Speaking of "watching Ozone" - one of my favorite pastimes - I'll relate a recent experience that's tangentially relevant to the topic at hand, as it's about metering.

Last week I downloaded a trial of the IK Multimedia Metering Suite, since it's currently on sale. This has all the usual metering tools, most of which any Ozone user does not need. But it also includes a novel "perceived loudness" meter and a range guide for different musical styles that I thought might be useful, or at least interesting.

I chose "pop ballad" from the menu of styles, one of the more conservative dynamic ranges. Then I brought up the song I'm working on now. According to the Perceived Loudness meter and the range guide, I was way too quiet - even though the song is not a ballad but an up-tempo rock/reggae tune. So I cranked it up (lowered Ozone's maximizer threshold) until I was within the meter's suggested range.

Following the meter's recommendations, Ozone's compression meter went from showing 2-3db maximum compression to peak reductions of 12-14db, way more than I would ever use. And it sounded just how I'd expect it to sound: loud but squashed, wimpy and dull. My nice thumping bass line lost all its attack, the snare crack turned into a snare scratch - just what you'd expect from a severely overcompressed tune.

I also noticed that the values did not seem to agree with Ozone's meters, which showed higher RMS readings. I brought up a few other meters for comparison and every one of them agreed with one another - all except the IKM product, which measured 3-4db lower on the RMS scale. It seemed to be in its own world. And even as I raised the RMS value the Perceived Loudness barely budged.

My conclusion: the IKM Metering Suite with its Perceived Loudness meter is at best useless and at worst harmful. If someone relied entirely on this meter they'd be doing their music a disservice.

---edited after Jeff's excellent post reminding me about threshold---

For me it's not so much a question about how many layers of compression as it is about answering two questions.

1.  Does this compressor contribute to the tone I want for this instrument? If not, use a different one.
2. Does this compressor make the instrument move or breathe the way I want it to? If not, a different compressor may be the answer if the one you're using is too fast or slow overall. Most likely, though, you'll just need to get the ratio, attack, release, and threshold set correctly.

The easiest way I've found to get the ratio, attack, and release correct is as follows:

1. Set the ratio as high as it goes and the Threshold pretty low
2. Set the attack as slow as it goes and bring it towards its fastest setting, listening to what you're allowing through before the compressor squishes the snot out of the sound. Find a setting where you feel the right amount of transient gets through before the compressor clamps down
3. Set the release as fast as it goes and start moving it to its slowest setting, listening to what you start to gain back after the compressor stops clamping down. Find a setting where the right amount of sustain is heard as the compressor releases
4. Raise the threshold until you feel like you're "squishing the sound at the right place
5. Back off on the ratio and until you feel the instrument or mix "moves" or "breathes" the correct amount with compression

In the above I use the terms Attack and Sustain. These are equivalent to the "A" and "S" controls on an old synth that has ADSR controls. Think about snare hit:

Attack- the first, "snappiest," loudest part of the sound
Decay- the part of the sound just behind the attack where the volume comes down to a lower level and then plateaus
Sustain- the plateau in the sound wave before it starts to trail off
Release- the part of the sound after sustain where the sound fades away.

FWIW I like to set the output gain of my compressor so the track sits right in the mix and looks good on my meters with my track fader set to unity (0.) I can then use the fader for fine adjustments.

---edited after Jeff's excellent post reminding me about threshold---

Thanks for that Dave. Just confirms some of my thoughts about such meters. I prefer to stick to a well known standard (of rms and peak levels) and you guessed it a nice chosen K system level. The BlueCat meter is very good and has all you need. Can be set for the K reference of your choice so rms values are hovering around 0 db Vu. I know I go on about this but I like having a set of good quality hardware VU's showing average rms level too.

The K-12 reference is pretty loud at the end of the day. It can be mastered up from there of course but it still is quite loud and they have adequate transients. He calls that the broadcast standard. The others are just getting quieter and have increased transients. I find K -14 is very good for a wide range of jobs and applications. K -20 is pristine for the ultimate sound of transients.

I have mastered a few albums and used the hardware VU's to show me what the rms level of the music is of course. I have found that in many cases just keeping tracks fairly uniform across an entire CD will put you in the ball park nicely. And yes you do alter the odd one here and there. Ones that just stand out a little for some reason and it is ususlly just 1 or 2 db added or subtracted.

In fact the BlueCat has got another graph that can be related to perceived loudness level. It shows you all this in real time over a long time range so you can see how loud tracks are in relation to each other. Studio One makes it easy to master tracks too.

tunekicker has got it right too in terms of what to adjust in what order. Attack, Release, Ratio. Lower ratios make the sound bigger. I would like to add Threshold last because then you have got just that last amount of fine control (after you have set all those parameters) over excatly how much that compressor is going to work on your music or not.

Great catch on the threshold Jeff! I sometimes use compressors that don't have thresholds (a la 1176) and spaced that one. I just edited my original post to add threshold in the workflow correctly.

Peace,

Tunes

Take a look for "Michael Brauer Compression" on Google- he's got a lot of info out there in articles as well.

Peace,

Tunes

Here is my uneducated opinion wrt to multiple layers of compression and the drums.  I use a compressor on each track but as of late, I do NOT put a compressor on the drum bus.

My cymbals are Paste, and I have been playing this brand of cymbal for more than 20 years because I like the way the sound and I know what they are supposed to sound like.  The cymbals in my home studio are captured w/ sm81’s.

I have been experimenting w/ compression on the individual drum tracks and on the drum bus for a long time now and to my ears … that additional layer of compression on the drum bus changes the timbre of the cymbals.

Also, I do not compress at the master bus for the mix.

When I do my own pseudo-pre-mastering I add one more layer of compression using a multiband compressor before the limiter.

I agree with quantumeffect  in that overhead cymbals can sound easily processed. What I do sometimes is assign all the drum sounds to the Drum buss where I might have some obvious compression. And that Drum buss ultimately goes to the Masterbuss. And I sometimes assign the OHead cymbals direct to the master buss instead or to another OHead buss only with some very gentle compression on it. That works too. The OHead cymbals don't have to be that loud either to be very effective.  Minimum voltage maximum illusion situation. I also find a single OHead mike works well too.

I will use compression on some of the drum tracks, but more sparing these days than I used to. I almost always compress toms, but kick and even snare are negotiable for me with compression depending on what sounds best. For a natural sound I rarely compress the overheads at all.

My preference lately for a more natural sound that still has the cojones for rock songs is using 3 busses for the main drum signal (more if adding reverb to snare, etc.) All overheads, high hat, and cymbals to buss A. All Kick, Snare, Toms to buss B. Buss A and B to Buss C as the overall drum buss.

I'll hit the Kick, Snare, and Toms pretty hard with a UAD Fairchild 670 and leave compression entirely off the overhead tracks and buss A. Then I mix output levels for buss A and B appropriately before using a compressor on buss C to "kiss" the overall drum sound. Usually this is with the UAD EL Fatso as it adds warmth and punch.

This gives great movement, energy, and volume without making it too compressed overall. The best benefit for me is that if I don't mess it up in the master buss I get solid drum hits without having the cymbal levels drop every time the drummer hits the kick. This allows the cymbals to be transient when hit, a wash when they're sustaining tone, and left alone otherwise.