Compression is designed to control dynamic range within a recording and will most likely exist in singular elements of a mix and over the mix itself in a mastering situation. Compression is achieved by making loud passages above a certain threshold quieter, therefore reducing the overall dynamic range. In modern commercial music we have become accustomed to the sound of compression and whether mixing or mastering it now an essential tool.
Let’s start by looking at the most common controls you will find on most compressors:
Threshold: measured in dB (decibels) the threshold allows you to select how much recorded signal you want compressed.
Ratio: The amount of compression applied over the threshold. For example 4:1 means every 4db over the threshold will be produce 1db of output.
Attack (ms sec): Allows you to select how much of the signals initial sound escapes the compression (once over the threshold). A higher attack setting will postpone the compression from kicking in. For example you might choose a slower attack to retain the natural snap and punch of a kick drum or a faster setting to get better control over a vocal track.
Release (ms sec): Allows you to select how quickly the recorded signal turns off the compression after it has fallen below the threshold, affecting the signals tail. When set to fast or too slow this can cause pumping effects. Release times can be the hardest thing to get right in compression so if you’re not sure use a compressor with an auto setting.
Knee: Causes the effects of compression to be applied gradually using an adaptive ratio, this is common especially in optical circuits; a smoothing tool.
Makeup gain/output: Once gain reduction is applied you will need to compensate for the loss in level. As a general rule of thumb you will have to increase level by the same amount of gain reduction you have applied.
How to judge your compression:
You will need to reduce the dynamic range to make an instrument or sound sit in a ‘mix’ of other sounds. Generally the larger the dynamic range (the difference between the quietest and loudest signal) the more compression needed in a modern mix.
When setting your threshold decide how much of the signal needs compression applied to it. For example with drum overheads the cymbals may be louder in the signal then the rest of the kit, so you may just want to bring down the volume of the cymbals and keep the rest of the sounds dynamics intact. So you would set the threshold to affect the loudest part of the signal, in this case the cymbals.
Once you have chosen what to compress you need to decide how hard to compress it, this is where the ratio comes into play. Start with 3.1 and adjust until the dynamic changes sit with everything else, generally the more signal being compressed the lower the ratio, for example in a mastering situation it’s common to set a fairly high threshold but a low ratio. If you find that your compressed instruments lack definition or impact or you wish to keep the attack of your notes intact increase your attack settings to suit.
To avoid pumping effects adjust the release to match the average note length in the program material where the notes lengths vary (e.g. vocals often have complex dynamics) try an auto release or set an average. If the effects of the compression are too obvious and out of context with the track adjust the knee, again think of the knee as a smooth tool.
Finally apply make up gain to adhere to amount of gain reduction achieved. Also remember different compressors have different sounds. Choose the correct compressors for the correct sound.
As you learn more and more about compression and how it sounds you will develop your ear and know what kind of settings work faster. There are some useful techniques with compressors that you can then start to look at. These include side chain compression, ducking, multiband compression, invisible compression and upward expansion.
Compressing a whole mix is often misunderstood and applied badly. Some high profile mix engineers famously use processors such as the SSL Buss compressor aggressively with up to -10dB of gain reduction achieved with a high threshold setting and low ratio. It’s also important to add that they “mix into” the compressor and use it to help build a good gain structure from the off. This also in theory means less compression on individual elements.
As a mastering engineer I’m often faced with over-compressed, lifeless mixes and wouldn’t recommend this approach unless you really know what you are listening for. Also ask yourself “does this music suite harder compression”.
My preferred technique with a mix compressor (I often use an SSL G SERIES for mixing) is to key the mix into the side chain input via a hi pass filter set to around 150-200hz, this way the compression isn’t dictated rhythmically by low frequency elements such as kick drums which tend to take up a lot of headroom in modern mixes.
When mastering I don’t often use a stereo compressor at all. Instead I prefer to use parallel compression via our TC System 6000 MD5 license. This effectively compresses a duplicate of the mix and combines with the original signal, the amount of which is user defined. This technique; often also referred to as invisible compression means you can tighten up a mix and raise its average level without compromising the sound of the original mix. I so often find straight up stereo compression slightly invasive even when using mastering grade compressors. The parallel compression technique results in a more natural and transparent compression.