Ok cool, thanks TK and Dirt . . . now I think we have a basic understanding of what it does.

When I said "settings to use", I meant that in a post-processing sense. Say I have a song with a lot of heavy distortion and some soft parts. And in order to get the heavy parts to fit in without clipping I had reduce the overall volume of the whole piece, thus making the soft parts barely distinguishable. How could I use compression to equalize the volumes of the soft and hard parts (or at least make the transitions less definable)?

So in other words, what settings would I use in my post-editing program (Sound Forge, Goldwave, whatever) to achieve this?