When do you use COMPRESSION?

[publicradio]

I struggle with this, too. For me, the answer is:

• On recorded items. Compression helps me turn up an electric bass so the bassline is clearly, tonally audible, without loud plosives in my low end overpowering the mix and competing with the drums.
• When I want to create apparent loudness. I'll put a reverb on a channel--sometimes, two reverbs, a tight and a big one---and then compress the signal after the reverb, which makes the recording seem louder.

That's basically it. I always put a tape simulator on my master channel, so I have high saturation from the beginning of the project. Other than that, I use compression very sparingly. I might not be doing it right, but I have never known when else to use it. I've never done ducking or sidechaining, for example.

With non-recorded channels (i.e. VST synths and drum machines) I have never felt the need to compress because I have such granular control over those sounds. If a kick drum is too plosive, I can just change the output in the VST. If a synth line has loud peaks, I can just bring down the velocity on individual notes. There might be some compression magic that I have yet to understand, though.

[jmm6486]

Basically if I have any sound that has loud parts that are too loud and quiet parts that are too quiet, I'll use compression to keep it more even throughout

[apan]

More experienced people may clarify this, but I read long ago about using compression to make some thing feel "closer" in the mix while reverbs tend to make things seem far away. Then panning gives you width. This all combines into a 3D image where you want a lead element that gives the listener a focal point. So along with everything else said, it can help bring a lead element forward.

Sometimes it is overused so much that you feel like you're in a closet with the singer.

I picked this up from the Mixing with Mike Youtube channel. Using different attack times you can affect the spacialization of sources in different ways. Longer attack times emphasize and shape the transient, with makes things sound upfront. Faster attack times csn soften the transient, which tends to sit things back. Similarly, the release time can be used to either stabilize a source, which tends to keep it back, or make it pump or groove, with tends to let it poke out. This approach can be subtle but quite effectice when used in conjunction with high end roll off for more distant sounds, and then of course delay and reverb