Like other dynamics processors, compressors generally take either a mono or stereo signal and connect to a conventional mixing console via an 'insert' socket. Doing so breaks the signal path in the channel, sending all of the audio to the compressor for processing, and then returning it back to the same point, before it reaches the EQ and fader. In simple terms, a compressor allows you to automatically adjust the gain of your audio, making the louder sections quieter, and the quieter sections (apparently) louder, thereby evening out the overall level of the track. A typical model will have a number of controls that allow you to modify how the signal is affected, in particular when and how quickly the compressor starts to change the dynamics, by how much, and also how quickly it stops processing.
Adjusting the threshold sets the level at which the compressor starts to reduce the gain of the audio being fed to its inputs. In the context of mastering, if you want to reduce only the peaks of a track without affecting the rest of the music, then you would set the threshold to slightly above the average level of the music. In this way, only the peaks are compressed, while the rest is left unaffected.
RMS or Peak
The part of the compression circuitry that 'listens' to the incoming audio is called the 'sidechain', and this controls the compression that is applied in conjunction with the settings on the front panel. Typically, the sidechain measures the audio detected in one of two ways:
- 'RMS' (Root Mean Square) determines the average signal level, in a similar manner to that experienced by the human ear.
- 'Peak' responds to brief signal peaks, ensuring that they are more accurately controlled.
Generally, RMS detection works best when you are using a compressor to raise overall apparent level, being more subtle in its action, while peak detection works best for applications where you are using limiting (see below) to prevent signal overloads. Most compressors only offer one type of detection, but some have a switch to let you choose which type of response you want.
This control determines how much compression is applied once the audio exceeds the threshold. The ratio knob is usually marked with values such as 4:1, 3:1, 2:1 etc, the typical range being from 1:1 (no compression at all) to infinity:1 (denoted as an '8' on its side), which means that the output level is never allowed to rise above the threshold setting - this is also known as limiting. An intermediate ratio such as 3:1 will cause the output signal to rise by only 1dB for every 3dB the input signal exceeds the threshold.
The attack time controls how long the compressor takes to pull the gain down once the input signal has exceeded the threshold level. Attack values are usually shown in milliseconds ranging from around 0.1ms to 200ms. Fast attack settings control the signal almost immediately, whereas slower attack times allow the start of a transient or percussive sound to pass through unchanged before the compressor starts acting. Allowing a deliberate overshoot by setting an attack time of several milliseconds is a much used way of enhancing the percussive characteristics of audio, making the attack stand out and giving it much more 'punch'. For most musical uses, an initial attack setting of between 1 and 20ms is typical. However, when treating less percussive sounds, vocals in particular, a fast attack time generally gives the best results, because it brings the level under control very quickly, producing a more natural sound.
This term applies to what happens when the input signal approaches the threshold. A hard knee compressor does nothing to the signal until it crosses the threshold, at which point compression is applied according to the attack time and other settings. Soft knee models give a gentler transition by starting to apply compression as the signal approaches the threshold, gradually raising the ratio to reach its full set value once the threshold is crossed. The attack of a hard knee compressor is less subtle, but can be useful creatively, and in peak limiting. As with RMS/peak detection, most compressors only offer one type, but some models have a switch to choose between the two.
The release control sets how long it takes for the side chain to start measuring the input signal again once the compressor has kicked into action (or how long it waits before telling the compressor what to do again). If the release time is too fast, the signal level may 'pump' - in other words, you are very aware of the level of the signal going rhythmically up and down (although this is sometimes desirable as an effect). If the release time is too long, a greater peak than the one on which the compressor is currently acting may go unprocessed, or compression may continue for too long when the signal has fallen back below the threshold. A good starting point for release time is between 0.2 and 0.6 seconds.
This disables the attack and release controls and instead automatically adjusts these settings depending on the characteristics of the incoming signal. This is especially useful when processing stereo mixed tracks, where the material and its dynamics may be too complex for fixed manual settings.
Gain Reduction Metering
Compressors usually have either a series of LEDs or a VU meter to indicate how much gain reduction is being applied at any given time. To show this, the metering moves 'backwards' - more LEDs light up or the VU meter needle moves from right to left to show how far the unit has turned your audio down.
