures phase of input audio, which is exported as metadata @code{lavfi.aphasemeter.phase},
24052 representing mean phase of current audio frame. A video output can also be produced and is
24053 enabled by default. The audio is passed through as first output.
24055 Audio will be rematrixed to stereo if it has a different channel layout. Phase value is in
24056 range @code{[-1, 1]} where @code{-1} means left and right channels are completely out of phase
24057 and @code{1} means channels are in phase.
24059 The filter accepts the following options, all related to its video output:
24063 Set the output frame rate. Default value is @code{25}.
24066 Set the video size for the output. For the syntax of this option, check the
24067 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24068 Default value is @code{800x400}.
24073 Specify the red, green, blue contrast. Default values are @code{2},
24074 @code{7} and @code{1}.
24075 Allowed range is @code{[0, 255]}.
24078 Set color which will be used for drawing median phase. If color is
24079 @code{none} which is default, no median phase value will be drawn.
24082 Enable video output. Default is enabled.
24085 @subsection phasing detection
24087 The filter also detects out of phase and mono sequences in stereo streams.
24088 It logs the sequence start, end and duration when it lasts longer or as long as the minimum set.
24090 The filter accepts the following options for this detection:
24094 Enable mono and out of phase detection. Default is disabled.
24097 Set phase tolerance for mono detection, in amplitude ratio. Default is @code{0}.
24098 Allowed range is @code{[0, 1]}.
24101 Set angle threshold for out of phase detection, in degree. Default is @code{170}.
24102 Allowed range is @code{[90, 180]}.
24105 Set mono or out of phase duration until notification, expressed in seconds. Default is @code{2}.
24108 @subsection Examples
24112 Complete example with @command{ffmpeg} to detect 1 second of mono with 0.001 phase tolerance:
24114 ffmpeg -i stereo.wav -af aphasemeter=video=0:phasing=1:duration=1:tolerance=0.001 -f null -
24118 @section avectorscope
24120 Convert input audio to a video output, representing the audio vector
24123 The filter is used to measure the difference between channels of stereo
24124 audio stream. A monaural signal, consisting of identical left and right
24125 signal, results in straight vertical line. Any stereo separation is visible
24126 as a deviation from this line, creating a Lissajous figure.
24127 If the straight (or deviation from it) but horizontal line appears this
24128 indicates that the left and right channels are out of phase.
24130 The filter accepts the following options:
24134 Set the vectorscope mode.
24136 Available values are:
24139 Lissajous rotated by 45 degrees.
24142 Same as above but not rotated.
24145 Shape resembling half of circle.
24148 Default value is @samp{lissajous}.
24151 Set the video size for the output. For the syntax of this option, check the
24152 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24153 Default value is @code{400x400}.
24156 Set the output frame rate. Default value is @code{25}.
24162 Specify the red, green, blue and alpha contrast. Default values are @code{40},
24163 @code{160}, @code{80} and @code{255}.
24164 Allowed range is @code{[0, 255]}.
24170 Specify the red, green, blue and alpha fade. Default values are @code{15},
24171 @code{10}, @code{5} and @code{5}.
24172 Allowed range is @code{[0, 255]}.
24175 Set the zoom factor. Default value is @code{1}. Allowed range is @code{[0, 10]}.
24176 Values lower than @var{1} will auto adjust zoom factor to maximal possible value.
24179 Set the vectorscope drawing mode.
24181 Available values are:
24184 Draw dot for each sample.
24187 Draw line between previous and current sample.
24190 Default value is @samp{dot}.
24193 Specify amplitude scale of audio samples.
24195 Available values are:
24211 Swap left channel axis with right channel axis.
24221 Mirror only x axis.
24224 Mirror only y axis.
24232 @subsection Examples
24236 Complete example using @command{ffplay}:
24238 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
24239 [a] avectorscope=zoom=1.3:rc=2:gc=200:bc=10:rf=1:gf=8:bf=7 [out0]'
24243 @section bench, abench
24245 Benchmark part of a filtergraph.
24247 The filter accepts the following options:
24251 Start or stop a timer.
24253 Available values are:
24256 Get the current time, set it as frame metadata (using the key
24257 @code{lavfi.bench.start_time}), and forward the frame to the next filter.
24260 Get the current time and fetch the @code{lavfi.bench.start_time} metadata from
24261 the input frame metadata to get the time difference. Time difference, average,
24262 maximum and minimum time (respectively @code{t}, @code{avg}, @code{max} and
24263 @code{min}) are then printed. The timestamps are expressed in seconds.
24267 @subsection Examples
24271 Benchmark @ref{selectivecolor} filter:
24273 bench=start,selectivecolor=reds=-.2 .12 -.49,bench=stop
24279 Concatenate audio and video streams, joining them together one after the
24282 The filter works on segments of synchronized video and audio streams. All
24283 segments must have the same number of streams of each type, and that will
24284 also be the number of streams at output.
24286 The filter accepts the following options:
24291 Set the number of segments. Default is 2.
24294 Set the number of output video streams, that is also the number of video
24295 streams in each segment. Default is 1.
24298 Set the number of output audio streams, that is also the number of audio
24299 streams in each segment. Default is 0.
24302 Activate unsafe mode: do not fail if segments have a different format.
24306 The filter has @var{v}+@var{a} outputs: first @var{v} video outputs, then
24307 @var{a} audio outputs.
24309 There are @var{n}x(@var{v}+@var{a}) inputs: first the inputs for the first
24310 segment, in the same order as the outputs, then the inputs for the second
24313 Related streams do not always have exactly the same duration, for various
24314 reasons including codec frame size or sloppy authoring. For that reason,
24315 related synchronized streams (e.g. a video and its audio track) should be
24316 concatenated at once. The concat filter will use the duration of the longest
24317 stream in each segment (except the last one), and if necessary pad shorter
24318 audio streams with silence.
24320 For this filter to work correctly, all segments must start at timestamp 0.
24322 All corresponding streams must have the same parameters in all segments; the
24323 filtering system will automatically select a common pixel format for video
24324 streams, and a common sample format, sample rate and channel layout for
24325 audio streams, but other settings, such as resolution, must be converted
24326 explicitly by the user.
24328 Different frame rates are acceptable but will result in variable frame rate
24329 at output; be sure to configure the output file to handle it.
24331 @subsection Examples
24335 Concatenate an opening, an episode and an ending, all in bilingual version
24336 (video in stream 0, audio in streams 1 and 2):
24338 ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex \
24339 '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2]
24340 concat=n=3:v=1:a=2 [v] [a1] [a2]' \
24341 -map '[v]' -map '[a1]' -map '[a2]' output.mkv
24345 Concatenate two parts, handling audio and video separately, using the
24346 (a)movie sources, and adjusting the resolution:
24348 movie=part1.mp4, scale=512:288 [v1] ; amovie=part1.mp4 [a1] ;
24349 movie=part2.mp4, scale=512:288 [v2] ; amovie=part2.mp4 [a2] ;
24350 [v1] [v2] concat [outv] ; [a1] [a2] concat=v=0:a=1 [outa]
24352 Note that a desync will happen at the stitch if the audio and video streams
24353 do not have exactly the same duration in the first file.
