avfilter: add audio spectral stats filter

1 files changed, 63 insertions, 0 deletions

diff --git a/doc/filters.texi b/doc/filters.texi
index a1b212d38a..9ddff533de 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -2695,6 +2695,69 @@ Set oversampling factor.
 
 This filter supports the all above options as @ref{commands}.
 
+@section aspectralstats
+
+Display frequency domain statistical information about the audio channels.
+Statistics are calculated and stored as metadata for each audio channel and for each audio frame.
+
+It accepts the following option:
+@table @option
+@item win_size
+Set the window length in samples. Default value is 2048.
+Allowed range is from 32 to 65536.
+
+@item win_func
+Set window function.
+
+It accepts the following values:
+@table @samp
+@item rect
+@item bartlett
+@item hann, hanning
+@item hamming
+@item blackman
+@item welch
+@item flattop
+@item bharris
+@item bnuttall
+@item bhann
+@item sine
+@item nuttall
+@item lanczos
+@item gauss
+@item tukey
+@item dolph
+@item cauchy
+@item parzen
+@item poisson
+@item bohman
+@end table
+Default is @code{hann}.
+
+@item overlap
+Set window overlap. Allowed range is from @code{0}
+to @code{1}. Default value is @code{0.5}.
+
+@end table
+
+A list of each metadata key follows:
+
+@table @option
+@item mean
+@item variance
+@item centroid
+@item spread
+@item skewness
+@item kurtosis
+@item entropy
+@item flatness
+@item crest
+@item flux
+@item slope
+@item decrease
+@item rolloff
+@end table
+
 @section asr
 Automatic Speech Recognition