Sidechain Compression Alternative - Phase Inversion on a Gated Group

Faction

Proto-col
Messages
21,788
Reaction score
963
Location
Bristol
I just got asked to link to my Isra tutorial from a few years ago, on an interesting alternative to compressor sidechaining. Isra search is crap and I couldn't find it on Google either, so I rewrote it here. Hope it's useful to someone :Smile3:


This note is a description of an alternative method of getting the 'sidechain pump' sound that is lower in CPU usage and cleaner-sounding than using most sidechained compressors. With this method it is also possible to get many sounds doing the same pumping action at once, to different degrees, without using any extra plugins. I'll be using Cubase terminology to describe it in the context of getting a pad pumping with the kick (and I'll be assuming you're fairly proficient with Cubase), but the technique translates to any DAW system and to any situation where this kind of gain control is needed. The only plugin you'll need is a gate that can be triggered either by a MIDI note or a sidechain input.

Ok so first some background. What happens when you flip the phase of a sound and mix it with the original un-inverted version? Total cancellation occurs and the result is silence. Taking this one step further, what happens when you flip the phase of a sound, reduce it's level by 6dB and then mix it with the original sound? Partial cancellation occurs, and the result is a quieter version of the original.

We can use this concept to achieve a sidechain compression-type sound as follows: first, create a new stereo group track and label it 'Pump Group'. Place your chosen gate plugin in the first insert slot on the Pump Group. Next, according to which kind of gate you've chosen, either:
  • set up an aux send from the kick's audio channel to the gate's sidechain input, with the send level at 0dB, OR
  • make a MIDI track that shadows the kick pattern and route its output to the MIDI gate's input.
Either way, the result you're aiming for is to trigger the Pump Group gate with the kick pattern. Set the gate's attack, release and hold (if available) to as short as they will go. We'll tweak these later.

Next, flip the phase on the group - in Cubase you can use the phase button at the top of the channel strip, and do something similar in Logic, but in Ableton the simplest way is with a Utility plugin. This will invert the signal coming out of the group.

Now we're ready to start pumping. Set up a loop of kick and pad, but before you press play, open up the editor window for the pad's audio channel and set up an aux send routed the Pump Group; make sure the send is pre-fader. In Cubase you can choose the destination for an aux send in the list box above its send level fader. Set the send level to 0dB. Now press play. If I've managed to explain this clearly enough, what you should hear is the pad disappearing each time the kick sounds.

Why is this happening? Well, going back to the theory I gave earlier, the gated pad is being inverted (phase button on the group), and because it's being sent the same level as the pad's actual output (pad's send level is 0dB) the gated signal is totally cancelling out the ungated pad output each time the gate is triggered. Try setting the pad's send to the Pump Group to -6dB, and you'll hear the pad come back a little when the kick plays; this is because the gate is now outputting the pad at half its original level, so the cancellation is no longer complete. Capice? Good. Now make sure the pad's send level is back at 0dB, and let's tweak the gate.

The gate's time-constants are responsible for shaping the pump effect, and you can set them up in the same way as you would if you were using a compressor. You generally want the attack to be very short so the kick ducks the pad immediately. The longer the release, the longer the pump effect will take to return the pad to its normal level so raise it slowly until you get a pump that fits with the bass.

And that's it! You can send any number of channels to the Pump Group and they will all duck with the kick according to how much of each one is being sent. I find send levels between 0dB and -6dB give the best range for more subtle sidechaining.

A few words of warning: the pumping effect can be destroyed if you're not careful. You'll lose the effect if you send more than 0dB to the Pump Group, or if you add effects or inserts to the Pump Group after the gate (as the gated sound will no longer just be an inverted version of the source), or if you use signal routing that means that the output of the sidechained channels get processed differently to the Pump Group. For this reason, if you're using this technique to sidechain a bass sound, and you subsequently want to process the kick and bass together (ie. a compressor) you should use a separate Pump Group for the bass, and either route both this and the bass channel through another group (I call this one Bass Sum) before grouping Bass Sum and Kick together for compression, or route all three channels/groups directly to the kick/bass subgroup. It sounds complicated but you'll get what I mean when you try it.

