Tracking with 2 compressors - how and why?

[Paul Frindle]

Originally Posted by dokushoka

I find it useful to use a combination of a compressor with a slow attack/release with a comp with fast attack/release. Sometimes I'll follow this with a limiter as well.

This works because you are effectively speeding up the compression (and increasing the ratio) in stages depending on how large the gain differences are - and how long they last. The slow one prescales the level based on longer term average, the intermediate (faster) one handles faster peaks that get through the slow one and the final limiter reshapes the very fast peaks that get through both the previous ones :-)

The distortion from the first one is fairly low because it's operating slowly (just gain scaling). The distortion from the second one is greater, but because it lasts less long it doesn't intrude too much (may even help by giving some bite to softer attacks). The distortion from the final one is large, but because it only handles very fast peaks you rarely get the chance to hear it before it's over.

The important thing is to ensure that each stage has enough headroom to avoid clipping before the next.

Although it's a drag setting all this up, you can get a much kinder compression than just using one single stage - for the reasons above.

The dual stage compression approach is used in several compressors around (including the GML8200, and the Oxford Dyn). The Oxford limiter in fact has two discrete signal paths in series (with the last one having two simultaneous reduction stages), to allow the overshoots to be processed much in the manner you mention above. The progressive compression is provided by a variable soft knee function that can operate over a wide range of input level. This has an advantage as the ratio curve is continuous and therefore avoids the hard break points produced by several conventional comps in series.