24357 @subsection Commands
24359 This filter supports the following commands:
24362 Close the current segment and step to the next one
24368 EBU R128 scanner filter. This filter takes an audio stream and analyzes its loudness
24369 level. By default, it logs a message at a frequency of 10Hz with the
24370 Momentary loudness (identified by @code{M}), Short-term loudness (@code{S}),
24371 Integrated loudness (@code{I}) and Loudness Range (@code{LRA}).
24373 The filter can only analyze streams which have a sampling rate of 48000 Hz and whose
24374 sample format is double-precision floating point. The input stream will be converted to
24375 this specification, if needed. Users may need to insert aformat and/or aresample filters
24376 after this filter to obtain the original parameters.
24378 The filter also has a video output (see the @var{video} option) with a real
24379 time graph to observe the loudness evolution. The graphic contains the logged
24380 message mentioned above, so it is not printed anymore when this option is set,
24381 unless the verbose logging is set. The main graphing area contains the
24382 short-term loudness (3 seconds of analysis), and the gauge on the right is for
24383 the momentary loudness (400 milliseconds), but can optionally be configured
24384 to instead display short-term loudness (see @var{gauge}).
24386 The green area marks a +/- 1LU target range around the target loudness
24387 (-23LUFS by default, unless modified through @var{target}).
24389 More information about the Loudness Recommendation EBU R128 on
24390 @url{http://tech.ebu.ch/loudness}.
24392 The filter accepts the following options:
24397 Activate the video output. The audio stream is passed unchanged whether this
24398 option is set or no. The video stream will be the first output stream if
24399 activated. Default is @code{0}.
24402 Set the video size. This option is for video only. For the syntax of this
24404 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
24405 Default and minimum resolution is @code{640x480}.
24408 Set the EBU scale meter. Default is @code{9}. Common values are @code{9} and
24409 @code{18}, respectively for EBU scale meter +9 and EBU scale meter +18. Any
24410 other integer value between this range is allowed.
24413 Set metadata injection. If set to @code{1}, the audio input will be segmented
24414 into 100ms output frames, each of them containing various loudness information
24415 in metadata. All the metadata keys are prefixed with @code{lavfi.r128.}.
24417 Default is @code{0}.
24420 Force the frame logging level.
24422 Available values are:
24425 information logging level
24427 verbose logging level
24430 By default, the logging level is set to @var{info}. If the @option{video} or
24431 the @option{metadata} options are set, it switches to @var{verbose}.
24436 Available modes can be cumulated (the option is a @code{flag} type). Possible
24440 Disable any peak mode (default).
24442 Enable sample-peak mode.
24444 Simple peak mode looking for the higher sample value. It logs a message
24445 for sample-peak (identified by @code{SPK}).
24447 Enable true-peak mode.
24449 If enabled, the peak lookup is done on an over-sampled version of the input
24450 stream for better peak accuracy. It logs a message for true-peak.
24451 (identified by @code{TPK}) and true-peak per frame (identified by @code{FTPK}).
24452 This mode requires a build with @code{libswresample}.
24456 Treat mono input files as "dual mono". If a mono file is intended for playback
24457 on a stereo system, its EBU R128 measurement will be perceptually incorrect.
24458 If set to @code{true}, this option will compensate for this effect.
24459 Multi-channel input files are not affected by this option.
24462 Set a specific pan law to be used for the measurement of dual mono files.
24463 This parameter is optional, and has a default value of -3.01dB.
24466 Set a specific target level (in LUFS) used as relative zero in the visualization.
24467 This parameter is optional and has a default value of -23LUFS as specified
24468 by EBU R128. However, material published online may prefer a level of -16LUFS
24469 (e.g. for use with podcasts or video platforms).
24472 Set the value displayed by the gauge. Valid values are @code{momentary} and s
24473 @code{shortterm}. By default the momentary value will be used, but in certain
24474 scenarios it may be more useful to observe the short term value instead (e.g.
24478 Sets the display scale for the loudness. Valid parameters are @code{absolute}
24479 (in LUFS) or @code{relative} (LU) relative to the target. This only affects the
24480 video output, not the summary or continuous log output.
24483 @subsection Examples
24487 Real-time graph using @command{ffplay}, with a EBU scale meter +18:
24489 ffplay -f lavfi -i "amovie=input.mp3,ebur128=video=1:meter=18 [out0][out1]"
24493 Run an analysis with @command{ffmpeg}:
24495 ffmpeg -nostats -i input.mp3 -filter_complex ebur128 -f null -
24499 @section interleave, ainterleave
24501 Temporally interleave frames from several inputs.
24503 @code{interleave} works with video inputs, @code{ainterleave} with audio.
24505 These filters read frames from several inputs and send the oldest
24506 queued frame to the output.
24508 Input streams must have well defined, monotonically increasing frame
24511 In order to submit one frame to output, these filters need to enqueue
24512 at least one frame for each input, so they cannot work in case one
24513 input is not yet terminated and will not receive incoming frames.
24515 For example consider the case when one input is a @code{select} filter
24516 which always drops input frames. The @code{interleave} filter will keep
24517 reading from that input, but it will never be able to send new frames
24518 to output until the input sends an end-of-stream signal.
24520 Also, depending on inputs synchronization, the filters will drop
24521 frames in case one input receives more frames than the other ones, and
24522 the queue is already filled.
24524 These filters accept the following options:
24528 Set the number of different inputs, it is 2 by default.
24531 How to determine the end-of-stream.
24535 The duration of the longest input. (default)
24538 The duration of the shortest input.
24541 The duration of the first input.
24546 @subsection Examples
24550 Interleave frames belonging to different streams using @command{ffmpeg}:
24552 ffmpeg -i bambi.avi -i pr0n.mkv -filter_complex "[0:v][1:v] interleave" out.avi
24556 Add flickering blur effect:
24558 select='if(gt(random(0), 0.2), 1, 2)':n=2 [tmp], boxblur=2:2, [tmp] interleave
24562 @section metadata, ametadata
24564 Manipulate frame metadata.
24566 This filter accepts the following options:
24570 Set mode of operation of the filter.
24572 Can be one of the following:
24576 If both @code{value} and @code{key} is set, select frames
24577 which have such metadata. If only @code{key} is set, select
24578 every frame that has such key in metadata.
24581 Add new metadata @code{key} and @code{value}. If key is already available
24585 Modify value of already present key.
24588 If @code{value} is set, delete only keys that have such value.
24589 Otherwise, delete key. If @code{key} is not set, delete all metadata values in
24593 Print key and its value if metadata was found. If @code{key} is not set print all
24594 metadata values available in frame.
24598 Set key used with all modes. Must be set for all modes except @code{print} and @code{delete}.
24601 Set metadata value which will be used. This option is mandatory for
24602 @code{modify} and @code{add} mode.
24605 Which function to use when comparing metadata value and @code{value}.
24607 Can be one of following:
24611 Values are interpreted as strings, returns true if metadata value is same as @code{value}.
24614 Values are interpreted as strings, returns true if metadata value starts with
24615 the @code{value} option string.
24618 Values are interpreted as floats, returns true if metadata value is less than @code{value}.
24621 Values are interpreted as floats, returns true if @code{value} is equal with metadata value.
24624 Values are interpreted as floats, returns true if metadata value is greater than @code{value}.
24627 Values are interpreted as floats, returns true if expression from option @code{expr}
24631 Values are interpreted as strings, returns true if metadata value ends with
24632 the @code{value} option string.
24636 Set expression which is used when @code{function} is set to @code{expr}.
24637 The expression is evaluated through the eval API and can contain the following
24642 Float representation of @code{value} from metadata key.
24645 Float representation of @code{value} as supplied by user in @code{value} option.
24649 If specified in @code{print} mode, output is written to the named file. Instead of
24650 plain filename any writable url can be specified. Filename ``-'' is a shorthand
24651 for standard output. If @code{file} option is not set, output is written to the log
24652 with AV_LOG_INFO loglevel.
