Techniques - Compression

Compression

Compressors work on the amplitude (gain) of a signal. A compressor or limiter 'turns down' any part of the signal that goes above the threshold. How much compression (gain reduction) is applied is determined by the ratio. The attack and release settings control how long the compression takes to fully kick in and fully end respectively.

For example, if the threshold is set to -6dBFS and the ratio set to 2:1, any signal that goes above -6dBFS will be attenuated by a factor of 2 therefore a peak of -4dBFS will be reduced to -5dBFS, i.e. the 2dB overshoot is halved. If the ratio was set to 4:1 then the 2dB overshoot would be reduced to a quarter (or 0.5 dB) resulting in a signal of -5.5dBFS, and so on.

If the attack is set to a low number (sometimes labelled as 'fast') the time between the threshold value being reached (or, actually, exceeded) and the compression being applied is quite short resulting in a 'hard knee' response. This is more efficient at catching transients and therefore can be more sonically obvious than a higher (slower) attack value. Similarly, a low/fast release value will disengage the compression quickly, which can lead to the 'pumping' effect often heard on club bangers. Obviously, if that isn't the effect one wants an increase in this value will eliminate that.

Because the overall effect of using a compressor is that the gain/amplitude (volume level) is reduced, most compressors also include a make-up gain control that sits between the output of the compressor circuit and the output of the whole device. By adding gain at this stage, everything (including the 'quiet' sections) is increased in volume.

Compressors, by definition, reduce the 'distance' between the 'loudest' and 'quietest' peaks. This distance is more commonly referred to as dynamic range and any reduction in dynamic range will have an effect on the overall sound and feel of a piece. Sometimes this effect is exactly what's desired/required by the engineer/client but in most cases the opposite is true, which is why most professionals and experienced amateurs only use compression sparingly.

However, besides dynamics control, compressors are also used to 'glue' tracks together, e.g. a bus compressor might have a fairly low ratio (say in the range of 1.1:1 to 1.5:1) and mid to slow-ish attack and release values, that conspire to make all of the tracks that are routed to it sound as if they belong together.

This technique can be very effective on the mix bus too provided that the compressor is sonically appropriate for all the tracks in the song. That usually means that the mix bus compressor is 'clean' and 'uncoloured' but, through experimentation and experience, one might find that there are two or three candidates for this job depending on the programme material. It's best to apply mix bus compression from the start of the mix process, i.e. that one mixes into the compressor as opposed to adding it later in the process, especially at the end - any compression added at that stage is best left to the mastering process (if indeed it's required at all).