I had a break for a few weeks while I was away. Now I am back at work, and back in geek mode. This week and next week I am teaching my students about dynamics processing. That means compression! The following are my notes from the first class. There will be a second part next week when we delve further into the application of these ideas….

Dynamics = loud and quiet

Dynamics processors control the amplitude of your audio, and enable you to shape sounds by manipulating their natural dynamics.
Compression

FACT – compression does not make sounds louder. A compressor attenuates (reduces) the signal. It is only when you use the Make Up Gain afterwards that a sound becomes louder.
Compressors can be used to reduce the dynamic range of an audio signal, which can then be boosted thereby creating an impression of a louder overall signal. This is what is meant when you hear that a compressor ‘makes things louder’.

There are 5 key features you will find on every compressor:
Threshold – this determines when the compressor starts working. When the amplitude of the signal goes over the threshold the compressor starts working.
Ratio – how hard the compressor works.
Attack – how quickly the compressor start working (grips the signal).
Release – how long the compressor takes to go back to zero compression (lets go).
Makeup Gain – this boosts the signal after the compression as taken place.

Other features you may find on your compressor.
Knee – the curve between compression and no compression
Mix – enables you to mix the compressed signal with the dry signal. This effectively enables you to use parallel compression to combine dry and wet signals.
Distortion – many hardware compressors naturally create harmonic distortion (the nice kind!). This feature on the Logic compressor enables you to emulate this natural distortion and effectively saturate the signal.
Saturation = squash = thicker signal.
Drive – similar to the drive on a filter, this will saturate the signal.

Applications of compression

Controlling signal with wide dynamic range
Some signal, such as vocals, will have a very wide dynamic range. This means that some of the quieter words may get lost in the mix, or the loudest parts are too powerful and overpower your mix. In this instance, you will use a compressor to reduce dynamic range and create a more level signal. Be careful though – completely removing dynamic range will suck the life out of the performance.

Look for the loudest part of your signal and set the threshold and ratio so that this section of the signal is controlled (without completely killing it). The attack and release will determine how aggressive the compression sounds. No attack will means the compressor acts immediately but also kills the natural transients of the signal. The following three examples show the ways we can use compression to control vocals.

In this example, the compressor has been set with no attack at all (0ms). This means that all of the transients are being lost, and as the compressor’s threshold was set pretty high, all the vocals are being squashed so much that there is practically no dynamic range. This vocal will cut through a mix, but there is no sense of the vocal building up and it will sound very unnatural. The louder parts are simply being squashed more.

In the second example, the attack time has been pushed up to around 80ms. This means that some transient information is getting through (although you can see some strange spikes which appear to have been caused by the compression), retaining some of the original delivery. The threshold was also reduced slightly so the compression was less aggressive, and the result is a more dynamic vocal. However these spikes are definitely a problem.

In this third example, the attack and release were left the same, but the Distortion feature of the Logic compressor has been engaged. This Soft Clipping has meant that the spikes in the previous example are gone, and the quieter moments are squeezed up a bit louder. This has resulted in more headroom, but still an impression of dynamic range. What you can’t hear from the screenshot is the warmth or thickness that harmonic distortion provides to an audio signal.

Shaping individual sounds
A compressor can be used to manipulate and accentuate the natural dynamics of a sound. For example drums… you can use a short attack (30-100ms) to allow the initial hit (transient) of a drum to hit as normal then the compressor grips the signal. You can shape the front of your drum hits to emphasise the first punch. This can result in very punchy tight drums, but also very spiky drums which will need to be controlled afterwards.

You can achieve similar effects by using a transient shaper to control the attack and release of a drum hit. Combining both tools give you maximum control over the shape of your drums.