Kim Lajoie's blog

What is sidechain compression?

by Kim Lajoie on June 30, 2014

Sidechain compression is a special variant of regular channel compression. A normal compressor adjusts the output level of the audio based on the input level. Sidechain compression, however, adjusts the output level of the audio based on the level of a different audio channel.

This means that the volume of a channel reacts to the volume of another channel. The audio that the compressor is reacting to is often referred to as the ‘key’ or the ‘sidechain’.

There are two common uses for this:

  • Kick drum ducking. This technique uses the kick drum for the sidechain signal. It’s set up so that the compressed channel (usually the bass) is briefly turned down when the kick drum is sounding. It was originally used to make the kick drum bigger – by reducing the level of some other tracks (usually the bassline), the kick punches through the mix with relatively more presence and power. It’s most commonly used to compress the bassline (either bass synth or bass guitar), but is also used to compress synth pads, vocals or even other drum and percussion tracks. It’s become a recognisable and characteristic sound in a lot of electronic dance music.
  • Vocal ducking. This technique uses the main vocal channel as the sidechain audio. It’s set up so that the compressed channel is turned down when the main vocal is sounding. It was originally used in radio broadcast so that the music would be automatically turned down when the announcer or DJ started speaking. It can useful when mixing a song that contains a prominent foreground part (such as a guitar or vocal harmonies) that should be pushed to the background when the lead vocals come in. Ideally, however, this situation is best avoided by careful composition and arrangement.

In day-to-day mixing, there’s usually not much need to use sidechain compression unless you’re aiming to create a certain effect such as a pumping bassline for a dance song.


2 thoughts on “What is sidechain compression?

  1. Kapitano says:

    A couple of thoughts about this:

    1) Having the kick drum trigger compression without makeup gain – in other words, effectively lowering the overall amplitude of the other sounds – is very familiar from “pumping” dance music.

    But doing the opposite – using the kick to trigger a temporary raised amplitude of another instrument, usually the bassline – is something I’ve also heard. Though not for years. It seems an out-of-fashion bit of dynamics trickery.

    Maybe a few posts on mixing procedures that aren’t used much now, but can be used for an authentic vintage flavour?

    2) What about compressing (or just lowering) certain frequency bands? I’m thinking of doing the following:

    * Isolate the dominant frequency bands of the vocal – say 150-300 and 3000-6000, for tone and sibilence respectively.
    * Set up an EQ on a bus of everything except the vocal, such that the greater the vocal amplitude, the more the two above frequency ranges of the instrument bus are reduced.

    In other words, the vocals carve out a space for themselves, but they control the size of that space from moment to moment, as they need it.

    Low bass, mid and high “air” ranges are unaffected – apart from any phase issues created by the filtering.

    Is this insane, or have I stumbled on one of the black arts of mixing?

  2. Kim Lajoie says:

    Hi Kapitano,

    Vintage ain’t vintage. It depends on what era, what genre, what subculture you’re interested in. In general, however, you’ll find that mixing has become *more* complex over time, not less (e.g. your second point). To go back in time, you need to look at simplifying your capabilities and looking at how to get the best results from that. It might mean limited tracks. It might mean limited microphones. It might mean limited outboard/processing. I think vintage flavour is more about the process and the tradeoffs than about specific techniques.

    Regarding your second point, you’re probably overthinking it. I’ve never had a mix that wasn’t working, where I thought the best solution is side-chained dynamic EQ. The only time I’ve ever used dynamic EQ was in mastering when I had a few tracks that were particularly sibilant (and had loud crashes) because they were mixed dark. And that’s a very specific technical situation that required that kind of control over the spectral and dynamic behaviour of the sound. And even then it wasn’t side-chained.


Leave a Reply

Your email address will not be published. Required fields are marked *