October 17th, 2010 – Imagine you’re a mix engineer and you get a song to mix recorded by another engineer. Unfortunately, it has a lot of noise: hissing, pops, background noise. When there’s something playing on that track, for example the vocalist is singing, you can barely hear the noise, but if the singer isn’t singing constantly, you can hear it.
Perhaps it will be hidden by other instruments that play during those gaps, but when you’re trying to do the best job possible, you want to take out as much of that noise as possible, so it doesn’t compound upon itself (unless of course, you’re going for a certain stylistic choice).
There are three main ways to deal with this noise: you can cut out the unnecessary spaces in the audio to minimize noise, you can automate the volume to either mute or turn down the audio, or you can use a noise gate.
A noise gate is a device that acts very similar to a compressor, with one huge difference; instead of turning the volume up in quiet parts, a noise gate turns the volume down. In the picture above (I should probably point out that I just opened and screenshotted the noise gate, it’s not set up for the project I’m working on), you can see that the main settings on the device are shared with compressors: threshold, reduction, attack, hold, and release.
Like on a compressor, the threshold is user settable point that triggers the device. On a compressor, when the source goes above the threshold, compression is triggered. On a noise gate, when the source goes below the threshold, the noise gate will clamp down on the audio. Just how much it will lower the volume depends on the amount of reduction that is set. Attack, hold, and release, as with compression, need to be set on a source by source basis.
I should also mention that noise gates aren’t always used as a tool to get rid of noise; sometimes they’re used as an effect. If you listen to a lot of electronic music, you will no doubt recognize stuttering vocals and synths that pulse in rhythm to the song. While sometimes the effect is managed manually through edits and automation, it has been traditionally created through the use of gates. Usually, a noise gate will be put on the desired track, but it will be triggered by something on a different track using the noise gate’s sidechain option. For more information on side-chaining check out this previous post.