I wrote this from scratch after not being able to find my original description on Isratrance. If anything needs clarifying or if you have any interesting uses or extensions to the technique just let me know, and if you know anyone who might benefit from this, please share it with them :Smile3: Let me know how you get on!
 
Ok! I had some difficulties at first. It took me a while to figure out what happened. I had the pad channel in the mixer already at -8db but was sending it out to the pumpgroup prefade at 0 db.
The results were ... hm .. confusing :Smile3:

In such a case you have to set the send level also to -8db. Then it works like a charm.
It seems to be better controllable than before (i had done it before with sidecahin-comp), the rythm is better tweakable now.
 
Ok! I had some difficulties at first. It took me a while to figure out what happened. I had the pad channel in the mixer already at -8db but was sending it out to the pumpgroup prefade at 0 db.
The results were ... hm .. confusing :Smile3:

In such a case you have to set the send level also to -8db. Then it works like a charm.
It seems to be better controllable than before (i had done it before with sidecahin-comp), the rythm is better tweakable now.
Easier fix, and something I should have mentioned in the main post: send to the Pump Group post-fader.
 
My impression is that it sounds cleaner, because it relies on phase inversion rather than explicit gain change by a compressor that might be designed to colour the sound. Also, it's often easier to compare the sound with/without the ducking effect this way, than with a compressor, as in my experience it's quite usual when using a compressor to not get the ducking shape you want unless the release phase doesn't totally complete, which results in the ducked sound becoming quieter as a whole. This can be compensated for by increasing the make-up gain, but then when the kick stops the ducked sound gets suddenly louder! Whereas with this technique you set up a good balance and then make it duck by the amount you want, and the balance doesn't change.
 
i have been using this tool here lately a lot:
http://www.xferrecords.com/products/lfo-tool
it does not give you the possibility to route various triggers to duck something and is thus not completely up to par with colins sophisticated routing scheme, but it allows you to make nice pumping sounds without any coloration or compression. its a very very fast approach and gives you also a versatile gate at hand!
 
My impression is that it sounds cleaner, because it relies on phase inversion rather than explicit gain change by a compressor that might be designed to colour the sound. Also, it's often easier to compare the sound with/without the ducking effect this way, than with a compressor, as in my experience it's quite usual when using a compressor to not get the ducking shape you want unless the release phase doesn't totally complete, which results in the ducked sound becoming quieter as a whole. This can be compensated for by increasing the make-up gain, but then when the kick stops the ducked sound gets suddenly louder! Whereas with this technique you set up a good balance and then make it duck by the amount you want, and the balance doesn't change.

Also you have more unity between the different ducking elements as they all share the same exact envelope on the gate. Still think separate comps ducking has things to offer it's a case by case thing like anything else.
 
Another use of this technique is to implement a gate with look-ahead (which the ableton gate, unlike the compressor, doesn't have).

Make a rack with two channels.
On the first channel, put a compressor followed by a utility inverting both left and right.
On the second channel, put a simple delay, set to 'link' and match the delay to the compressor lookahead.

Voila, a gate but with all the controls you get on compressor.
 
Been lurking through posts of the last 15 years ,good thread to say nice one and to clarify on your post a post that make me register

Sideshow Bob mentions to implement a gate but says place a compressor on the channel , assuming he means a gate abit confused bare with but if that’s the case yes exploring so appreciating all the pointers here
 
Last edited:
Ahh yes I know I am very late to the party , and just to add to matters it’s been predicted the end of civilised life as we know it by 2045....

As you were.....oh yes there’s no one here
‘Climbs back with a joint reading up of the last 15 years here’

Had to edit this turns out I can’t count
 

Attachments

  • DE7E04D1-6F62-41AD-B21C-6F93C39A4904.gif
    DE7E04D1-6F62-41AD-B21C-6F93C39A4904.gif
    23.9 KB · Views: 3
Last edited:

Here is a fairly new plugin that uses a new technique - instead of ducking the signal using amplitude it uses frequencies to achieve the sidechain.
Great stuff :Smile3:
 
I could totally use that for making room in the mix for a lead, dynamically. Cheers bruv.
 
Back
Top