Chances are compression is used on every single piece of commercially released music in your collection, but can you hear it and can you do it, too? Studio engineer Mick Williams explains the hows and whys...
Compression has got to be the most common processing effect applied during recording and mixing, but it's also perhaps the most difficult technique for anyone new to the game to get their head around. The fact is, most processing techniques produce an immediately audible effect, but compression, in many cases, is an altogether more subtle creature, which definitely does something to the sound, but in a way that might not always jump out and grab you. So how do you find out what compression does and how and when to use it? Simple... just sit back and read our pro's guide.

Compression basics
At a fundamental level, the function of a compressor is to decrease the dynamic range of a signal, that is to reduce the difference between the quietest and the loudest sounds, the end result being to effectively make the loud sounds quieter and the quiet sounds louder. The point of compressing a sound is to boost its average loudness and to give it a more consistent level with less fluctuations, which translates to more presence, punch or prominence in a mix.

Take a vocal, for example, where the performance may be inconsistent, with some phrases or words sounding quieter than others. This wouldn't sit terribly well in a mix and would get swamped in places by other instruments. If you compress the vocal, the quieter phrases will be brought up, the louder ones tamed, and the whole thing will sit better and be consistently audible throughout the song. And it's not just individual sounds that can benefit from compression; compressing a whole mix can make it sound louder and perhaps more exciting, and can help it to stand out when played on the radio.

Can you hear it
Although all compressors are basically designed to do a similar job, the various models, makes and designs can all sound different from each other. Some units, usually dedicated compressors, are quite subtle and transparent in action, whereas others often have additional functions alongside their compressing abilities. Some compressors can add a little distortion to the sound, while others are perceived to fatten up sounds and add a little warmth

And then there is the question of whether or not compression should be audible. While there are some musicians and engineers who think compressors shouldn't be heard to be working, other people make the most of the pumping, sucking or breathing effects that can be coaxed from a compressor by setting a too short release time. However they are used or abused, compressors are an indis-pensable part of modern music production.

Know your knees
A compressor works by effecting the level of a signal once it rises above a threshold level, set by the user. The amount of effect or 'attenuation' is known as gain reduction, and is directly dependent on the compression ratio; the higher the ratio, the heavier the compression. A 2:1 ratio means that when the input signal exceeds the threshold by, say, 2dB, the resultant output signal will only increase by 1dB.

Compression at really high ratios of 20:1 and above is known as limiting. The highest ratio is infinity:1, known as hard limiting or brick wall limiting, where the signal can get no louder once it has reached the threshold, or its limit, in other words. While some compressors have separate controls to set both ratio and threshold, others simply have a single control to turn up the compression and all the mathematics will be done internally without you noticing.

The point where the input signal hits the threshold and starts to get compressed is known as the 'knee'. A compressor will incorporate either hard or soft knee operation, although some compressors incorporate a choice of both modes. With a hard knee, compression kicks in once the input signal exceeds the threshold. Soft knee operation sees a more gradual onset of compression as the input signal approaches the threshold, resulting in a gentler and less obtrusive effect.

Compressor controls
Envelope control, in the form of attack and release knobs, is a feature of some compressors, although some have automatic envelope control. Attack is the time it takes for compression to commence after the threshold has been reached and release is the time it takes for compression to stop after the signal has fallen below the threshold.

The final parameter on a compressor is the gain control, known as make-up gain. Because a compressor reduces the dynamic range of a signal by turning down the loud sounds, the overall peak signal is reduced. The make-up gain is used to turn up the overall output of the device so the peak signal level of the compressed sound is turned back up.

One other feature worth a mention is the sidechain. This is basically the sensing circuit that detects the incoming input signal. In some compressors it can be accessed via rear panel connectors allowing an external input to affect the compressor's operation; more on this later.

Practical uses
Compression can be used on individual instruments when recording or mixing, and it can be applied overall to a final stereo mix. There is little difference between the way compression is used in top professional studios and our home studios; the choice of compressors may be restricted by cost (check out some compressor options in our Buying guide box on the right), but the techniques and applications apply universally.

Vocals are always a prime candidate for compression. The human voice has a wide dynamic range, 'from a whisper to a scream' as the old saying goes, and compression is used to effectively reduce this range so a consistently high level of vocal can be recorded and so the vocal stays consistent throughout a track.

The usual method with vocals is to compress them while recording and again during the mixing process, if more compression is necessary. The important thing to remember about compression is once a sound is recorded with compression, the compression becomes an integral part of that sound and cannot be removed later... you are stuck with it.

The advice is to err on the side of caution. Don't pile on the compression while recording unless you're absolutely sure that's the sound you want; you can always experiment with more compression at the mix stage. For vocals, try a ratio of 4:1 while recording and, perhaps, up to about 5dB of gain reduction on the loudest bits. The actual amount of gain reduction needed will vary from singer to singer; let your ears be the judge.

