I still think it is good to have a structured approach to things sometimes and compressors are in this category. I do have teaching as part of my income as well as composing and producing. And as a sound engineering teacher every day I am faced with the concepts of explaining how things like compressors work to people. Could you imagine me saying Oh just play around with the controls and see what you come up with! It would not work. So we think about structured approaches especially as a starting point and then let people bend the rules here and there. As teachers we are talking all the time about how to teach people how to sound engineer things well. This morning we were getting into how to record piano and vocals at the same time (one person) and get great separation on the two. (Any ideas!)

It is good that Danny has brought up the concept of using a compressor for more than one job as I have only explained the concept of using a slower attack to achieve a transient sound and as Danny also correctly says can add a nice pop to a snare for example. That is a great use for a compressor. 

But another great job of course is levelling. Faster attacks are well suited in this role. You want the compressor to act quickly and do its thing. (notice how if you speed up the attack setting how much more gain reduction occurs without changing threshold. Makes a huge difference. Attack effects many things and is better sorted out early.) 

But notice one can still think of the Attack setting first for the type of compressor job required. Release is also a good second thing to get around for this role too. How fast do you want this to come back? When using a compressor in limiting roles fast Attacks and fast Releases can be very effective. 

Ratio is still a good one to choose in any role next. Limiting we go for high ratios but just some over easy leveling and we might be into much smaller ratios. And of course Threshold can still be last to tidy up how much gain reduction we want to hear and see. For limiting we are using high Thresholds and that is when the compressor/limiter is just catching the odd out of context peak. For more even leveling duties Threshold can come down way lower (with quite high gain reduction) but as the Ratio is low the effect is much more transparent. This is when the lower ratio makes things sound bigger. In mastering the structured ARRT approach also definitely works well every time. No time to wander around setting compressors randomly.

Two compressors in series can be fantastic too but the first one set for limiting with a very high threshold just catching peaks. Followed by a more levelling compressor with a much lower ratio. You can get the best of both worlds. Very even but big at the same time. The Compressor > Limiter path can be wrong in some cases. Limiter > Compressor path can sound much better at times. It is not done often enough. 

Just out of interest in the other Techniques thread Mixing lessons w/ Bobby Owsinski   the man himself chimes in at one point to help clarify some points. Here is what he said about compression.

2) Wow, the compressor thing really got blown out of proportion. Forget the numbers on the things and use your ears. It doesn't matter if you start with the attack time all the way short or all the way long. What counts is that it's set so you can hear some of the attack getting through so the sound isn't too dull. As for the release time, you set it to where you don't hear it breathe, which usually works best if it's timed to the track. As far as the ratio, it's great if you can set it low and it works for you, but another way is the more dynamics, the higher you set the ratio. Again, use your ears. If you can hear it working, maybe that's perfect for the song and maybe it's not.   

Interesting he mentions these parameters in the exact order I am suggesting. Maybe I am onto something.

Awesome thread guys. Thanks for posting it, Jeff. I'll be trying it out soon... and then I'll fiddle with the knobs from there if necessary. Cheers.

One thing that really helped me when I first started using compression was to experiment with the volume really low. When you first start out and your ears don't know what to listen for, it can be hard to hear what the compressor is actually doing. But when you listen at a real low volume, it becomes far more apparent. Worked for me anyway. 

And the other thing is, when you are first experimenting, don't be subtle initially. Let it do quite a bit of gain reduction, so that you can really hear what it's doing. Once you get your head around what it's doing, then you'll be able to listen to it being done more subtly.

And, though you should always mix with your ears, it may not be immediately apparent to you why what you are hearing is happening. So you may want to bounce down some of those unsubtly compressed tracks and compare the original and compressed wave forms, to see how the shape has changed. Has it knockd the peak all the way off, has it made the peak thinner, has it brought up the trailing part relative to the peak, Has it made a dip on the trailng end of the peak and then held down the trailing part, etc...

 I'm sure most folks here have been twisting knobs on their software compressors for some time now and have already learned how to mangle audio with a compressor. I don't plan on getting set in any habit that my ears tell me is a bad idea but after trying Jeff's ARRT method I find I will probably use it for now...seems to have taken me to that same point in the road that my ears would have in a bit shorter time.