24655 Reduces buffering in print mode when output is written to a URL set using @var{file}.
24659 @subsection Examples
24663 Print all metadata values for frames with key @code{lavfi.signalstats.YDIF} with values
24666 signalstats,metadata=print:key=lavfi.signalstats.YDIF:value=0:function=expr:expr='between(VALUE1,0,1)'
24669 Print silencedetect output to file @file{metadata.txt}.
24671 silencedetect,ametadata=mode=print:file=metadata.txt
24674 Direct all metadata to a pipe with file descriptor 4.
24676 metadata=mode=print:file='pipe\:4'
24680 @section perms, aperms
24682 Set read/write permissions for the output frames.
24684 These filters are mainly aimed at developers to test direct path in the
24685 following filter in the filtergraph.
24687 The filters accept the following options:
24691 Select the permissions mode.
24693 It accepts the following values:
24696 Do nothing. This is the default.
24698 Set all the output frames read-only.
24700 Set all the output frames directly writable.
24702 Make the frame read-only if writable, and writable if read-only.
24704 Set each output frame read-only or writable randomly.
24708 Set the seed for the @var{random} mode, must be an integer included between
24709 @code{0} and @code{UINT32_MAX}. If not specified, or if explicitly set to
24710 @code{-1}, the filter will try to use a good random seed on a best effort
24714 Note: in case of auto-inserted filter between the permission filter and the
24715 following one, the permission might not be received as expected in that
24716 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
24717 perms/aperms filter can avoid this problem.
24719 @section realtime, arealtime
24721 Slow down filtering to match real time approximately.
24723 These filters will pause the filtering for a variable amount of time to
24724 match the output rate with the input timestamps.
24725 They are similar to the @option{re} option to @code{ffmpeg}.
24727 They accept the following options:
24731 Time limit for the pauses. Any pause longer than that will be considered
24732 a timestamp discontinuity and reset the timer. Default is 2 seconds.
24734 Speed factor for processing. The value must be a float larger than zero.
24735 Values larger than 1.0 will result in faster than realtime processing,
24736 smaller will slow processing down. The @var{limit} is automatically adapted
24737 accordingly. Default is 1.0.
24739 A processing speed faster than what is possible without these filters cannot
24744 @section select, aselect
24746 Select frames to pass in output.
24748 This filter accepts the following options:
24753 Set expression, which is evaluated for each input frame.
24755 If the expression is evaluated to zero, the frame is discarded.
24757 If the evaluation result is negative or NaN, the frame is sent to the
24758 first output; otherwise it is sent to the output with index
24759 @code{ceil(val)-1}, assuming that the input index starts from 0.
24761 For example a value of @code{1.2} corresponds to the output with index
24762 @code{ceil(1.2)-1 = 2-1 = 1}, that is the second output.
24765 Set the number of outputs. The output to which to send the selected
24766 frame is based on the result of the evaluation. Default value is 1.
24769 The expression can contain the following constants:
24773 The (sequential) number of the filtered frame, starting from 0.
24776 The (sequential) number of the selected frame, starting from 0.
24778 @item prev_selected_n
24779 The sequential number of the last selected frame. It's NAN if undefined.
24782 The timebase of the input timestamps.
24785 The PTS (Presentation TimeStamp) of the filtered video frame,
24786 expressed in @var{TB} units. It's NAN if undefined.
24789 The PTS of the filtered video frame,
24790 expressed in seconds. It's NAN if undefined.
24793 The PTS of the previously filtered video frame. It's NAN if undefined.
24795 @item prev_selected_pts
24796 The PTS of the last previously filtered video frame. It's NAN if undefined.
24798 @item prev_selected_t
24799 The PTS of the last previously selected video frame, expressed in seconds. It's NAN if undefined.
24802 The PTS of the first video frame in the video. It's NAN if undefined.
24805 The time of the first video frame in the video. It's NAN if undefined.
24807 @item pict_type @emph{(video only)}
24808 The type of the filtered frame. It can assume one of the following
24820 @item interlace_type @emph{(video only)}
24821 The frame interlace type. It can assume one of the following values:
24824 The frame is progressive (not interlaced).
24826 The frame is top-field-first.
24828 The frame is bottom-field-first.
24831 @item consumed_sample_n @emph{(audio only)}
24832 the number of selected samples before the current frame
24834 @item samples_n @emph{(audio only)}
24835 the number of samples in the current frame
24837 @item sample_rate @emph{(audio only)}
24838 the input sample rate
24841 This is 1 if the filtered frame is a key-frame, 0 otherwise.
24844 the position in the file of the filtered frame, -1 if the information
24845 is not available (e.g. for synthetic video)
24847 @item scene @emph{(video only)}
24848 value between 0 and 1 to indicate a new scene; a low value reflects a low
24849 probability for the current frame to introduce a new scene, while a higher
24850 value means the current frame is more likely to be one (see the example below)
24852 @item concatdec_select
24853 The concat demuxer can select only part of a concat input file by setting an
24854 inpoint and an outpoint, but the output packets may not be entirely contained
24855 in the selected interval. By using this variable, it is possible to skip frames
24856 generated by the concat demuxer which are not exactly contained in the selected
24859 This works by comparing the frame pts against the @var{lavf.concat.start_time}
24860 and the @var{lavf.concat.duration} packet metadata values which are also
24861 present in the decoded frames.
24863 The @var{concatdec_select} variable is -1 if the frame pts is at least
24864 start_time and either the duration metadata is missing or the frame pts is less
24865 than start_time + duration, 0 otherwise, and NaN if the start_time metadata is
24868 That basically means that an input frame is selected if its pts is within the
24869 interval set by the concat demuxer.
24873 The default value of the select expression is "1".
24875 @subsection Examples
24879 Select all frames in input:
24884 The example above is the same as:
24896 Select only I-frames:
24898 select='eq(pict_type\,I)'
24902 Select one frame every 100:
24904 select='not(mod(n\,100))'
24908 Select only frames contained in the 10-20 time interval:
24910 select=between(t\,10\,20)
24914 Select only I-frames contained in the 10-20 time interval:
24916 select=between(t\,10\,20)*eq(pict_type\,I)
24920 Select frames with a minimum distance of 10 seconds:
24922 select='isnan(prev_selected_t)+gte(t-prev_selected_t\,10)'
24926 Use aselect to select only audio frames with samples number > 100:
24928 aselect='gt(samples_n\,100)'
24932 Create a mosaic of the first scenes:
24934 ffmpeg -i video.avi -vf select='gt(scene\,0.4)',scale=160:120,tile -frames:v 1 preview.png
24937 Comparing @var{scene} against a value between 0.3 and 0.5 is generally a sane
24941 Send even and odd frames to separate outputs, and compose them:
24943 select=n=2:e='mod(n, 2)+1' [odd][even]; [odd] pad=h=2*ih [tmp]; [tmp][even] overlay=y=h
24947 Select useful frames from an ffconcat file which is using inpoints and
24948 outpoints but where the source files are not intra frame only.
24950 ffmpeg -copyts -vsync 0 -segment_time_metadata 1 -i input.ffconcat -vf select=concatdec_select -af aselect=concatdec_select output.avi
24954 @section sendcmd, asendcmd
24956 Send commands to filters in the filtergraph.
24958 These filters read commands to be sent to other filters in the
24961 @code{sendcmd} must be inserted between two video filters,
24962 @code{asendcmd} must be inserted between two audio filters, but apart
24963 from that they act the same way.