Bass guitar is another instrument that benefits from compression. It gives it a more even and, perhaps, fatter sound; try ratios up to about 10:1. Compression can also be used to smooth out the sound of both acoustic and electric guitars, although distorted or overdriven electric guitars played through valve guitar amps are naturally compressed anyway. With cleaner guitars the compressor's attack control can be adjusted to a slower setting to allow the transient or front end of the note through before the compression kicks in for a slightly different sound.

The same technique works well on percussive instruments. If a more defined attack is needed, roll back the attack time slightly to check the effect and keep a relatively short release time so the compressor can recover between beats. Compression also works on more electronic sounds, but be aware the sounds in some synths and samplers may already be compressed.

Mental experiments
While all the above examples involve the whole sound being compressed, there is an alternative technique that can lead to some interesting sounds. The idea is to split a sound so it comes up two desk channels and to compress just one of those channels. Mixing the two sounds - compressed and uncompressed - makes the sound more exciting and dynamic, without compromising the original sound. This technique works well with drums, whether it be a single snare or a submix of several drums. Try using heavy compression and tweaking the attack and release controls so the compressed sound is pumping away rhythmically, then gradually mix the compressed sound in with the uncompressed one to get a larger than life drum sound.

As previously mentioned, external access to the sidechain allows the compressor's action to be influenced by whatever is connected. Equalisers are commonly connected in this manner for a type of frequency-conscious compression. The compressor will respond more vigorously to any frequencies emphasised by the EQ. In this way a compressor can be used as a de-esser to clamp down on sibilant sounds by boosting up the EQ for those frequencies where sibilance occurs (usually between 4kHz and 9kHz).

Another use of the sidechain is for 'ducking', whereby one signal can control the gain of another. This technique can be used to tighten up the sound of a rhythm section or, more specifically, the way a kick drum and a bass sound mesh together. If the bass is being compressed, a split signal from the kick drum can be fed into the sidechain input. This means every time the kick drum plays, the bass is compressed a bit more and reduced in volume for the duration of the kick drum hit, resulting in a tighter rhythm and smoother bass end. Setting the release control is crucial in this instance to bring the sound of the bass back up at exactly the right time.

Finally, compressing a whole stereo mix can give a professional sheen, and can make it sound louder, tighter and punchier. Use a stereo compressor or a dual mono one in Link mode, so there are no undue shifts in the stereo image, and use quite a low ratio. Alternatively, multi-band compressors allow you to apply differing amounts of compression to different frequency bands of the track, so you can, say, tighten up the bottom end in relation to the top.

The one thing to bear in mind is it can be easy to overdo things. If your track is being mastered for release it may be wise to leave the job to an experienced studio engineer for optimum results

Step-by-step: compress a live vocal
1.Plug the compressor across the insert point of the desk channel through which the vocal is being recorded, or, if you are using a separate preamp, connect the compressor between the preamp output and the recorders input. If you are using a compressor which allows you to set a ratio like this dbx 160x, use a lower ratio like 3:1 or 4:1.

2.Gradually turn the Threshold knob until the compressor is working on the signal. The amount of compression or gain reduction is usually indicated on a meter (in this case it's a red LED ladder) and you can use this as a guide to set the gain reduction to be between 3dB and 5dB on the loudest notes.

3.As compression effectively decreases the signal level, compressors have an Output or Make-up Gain knob to compensate for the level drop. Turn the knob until you have sufficient signal to feed the next piece of equipment in the chain.

Step-by-step: compress a drum loop
1.Connect the compressor across the insert point of the desk channel that is handling the drum loop, or plug it in line with the output of your sampler. For this example we'll be using a Drawmer LX20 compressor.

2.The LX20 is a dual mono unit where the two channels can be linked. In this case the drum loop is mono so take the machine out of Link mode and just the left channel is used. The expander is a rudimentary gating device that we don't need so should be turned off.

3.Some compressors have separate ratio and threshold controls, whereas others, like this unit, combine them into a single compression knob. Turn the knob up until the signal is heavily compressed, with a large amount of gain reduction indicated on the meters.

4.The pumping or breathing effect is dependent on the attack and release controls. Set the attack to the fastest position and release to the slowest. Now move the release control through its range to its fastest position and note how the sound changes. You'll hear rhythmic compression effects at fast release settings

5.Now, with the release control on a fast setting, move the attack control for slower attack times and note how the sound changes. The relationship between the two controls will give many sound variations, and it may be possible to get the compression turning on and off in time with the track.

6.Once you have got a sound that you like, use the Output or Make-Up gain knob to set the output level to match the next piece of equipment.

Mick Williams 01/01