Because the compressor reduces the levels of the audio exceeding the threshold, the peak output level will often be lower than the input level. Most compressors provide an 'Output' or 'Make-Up' gain control to allow you to match the overall apparent level of the compressed signal with the original uncompressed input signal. This of course increases the level of any audio below the threshold, which is why quieter signals can seem much louder once the signal has been compressed, and gives rise to the popular misconception that compressors make quiet sections of audio louder as well as making the loud sections quieter. As a result, a vocal (for instance) will stand out more in relation to the rest of a track. This phenomenon is often used to advantage on music played on the radio and television, where the competitive advantage of having your audio stand out by being louder than other's is obvious (see 'The Loudness Wars'). Increased loudness is also an important factor at the recording stage - when recording to analogue tape, compression can help to raise the level of the signal being recorded to an optimum level above the noise floor of background tape hiss, and when recording to a digital medium, compression can ensure that the signal is encoded at the highest possible level where more bits are used, so that better signal definition is achieved.
When processing stereo signals, it is important that both channels are treated exactly equally - the stereo image will 'wander' if one channel receives more compression than the other, or is compressed at a different time. The stereo link switch of a dual channel compressor simply forces both channels to work together, based either on an average of the two input signals, or whichever has the highest level at any one time. When the two channels are switched to stereo, one set of controls usually becomes the master for both channels - though some manufacturers opt for averaging the channels' control settings, or for reacting to whichever channel's controls are set to the highest value.
Mastering with a Compressor
Although you might use compressors most frequently on individual instruments such as bass guitar or vocals to even out the dynamics and make them sit better with other tracks, applying compression to an overall mix is a rather different matter. Here you must consider that all of the individual parts have come together with their own dynamic differences - extreme compression settings might work well to control some tracks, but the knock-on effect on the rest of the mix is unlikely to be good. A general rule is to start by applying a low ratio with medium attack and release settings, and then setting the threshold and makeup gain so that the compressed signal has a similar perceived loudness to the original, allowing you to make more objective judgements when comparing or 'A/Bing' the two. When completing an album, a mix engineer will often leave this process to the mastering suite, as applying overall compression will affect each track's individual level, and therefore the artistic decisions relating to song order, track to track contrast and the overall album concept.
The compressor that we've considered so far is termed a 'full range' or 'broadband' dynamics processor as it operates over the entire audible frequency range - it responds to the incoming signal regardless of where in the frequency spectrum peaks may be occurring. Multiband devices make it possible to control separately the dynamics of individually defined frequency areas by splitting the audio up into independent bands. The end result is that separate frequency areas of a mixed track, the bottom end for example, can be compressed more heavily than if a single, broadband compressor was used.
In a multiband compressor, a crossover splits the audio into different frequency bands, then separate compressors process these individually before they are summed back together to provide a full range signal. Because these bands contain different amounts of energy, they can be compressed using substantially different settings. A clear advantage of this system is that a loud, low-frequency sound will only trigger gain reduction in the low-frequency compression band, leaving any mid-range and high-frequency sounds occurring simultaneously unaffected. In contrast, a conventional full band compressor acting on a loud kick drum for instance, will pull down the level of any simultaneous hi-hat and snare beats.
Relatively inexpensive dedicated mastering products such as the TC Finalizer are now available which not only give incredibly accurate control over a stereo mix's dynamics, but also offer a range of instant presets that can be used as starting points for specific types of material such as 'Classical', 'Commercial' and 'Pop', as well as a built-in 'assistant' to help you find the best settings for the job in hand. The digital technology employed means that a whole range of tools aimed at the mastering process can now be provided by a single device - EQ, limiting, compression (usually both broadband and multiband), gating, and stereo width and balance adjustment. These products have allowed smaller studios and home-based musicians to add 'polish' to their recordings without the expense of using a professional mastering house.
An additional and unique digital feature is the facility known as 'look ahead', where the processor is able to detect the incoming signal a fraction before it arrives at the gain control section of the compressor, by 'buffering' the audio - this allows oncoming peaks to be dealt with faster and more intelligently. One note of caution - although it's very satisfying to hear your music mastered at home, if you think there is any chance of it being released professionally, always keep an 'unmastered' copy of the mix, so that a professional mastering engineer has some dynamics left to work with!