Firstly AYPK don't confuse compression with limiting.
Limiting catches any peaks and should only light occassionaly.
Personally I'm not a great fan for compressing final mixes, but when/if I do, I use between 2:1 and 4:1.
(Ccertainly never mix into a compressor. Get the mx first , then try the compression)
Depending on the sound you want, what I find works better is not increasing the ratio, but lowering the threshold.
ie 20dB gr at 2:1 (-10) "sounds" very different to 5dB gr at 5:1 (-6)
(sorry rubbish example figures but it's 2am).
Try it and tell me which you think is louder, has more energy an dwhichproduces the better mix. Don't be surprised of they don't follow each other ;-)
I prefer the more gentle but global approach.
But JMTCW.
Personally I aim "not" to global compress at all. I much prefer to get the individual tracks cooking and just mix it.
I might tickle a 2:1 at 4dB and use a limiter at -0.5dB to catch any remaining peaks.
For me, I find this method retains dynamics, often improves perceived loudness and doens't sound squashed to f... but has high energy.
