Frequency Specific Side-Chain Compressor

  • Thread starter Thread starter The Cancers
  • Start date Start date
The Cancers

The Cancers

New member
I had an idea for a particular kind of compressor and wondered if it already exists.
To start, some terms as I understand them:
Side-chain: In a compressor circuit, a way of compressing a signal using a source besides the signal itself.
"EQ in the side-chain" : A way of ensuring compression does or does not occur when certain events happen - ie, A de-esser uses the same signal being compressed for the side-chain input, only with an eq in place to emphasize sibilant sounds so that they are loud enough to trigger the compressor while other elements are not. Another place an eq might be placed in the side chain is when one is compressing program material but doesn't want heavy compression with every bass note or kick hit - filter out some low end in the side-chain to ensure more even compression.
Multi-Band Compressor: A compressor that acts differently on different parts of the frequency spectrum.

OK, assuming all that is basically right, I'm wondering whether there exists a frequency specific side-chain multiband compressor.
My best way of describing this would be simply to describe an application:
Say you've got some sort of giant, thick, full spectrum pad going on throughout a song. Say it's burying the kick. One solution might be to compress the pad with a side-chain keyed from the kick. However, as I understand it, doing this would cause full spectrum compression of the pad.
The tool I'm envisioning would cause frequency specific compression across the spectrum of the pad. That is, with each kick hit, those areas of the frequency spectrum where the kick displays higher amounts of energy (generally, the low end and, depending on the kind of recording, the mid/high-mid where the click of the kick lives) would be compressed more heavily on the big pad track than those with less energy (say, generally, those "bouncing ball" sounding low mids you eq'ed out or the highs that weren't very present in the first place), leaving more of the sound of the pad intact and creating a less noticeable sounding compression than a full spectrum dip.
Now, as I understand it (and this is why I went to the trouble of defining it), an eq in the side-chain does not get at what I'm talking about here, because it only effects the detection circuit; the compressor itself still acts on the entire spectrum. Nor is a multi-band compressor with side chain inputs quite what I'm getting at either. It could be made to emulate what I'm considering (and probably is), and it's the closest thing I've heard of to the tool I'm considering, but it's a bit like the difference between trying to determine the area under a curve by cutting that area into rectangular segments and estimating (s-c multiband compression) and using calculus to find that exact area (what I'm picturing).

So, that's about it. Wondering if anyone knows of a tool that works like this. I'd be happy to try and better explain it if it doesn't make sense as I've described it, so please let me know. Also, if you understand what I'm describing but it seems unnecessary because there's some other tool that does something similar but in a simpler way, let me know.

Finally, I'd like to say that, though I haven't been active on the forums lately (thankfully because I've been busy out here in the real world :p), this site remains a constant source of great information and interesting reading, and I hope to continue finding great people and great opinions here for years to come.
Thanks.
 
You can (and I do) do what you describe with the waves c1 which will compress the signal within specified EQ from very narrow band to the complete frequency spectrum sidechained to to whatever source you like

You could do this to a degree with the c4 multiband which gives you variable Q and variable frequency centers for each band but it's not as surgical as the C1
 
Gotcha. I haven't used the C1, but looking at it, it seems like it does more what I was talking about as a workaround with a multiband compressor. What I'm picturing would be settable as far as your normal compressor settings (attack, ratio, etc...), but these settings would be applied along a frequency gradient. Like, you'd take the fft graph of the signal you're using to key the sidechain, and the compression would be applied along a gradient that matches the major peaks and troughs in that graph. So, basically, there wouldn't be any eq style settings (q size, frequency, etc...), because this information would all automatically be determined by the information contained in the keying signal.
Thanks for the response, and if I've got it wrong, please let me know. Meanwhile, I'll have to figure out a way to try out that C1.
 
Back
Top