24965 The specification of commands can be provided in the filter arguments
24966 with the @var{commands} option, or in a file specified by the
24967 @var{filename} option.
24969 These filters accept the following options:
24972 Set the commands to be read and sent to the other filters.
24974 Set the filename of the commands to be read and sent to the other
24978 @subsection Commands syntax
24980 A commands description consists of a sequence of interval
24981 specifications, comprising a list of commands to be executed when a
24982 particular event related to that interval occurs. The occurring event
24983 is typically the current frame time entering or leaving a given time
24986 An interval is specified by the following syntax:
24988 @var{START}[-@var{END}] @var{COMMANDS};
24991 The time interval is specified by the @var{START} and @var{END} times.
24992 @var{END} is optional and defaults to the maximum time.
24994 The current frame time is considered within the specified interval if
24995 it is included in the interval [@var{START}, @var{END}), that is when
24996 the time is greater or equal to @var{START} and is lesser than
24999 @var{COMMANDS} consists of a sequence of one or more command
25000 specifications, separated by ",", relating to that interval. The
25001 syntax of a command specification is given by:
25003 [@var{FLAGS}] @var{TARGET} @var{COMMAND} @var{ARG}
25006 @var{FLAGS} is optional and specifies the type of events relating to
25007 the time interval which enable sending the specified command, and must
25008 be a non-null sequence of identifier flags separated by "+" or "|" and
25009 enclosed between "[" and "]".
25011 The following flags are recognized:
25014 The command is sent when the current frame timestamp enters the
25015 specified interval. In other words, the command is sent when the
25016 previous frame timestamp was not in the given interval, and the
25020 The command is sent when the current frame timestamp leaves the
25021 specified interval. In other words, the command is sent when the
25022 previous frame timestamp was in the given interval, and the
25026 The command @var{ARG} is interpreted as expression and result of
25027 expression is passed as @var{ARG}.
25029 The expression is evaluated through the eval API and can contain the following
25034 Original position in the file of the frame, or undefined if undefined
25035 for the current frame.
25038 The presentation timestamp in input.
25041 The count of the input frame for video or audio, starting from 0.
25044 The time in seconds of the current frame.
25047 The start time in seconds of the current command interval.
25050 The end time in seconds of the current command interval.
25053 The interpolated time of the current command interval, TI = (T - TS) / (TE - TS).
25058 If @var{FLAGS} is not specified, a default value of @code{[enter]} is
25061 @var{TARGET} specifies the target of the command, usually the name of
25062 the filter class or a specific filter instance name.
25064 @var{COMMAND} specifies the name of the command for the target filter.
25066 @var{ARG} is optional and specifies the optional list of argument for
25067 the given @var{COMMAND}.
25069 Between one interval specification and another, whitespaces, or
25070 sequences of characters starting with @code{#} until the end of line,
25071 are ignored and can be used to annotate comments.
25073 A simplified BNF description of the commands specification syntax
25076 @var{COMMAND_FLAG} ::= "enter" | "leave"
25077 @var{COMMAND_FLAGS} ::= @var{COMMAND_FLAG} [(+|"|")@var{COMMAND_FLAG}]
25078 @var{COMMAND} ::= ["[" @var{COMMAND_FLAGS} "]"] @var{TARGET} @var{COMMAND} [@var{ARG}]
25079 @var{COMMANDS} ::= @var{COMMAND} [,@var{COMMANDS}]
25080 @var{INTERVAL} ::= @var{START}[-@var{END}] @var{COMMANDS}
25081 @var{INTERVALS} ::= @var{INTERVAL}[;@var{INTERVALS}]
25084 @subsection Examples
25088 Specify audio tempo change at second 4:
25090 asendcmd=c='4.0 atempo tempo 1.5',atempo
25094 Target a specific filter instance:
25096 asendcmd=c='4.0 atempo@@my tempo 1.5',atempo@@my
25100 Specify a list of drawtext and hue commands in a file.
25102 # show text in the interval 5-10
25103 5.0-10.0 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=hello world',
25104 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=';
25106 # desaturate the image in the interval 15-20
25107 15.0-20.0 [enter] hue s 0,
25108 [enter] drawtext reinit 'fontfile=FreeSerif.ttf:text=nocolor',
25110 [leave] drawtext reinit 'fontfile=FreeSerif.ttf:text=color';
25112 # apply an exponential saturation fade-out effect, starting from time 25
25113 25 [enter] hue s exp(25-t)
25116 A filtergraph allowing to read and process the above command list
25117 stored in a file @file{test.cmd}, can be specified with:
25119 sendcmd=f=test.cmd,drawtext=fontfile=FreeSerif.ttf:text='',hue
25124 @section setpts, asetpts
25126 Change the PTS (presentation timestamp) of the input frames.
25128 @code{setpts} works on video frames, @code{asetpts} on audio frames.
25130 This filter accepts the following options:
25135 The expression which is evaluated for each frame to construct its timestamp.
25139 The expression is evaluated through the eval API and can contain the following
25143 @item FRAME_RATE, FR
25144 frame rate, only defined for constant frame-rate video
25147 The presentation timestamp in input
25150 The count of the input frame for video or the number of consumed samples,
25151 not including the current frame for audio, starting from 0.
25153 @item NB_CONSUMED_SAMPLES
25154 The number of consumed samples, not including the current frame (only
25157 @item NB_SAMPLES, S
25158 The number of samples in the current frame (only audio)
25160 @item SAMPLE_RATE, SR
25161 The audio sample rate.
25164 The PTS of the first frame.
25167 the time in seconds of the first frame
25170 State whether the current frame is interlaced.
25173 the time in seconds of the current frame
25176 original position in the file of the frame, or undefined if undefined
25177 for the current frame
25180 The previous input PTS.
25183 previous input time in seconds
25186 The previous output PTS.
25189 previous output time in seconds
25192 The wallclock (RTC) time in microseconds. This is deprecated, use time(0)
25196 The wallclock (RTC) time at the start of the movie in microseconds.
25199 The timebase of the input timestamps.
25203 @subsection Examples
25207 Start counting PTS from zero
25209 setpts=PTS-STARTPTS
25213 Apply fast motion effect:
25219 Apply slow motion effect:
25225 Set fixed rate of 25 frames per second:
25231 Set fixed rate 25 fps with some jitter:
25233 setpts='1/(25*TB) * (N + 0.05 * sin(N*2*PI/25))'
25237 Apply an offset of 10 seconds to the input PTS:
25243 Generate timestamps from a "live source" and rebase onto the current timebase:
25245 setpts='(RTCTIME - RTCSTART) / (TB * 1000000)'
25249 Generate timestamps by counting samples:
25258 Force color range for the output video frame.
25260 The @code{setrange} filter marks the color range property for the
25261 output frames. It does not change the input frame, but only sets the
25262 corresponding property, which affects how the frame is treated by
25265 The filter accepts the following options:
25270 Available values are:
25274 Keep the same color range property.
25276 @item unspecified, unknown
25277 Set the color range as unspecified.
25279 @item limited, tv, mpeg
25280 Set the color range as limited.
25282 @item full, pc, jpeg
25283 Set the color range as full.
25287 @section settb, asettb
25289 Set the timebase to use for the output frames timestamps.
25290 It is mainly useful for testing timebase configuration.
25292 It accepts the following parameters:
25297 The expression which is evaluated into the output timebase.
25301 The value for @option{tb} is an arithmetic expression representing a
25302 rational. The expression can contain the constants "AVTB" (the default
25303 timebase), "intb" (the input timebase) and "sr" (the sample rate,
25304 audio only). Default value is "intb".
