|
Indeed the basics of a compressor are very easy, yet it's the most misused device in today's music production.
First of all, if you want a thorough explanation, visit http://www.saecollege.de/reference_material/index.html (compression is under the effects section). It will give you a graphical representation also of what happens. The SAE reference material site is an excellent read for anyone doing something with audio. I suggest you read through it (yes, it's free, and not only for SAE students, so take profit).
The explanation of Dezma is right, but there's more to it than that (otherwise, everyone would use a compressor well, which certainly isn't the case).
You must understand first of all that a compressor is a gain reduction device. It DOESN'T make things louder like everyone is saying. It's just the opposite. It's the output (or make up) gain that is responsible for this.
Threshold and ratio are pretty easy to understand. The threshold is the level where the compressor will start working. Setting it for example at -20 dB, every signal louder than 20 dB will be compressed.
How much it compresses is dictated by the ratio.
The ratio is just the relation between what goes in and what comes out.
To use Dezma's example, 8:1. For every extra 8 dB at the input, only 1 dB extra will show up at the output.
Above example, threshold -20 dB, ratio 8:1. If your input signal is -12 dB, the output will be -19 dB. The compressor will start working from -20 dB on, -12 dB is 8 dB higher than -20 dB. So with our ratio, only 1 extra dB will be presented at the output. -20 dB + 1 dB = -19 dB.
You see here that the compressor actually reduced signal. So it will appear less loud. To counter this, you use the make up gain, to get about the same peak level again (so for example, if you had about 7 - 8 dB of gain reduction, you'll increase the make up gain by 7-8 dB also). After this action, the average loudness will seem louder, as you have less dynamics (the difference between the loudest and softest sound).
Some compressors let you adjust the knee also. A normal compressor usually uses a hard knee, which means it'll start to compress at your set ratio beginning from the threshold (so in our example, 8:1 at -20 dB and over). Some compressors let you select a soft knee, which means the compressor won't start to compress with the full ratio directly, but gradually use a higher ratio, the louder the input, the higher the ratio, until it reaches your set value...
Easy isn't it?
Indeed, but the most overlooked, and in fact most important controls are the attack and release time. Just picture the compressor as a guy sitting behind a mixer. He tries to keep a constant level. He hears a loud peak, so he reduces the fader. How fast he does that, that's the attack. When the input gets softer again, he increases his fader back to nominal level. The speed at which he does that is the release time.
Same for a compressor. But it's those who can make compression sound good or sound bad...
With attack you decide if you leave transients passing or not. A very popular example is a kickdrum. Basically it has two components. A click, and an, erm, oomf. If you use a short attack, the compressor will start compressing as soon as the click is heard. If you use a slower attack, the click will be able to pass unaffected, the only thing compressed will be the body of the kick. In the first example, the click will be made duller.
So you see, the attack can shape your sound.
The release, another dangerous control, you need to pay attention. If it's set too short, you'll hear the compression working. You'll hear a very distinct sound called pumping. Picture that sound guy moving the fader up and down in an obvious way, well that's the sound you'll hear, level going up and down audibly.
Most of the time you don't want pumping and you'll use a longer release, but it can be used creatively too (Satisfaction of Benny Benassi uses that, a little bit too much imo, but just listen to the synth when the kick comes in).
The problem is, you'll probably ask, ok why don't I set a long release time, then the pumping will go.
True, but then the compressor might compress sounds it shouldn't, as it's still raising it's level from a previous peak.
So you need to find a good compromise. Modern compressors have auto release times sometimes, and they are not too bad.
Nowadays, digital compressors and plugins have a very useful feature, called lookahead. Due to the nature of digital audio, the compressor can look at what signal is coming before it actually hits the compressor, and choose the appropriate timings to get the best results. Though, experienced engineers will still prefer doing it manually, as they want total control over everything.
So, a very simple device, but it can save a mix, or wreck it (and when you don't know what you're doing, mostly the second option).
Don't overdo it, watch your gain reduction meters carefully. Unless the compressor is used as creative effect, you'll usually want it to work on the highest peaks only (and then maybe shave off about 8 dB maximum I'd say), not all the time. You use it to get a more constant level of an instrument, so you can sit it better in the mix.
|
) don't go compress it
