Compression Basics

Compression. It’s likely the most misunderstood and misused tool in all of live sound and recording. From the magazine and forums we all read, it seems to be an indispensable tool for making our mixes “big” and “punchy” and “glued”, and yet we hear constant warnings not to over-compress our mixes. Where is that magical middle ground that will make the elements in our mixes sound big, but not squashed? For that matter, what exactly is compression? Let’s take a look at some of those questions.

First, what is it? What we’re typically talking about when we use the term compression is “Dynamic Range Compression”…which still sounds like something we’d need a doctorate to understand, and so, in the doctoral spirit, let us dissect: first is the term “dynamic range”. This refers to the variance or difference between the loud parts and soft parts in a given sound. On an waveform (electrical or physical) level, it’s the comparative change in amplitude over a given time…ok, that sounded a little doctor-y, let me rephrase that: if you were looking at a waveform on your favorite recording software or on an oscilloscope , the dynamic range would be the difference from where the waveform is small (aka a quiet part) to where the waveform is big (aka a loud part). That difference/change is the dynamic range. Now we move to the second part of the definition “Compression”. Compression simply means to make smaller, or to squeeze something so it fits in a smaller area. So if we put all that together, we can define dynamic range compression as “a process to reduce the difference between the loud and soft sections of a sound”. What does that mean for music and for the ears of our congregation? It means after compression, things will be a more even volume.

Next, how does this process happen? Typically, it happens in one of two ways, the first is called downward compression and it’s what we usually think of when we talk about an audio compressor. Its function is to make louder sounds softer and leave soft sounds alone. The next, slightly less common way, is through upward compression. This has the function of taking sections of audio that are soft and making them louder, while leaving loud sections alone. In short, it will either make loud sounds softer, or soft sounds louder, but the end result is always that there is less difference between soft and loud in a song.

Now, lets look at some of the features on a few typical compressors. I like to start by talking about “Threshold” because all the other functions either change or depend on it, as we’ll see shortly. The threshold of a compressor is the volume level (or amplitude level if you like science and stuff) that a signal must cross in order for the compressor to turn on. In other words, how loud does a signal need to be before my compressor kicks in and says “ok, I’m going to turn this section down”. Once the signal crosses the threshold the next function is for the compressor to actually react and turn down (or compress) the signal, but how does it know how much compress? That’s where the ratio control comes in. The ratio tells the compressor how much to attenuate (or turn down) the signal once the threshold has been reached. Before the threshold, no gain reduction occurs, once the threshold is crossed, the ratio prescribes how much reduction will take place. You’ll see ratios listed as numbers like 2:1 or 4:1 or 20:1 or something like that. What those number indicate is the ratio of input to output. I don’t want to get too math-y cause that’s not fun for anyone, but bear with me for a minute. Lets start with the ratio of 1:1. That ratio means that no compression is happening. If you put 100db of sound in, you’ll get 100db of sound out. No change at all. Now, lets go to 2:1. That means that for every 2 decibels you go over the threshold, you get only 1 decibel in volume increase. You see what happened there? We went 2db over the threshold and we only got 1db out because the compressor turned on and reduced the signal by the amount we specified. 2db in, 1db out…2:1…got it? Ok, lets apply that math to some more numbers. Let’s say our threshold is 100db (meaning the compressor will turn on and effect any signal over 100db) and we put a sound through the compressor set on 2:1 that’s at 104db…what will come out? The answer is 102db, because we went 4db over our 100db threshold and that extra 4db change will be compressed to only a 2db change because of our 2:1 ratio. Ok are anyone’s brains melted yet?? No? Ok, we continue! Lets change up our ratio just a little bit by going to 4:1. Now, we’ve still got our 100db threshold and we’re still introducing a 104db signal. Now what comes out? It’s 101db! This is because our ratio being 4:1 says that 4db must be introduced above the threshold to get 1db of actual volume increase after compression. So, with our 100db threshold, if we put in 108db, we’d get 102db out, if we put in 112db we’d get 103db out, if we put in 116db we’d get 104 db out. You all see the pattern? Let’s do a one more at 8:1 (8db over the threshold in equals only 1db of actual volume increase) so now, with our threshold still set at 100db, 108db in gives us 101db out, 116db in gives us 102db, and 124db gives us 103db out. Ok, enough math for now! Let’s talk about the last two common controls and then we’ll call it a day.

Attack and release. These are really easy concepts, but they’re REALLY important to your sound. In short, the attack time is how quickly the compressor reaches a certain level of gain reduction (usually manufactures set this at 9db) once the threshold has been crossed, and the release time denotes how long it takes the compressor to stop reducing gain (sometimes called the recovery time) once the loud part ends and the sound crosses back below the threshold. So why would we want to change that? For attack time, a common reason would be to allow the transient (the loud beginning part of a sound, like the initial impact of a drum hit) to pass through the compressor at basically its original, but to still get some gain reduction happening on a later part of the sound. Lengthening the attack time is common on any sound source where you want to control the dynamic range of one part of the sound, but not necessarily the beginning of it. Alternatively, you may find some sounds that you want the compression to begin as soon as the threshold is crossed, which would mean you would shorten the attack time so that compression fully engages sooner. The release time could be changed for a number of different reasons. For example, you could smooth out sustained vocal notes by using a longer release time that will reduce the initial loud volume, and then return to normal as the note dies off, giving the impression of a longer more evenly sustained note. Also, if you’re running compression on a master bus or a drum bus, you can even sync the release up with the beat for a really cool vibe/energy as the compressor attacks and releases in time with the music.

If you don’t follow this blog already, I’d suggest you do so you can keep up with articles like this and updates!