MrFigg wrote: ↑04 Oct 2018
So I can hear the differences, both subtle and obvious, in different effects. Dist, chorus, phaser, flanger etc. etc. I can hear different tones in different EQs. All of these effects I use, selecting carefully to get exactly the sound I want. But I’m really still not getting compressors. I’m sure it must be the same but I can’t hear it. Well, I can hear a marked difference between say the FET and the RE-2A. Got them both and use them on different stuff depending. Also got Curve, couple of DCam and probably a couple more I got bundled but,that said, I don’t get it. There’s so many in the shop as well. If I go ahead and buy the Mastering Rig upgrade from 2 to 3 I’m going to get another three compressors there too. I guess I could maybe be dubbed a philistine for this but could someone please help me to understand? Thanks
I can tell you this - you're not alone. Compression took me the longest to learn of any audio effect. I can honestly say it was at least 5-10 years into my engineering career before I could say I "knew" compressors well.
Why is compression so difficult to master?
Audio level is probably one of the most difficult qualities to hear well, and pitch is probably one of the easiest. We can tell when two sounds are a semitone apart quite easily. There are 120 semitones across the audible range. Most of use can hear when two pitches are a half a semitone apart, or even closer. Let's use the 120 value and compare that to level, using decibels. There is also roughly 120 decibels across the audible spectrum, from the threshold of hearing to the threshold of pain. It's actually more like 140, but that requires a dead silent room to hear the bottom 20 dB. So how many of us can clearly identify a change in level of 1dB as obviously as a change in pitch of 1 semitone? Sure, we will probably be able to tell one sound is softer or louder than the other, but by how much?
And as EQ is basically gain at a specific frequency, we can hear the difference between an EQ tuned to different frequencies more easily than a difference between different boost/cut amounts (again, pitch vs level).
Of course, delays are also easy to hear the difference between a short slap delay and a long repeating delay, or a short room reverb vs a long hall. Even distortion is easier to identify once it's past the super subtle level (again, because "level").
Plus, with the except of distortion, all of these other effects are linear and static - they are the same every time regardless of what you feed into them. Compression (and distortion) is non-linear, which means if you play louder you'll hear a different effect than if you play softer. And it's not a static effect either, constantly changing when it's "working", making it even more difficult to pinpoint exactly what's going on.
So compression, which is often only micro changes to level over time, is the most difficult to learn to hear IMO. But there is a path forward, if you're willing and patient!
Learning to hear compression
My only advice would be to start simple. Learn which compressors are favored for which instruments, and start using those compressors in that way (until you get more confidence). This can be as simple as an opto compressor (LA-2a and clones) on vocals or bass guitar, FET (1176 and clones) on drums or more aggressive instruments such as harder vocals or guitars or harder bass sounds, over-easy (soft knee) compression for more subtle control of level such as wild background vocals than need to be smoothed out, and VCA based compressors (SSL) when you need more versatility for things like punch or glue. And you'll probably find folks that don't agree with my categories, so take them with a grain of salt and gather multiple opinions as always! And after you're familiar with the classics, try using a "smooth" compressor on a source you want to make "harder" and see what happens - you never know what will work until you try. And as you get more familiar with what each control actually does, you'll begin to hear the difference between different types and brands of compressors. How do you learn? Read on!
The next step is to learn the basic controls: Threshold, Ratio, Attack, Release. Work with these by exaggerating them so you can better hear what they're doing. Everything is easier to hear when it's exaggerated, and with only a little practice once you return the controls to more subtle settings you'll still be able to hear the "effect".
Start with Ratio, which is basically a depth control for the whole compression effect. A ratio of 1:1 means no compression no matter how the other controls are set, while a ratio of ∞:1 (or the highest ratio provided) is maximum effect. When trying to hear what the other controls do, use the highest ratio possible. Next up is threshold, which is a little easier to understand since it's the point where everything kicks in. If a signal is below the threshold, nothing can happen. And so, like with ratio, there is a setting where it's possible for nothing to happen (when the threshold is above the loudest level coming into the compressor). And so the threshold setting that has the greatest effect is when it is set the lowest. Combine this with a high ratio to exaggerate the next two controls, known collectively as the "time constants".
Attack and release are familiar to any synth head, but the do a slightly different thing with compressors. Assuming a signal above the threshold, attack affects the signal any time it increases, not just at the point it crossed the threshold. Same for release. So again assuming a signal above the threshold, attack affects any positive going signal, release affects any negative going signal. Even if the signal doesn't fall below the threshold between musical "hits", attack time will still be relevant.
Examples
Try this: use a long release time, and play a kick fast enough so the signal never goes below the threshold (use ReDrum to make it easy, and the MClass Comp with threshold all the way down and ratio all the way up). With the release time at max and ReDrum playing 16ths on a kick sound, turn the attack up/down and hear the difference. Even though the release time keeps the signal above the threshold, the attack control still affects each new "hit".
There's a relationship you can hear: slow attack sounds louder (because the attack is allow to pass), fast attack sounds softer (you can always increase makeup gain (or output level) to compensate.
Now with ReDrum still playing, set attack to the middle and turn release up/down. The relationship with release and level is the opposite - fast releases sound louder, slow releases sound softer. Remember this when setting up compression and compensate with makeup gain!
Now this isn't a hard and fast rule. For one, it depends on the range of time constants the compressor provides (if it provides them at all - notice the LA-2a family has no time constant controls). But it can be helpful when you wonder why things suddenly got softer/louder after adjusting a time constant!
Another more obvious relationship is that a lower threshold or higher ratio will also result in a decrease in level in most cases, with exceptions for when there is an "auto-gain" circuit in play (such as the SSL Channel Comp).
Starting points
Finally, there are a few basic ways to use compressors that are worth noting and can serve as starting points, especially when trying to "audition" or learn a new compressor.
One approach is gentle compression, which starts with a low 2:1 ratio and a low threshold. Set the threshold so the signal is mostly above it, which gives you gentle compression most of the time. Try medium time constants to start, adjusting faster attack for more gain reduction, or slower release for more transparency. This is a handy setting for bus compression if you don't hit it too hard. Sometimes all you need is 3-6 dB gain reduction to do the trick.
Another common approach is limiting (aka "peak reduction"), which uses a high threshold and a high ratio, and typically fast time constants (especially attack, for maximum control of the level). Set the threshold so compression only kicks in on the loudest peaks. A faster release will sound "louder" but at some point if the release control goes fast enough you may hear distortion, especially on lower frequency sources. This is not always a bad thing, just be aware of it and use it consciously!
Finally there is pumping (obvious) compression. Use this when you're not being shy with the effect. Try a high ratio and a medium threshold (adjust to taste from there), and a slower attack with a faster release. You'll be letting the peaks through just long enough to punch, then grabbing them and yanking them down for maximum effect. This approach can impart a "hard" attack to softer sounds. I use an 1176 or the SSL channel comp effectively for this. On the SSL use the defaults and set the ratio to max, the release to min, and lower the threshold until you see one or two LEDs light up. On the 1176, use the slowest attack and fastest release and the highest ratio, and aim to deflect the needle by a good amount (at least half way down). On the 1176, you have a fixed threshold and get more compression by increasing the input, which sometimes requires decreasing the output to compensate (the opposite of threshold based compressors). From there you can adjust the release time for a "breathing" effect, timing it with the music so the signal rises back up just in time for the next smack down.
Didn't intend to write a wall of text, hopefully something here will be helpful and hopefully others will also respond - I learned a lot by hearing many DIFFERENT opinions (and choosing what worked for me from among them), and as such my goal here is to provide but one voice of many.