Compress Yourself: Parallel Processing in Ableton Live

Parallel processing is a technique for mixing an effected version of a track with a dry version. This is particularly helpful when you want to process a sound but the plug-in doesn’t have a dry/wet knob or you want to be able to control the effected signal with additional effects, such as EQ, reverb etc. etc.

Compressed to Impress

One of the most common examples is ‘New York Compression’, which is achieved by sending a copy of your track to a compressor on a send. The net result is a (usually) heavily compressed copy that can be blended in with the original but where the dynamics and transients are preserved.

I’m going to demonstrate this with a break beat. This is The Soul Lifters’ Hot, Funky and Sweaty. It kicks in at 2.21. I’m using Ableton Live 9 but this can be done on all DAWs. Here’s the dry version of the beat:

I’m going to create a send (alt+cmd+t on a Mac, ctrl+alt+t on a PC) and add a compressor to it. I’m going to use the Audio Damage Rough Rider, which can be downloaded for free on their website. Be sure the Show/Hide Mixer section is enabled and send your track to the return.

Now solo the return so we can hear it in isolation. I’m going to crank the settings to really show this effect off. I’m using an extremely high ratio with fast attack and release times, the sensitivity is at about 2 o’clock and the makeup gain is adjusted to bring the level up.

Now drop the volume of your return to zero and unsolo it. Blend the return in with your dry signal and hey, presto: you’ve just parallel-processed your break.

In the example below are 2 bars of the dry break and 2 bars with the return blended in. It really toughens it up, but without completely squashing it and killing any dynamic range/transient information.

Next Steps

Next we’re going to take the effect a step further. Using the same break, I’ve created two sends (deleting the one I used in the previous example) and set them to pre-fader. Next reduce the volume of the break to zero.

Soloing the first return, I’ve added SPL Transient Designer to it though any transient designer will do. I’ve dropped the sustain, increased the attack and dropped the volume.

Moving on to our next return, I’ve added Dada Life’s Sausage Fattener. This is a great, cheap plug-in for obliterating sounds. It’s really a mastering tool, but I’ve found it can be used anywhere. I’ve turned both knobs to about 12 o’clock.

What we’re left with is one return with a concentrated transient and another with no transients whatsoever, but a healthy RMS signal. Balance the volumes so both are peaking at around -6dB. Next add Live’s Compressor to our Sausage Fattener return, and sidechain it to the SPL transient designer return.

Use attack values of zero and a quick release with ratio and threshold values to taste. Now you have independent control over the transient and body of the sounds and they’re routed to a new track that can be processed and resampled.

Conclusion

Another advantage of using this technique, is that it allows us to further process the compressed signal removing bass, de-easing, using mutli-band effects and so on. Balancing different frequency bands with an dry signal can introduce phasing issues, so be careful to check things in mono and/or with a phase correlation meter or similar.

Parallel processing doesn’t have to just stop with compression or gain regulating effects; it can work well with distortions, EQs, reverbs; anything where you want to have independent control over the balance of dry and wet signal.

***EDIT***

A number of people have mentioned that latency can be an issue with adding effects to a return track. Of course be careful of any delays introduced and A/B the sound with and without to be sure you’re not better off without the effects.

Also, it’s been pointed out that it’s possible to achieve this effect within Ableton’s Audio Rack Effect. This is certainly possible, but I want this demonstration to be non-DAW specific. Different DAWs handle send/returns differently, some allow for plug-in delay compensation and other don’t. So just be careful about how much processing you do. If in doubt, bounce it as a separate audio file and align it manually.