Audio Engineering: Compression for Music Explained
Dynamics is how loud or quiet a signal is. Dynamics processors typically are either compressor, limiters, gates, and expanders or some combination of the those 4 things. They typically have several knobs which can be set to change the signal level up or down in different ways.
A threshold, a ratio and possibly an attack and release setting. The threshold is the level at which the signal will be affected.
The gain is the overall level going out or in.
The attack is the rate at which the change kicks in
The release is the amount of time after the change happens, that it can release and return to normal.
Ratios 1.5:1, 4:1, ∞:1
The ratio or is the amount of reduction. Ratios are shown like 2:1 4:1 10:1 and are pronounced as “Two to one, four to one, ten to one respectively. They work like this: for 2:1 every 1dB over the threshold the compressor kicks in 2dB of reduction. For 4:1, it reduces 4dB for every 1 dB over. 10:1 ratio is 10 db of gain reduction for every 1 dB over the threshold the signal goes. The ratio relates to the steepness or strength of the compressor.
A very aggressive steep compressor is a limiter and is meant to stop the signal from going over completely. Any compressor over 10:1 ish is a limiter in my book! a ratio of ∞:1 is called a brick-wall limiter. This means that for every 1 db the signal goes over the threshold it reduces it by 100%. Basically a brick wall limiter just stops the signal completely from going over the threshold. Different types of brick wall limiters will affect the signal differently.
Some modern digital limiters can use complex algorithms and look-ahead predictive threshold type functions to make very transparent reduction. Almost all analog devices have a limiting type function when pushed hard (going into the red) where the electronics just can’t provide any more and the signal tops out at max volume.
In a typical compressor the signal goes in, and if it’s under the threshold, it does nothing. If it goes over the threshold, the signal is reduced by the ratio amount, after the amount of attack time has passed, and then returns to normal after the release time has passed and if the the signal is below the threshold.
The metering usually shows the gain reduction. A rate of 1-6dB gain reduction is a light amount, with higher amounts of reduction like 10-20 dB being usually considered aggressive compression. By looking at the meter you can get a feel for how the compressor is working. You want to get the compressor to gently lay into the sound at first, so be careful that the threshold is not set too low or the ratio too high. Usually, you can kind of determine what is a good rate and nice sounding compression by playing with the knobs a bit to get the meter to work. By tweaking the attack and release settings you can affect the transient response, and decay of drums, or even out the delivery of vocals. Some sounds like guitar and bass sound okay with lots of compression. Most guitar amplifiers are already highly compressed by the tube circuitry.
Push Pull Compressor and Static Threshold Compressors/Limiters
Some compressors operate on a slightly different method. The threshold is constant, and instead the user drives the input hotter to push the signal over the threshold. By controlling the input and output (with the threshold set in the middle somewhere) you can turn one up and the other down to get the desired effect. One of the most famous compressors the 1176 operates like this. There’s an input and and output, and a ratio control, but no threshold level, so you just have to control it with the input and output knobs. Whatever, just play with it and get used to it….. Some mastering plugins work like this too. There’s a static threshold at 0db, and the user can turn the input up to make the limiter work, and it just always puts out the same output of 0 dB.
Gates and expanders
A gate or expander works the same way, except the threshold works in the opposite direction. If the signal is below the threshold, it reduces the signal by however much the ratio or reduction amount is set to. For example, a snare drum gate can be set to cut the signal out, but whenever the drummer hits the drum, and the signal goes over the threshold, the gate opens up and you can hear the snare.
Some compression for music can be frequency dependent. A side chain is where the threshold gets a separate signal from the actual audio signal. A high pass side chain can control a compressor from pumping too much, or a frequency dependent gate can be used to tighten up toms or snare drum.
Compression for music can compress different frequencies at different ratios and with different thresholds. This is called multiband compression. This could be used to compress just the bass frequencies of a vocal, or to tame loud high frequency sibilance and not affect the other frequencies. This is called de-essing
A transient is a sharp spike in frequency at the beginning of the signal. Think of the sharp report of a snare drum where there’s a large spike, then a trailing wave afterwards as the sound resonates. Transient response is the character of whether the spike is sharp, or soft.
Compression can be used to make music sound punchier by having a longer attack time. This lets the spike of the transient push through, then reduces the volume of the rest with some gain reduction. This could also be explained as making the sound sharper sounding.
The opposite effect would be to soften up a sound with a faster attack which reduces the initial transient, and then a faster release to allow the sound to resonate evenly. This could be explained as making the sound rounder.
Compression can be used to manipulate the tone in a similar way. While compressors are generally supposed to have a flat frequency response, sometimes taming back the transient will make the sound rounder and warmer. For example an acoustic guitar with a sharp attack will sound warmer once that attack is leveled off with a compressor. Other analog compressors like 1176 or LA2A may introduce harmonics and distortion from their own internal transformers and vintage circuitry.
Listening to music carefully you can determine when it gets loud and when its gets quieter. For example, if a singer has some words that are stronger than others, a little compression can help bring up the quiet sections of the music and turn down the sections of the music that are too loud. Generally, with almost any DAW you can look at the waveforms and see how much dynamic range is in the performance and apply compression by ear to bring the quiet sections up to an acceptable level, or see how loud the loud sections are.