You are here


The A-Zed of Audio

Mouse Trap gameIf you've ever played the board game Mouse Trap, you can easily understand one of the most interesting digital audio techniques: sidechaining. You might remember that Mouse Trap involves a scheme for capturing mice that's far more complex than your standard cheese/spring/snap design available at the dollar store; Mouse Trap was a child's introduction to the incredible world of Rube Goldberg machines. 

The fact that there was even a board game connected to this incredible scheme was a surprise to me as I went back to do a bit of research, because all I remember was turning a crank which makes a red hexagon hit a boot that kicks a bucket that sends a marble down a ramp...until finally a small cage descends on whatever poor mouse happens to be waiting below it. I don't ever remember rolling a die or moving my pieces around the board – I'm pretty sure we'd just set up the machine and run it again and again until we got bored.

In a similar way that fascination is at play when you sidechain a compressor on your bass track to your kick drum inside a DAW. The basic scheme is similar to Mouse Trap because the end result (either compression or the mouse getting trapped) is triggered by something that's distantly related. The biggest difference is that instead of seeing the result, you hear it.

There are a few common uses of sidechaining and they involve manipulating the volume of one track based on the activity of another track. Remember, a compressor's job is to manage the dynamics of the track on which it's placed; if we put it on a bass track and set the threshold at -25 dB, every time the bass level exceeds -25 dB, the compressor springs into action, reducing the level of whatever is over that threshold. Normally a compressor is 'looking' at the track on which it’s placed, so it monitors that track and affects that track. But when a compressor is placed on a bass and sidechained to a kick drum it 'looks'at the kick drum's activity, and then impacts the bass.

Loudspeaker cross-sectionThis is a common arrangement for a reason that gets at the heart of what distinguishes recording from live performance. A loudspeaker needs to produce every frequency you're hearing, and in this way it functions like a small economy – available is a finite amount of frequencies, and and sometimes demand exceeds supply. A bass synth and a kick drum occupy a similar frequency range and if both play together the result can be muddy and undefined. To deal with this a producer might set up the arrangement I’ve described above, where the kick drum will trigger the compressor to reduce the volume on the bass. It frees up some low frequency resources in the loudspeaker which can then be devoted to crushing your chest with an enormous kick drum. As soon as the kick is done (which is very quickly) the compressor stops working on the bass, and the bass returns to its full level.

Depending on the settings used, this quick drop in bass volume may be directly imperceptible, the only noticeable effect being that the kick drum stands out more clearly than it otherwise would have. A bit less than a decade ago, producers including Eric Prydz and Benny Benassi began to experiment in such a way that this usually discrete effect became extremely noticeable, so much that the "pumping" sound this sidechain arrangement creates is one of the signature sounds of dance music.

Here's a technique that emerged from producers who clearly recognized one of recorded music' unique features, in this case observing how loudspeakers work. Prydz and Benassi didn't pretend that they were simulating a live band; instead they embraced a seeming technological limitation and turned it into a staple in the sound of the music they make.

Rubber duckyThere are also uses of sidechaining that extend beyond dance music. When you hear commercials or documentaries that have a voice actor speaking on top of background music, a sidechain arrangement is being used that’s referred to as "ducking". With ducking there’s a compressor set up on the background music track, and it's sidechained to the actor's voice. When the actor isn't speaking the music plays at its full volume, yet as soon as the actor's voice comes in, it crosses the [very low] threshold of the compressor, and the compressor then starts working: it "ducks" the background music out of the way so the actor's voice is clearly audible. Then when the actor stops speaking, the music comes back up in level.
Both in the case of pumping and ducking, there are two parameters that set apart good uses from lousy ones: attack and release. These control how suddenly the effect comes on once the threshold is crossed, and how long it takes for the effect to stop working after the signal drops below the threshold again, respectively.

You can understand these in terms of a person's temperament. Having a short attack time is like a person who immediately freaks out the moment they're cut off by a bad driver or someone says something rude. On the other hand, having a long attack time is like a person who can put his temper on hold to make sure that what his lizard brain thinks just happened, actually did happen. The release time is the opposite, and dictates how long a person will hold onto something after it's finished. There are those who let things go immediately after a conflict is resolved, and those who – at the age of 35 – are still upset about something that happened in grade three. The former has a short release time, and the latter has a long release time.

Voice over microphoneIn the case of a ducking compressor, it makes sense to set a medium attack time so the background music doesn’t decrease in level so suddenly that it’s obvious. And a medium or long release is a good idea so the volume doesn’t jump back up to its full level in between the voice actor’s sentences.

The first, more traditional, example with the kick drum and bass would use extremely short attack and release times on the compressor to make the effect as functional as possible, yet pumping becomes an effect in its own right because it's audible, obvious, and downright cool. The traditional sidechain effect is being intentionally misused, and dance producers set a medium or even long release time to draw attention to something that’s usually supposed to be inaudible.

The thing that makes sidechaining and Mouse Trap so great is that we get the sense of cooperation amongst a number of seemingly unrelated parts, all in service of a final goal. I think deep down we love to see – and in the case of sidechaining, hear – how things can affect each other which on their own might seem disconnected. They build something bigger than themselves and do something no part on its own could have accomplished. It's teamwork, it's a well-oiled machine, it's a smooth society, and when it's really working, it seems like magic.

Jordan Mandel is a Digital Media Lab Instructor at the UW Stratford Campus, and writes for this blog regularly. His hobbies include sandblasting, Rube Goldberg machines, and finding rhymes for the world purple. More of his work can be found at, which is home to the award-winning satire rag, The Outa Times.

Add new comment

Blog topics

  1. 2017 (8)
    1. March (1)
    2. February (3)
    3. January (4)
  2. 2016 (24)
    1. December (1)
    2. November (4)
    3. October (2)
    4. September (4)
    5. June (1)
    6. March (4)
    7. February (3)
    8. January (5)
  3. 2015 (79)
    1. December (1)
    2. November (8)
    3. October (6)
    4. September (4)
    5. August (1)
    6. July (2)
    7. June (10)
    8. May (10)
    9. April (10)
    10. March (15)
    11. February (10)
    12. January (2)
  4. 2014 (83)
    1. December (6)
    2. November (7)
    3. October (10)
    4. September (10)
    5. August (8)
    6. July (16)
    7. June (11)
    8. May (4)
    9. April (2)
    10. March (7)
    11. February (2)
  5. 2013 (25)
    1. December (3)
    2. November (3)
    3. October (3)
    4. June (5)
    5. May (6)
    6. April (1)
    7. March (2)
    8. January (2)
  6. 2012 (26)
    1. December (2)
    2. November (1)
    3. April (5)
    4. March (4)
    5. February (7)
    6. January (7)
  7. 2011 (14)
  8. (0)