The aim of a compressor in recording is to reduce the range of dynamics of an audio signal. (Don’t confuse this with file compression, which is used to make a computer file smaller). Any parts of the signal louder than a certain THRESHOLD are reduced. The amount of reduction is relative to the level of the signal and expressed as a RATIO. If the ratio is set at 2:1 the signal above the threshold is reduced to half its original, if the ratio is set to 5:1 the signal above the threshold is reduced to one fifth of its original.
(NB if the ratio is set as high as possible, usually infinity:1, the signal above the threshold is reduced to practically the same as the threshold and is referred to as LIMITING.
In this graph a 45o line represents equal input and output levels, ie no compression (1:1) ratio.
The amount of compression is based on a relationship between the threshold and the ratio. Low thresholds and high ratios give you more gain reduction, i.e. more compression, but neither threshold nor ratio on their own will determine the amount of gain reduction. You can achieve the same gain reduction by a high ratio and a high threshold as you can with a low ratio and a low threshold:
As you can see from the above, the amount of compression with 7:1 ratio and high threshold can be similar to the amount of compression with a 2:1 ratio and lower threshold, and is actually the same at one point. Whether to use high ratios or low threshold or both is a subjective decision and usually based on trial and error and taste.
This is not the end of the story. Usually the point of compressing a signal is to get the whole signal louder. Most compressors have a gain reduction meter. Sometimes there is a switch to change an input gain meter into a gain reduction (GR) meter. You can tell from this how much compression has been applied. Having reduced the loud bits by a certain amount, you can then adjust the gain make up by the same amount. This effectively brings the loud bits back to their original level, but at the same time brings the quiet bits up to a level higher than they were originally, with the end result of reduced dynamic range and higher overall average level. (Many software compressors can do this automatically)
A wide dynamic range is often ideal in a live acoustic performance, however recording and broadcast media generally have limitations. Peaks of audio need to be below the level at which distortion occurs. The grooves of a vinyl record can only cope with so much level until the needle jumps. Digital media only allow levels up to 0dB. Analogue tape causes distortion when the level is too high. With very dynamic music, once the peak level is set at a practical maximum, the low levels may well be close to the level of unwanted noise, e.g. tape hiss, vinyl surface noise, background sounds. Many radio stations compress the output signal as very quiet passages of classical or acoustic music can become totally inaudible on car radios that have to compete with the car’s engine noise and the sounds of other traffic.
- Compressing a final mix will make it louder, and therefore possibly sound better when compared with less compressed tracks, e.g. on radio or on compilation albums.
- It can be useful to compress single tracks of a multitrack recording to avoid quiet words of a vocal, or quiet notes of a solo instrument getting swamped by the backing.
- A combination of compression and automatic gain make-up on instruments with a natural decay will create a sustained sound.
Hard or Soft?
In the graphs above, compression occurs exactly as the signal hits the threshold. Some compressors gradually apply compression as the signal approaches and after it has crossed the threshold. This is often described as soft knee compression. The graph can be seen as the thigh, knee and calf of a leg:
There are often two more controls on a compressor, ATTACK and RELEASE. The speed at which the level decreases as the signal crosses the threshold is determined by the Attack control, and the speed at which the level increases after the signal drops below the threshold is determined by the Release control.
A fast attack means that the full amount of compression kicks in almost immediately. This can sometimes cause distortion or have an adverse effect on the tone of an instrument or the amount of punch, especially bass drums. The best settings are found by trial and error, but a good starting point is a medium to fast such as 100-200 ms. With a slower attack, the compression gradually kicks in.
The next two pictures show what happens when a signal is is passed through the digital compressor in Apple Logic Pro, with a threshold of -10dB and a ratio of 2:1. The signal starts just below the threshold at -11dB, then suddenly goes up to 0dB and back again.(A) shows the signal with no compression. (B) shows the same signal when compressed with attack of 200ms and release of 500ms.
[A] No compression:
[B] With compression:
Tips & Hints
- Start with a ratio of between 2:1 and 7:1, medium-fast attack and medium release then gradually lower the threshold until you get gain reduction of about 5 dB. You can then set the output gain make up to compensate, eg if your gain reduction is 5 dB, set the gain make up the same to bring the peak level back to its original. Then gradually speed up attack until it gets noticeable and back it off slightly.
- Listen carefully. Compression can affect the timbre of an instrument, either because of the inherent sound of the compressor, or because the peaks of an instrument may have a different tone to the troughs, so reduction in level of the peaks relative to the troughs changes the overall tone. This can be especially true when applying fast compression to instruments with broad vibrato.
- Be careful when compressing an entire mix. Very often in pop music there is a bass line which is generally constant in level. If there are sudden very loud peaks such as brass stabs, the compressor will lower the whole track at that point, with the result that although you may get an overall levelling of dynamics and increased loudness, the bass line will dip at this point and lose its flow. This is sometimes called pumping. This problem can be overcome by using a multiband compressor, which splits the signal into different frequency ranges, and compresses them individually. In the above case the brass stab would be compressed, but the lower frequencies can pass through a lower band of the compressor with no or very little compression. It is usually best to mix a track with no, or very gentle, compression, then apply compression at a later stage (final mastering).
- If you are using compression on a vocalist, ask them not to back off the microphone for loud bits. Singers often do this when singing live, but usually a constant distance from the microphone is better in a studio. If they suddenly back off, the vocal sound will suddenly get more ambient which may not be good if you want an “in your face” sound. A good compressor will cope with a large range of dynamics without changing the sound drastically.