Compression….again!!??

MLA-2

“What on earth can you tell us, that we haven’t read somewhere already Tone?”. You can let me know later :)

I think maybe the biggest thing I can do, is give you some tips on what doesn’t work in compression.

I’ve been recording for about 25 years, but it wasn’t until I got into mastering that I could actually start to make a compressor do what I really wanted. I know that’s a bit of a mad statement, but I’ll explain.

In tracking and mixing you can get away with minor “creative mistakes”. The pump of the compressor on the drums/bass/guitar or whatever can add character to the song, even if you got it slightly wrong. Not the case in mastering, where  there’s actually very little room for creative compression, beyond the natural attack/gain reduction and release character of the model used. This may seem like an obvious statement to some, but new info for others.

I’ve been racking my brains to find a way of doing good/bad audio examples, but the fact is, if you can hear the differences, you’re probably getting it right already :) .

Pay close attention to the image!

Ok, so you probably all know that the ratio adjusts the amount of compression, but what about a knee setting?.

kneeYou may not have seen this on too many comp  models, but it can be extremely useful in mastering.

The knee control adjusts the speed the compression sets in, giving far greater dynamic control. The harder the knee, the faster and closer to the threshold setting the compression starts to act. Soft knee starts to act further back from threshold, and gives a smoother sound as it gently reaches full ratio.

“How do I tell what knee setting I need?”

Listen to the image!. In mastering, I’m usually using a setting of between 10-14  db slope. If you set the knee too hard(lower db number), you should hear the image narrow, almost as if its pulling into mono as the compression sets in. If it’s too soft, the compression wont happen fast enough, and you may find yourself trying to compensate with threshold and ratio.

The attack=pressure release valve

“He’s finally flipped and started talking about plumbing”

Have a look at this (no laughing at my photoshop work)

compression limiting diag copy

Humour me and imagine the shape of your material looks like this. If you try to compress that peak (shown in red) it’s most unlikely you will get the desired result. That peak could be quite a few db over the level of the body of the sound, so the compressor wont make a great job of dealing with it. This is where the attack comes into play, it allows the untreatable peak through untouched, for the limiter to deal with. So for most dynamic material, it’s best to setup using at least a short attack.

Initial setup is a case of picking a ratio and find the right threshold for it. If you raise the ratio, raise the threshold. Leaving it static and changing ratio isn’t going to work!. Forget the input meter for setting threshold!. Pick a ratio (start lower rather than higher)then drag the threshold down until the sound starts to blend. Not enough compression…raise the ratio then re set the threshold. If the peaks start to actually sound louder, open the attack time. The thing to pay the most attention to is the image. If the image is being damaged, the settings are wrong.

Hearing compression can take some experience. I read an interview with top mix engineer Michael Brauer(Here), where he admits it wasn’t an instant thing, even for him. It wasn’t until I mainly read the image that my compression started to work. I now have a constant print of the tonal curve in my head. If it starts to get squashed by anything I’m doing, I know I’ve done something wrong!.

As I said in one of my previous articles “Get a limiter in there….Fast!!!”, that peak that’s been allowed through the attack still needs dealing with, and sooner rather than later, before the overshoots numb your ears to the offending freq’s within them.

How much gain reduction?

If the material you’re working on is particularly dynamic, trying to bring it under control with one comp probably isn’t going to happen. 4db reduction is starting to push your luck, before it becomes very audible. Don’t be afraid to use one model to do most of the work, and then use another to finish the job. Many have different characteristics, finding their strengths and weaknesses, and where to use them is all part of the fun. Multiband compression can be an extremely powerful tool, but hearing clearly enough to get all the bands right, so the curve goes through undamaged requires some skill. I’ll cover MB comp in a future article.

The most misunderstood control ever

The release control…….

Just noticed how much I’ve written already. Will finish this off in the next post

Thanks for reading people

Tone