adding a sound volume property for the postprocessor

[Techbot121]

This is something I've been thinking for a while.
When the preprocessor runs it gets the length of the soundsample, wouldn't it be also useful to have other information as well? For example sample-volume (or similar) so you could filter all sounds by volume and potentially, for instance, filtering out loud sounds locally.

[PotcFdk]

Having an existing setup (i.e. the preprocessor) that inspects all files makes it easy(-ish) to further enhance it.
I've always thought about what else might be useful other than just the duration itself.
I'll look into volume/gain analysis and report back as to how easy it would be.

[PotcFdk]

Ok, apparently it might be possible with just ffmpeg/libav.
To be specific, libavfilter's loudnorm filter seems to be able to do what we'd be interested in, but I have no idea how to use it.
Using a separate lib would be an idea, or just slapping standard ReplayGain on all files and transferring that to the lists.
Needs more investigation.

[Techbot121]

or eventually calculate it yourself with a RMS measurement.

[edunad]

can we not cap the ^ to the listening client? Would not remove loud annoying sounds but it would at least prevent my ears from exploding