25306 @subsection Examples
25310 Set the timebase to 1/25:
25316 Set the timebase to 1/10:
25322 Set the timebase to 1001/1000:
25328 Set the timebase to 2*intb:
25334 Set the default timebase value:
25341 Convert input audio to a video output representing frequency spectrum
25342 logarithmically using Brown-Puckette constant Q transform algorithm with
25343 direct frequency domain coefficient calculation (but the transform itself
25344 is not really constant Q, instead the Q factor is actually variable/clamped),
25345 with musical tone scale, from E0 to D#10.
25347 The filter accepts the following options:
25351 Specify the video size for the output. It must be even. For the syntax of this option,
25352 check the @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25353 Default value is @code{1920x1080}.
25356 Set the output frame rate. Default value is @code{25}.
25359 Set the bargraph height. It must be even. Default value is @code{-1} which
25360 computes the bargraph height automatically.
25363 Set the axis height. It must be even. Default value is @code{-1} which computes
25364 the axis height automatically.
25367 Set the sonogram height. It must be even. Default value is @code{-1} which
25368 computes the sonogram height automatically.
25371 Set the fullhd resolution. This option is deprecated, use @var{size}, @var{s}
25372 instead. Default value is @code{1}.
25374 @item sono_v, volume
25375 Specify the sonogram volume expression. It can contain variables:
25378 the @var{bar_v} evaluated expression
25379 @item frequency, freq, f
25380 the frequency where it is evaluated
25381 @item timeclamp, tc
25382 the value of @var{timeclamp} option
25386 @item a_weighting(f)
25387 A-weighting of equal loudness
25388 @item b_weighting(f)
25389 B-weighting of equal loudness
25390 @item c_weighting(f)
25391 C-weighting of equal loudness.
25393 Default value is @code{16}.
25395 @item bar_v, volume2
25396 Specify the bargraph volume expression. It can contain variables:
25399 the @var{sono_v} evaluated expression
25400 @item frequency, freq, f
25401 the frequency where it is evaluated
25402 @item timeclamp, tc
25403 the value of @var{timeclamp} option
25407 @item a_weighting(f)
25408 A-weighting of equal loudness
25409 @item b_weighting(f)
25410 B-weighting of equal loudness
25411 @item c_weighting(f)
25412 C-weighting of equal loudness.
25414 Default value is @code{sono_v}.
25416 @item sono_g, gamma
25417 Specify the sonogram gamma. Lower gamma makes the spectrum more contrast,
25418 higher gamma makes the spectrum having more range. Default value is @code{3}.
25419 Acceptable range is @code{[1, 7]}.
25421 @item bar_g, gamma2
25422 Specify the bargraph gamma. Default value is @code{1}. Acceptable range is
25426 Specify the bargraph transparency level. Lower value makes the bargraph sharper.
25427 Default value is @code{1}. Acceptable range is @code{[0, 1]}.
25429 @item timeclamp, tc
25430 Specify the transform timeclamp. At low frequency, there is trade-off between
25431 accuracy in time domain and frequency domain. If timeclamp is lower,
25432 event in time domain is represented more accurately (such as fast bass drum),
25433 otherwise event in frequency domain is represented more accurately
25434 (such as bass guitar). Acceptable range is @code{[0.002, 1]}. Default value is @code{0.17}.
25437 Set attack time in seconds. The default is @code{0} (disabled). Otherwise, it
25438 limits future samples by applying asymmetric windowing in time domain, useful
25439 when low latency is required. Accepted range is @code{[0, 1]}.
25442 Specify the transform base frequency. Default value is @code{20.01523126408007475},
25443 which is frequency 50 cents below E0. Acceptable range is @code{[10, 100000]}.
25446 Specify the transform end frequency. Default value is @code{20495.59681441799654},
25447 which is frequency 50 cents above D#10. Acceptable range is @code{[10, 100000]}.
25450 This option is deprecated and ignored.
25453 Specify the transform length in time domain. Use this option to control accuracy
25454 trade-off between time domain and frequency domain at every frequency sample.
25455 It can contain variables:
25457 @item frequency, freq, f
25458 the frequency where it is evaluated
25459 @item timeclamp, tc
25460 the value of @var{timeclamp} option.
25462 Default value is @code{384*tc/(384+tc*f)}.
25465 Specify the transform count for every video frame. Default value is @code{6}.
25466 Acceptable range is @code{[1, 30]}.
25469 Specify the transform count for every single pixel. Default value is @code{0},
25470 which makes it computed automatically. Acceptable range is @code{[0, 10]}.
25473 Specify font file for use with freetype to draw the axis. If not specified,
25474 use embedded font. Note that drawing with font file or embedded font is not
25475 implemented with custom @var{basefreq} and @var{endfreq}, use @var{axisfile}
25479 Specify fontconfig pattern. This has lower priority than @var{fontfile}. The
25480 @code{:} in the pattern may be replaced by @code{|} to avoid unnecessary
25484 Specify font color expression. This is arithmetic expression that should return
25485 integer value 0xRRGGBB. It can contain variables:
25487 @item frequency, freq, f
25488 the frequency where it is evaluated
25489 @item timeclamp, tc
25490 the value of @var{timeclamp} option
25495 midi number of frequency f, some midi numbers: E0(16), C1(24), C2(36), A4(69)
25496 @item r(x), g(x), b(x)
25497 red, green, and blue value of intensity x.
25499 Default value is @code{st(0, (midi(f)-59.5)/12);
25500 st(1, if(between(ld(0),0,1), 0.5-0.5*cos(2*PI*ld(0)), 0));
25501 r(1-ld(1)) + b(ld(1))}.
25504 Specify image file to draw the axis. This option override @var{fontfile} and
25505 @var{fontcolor} option.
25508 Enable/disable drawing text to the axis. If it is set to @code{0}, drawing to
25509 the axis is disabled, ignoring @var{fontfile} and @var{axisfile} option.
25510 Default value is @code{1}.
25513 Set colorspace. The accepted values are:
25516 Unspecified (default)
25525 BT.470BG or BT.601-6 625
25528 SMPTE-170M or BT.601-6 525
25534 BT.2020 with non-constant luminance
25539 Set spectrogram color scheme. This is list of floating point values with format
25540 @code{left_r|left_g|left_b|right_r|right_g|right_b}.
