3. Improper Attack and Release
This common problem affects even experienced engineers: I’ve received masters where the compressor’s attack and release fought against the groove of my mix on more than one or six occasions.
It’s a pity, because improper use of attack and release can take a glorious sound and smother it to death, while proper settings can pleasantly emphasize the rhythm as they tame the peaks.
Years ago, after much research and experimentation, I went with the following technique—and I've been happy with the results:
With a quick release, a high ratio, and a low threshold (not too low—you should still see some bounce-back on the meters), I finetune the attack until I like how it's clamping down. Then, I work on getting the release to a place where I like how the signal returns to its original state.
The trick here is that I don't care about pumping at this stage. In fact, I want pumping. I want to love how it's pumping. If I'm working on a drum loop, perhaps the way the compressor clamps down on the snare lends a nice bounce to the hi-hats. If it's a chicken-pickin' guitar part, maybe the way the downbeat hits the compressor spins a nice swing on the upbeat. The point is, I listen for musical feels the compressor imparts in this exaggerated state, and once I hear something I like, I bring the ratio lower and lower—past where I’d want it, just to make doubly sure. Finally, I swing back to the sweet spot. Threshold comes next, with the following tip:
4. Improper Threshold
With the threshold point set too high, you're basically using a compressor for its tonal characteristics (which is fine, if that's your intent—but often it isn’t). Set it too low, and you run the risk of squashing too much dynamic movement.
There is a goldilocks point for every audio source, and it can be categorized in a couple of ways:
A) The action point, around which the audio begins to bounce in a pleasant manner.
B) The ceiling point, above which all sudden spikes come down to a reasonable strength.
The former conceptualization works better across groups of instruments, while the latter seems better for individual, “poky” sound-sources—vocals, basses with sudden spikes, et cetera.
To find this sweet spot with surgical precision, first mark down the attack, release, and ratio you’ve secured with the technique mentioned previously. Now, dial in a medium attack, a medium release, and a higher ratio, say between 4:1 and 6:1. As you adjust the threshold, it’ll be easier to hear where this “action” or “ceiling” point is located. Don’t forget to return to your original attack, release, and ratio settings when you’re done.
After a while, you won’t need to go through this labor-intensive process—you’ll be able to judge the action/ceiling point with the other settings already dialed in.
5. Compressing an Already-Compressed Sound
Remember, the sound in your session might’ve been compressed already—and not just by the engineer. A distorted electric guitar is compressed by virtue of its overdrive. Likewise, most synth patches have already been treated by the producer. They might come to you with compression, but they definitely arrive at your digital doorstep with their envelopes and LFOs finely-tuned for maximum impact. Your compression could negatively impact these parameters. If you're working from a sample, well, it’s probably been processed too.
You might be noticing a theme here—yes, the “don't overdo it on the compression” mantra might be wearing thin—but it’s simply biggest trap when learning how to tame dynamics. I know I certainly fell victim to this practice in my engineering infancy.
So here’s an exercise to get you speedily over the hump: Take a track (any track will do) and squash it to the point where you can easily hear that it's too squashed. Now study the aspects of its timbre, so that you drill down on what too much compression sounds like. Home in, specifically, on resulting tonal changes that might be unfavorable, or unmusical changes in the feel/groove. Now dial all the settings back halfway or so, and listen again. This time, switch between bypassed signal and instantiated sound when listening. Upon hearing the compressed signal, do you recognize any of tell-tales of over-compression? Keep playing around with the ratio of these settings until you start to notice when things sound just right. Then, when you do make the call to compress any instrument, you’ll know exactly why.
By all means, compress. But do so smartly: consider all these variables before compressing, because you could end up fighting against the quintessence of the sound.
6. Placing the Compressor Improperly in the Chain
Pick an engineer’s brain and you might very well hear an absolute statement like, “I always EQ into my compressor," or, "I always compress into an EQ,” or occasionally, “I do all my cuts, then compress, then do all my boosts.”
All of these tactics are fine, provided the sound calls for the tactic. You have to be the judge of which tactic is the best. The big sin here—and consequently, the big mistake—is to subscribe in totem to one process over the other.
It’s the sound that matters, not your approach, not your ideology.
7. Using Multiband as a Shortcut to “Loudness”
In this instant gratification world, where so-called “radio ready” mixes are expected quickly, multiband compression can cut a quick path up the mountain of loud. It allows you, in theory, to skip massaging individual tracks in favor of processing whole busses, dynamically taming wide swaths of frequencies in the service of pushing the level.
In my earlier days, I certainly fell victim to this practice. I had a partner in top-line crime out in LA, and on our mix-buss, we would slap on a round of EQ, compression, and multiband, followed by another round of EQ, compression, and multiband—all of it driving a limiter.
Was it loud? Absolutely. Was it good? Far from it: Every sound was constricted in an unmusical box. And the groove? Decimated. The loudness only served to make people turn down the mixes, till our petty sounds screamed nasally at even the quietest settings.
This is a pitfall of multiband dynamics processing, be it of the traditional kind, or the dynamic EQs more prevalent these days. I’m not saying avoid multiband like the plague. Instead, learn the intrinsic strengths, weaknesses, and rules. For instance, if I have a vocal with too much meat in it, EQ might not do the trick. Knocking out the low midrange could kill all the body outright. A multiband compressor with a slow attack time on the offending frequency band might be just the ticket. Now the transient of the warmth can peak through, giving you the illusion that the frequency is louder than it is. You'd get all the warmth and none of the mud.
The takeaway: Pay attention to what multiband processing is doing to your signal—particularly at the crossover points—and you’ll get more mileage out of the technique, rather than complete sonic annihilation.