A very common question I see is “should I use compression in audio recording?” The best answer to this question is, as is so often the case, “Do you acim compression in your recordings?” Let’s start with a workable definition of compression in the audio recording sense. When you lower the volume of only SOME of your audio, usually the bits that are clearly louder than most of the rest of the audio you’re working with, you are “compressing” that audio. This is usually done in order to allow you to raise the average loudness of the entire audio file.
Now, this can be done manually, by which I mean you could open your audio in an editor, seek out all the areas where the wave forms (I like to use the term “blobs” instead) are loudest, then turn those bits down. But that can get REALLY tedious and time consuming. So to automate this process, a machine (nowadays done with software) called a “compressor” was invented. This allowed folks who really knew what they were doing to more quickly manipulate volume and loudness dynamics. The dark side of the situation, though, was that it allowed folks who were less expert to mess up their audio, and do it much faster and more efficiently than ever.
There are all kinds of settings on a compressor that are better discussed in other articles. For the moment I’d like to focus on an explanation of basic compression, since I strongly believe that those who use compressors to mess up their audio (usually without actually WANTING to mess up their audio) do it because they don’t have a really good grasp on what compression really is. This should help.
Let’s say you have a voice narration file open in your audio editor. You’ve got your familiar blobs going horizontally across the screen (in its “swim lane”). Now imagine that at two points in the audio there are very loud and quick peaks, perhaps caused by a cough or an overly-excited consonant. In this example, most of the blobs top out about half way between the center line (total silence, remember?) and the top of the swim lane. However, the two loud peaks I mentioned go most of the way up to the top. If I tried to turn up the volume of this audio, all of the audio, including the 2 peaks, would “get bigger.” But there is a problem here. Do you see it? If any of the audio wave forms (blobs) get pushed beyond the top border line, you get nasty, awful digital distortion. Suffice it to say that the border line represents the upper boundary you must not cross.
Knowing that, it should be easier to see why you can’t turn the loudness up very far before the peaks hit that boundary. And when that happens, none of the rest of the audio can get any louder either. Oh no, whatever shall we do?
How about if we turn DOWN the level of JUST those two peaks? Well, the first thing that would happen is that the average level of the audio blobs would be much more even. THAT is what compression is for. Now that we don’t have just a couple of pesky peaks preventing us from turning our audio up without it distorting, we can raise the average level much higher, resulting in louder audio across the entire file. If we now tell our editor to raise the level of the loudest peak to maximum loudness (raising the rest of the audio by the same amount), you will notice that blobs are much wider/bigger across the board. Turning down only the few offending bits of audio was the act of compressing the audio file. Once that was done, and we were left with more even levels, we were able to turn the entire thing up without distortion, which we couldn’t do before.