25541 The default is @code{1|0.5|0|0|0.5|1}.
25545 @subsection Examples
25549 Playing audio while showing the spectrum:
25551 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt [out0]'
25555 Same as above, but with frame rate 30 fps:
25557 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=fps=30:count=5 [out0]'
25561 Playing at 1280x720:
25563 ffplay -f lavfi 'amovie=a.mp3, asplit [a][out1]; [a] showcqt=s=1280x720:count=4 [out0]'
25567 Disable sonogram display:
25573 A1 and its harmonics: A1, A2, (near)E3, A3:
25575 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),
25576 asplit[a][out1]; [a] showcqt [out0]'
25580 Same as above, but with more accuracy in frequency domain:
25582 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),
25583 asplit[a][out1]; [a] showcqt=timeclamp=0.5 [out0]'
25589 bar_v=10:sono_v=bar_v*a_weighting(f)
25593 Custom gamma, now spectrum is linear to the amplitude.
25599 Custom tlength equation:
25601 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)))'
25605 Custom fontcolor and fontfile, C-note is colored green, others are colored blue:
25607 fontcolor='if(mod(floor(midi(f)+0.5),12), 0x0000FF, g(1))':fontfile=myfont.ttf
25611 Custom font using fontconfig:
25613 font='Courier New,Monospace,mono|bold'
25617 Custom frequency range with custom axis using image file:
25619 axisfile=myaxis.png:basefreq=40:endfreq=10000
25625 Convert input audio to video output representing the audio power spectrum.
25626 Audio amplitude is on Y-axis while frequency is on X-axis.
25628 The filter accepts the following options:
25632 Specify size of video. For the syntax of this option, check the
25633 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25634 Default is @code{1024x512}.
25638 This set how each frequency bin will be represented.
25640 It accepts the following values:
25646 Default is @code{bar}.
25649 Set amplitude scale.
25651 It accepts the following values:
25665 Default is @code{log}.
25668 Set frequency scale.
25670 It accepts the following values:
25679 Reverse logarithmic scale.
25681 Default is @code{lin}.
25684 Set window size. Allowed range is from 16 to 65536.
25686 Default is @code{2048}
25689 Set windowing function.
25691 It accepts the following values:
25714 Default is @code{hanning}.
25717 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
25718 which means optimal overlap for selected window function will be picked.
25721 Set time averaging. Setting this to 0 will display current maximal peaks.
25722 Default is @code{1}, which means time averaging is disabled.
25725 Specify list of colors separated by space or by '|' which will be used to
25726 draw channel frequencies. Unrecognized or missing colors will be replaced
25730 Set channel display mode.
25732 It accepts the following values:
25737 Default is @code{combined}.
25740 Set minimum amplitude used in @code{log} amplitude scaler.
25743 Set data display mode.
25745 It accepts the following values:
25751 Default is @code{magnitude}.
25754 @section showspatial
25756 Convert stereo input audio to a video output, representing the spatial relationship
25757 between two channels.
25759 The filter accepts the following options:
25763 Specify the video size for the output. For the syntax of this option, check the
25764 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25765 Default value is @code{512x512}.
25768 Set window size. Allowed range is from @var{1024} to @var{65536}. Default size is @var{4096}.
25771 Set window function.
25773 It accepts the following values:
25798 Default value is @code{hann}.
25801 Set ratio of overlap window. Default value is @code{0.5}.
25802 When value is @code{1} overlap is set to recommended size for specific
25803 window function currently used.
25806 @anchor{showspectrum}
25807 @section showspectrum
25809 Convert input audio to a video output, representing the audio frequency
25812 The filter accepts the following options:
25816 Specify the video size for the output. For the syntax of this option, check the
25817 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
25818 Default value is @code{640x512}.
25821 Specify how the spectrum should slide along the window.
25823 It accepts the following values:
25826 the samples start again on the left when they reach the right
25828 the samples scroll from right to left
25830 frames are only produced when the samples reach the right
25832 the samples scroll from left to right
25835 Default value is @code{replace}.
25838 Specify display mode.
25840 It accepts the following values:
25843 all channels are displayed in the same row
25845 all channels are displayed in separate rows
25848 Default value is @samp{combined}.
25851 Specify display color mode.
25853 It accepts the following values:
25856 each channel is displayed in a separate color
25858 each channel is displayed using the same color scheme
25860 each channel is displayed using the rainbow color scheme
25862 each channel is displayed using the moreland color scheme
25864 each channel is displayed using the nebulae color scheme
25866 each channel is displayed using the fire color scheme
25868 each channel is displayed using the fiery color scheme
25870 each channel is displayed using the fruit color scheme
25872 each channel is displayed using the cool color scheme
25874 each channel is displayed using the magma color scheme
25876 each channel is displayed using the green color scheme
25878 each channel is displayed using the viridis color scheme
25880 each channel is displayed using the plasma color scheme
25882 each channel is displayed using the cividis color scheme
25884 each channel is displayed using the terrain color scheme
25887 Default value is @samp{channel}.
25890 Specify scale used for calculating intensity color values.
25892 It accepts the following values:
25897 square root, default
25908 Default value is @samp{sqrt}.
25911 Specify frequency scale.
25913 It accepts the following values:
25921 Default value is @samp{lin}.
25924 Set saturation modifier for displayed colors. Negative values provide
25925 alternative color scheme. @code{0} is no saturation at all.
25926 Saturation must be in [-10.0, 10.0] range.
25927 Default value is @code{1}.
25930 Set window function.
25932 It accepts the following values:
25957 Default value is @code{hann}.
25960 Set orientation of time vs frequency axis. Can be @code{vertical} or
25961 @code{horizontal}. Default is @code{vertical}.
25964 Set ratio of overlap window. Default value is @code{0}.
25965 When value is @code{1} overlap is set to recommended size for specific
25966 window function currently used.
25969 Set scale gain for calculating intensity color values.
25970 Default value is @code{1}.
25973 Set which data to display. Can be @code{magnitude}, default or @code{phase}.
25976 Set color rotation, must be in [-1.0, 1.0] range.
25977 Default value is @code{0}.
25980 Set start frequency from which to display spectrogram. Default is @code{0}.
25983 Set stop frequency to which to display spectrogram. Default is @code{0}.
25986 Set upper frame rate limit. Default is @code{auto}, unlimited.
25989 Draw time and frequency axes and legends. Default is disabled.
25992 The usage is very similar to the showwaves filter; see the examples in that
25995 @subsection Examples
25999 Large window with logarithmic color scaling:
26001 showspectrum=s=1280x480:scale=log
26005 Complete example for a colored and sliding spectrum per channel using @command{ffplay}:
26007 ffplay -f lavfi 'amovie=input.mp3, asplit [a][out1];
26008 [a] showspectrum=mode=separate:color=intensity:slide=1:scale=cbrt [out0]'
26012 @section showspectrumpic
26014 Convert input audio to a single video frame, representing the audio frequency
26017 The filter accepts the following options:
26021 Specify the video size for the output. For the syntax of this option, check the
26022 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26023 Default value is @code{4096x2048}.
26026 Specify display mode.
26028 It accepts the following values:
26031 all channels are displayed in the same row
26033 all channels are displayed in separate rows
26035 Default value is @samp{combined}.
26038 Specify display color mode.
26040 It accepts the following values:
26043 each channel is displayed in a separate color
26045 each channel is displayed using the same color scheme
26047 each channel is displayed using the rainbow color scheme
26049 each channel is displayed using the moreland color scheme
26051 each channel is displayed using the nebulae color scheme
26053 each channel is displayed using the fire color scheme
26055 each channel is displayed using the fiery color scheme
26057 each channel is displayed using the fruit color scheme
26059 each channel is displayed using the cool color scheme
26061 each channel is displayed using the magma color scheme
26063 each channel is displayed using the green color scheme
26065 each channel is displayed using the viridis color scheme
26067 each channel is displayed using the plasma color scheme
26069 each channel is displayed using the cividis color scheme
26071 each channel is displayed using the terrain color scheme
26073 Default value is @samp{intensity}.
26076 Specify scale used for calculating intensity color values.
26078 It accepts the following values:
26083 square root, default
26093 Default value is @samp{log}.
26096 Specify frequency scale.
26098 It accepts the following values:
26106 Default value is @samp{lin}.
26109 Set saturation modifier for displayed colors. Negative values provide
26110 alternative color scheme. @code{0} is no saturation at all.
