I threw together a quick video over the weekend, explaining an audio processing technique I use to split audio into “high” and “low” frequency bands, and apply effects separately to each band. The cool part about this (as opposed to using high and lowpass filters) is that it’s a lossless separation: so it doesn’t change the audio stream in any way (until you apply effects, of course).

I’ve found this technique incredibly helpful when building DJing effects, as it allows me to utterly mutilate/phase/beat-repeat/flange the high frequencies of a track, without affecting the bass and kick in any way. One of the limitations of classic effects is that they apply to the entire audio stream, so they can make the kick messy or otherwise cause the energy of the track to drop. Isolating the bass and leaving it untouched avoids this problem entirely.

The other cool aspect of this technique is that I can sweep DJ effects over a frequency band. I’m quite fond of starting a flanger running on the very highest frequencies of my audio stream, and slowly sweeping the high/low cutoff down, until the flanger is affecting everything except the kick. This produces a sound not unlike a classic filter sweep (Xone92 or DJM-800 style), but without changing the overall frequency footprint of the track, lowering the volume or otherwise causing grief.

Lastly, it lets you do weird glitch effects that de-sync the audio streams. The simplest way of doing this is to apply a beat-repeat to the high channel, so the melody stutters while the bassline keeps going.

The video below explains the theory behind this technique, and shows how to set it up inside Ableton in about 30 seconds. Please don’t shoot me because my signal theory is grossly oversimplified.

You can download the Ableton Rack shown in this video from http://marshall-law.co.nz/audio/boywonder/Splitter.adg.zip.

If you want to see some example DJ effects built on top of this rack, take a look on the Master channel of the latest version of my DJing Template..