Re: [SoX-users] trying to mimic audacity's dynamic compression

[Jan S. <ha...@st...>]

> >> Threshold: -12 dB
> >> Noise Floor: -40 dB
> >> Ratio: 2:1
> >> Attack Time: 0.1
> >> Decay Time: 1.0
> > 
> > SoX has a 'compand' effect (as in compress/expand), which allows you
> > to prescribe the companding transfer function with arbitrary precision.
> > The above seems to only specify the noise floor and a one-knee compander.
> > 
> > Also, the values of -40dB, -12dB, and 2:1 might be good for one
> > particular recording, but completely unusable for another
> > (if that's what "scripting your audio processing" is aimed at).
> 
> I have a "final" processor in my audio chain that makes most recordings relatively the same.  So applying the same compression settings is usually sufficient.
> 
> > 
> >> Make-up gain for 0dB after compressing is checked
> > 
> > Not sure what that means, so I would guess normalize after companding.
> > There is an effect for that on SoX, yes.
> 
> Here's the command I was playing with:
> 
> sox 110925_180231_SRS001.wav OutgoingFile.wav -S compand 0.3,1 6:-70,-60,-20 -5 -90 norm
> 
> but everything is coming out very clipped.

That is IMHO because the input to 'norm' is already clipped.

You might want to use the global -G option to prevent clipping,
and -V to see what is happening in detail.

Also, your parameters are apparently different to what you used in
audacity: different attack (0.3 vs 0.1), different noise floor (-90 vs -40),
different compression (-60:-20 = 3:1, not 2:1).

(The parameters '0.3,1 6:-70,-60,-20 -5 -90' seem to come right from
the manpage example for compand, and don't seem to be suitable
for your particular sound.)