Compression can be hard to understand but is a vital effect if you are making modern dance music, or nearly any type of music that needs to sound loud. At the simplest level a  compressor works like an automatic fader that lowers the loudest parts of the signal so the overall level of the sound can be raised. Well applied compression will make a good mix sound great but it won’t fix a bad mix and poorly applied compression will sound terrible.

If you don’t know what compression sounds like – think about when you are watching the TV and the adverts come on – they appear to be louder than the actual programme! If you were to record the audio you would see they are actually the same volume – but the adverts have been compressed aggressively to make them seem louder than the programme you are watching. If programmes were compressed this much you would find them hard to watch as the sound would quickly tire your ears but for the sneaky advertisers this bit of ear-grabbing extra perceived volume is a great tool to get the audience’s attention.

You operate a  compressor by telling it at what level it should start to turn the volume down, how quickly it should turn the volume down, and then how quickly it should turn it back up again, how much it should turn the volume down by and then finally you can turn the volume up again to make up the lost level.

You should grab these plugins before you start – the Kjaerhus Classic compressor which is exactly as it sounds, the Bram Smexoscope which is a waveform display which can be really handy for seeing how the compressor effects the sound and finally GComp which is a compressor that includes a waveform display.

Anatomy of a Compressor

• Threshold – This is where you dictate at which level the compressor should start to reduce the signal.  The lower the threshold the more dramatic the compression – if the signal only occasionally goes over the threshold you set then the signal will only be compressed at that point. If the threshold is set very low you compressor will work more of the time, in effect a high threshold will mean less of the effect and a low threshold will mean more of the effect.
• Ratio – This dictates how much compression is applied when the signal goes above the threshold. No compression is expressed as 1:1 – for every decibel of volume over the threshold the amount allowed through is one decibel, i.e. the signal is uncompressed. A ratio of 4:1 means that for four dB that is over the threshold only one dB is allowed through. The higher the ratio, the more the sound is compressed. Ratios of 20:1 or over are extremely high and get referred to as “Limiting”.
• Knee –   Hard knee is a sudden and dramatic change (good for percussion elements) and a softer knee is more subtle and would be better for more gentle vocal compression. A hard knee is brittle – the compressor will start to work as soon as the threshold is crossed, soft knee broadens the range of threshold volume needed.
• Attack – How quickly the compressor starts to work, once the threshold has been reached. With an attack time of 0ms the compressor would work as soon as the threshold is hit, with an attack of 20ms the compressor would start to work 20ms after the threshold has been reached.  An attack that’s too fast will mean you are compressing the initial transients too much, conversely an attack that’s too slow will mean that parts of the signal that should be compressed aren’t compressed.
• Release – How quickly the compressor stops compressing the signal once it’s dropped under the threshold.  With a release of 1ms then 1ms after the signal dropped below the threshold the compression will stop. With a release of 20ms then the compressor will stop compressing the signal 20ms after the signal has dropped below the threshold. Using a release time that’s too fast will result in a pumping sound.
• Output – Sometimes called “make up gain” this is where you can turn the whole signal up to make up for the peaks you’ve compressed down in volume.

You can be forgiven for finding this all a bit difficult to understand, but that’s ok: There’s a  practical method for you to use when compressing sounds. It’s very simple and helps you to train your ear to hear what each parameter on the compressor does and the best bit is it will work on any sound.

A.R.R.T

This is a technique that’s used by a lot of people as it is really easy to get good results. I learnt this from reading the book ‘mixing with your mind’ by Micheal Stavrou – it’s a great book as it avoids getting bogged down in technical stuff and concentrates on cunning mixing techniques. I’ve included the A.R.R.T. trick as it’s too good not to, but Stavrou does go into far greater detail in his book and it includes dozens of other great ideas about mixing.

First set up the compressor with a really high ratio, as short decay as possible and the threshold quite low so the compressor is working a lot. This will initially sound really terrible, but it’s good for hearing what each part of the compressor does to the sound.

• Attack – Set the attack, listening only to the attack – the initial part of the sound. With a really fast attack on a drum sound it will sound as if the stick is really thin and feeble. With a slower attack it will sound thicker and chunkier. Focusing only on the attack try to get it set where you think it sounds best.
• Release – Set the release listening only to the release portion of the sound. A release that’s too short will sound sudden and jarring, a release that’s too long will sound a bit sloppy, play with it until it’s in the optimum position.
• Ratio – Currently the ratio is far too high and is crushing the sound down far too much. Slowly reduce the ratio down until you can’t hear the effect working, then raise it up a little bit so the effect it still happening.  The higher the ratio the more controlled the sound is, but the smaller it will sound – use the ratio to get the balance between size and control right.
• Threshold – Now turn the threshold up. It’s important to have the threshold high enough so that the whole sound isn’t compressed – you want the light or VU meters to move with the sound, with the quiet parts being allowed to pass uncompressed but with the louder parts being reduced.
• Gain reduction  – this is shown as a single LED in the classic compressor (lit = compressed, unlit = uncompressed) but other compressors often show how much the signal is being compressed by.

You should have now compressed the sound subtly but with a pleasing effect. You can now turn the Output amount up a bit – if the sound is compressed by 2dB you should turn it up by the same amount. It’s worth bypassing the compressor to listen to the effect is has on the sound compared to the clean signal – the compressed signal should sound slightly bigger and more consistent and firm. Don’t try to get a massive change to the sound – the best application of most effects are subtle and lots of small changes will build up to make a solid and ear-pleasing mix, over compressed (or badly compressed!) mixes are tiring on the ear.

Using this method will not only get you good results but should help you get your head around how the compressor works, what each parameter does and how they fit together.

There are lots of uses for compression from getting vocals more upfront and level in a mix, accentuating the transient of a kick or a snare drum, squeezing more volume from a drum bus via parallel compression or using a compressor to sidechain a signal for a pumping dance effect.

If you are still having trouble applying compression – use the Smexoscope or GComp to see how the compression is changing the shape of the sound while hearing the effect.