26111 Saturation must be in [-10.0, 10.0] range.
26112 Default value is @code{1}.
26115 Set window function.
26117 It accepts the following values:
26141 Default value is @code{hann}.
26144 Set orientation of time vs frequency axis. Can be @code{vertical} or
26145 @code{horizontal}. Default is @code{vertical}.
26148 Set scale gain for calculating intensity color values.
26149 Default value is @code{1}.
26152 Draw time and frequency axes and legends. Default is enabled.
26155 Set color rotation, must be in [-1.0, 1.0] range.
26156 Default value is @code{0}.
26159 Set start frequency from which to display spectrogram. Default is @code{0}.
26162 Set stop frequency to which to display spectrogram. Default is @code{0}.
26165 @subsection Examples
26169 Extract an audio spectrogram of a whole audio track
26170 in a 1024x1024 picture using @command{ffmpeg}:
26172 ffmpeg -i audio.flac -lavfi showspectrumpic=s=1024x1024 spectrogram.png
26176 @section showvolume
26178 Convert input audio volume to a video output.
26180 The filter accepts the following options:
26187 Set border width, allowed range is [0, 5]. Default is 1.
26190 Set channel width, allowed range is [80, 8192]. Default is 400.
26193 Set channel height, allowed range is [1, 900]. Default is 20.
26196 Set fade, allowed range is [0, 1]. Default is 0.95.
26199 Set volume color expression.
26201 The expression can use the following variables:
26205 Current max volume of channel in dB.
26211 Current channel number, starting from 0.
26215 If set, displays channel names. Default is enabled.
26218 If set, displays volume values. Default is enabled.
26221 Set orientation, can be horizontal: @code{h} or vertical: @code{v},
26222 default is @code{h}.
26225 Set step size, allowed range is [0, 5]. Default is 0, which means
26229 Set background opacity, allowed range is [0, 1]. Default is 0.
26232 Set metering mode, can be peak: @code{p} or rms: @code{r},
26233 default is @code{p}.
26236 Set display scale, can be linear: @code{lin} or log: @code{log},
26237 default is @code{lin}.
26241 If set to > 0., display a line for the max level
26242 in the previous seconds.
26243 default is disabled: @code{0.}
26246 The color of the max line. Use when @code{dm} option is set to > 0.
26247 default is: @code{orange}
26252 Convert input audio to a video output, representing the samples waves.
26254 The filter accepts the following options:
26258 Specify the video size for the output. For the syntax of this option, check the
26259 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26260 Default value is @code{600x240}.
26265 Available values are:
26268 Draw a point for each sample.
26271 Draw a vertical line for each sample.
26274 Draw a point for each sample and a line between them.
26277 Draw a centered vertical line for each sample.
26280 Default value is @code{point}.
26283 Set the number of samples which are printed on the same column. A
26284 larger value will decrease the frame rate. Must be a positive
26285 integer. This option can be set only if the value for @var{rate}
26286 is not explicitly specified.
26289 Set the (approximate) output frame rate. This is done by setting the
26290 option @var{n}. Default value is "25".
26292 @item split_channels
26293 Set if channels should be drawn separately or overlap. Default value is 0.
26296 Set colors separated by '|' which are going to be used for drawing of each channel.
26299 Set amplitude scale.
26301 Available values are:
26319 Set the draw mode. This is mostly useful to set for high @var{n}.
26321 Available values are:
26324 Scale pixel values for each drawn sample.
26327 Draw every sample directly.
26330 Default value is @code{scale}.
26333 @subsection Examples
26337 Output the input file audio and the corresponding video representation
26340 amovie=a.mp3,asplit[out0],showwaves[out1]
26344 Create a synthetic signal and show it with showwaves, forcing a
26345 frame rate of 30 frames per second:
26347 aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=30[out1]
26351 @section showwavespic
26353 Convert input audio to a single video frame, representing the samples waves.
26355 The filter accepts the following options:
26359 Specify the video size for the output. For the syntax of this option, check the
26360 @ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
26361 Default value is @code{600x240}.
26363 @item split_channels
26364 Set if channels should be drawn separately or overlap. Default value is 0.
26367 Set colors separated by '|' which are going to be used for drawing of each channel.
26370 Set amplitude scale.
26372 Available values are:
26392 Available values are:
26395 Scale pixel values for each drawn sample.
26398 Draw every sample directly.
26401 Default value is @code{scale}.
26404 Set the filter mode.
26406 Available values are:
26409 Use average samples values for each drawn sample.
26412 Use peak samples values for each drawn sample.
26415 Default value is @code{average}.
26418 @subsection Examples
26422 Extract a channel split representation of the wave form of a whole audio track
26423 in a 1024x800 picture using @command{ffmpeg}:
26425 ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
26429 @section sidedata, asidedata
26431 Delete frame side data, or select frames based on it.
26433 This filter accepts the following options:
26437 Set mode of operation of the filter.
26439 Can be one of the following:
26443 Select every frame with side data of @code{type}.
26446 Delete side data of @code{type}. If @code{type} is not set, delete all side
26452 Set side data type used with all modes. Must be set for @code{select} mode. For
26453 the list of frame side data types, refer to the @code{AVFrameSideDataType} enum
26454 in @file{libavutil/frame.h}. For example, to choose
26455 @code{AV_FRAME_DATA_PANSCAN} side data, you must specify @code{PANSCAN}.
26459 @section spectrumsynth
26461 Synthesize audio from 2 input video spectrums, first input stream represents
26462 magnitude across time and second represents phase across time.
26463 The filter will transform from frequency domain as displayed in videos back
26464 to time domain as presented in audio output.
26466 This filter is primarily created for reversing processed @ref{showspectrum}
26467 filter outputs, but can synthesize sound from other spectrograms too.
26468 But in such case results are going to be poor if the phase data is not
26469 available, because in such cases phase data need to be recreated, usually
26470 it's just recreated from random noise.
26471 For best results use gray only output (@code{channel} color mode in
26472 @ref{showspectrum} filter) and @code{log} scale for magnitude video and
26473 @code{lin} scale for phase video. To produce phase, for 2nd video, use
26474 @code{data} option. Inputs videos should generally use @code{fullframe}
26475 slide mode as that saves resources needed for decoding video.
26477 The filter accepts the following options:
26481 Specify sample rate of output audio, the sample rate of audio from which
26482 spectrum was generated may differ.
26485 Set number of channels represented in input video spectrums.
26488 Set scale which was used when generating magnitude input spectrum.
26489 Can be @code{lin} or @code{log}. Default is @code{log}.
26492 Set slide which was used when generating inputs spectrums.
26493 Can be @code{replace}, @code{scroll}, @code{fullframe} or @code{rscroll}.
26494 Default is @code{fullframe}.
26497 Set window function used for resynthesis.
26500 Set window overlap. In range @code{[0, 1]}. Default is @code{1},
26501 which means optimal overlap for selected window function will be picked.
26504 Set orientation of input videos. Can be @code{vertical} or @code{horizontal}.
26505 Default is @code{vertical}.
26508 @subsection Examples
26512 First create magnitude and phase videos from audio, assuming audio is stereo with 44100 sample rate,
26513 then resynthesize videos back to audio with spectrumsynth:
26515 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
26516 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
26517 ffmpeg -i magnitude.nut -i phase.nut -lavfi spectrumsynth=channels=2:sample_rate=44100:win_func=hann:overlap=0.875:slide=fullframe output.flac
26521 @section split, asplit
26523 Split input into several identical outputs.
26525 @code{asplit} works with audio input, @code{split} with video.
