Opened 17 months ago
Last modified 16 months ago
#9613 new enhancement
add volumedetect filter support for >16bit audio
|Reported by:||chconnor||Owned by:|
|Version:||unspecified||Keywords:||normalize, loudness, normalization, filter, filtering, filters, volume, volumedetect|
|Cc:||chconnor, slhck||Blocked By:|
|Blocking:||Reproduced by developer:||no|
|Analyzed by developer:||no|
Running the volume detect filter on e.g. 32bit float audio does not work correctly because it converts internally to 16bit integer format (see docs: https://ffmpeg.org/ffmpeg-filters.html#volumedetect)
It would be great if the filter supported such audio files.
Change History (5)
comment:1 by , 17 months ago
follow-up: 3 comment:2 by , 16 months ago
comment:3 by , 16 months ago
Yeah. In my case I was normalizing ~1700 files (convolution reverb impulse responses) most of which had peaks over 0 dBFS.
Floating point audio isn't too uncommon a thing to want normalize, and the reason to normalize it is often precisely because it can have peaks over 0 dBFS.
comment:4 by , 16 months ago
I'd agree that this would be useful to have.
comment:5 by , 16 months ago
It is already available in astats filter, -af astats=measure_overall=Peak_level:measure_perchannel=0
It would give no real benefits, unless you care for >0 dB samples.