Some of the most difficult to understand effects processors are dynamics processors. These include compressors, expanders, noise gates and limiters. We all know effects like distortion, equalisers and reverbs. Even if you don't know exactly what they do, you probably know what they sound like. Dynamics processors are different because they work on the... well... dynamics of the sound, and are not usually detectable unless they are used heavily, and even then, often only by a trained ear.

What are "dynamics"?
Dynamics are the differences in amplitude (level) that a signal has. Most sounds have components which are louder than others - parts of the sound which stand out. A piece of music is "dynamic" if there are large differences between the quieter and louder components. The difference between the quitest and the loudest parts of a sound is called the dynamic range.

Take a look at the wave files below of: a distorted electric guitar on the left, and an acoustic guitar track on the right:

While they both have the same peak level, the electric guitar track is denser, with less dynamics, while the acoustic has more dynamics. The electric guitar has a higher average amplitude (level).

Sometimes, you want to make a track more dynamic because a dynamically squashed track doesn't always cut through, while one with good dynamics peeks through the mix. Other times you may want a track less dynamic - if the overall track is too quiet and you cannot raise the level without clipping. This is where compressors come in.

What Compressors Do
The basic function of a compressor is to alter the dynamic content of the program material. It does this by boosting the amplitude of the quiet parts and reducing the amplitude of the loud parts. The result is that the difference between the loud and quiet parts is lessened, and that means the track's dynamics are reduced. By lowering the dynamics, the overall track level can be boosted, and thus the track is made louder. Using compression on a track can help it to "sit" in the mix - always audible, never too loud.

Compressor Controls

Threshold
The level at which the compressor starts reducing the volume. When the threshold set at it's highest point, the compressor will not change the sound's volume. As the threshold is lowered, lower peaks will trigger the volume reduction effect.

Ratio
Ratio is how much the amplitude will be reduced. A ratio of 2:1 will reduce the amplitude of peaks crossing the threshold by a factor of two - this means that a peak 5dB above the threshold level will be reduced to only 2.5dB above the threshold level. A ratio of 5:1 and the peaks will have their amplitude reduced by a factor of five - the same peak will now be only 1dB above the threshold level.

Attack
Attack is the length of time before the compressor reduces the volume. Setting a higher attack value lets more of the peak through before compression starts.

Release
Is the time it takes before the compressor allows the volume to return to normal after the peak has fallen below the threshold.

Basic Compressor Useage

Evening Levels
Sometimes you have a track which has an occasional peak which jumps out at you - way louder than the rest of the wave. If you mix this track loud enough to hear the quieter sections, the peaks are too loud. Mixing it lower makes the bulk of the track disappear. The levels are just too uneven to use - something like my bass playing... In short it looks like this:

Ummm... VERY dynamic (as it were) and unfortunately not what I want. So I use compression and get this:

I used (a lot) more compression than I normally would to make the effect obvious: threshold -12 db, ratio 10:1, attack 10 ms, release 40 ms. This fixes the problem, and gives me a smoother bass sound. It can be a bit obvious when done to this extreme, but sometimes that's the price you pay for having - or being - a lousy bassist.

Controlling Peaks
Sometimes the track will be fine apart from a few peaks which jump out, but otherwise the rest of the track is fine. Fixing this is easy - use the compressor as you did to even levels above, but set the threshold carefully so it clamps down on the peaks, nothing else.

Accentuating Attack
You can also use compression to make a track "bite" (jump out of a mix). If a track doesn't bite, it's because it doesn't have enough attack (where the initial part of the wave is louder than the sustained part). As mentioned earlier, the attack control on a compressor lets a certain amount of signal pass through uncompressed before compression kicks in.

Here's a track that had almost no attack before compression:

Note that there is no difference in volume between the attack portion and the rest of the note. So I used a compressor with a 5:1 ratio, threshold of -16 db, attack of 100 ms, and release of 10 ms to produce this:

You can see that the attack portion of each note stands out more from the sustained portion. Now, when this track is mixed, you can hear the guitarist hitting the strings, instead of getting just a wash of sound. We just used our compressor to increase the dynamics of the track instead of decreasing the dynamics!

Beware
Compressors can leave some undisirable artifacts by making something which was relatively quiet before compression stand out more: the string squeaks from a guitarist or the lip and breath noise from a vocalist. Of course, that could be the effect you are looking for!