26527 The filter accepts a single parameter which specifies the number of outputs. If
26528 unspecified, it defaults to 2.
26530 @subsection Examples
26534 Create two separate outputs from the same input:
26536 [in] split [out0][out1]
26540 To create 3 or more outputs, you need to specify the number of
26543 [in] asplit=3 [out0][out1][out2]
26547 Create two separate outputs from the same input, one cropped and
26550 [in] split [splitout1][splitout2];
26551 [splitout1] crop=100:100:0:0 [cropout];
26552 [splitout2] pad=200:200:100:100 [padout];
26556 Create 5 copies of the input audio with @command{ffmpeg}:
26558 ffmpeg -i INPUT -filter_complex asplit=5 OUTPUT
26564 Receive commands sent through a libzmq client, and forward them to
26565 filters in the filtergraph.
26567 @code{zmq} and @code{azmq} work as a pass-through filters. @code{zmq}
26568 must be inserted between two video filters, @code{azmq} between two
26569 audio filters. Both are capable to send messages to any filter type.
26571 To enable these filters you need to install the libzmq library and
26572 headers and configure FFmpeg with @code{--enable-libzmq}.
26574 For more information about libzmq see:
26575 @url{http://www.zeromq.org/}
26577 The @code{zmq} and @code{azmq} filters work as a libzmq server, which
26578 receives messages sent through a network interface defined by the
26579 @option{bind_address} (or the abbreviation "@option{b}") option.
26580 Default value of this option is @file{tcp://localhost:5555}. You may
26581 want to alter this value to your needs, but do not forget to escape any
26582 ':' signs (see @ref{filtergraph escaping}).
26584 The received message must be in the form:
26586 @var{TARGET} @var{COMMAND} [@var{ARG}]
26589 @var{TARGET} specifies the target of the command, usually the name of
26590 the filter class or a specific filter instance name. The default
26591 filter instance name uses the pattern @samp{Parsed_<filter_name>_<index>},
26592 but you can override this by using the @samp{filter_name@@id} syntax
26593 (see @ref{Filtergraph syntax}).
26595 @var{COMMAND} specifies the name of the command for the target filter.
26597 @var{ARG} is optional and specifies the optional argument list for the
26598 given @var{COMMAND}.
26600 Upon reception, the message is processed and the corresponding command
26601 is injected into the filtergraph. Depending on the result, the filter
26602 will send a reply to the client, adopting the format:
26604 @var{ERROR_CODE} @var{ERROR_REASON}
26608 @var{MESSAGE} is optional.
26610 @subsection Examples
26612 Look at @file{tools/zmqsend} for an example of a zmq client which can
26613 be used to send commands processed by these filters.
26615 Consider the following filtergraph generated by @command{ffplay}.
26616 In this example the last overlay filter has an instance name. All other
26617 filters will have default instance names.
26620 ffplay -dumpgraph 1 -f lavfi "
26621 color=s=100x100:c=red [l];
26622 color=s=100x100:c=blue [r];
26623 nullsrc=s=200x100, zmq [bg];
26624 [bg][l] overlay [bg+l];
26625 [bg+l][r] overlay@@my=x=100 "
26628 To change the color of the left side of the video, the following
26629 command can be used:
26631 echo Parsed_color_0 c yellow | tools/zmqsend
26634 To change the right side:
26636 echo Parsed_color_1 c pink | tools/zmqsend
26639 To change the position of the right side:
26641 echo overlay@@my x 150 | tools/zmqsend
26645 @c man end MULTIMEDIA FILTERS
26647 @chapter Multimedia Sources
26648 @c man begin MULTIMEDIA SOURCES
26650 Below is a description of the currently available multimedia sources.
26654 This is the same as @ref{movie} source, except it selects an audio
26660 Read audio and/or video stream(s) from a movie container.
26662 It accepts the following parameters:
26666 The name of the resource to read (not necessarily a file; it can also be a
26667 device or a stream accessed through some protocol).
26669 @item format_name, f
26670 Specifies the format assumed for the movie to read, and can be either
26671 the name of a container or an input device. If not specified, the
26672 format is guessed from @var{movie_name} or by probing.
26674 @item seek_point, sp
26675 Specifies the seek point in seconds. The frames will be output
26676 starting from this seek point. The parameter is evaluated with
26677 @code{av_strtod}, so the numerical value may be suffixed by an IS
26678 postfix. The default value is "0".
26681 Specifies the streams to read. Several streams can be specified,
26682 separated by "+". The source will then have as many outputs, in the
26683 same order. The syntax is explained in the @ref{Stream specifiers,,"Stream specifiers"
26684 section in the ffmpeg manual,ffmpeg}. Two special names, "dv" and "da" specify
26685 respectively the default (best suited) video and audio stream. Default
26686 is "dv", or "da" if the filter is called as "amovie".
26688 @item stream_index, si
26689 Specifies the index of the video stream to read. If the value is -1,
26690 the most suitable video stream will be automatically selected. The default
26691 value is "-1". Deprecated. If the filter is called "amovie", it will select
26692 audio instead of video.
26695 Specifies how many times to read the stream in sequence.
26696 If the value is 0, the stream will be looped infinitely.
26697 Default value is "1".
26699 Note that when the movie is looped the source timestamps are not
26700 changed, so it will generate non monotonically increasing timestamps.
26702 @item discontinuity
26703 Specifies the time difference between frames above which the point is
26704 considered a timestamp discontinuity which is removed by adjusting the later
26708 It allows overlaying a second video on top of the main input of
26709 a filtergraph, as shown in this graph:
26711 input -----------> deltapts0 --> overlay --> output
26714 movie --> scale--> deltapts1 -------+
26716 @subsection Examples
26720 Skip 3.2 seconds from the start of the AVI file in.avi, and overlay it
26721 on top of the input labelled "in":
26723 movie=in.avi:seek_point=3.2, scale=180:-1, setpts=PTS-STARTPTS [over];
26724 [in] setpts=PTS-STARTPTS [main];
26725 [main][over] overlay=16:16 [out]
26729 Read from a video4linux2 device, and overlay it on top of the input
26732 movie=/dev/video0:f=video4linux2, scale=180:-1, setpts=PTS-STARTPTS [over];
26733 [in] setpts=PTS-STARTPTS [main];
26734 [main][over] overlay=16:16 [out]
26738 Read the first video stream and the audio stream with id 0x81 from
26739 dvd.vob; the video is connected to the pad named "video" and the audio is
26740 connected to the pad named "audio":
26742 movie=dvd.vob:s=v:0+#0x81 [video] [audio]
26746 @subsection Commands
26748 Both movie and amovie support the following commands:
26751 Perform seek using "av_seek_frame".
26752 The syntax is: seek @var{stream_index}|@var{timestamp}|@var{flags}
26755 @var{stream_index}: If stream_index is -1, a default
26756 stream is selected, and @var{timestamp} is automatically converted
26757 from AV_TIME_BASE units to the stream specific time_base.
26759 @var{timestamp}: Timestamp in AVStream.time_base units
26760 or, if no stream is specified, in AV_TIME_BASE units.
26762 @var{flags}: Flags which select direction and seeking mode.
26766 Get movie duration in AV_TIME_BASE units.
26770 @c man end MULTIMEDIA SOURCES