A Guide to Compressors in PA Systems
A compressor can be extremely useful - even indispensable - in live sound applications. Some performers don't like them, and if the choice is between over-compressed or no compression at all, no compression at all would get most people's vote in most situations. Also, compression can have other undesirable effects (greater sensitivity to feedback for one, and noise for another), so compressors are certainly not one-way tickets to a better sound.
Some people have an almost pathological aversion to compressors, but the fact is that even the best live sound systems in the world struggle to achieve a dynamic range of much more than around 60dB (from audience and other background noise to painfully loud). If your source material ranges from the rustle of a paper bag to the roar of a jet engine, your system won't cope with it unless you apply some compression.
If, on the other hand, compression is readily noticeable then there is probably too much (the same also applies to reverb). However, if it is used effectively and in moderation, a compressor can turn a ragged noise into a manageable mix, and make an inexperienced vocalist more intelligible.
Physically a compressor is usually a 19″ 1U box with knobs or buttons (and more often than not a couple of LED meters) on the front. However, they come in different guises: guitarists may use pedal versions, and compression (accessed from the console's front panel controls) is a common channel facility on modern digital mixers.
A compressor reduces the dynamic range (the difference between the quietest and loudest parts) of a signal.
It uses a voltage controlled amplifier (VCA) to vary the amount of gain the signal receives according to the size of the signal itself: the signal voltage controls the gain of the VCA. As the signal gets bigger, the gain is reduced. As the signal gets smaller, more gain is applied.
Usually, fixed gain is applied to the signal as long as the signal is below a user-defined threshold. Above the threshold, the gain is reduced (usually by a ratio, again defined by the user).
It can have controls for:
If the signal isn't big enough when it reaches the compressor, it may not be possible to activate compression even if the threshold is set at its lowest level. The Input Sensitivity/Gain control is used to match the input signal to the compressor's normal operating level.
Instead of this (or in some cases as well as this), some models have a switch - often on the rear panel - allowing the user to select a nominal operating level of -10dBV or +4dBu.
This sets the level above which gain will be reduced. A low threshold means that most of the signal will be compressed, while a high threshold will only apply compression to signal peaks, and allow most of the signal to pass unchanged.
This allows the user to control the proportion (usually in decibels) by which signal exceeding the threshold will be reduced.
How fast gain reduction takes effect. Many instruments (particularly percussion, but also guitars, including bass guitar) are loudest at the beginning of each individual sound, and the beginning of each sound also includes much of its high frequency content (e.g. the strike noise of the drumstick, or the pick sound of the guitar). If gain reduction is immediate (or too fast), an audible dulling of the sound can result. If it is too slow, sounds that are too loud may escape compression.
If the gain applied to the signal returns to normal too quickly (the instant the signal falls below the threshold), the effect can be an audible ‘pumping’ or ‘breathing’ as the gain changes. If it returns to normal too slowly, quieter sounds following the sound that initiated compression can receive insufficient gain. Again, the result is audible. If attack and release times are both set too fast, the rapid on-and-off effect can make the signal sound distorted.
Hard knee compression reduces gain abruptly on any signal exceeding the threshold (and not at all on signals below it). Soft knee compression introduces gain reduction gradually on signals approaching the threshold, and progressively (up to the ratio set by the user) as they exceed it. Most compressors with an ‘auto’ function apply soft knee compression when ‘auto’ is selected. ‘Auto’ will usually also override attack and release settings.
When you apply compression, the loudest sounds get quieter. In order to achieve full output from a channel or system in which compression is applied, the gain may need to be increased. In effect this makes the loudest sounds as loud as they were before, while making the quietest sounds louder.
Most multi-channel compressors allow coupling of each pair of channels for stereo use. If left and right channels are compressed independently, stereo positioning will appear to wander whenever one channel is being compressed more than the other.
Most compressors have indicators (normally LED bar-graphs or something similar) to show - in decibels - how big the signal is, and how much the loudest part of the signal has been reduced by the effect of the compressor.
Many also include a basic noise gate.
If all else fails, read the manual!
The most common place to use a compressor is on channel, group or mix inserts (this places the compressor directly in the signal path after the signal has been raised to line-level by the channel preamplifiers).
In an under-powered system, compression across the whole mix (either in-line between the mixer outputs and the amplifier, or on mix inserts) allows you to raise the average level without increasing peak levels, making the system seem louder. Using a fairly high threshold and compressing the whole mix can also help vocals to stand out, as the vocals - which should be louder in the mix than anything else - will trigger compression, reducing the level of backing instruments during vocal sections.
Good candidates for individual compression are:
If you have enough compressor channels, they can also be useful on guitars, and (if you still have spare channels) almost everything or anything else.
If in doubt (and if the option is available), go for soft-knee rather than hard-knee compression. Start with Input Sensitivity at 0dB (unity gain), the Threshold at 0dB or higher (see Gain Structure on the Mixers page for channel settings), and the Ratio at 1:1 (no compression).
Increase the Ratio to apply modest compression (between 2:1 & 3:1). Gradually reduce the Threshold until the indicators show a few dB (between 1dB & 6dB) of compression. The only audible effect should be a slight reduction in overall level. Increase the compressor's Gain (or Output) to restore the level at the mixer (use the mixer's meters to compare levels).
If the signal still exceeds 0dB on peaks (according to the mixer's meters), increase the Ratio and/or lower the Threshold.
Gentle compression (e.g. 2:1 or lower) throughout a signal's range will not be greatly noticed. Hard compression (i.e. greater than 6:1) acting only on heavy peaks will not be greatly noticed. However, hard compression will be audible if the threshold is too low. Even quite modest (e.g. 3:1) compression can be intrusive if hard knee compression with inappropriate attack and release times is used.
Compression is a tool. Where the performers have good control over dynamics, you won't much miss a compressor. Where the singer's microphone technique is poor, or an acoustic guitarist's technique varies between quiet finger-picking and wild strumming, compression will help. In the latter case, however, it may be difficult to select settings that avoid over-compression of all the strummed material. Generally, however, a system with lots of compressors is easier to drive than one with none.
Any compressor that includes all the controls described above will probably be OK. Extra useful facilities are:
Plenty of channels are a bonus, so several compressors can be useful.
Every compressor has its own ‘sound’, so if possible try before you hire or buy.