compress Specific frequency

trogluddite · 26 Apr 2012

Three ways come to mind - though the first is probably a bit crude these days!

1) Filtering the 'side chain' of the compressor. This is the old traditional way - you use an EQ to boost the sibilant frequencies - but only in the side chain, not the main audio path. Because the sibilany frequencies will now be louder, the compressor will use more gain reduction when it hears them. But it is still the whole signal that get 'gain reduced', so it sounds pretty bad if you need a lot of de-essing.

2) Multi-Band compressor. Split the audio into bands (there's a frequency splitter module on the Wiki IIRC) - and then selectively compress whichever bands you need to. You don't necessarily need to make a full on multiband compressor - you can just compress the sibilance area and leave the rest untouched. A slight downside is that the splitting and recombining of the frequency bands might change the tone a tiny bit due to phase shifts in the filters.

3) Dynamic EQ For this one, you again filter the side-chain to find the frequencies you want to compress. But instead of using normal gain reduction, the side-chain controls the gain of a peaking EQ filter set to the same frequency - cutting those frequencies out more as the side-chain hears more sibilance. Again, you might get some slight tone change from the filters when gain reduction is happening - but when there is no sibilance it should be possible to pass the audio without changing it.

RJHollins · 26 Apr 2012

A 4th concept might be ...

utilizing a 'phase cancellation' technique.

where a particular sibilance band is phase inverted, and then recombined with the original signal, along with a level control that can be used to 'adjust' the degree of 'cancellation'.

aviadderi1 · 27 Apr 2012

Wow thanks you all

I need a small example of osm file to understand the way for create the compressor of specific frequency

Can you help me with this ?

Thanks agin Aviad