What's new

"Smart" Sidechain Compressor for Stem Mastering?

Jeremy Gillam

Senior Member
Hey everyone,

I'm looking for a compressor/limiter plugin that can be inserted across a series of stems and will "listen" to the other instances of the plugin that are instantiated in the session, sum that, then apply dynamics processing based on the summed signal rather than on the dynamics of each track. In a nutshell, sidechain compression without setting up complex routing, where the pre/post signal is handled inside the the plugin rather than via busses. Does anyone know if such a plugin exists? If not, can someone make it please?

Thanks,

JG
 
Don't know of anything like that which exists. Should ask Fabfilter to implement it. I just run a separate summing bus to use for sidechaining the stems.
 
I use LFO tool now for sidechain type of stuff. Waaayy easier than routing stuff with compressors.
 
Don't know of anything like that which exists. Should ask Fabfilter to implement it. I just run a separate summing bus to use for sidechaining the stems.
Thanks. I fear you're right that it doesn't exist. Fabfilter and iZotope seem like top devs to suggest the idea to. Perhaps the next version of Ozone...
 
But you realize you can setup a dummy master bus ? Feed that signal to each side chain so each stem comp hears the whole mix right?

And FWIW, and IMHO, sidechained stems don’t really work out anyway. You end up with ghosting dynamic effects on the stems (depending on how hard you hit the dynamics).

After much deliberation on this, I’ve come to the conclusion that it’s better ‘FOR ME’ to just run each stem out and put up the master channel again on the edited stems. But I’m talking about stems in an underscore situation.

You can always run out the stems through the master and compile in another session. What I find is that you need to pull down the tracks, but it sounds pretty close to the stereo master, minus some comp and limiting (even though you went through your master comp and limiter)

When you output ducked stems-

It really is strange to hear a pad sound being hit with ducking from an absent big perc drum. Unless that’s a creative choice. In which case you can set that up before the stems anyways.

just some thoughts on the matter
 
But you realize you can setup a dummy master bus ? Feed that signal to each side chain so each stem comp hears the whole mix right?

And FWIW, and IMHO, sidechained stems don’t really work out anyway. You end up with ghosting dynamic effects on the stems (depending on how hard you hit the dynamics).

After much deliberation on this, I’ve come to the conclusion that it’s better ‘FOR ME’ to just run each stem out and put up the master channel again on the edited stems. But I’m talking about stems in an underscore situation.

You can always run out the stems through the master and compile in another session. What I find is that you need to pull down the tracks, but it sounds pretty close to the stereo master, minus some comp and limiting (even though you went through your master comp and limiter)

When you output ducked stems-

It really is strange to hear a pad sound being hit with ducking from an absent big perc drum. Unless that’s a creative choice. In which case you can set that up before the stems anyways.

just some thoughts on the matter
Problem I've run into is with the master bus clipping. Even limiting all stems to -4dBFS you can easily get clipping on loud cues where multiple stems are close to hitting the limiter on each stem. That's usually when I have to sidechain so the master doesn't clip. Most people don't care if the full mix matches the stems exactly so the easy solution is to just put an additional limiter on the full mix so that it doesn't clip. At the end of the day it sounds pretty much the same as the stems but just doesn't clip and you don't have to sidechain.
 
For Dancey Prancy Synth stuff my OptiMaster and OctaSC do wonders.
I don’t like blending tracks and SC’ing more than once if possible.

I never usually need all 8 sections, usually 4, then do a parallel OptiMaster pass which is the only blending I might consider.
Usually don’t even need that if I set my OctaSC up right.

It’s really old DSP tech I used for CrossFit sessions years ago.
It’s great for HipHip and Freestyle too where Kick and HeavySine Bass are prevelant.

If no VST plug in like this is available it’s easy to set up several SC Comps and just tweak away.

Don’t see how this benefits Orchestral Music unless it’s synth and bass heavy.
Hey maybe SineWave Rap Timpanis using portamento, something different.
All instrumental, like Violins spitting rap rhythms, Cellos answer, triangle chicky chicky chaw high hat licks, Tympani Sine Bass..
 
Most people don't care if the full mix matches the stems exactly so the easy solution is to just put an additional limiter on the full mix so that it doesn't clip. At the end of the day it sounds pretty much the same as the stems but just doesn't clip and you don't have to sidechain.

This approach is playing with fire. If you're mixing music for film, tv, trailer, library etc (that's an assumption) the stems are as important as the full mix - in many cases more so. So the full and the stems need to match 100% - you might get away with it for a while, but at some point there'll be a panicked call from a composer at the dub saying that what they have up on the stage doesn't sound like the ref mix they remember approving, and explanations of "well, it's almost the same..." won't cut it.

In answer to OP, I'm not aware of any smart compressor that'll do this, but it's not especially complicated to set up the routing for it if you work using templates, particularly if you're already bussing instruments together in groups before they get to the stem processing stage. It's not really critical that every element of your mix is perfectly represented in the sidechain signal, just that overall shape and dynamics are there, and it can be very useful being able to tailor exactly what makes it to the sidechain and what is omitted.
 
This approach is playing with fire. If you're mixing music for film, tv, trailer, library etc (that's an assumption) the stems are as important as the full mix - in many cases more so. So the full and the stems need to match 100% - you might get away with it for a while, but at some point there'll be a panicked call from a composer at the dub saying that what they have up on the stage doesn't sound like the ref mix they remember approving, and explanations of "well, it's almost the same..." won't cut it.

Fabfilter limiting a single spike by 1dB is something that no one will notice. Even directly trying to A/B it most people probably wouldn't be able to notice it. Everything else in the mix is exactly the same. It's not like I'm limiting the entire mix. Maybe hitting the limiter once or twice in an entire cue. Regardless, I always ask for approval before doing this and no one has ever had an issue with it.
 
Fabfilter limiting a single spike by 1dB is something that no one will notice. Even directly trying to A/B it most people probably wouldn't be able to notice it. Everything else in the mix is exactly the same. It's not like I'm limiting the entire mix. Maybe hitting the limiter once or twice in an entire cue. Regardless, I always ask for approval before doing this and no one has ever had an issue with it.

Fair enough, that I agree with.
 
Thanks for chiming in everyone. In addition to use for stem mastering I was thinking there might be some interesting creative uses for such a plugin, working with drum and perc tracks etc. I don't like getting bogged down thinking about signal flow while producing/writing and I like setting up my sessions so I don't have auxes or empty tracks so if I want to send a multitrack to a collaborator using another DAW I can export it quickly with just a few simple clicks and no confusing routing to worry about. Obviously the traditional ways of doing things work perfectly well though!
 
Hey everyone,

I'm looking for a compressor/limiter plugin that can be inserted across a series of stems and will "listen" to the other instances of the plugin that are instantiated in the session, sum that, then apply dynamics processing based on the summed signal rather than on the dynamics of each track. In a nutshell, sidechain compression without setting up complex routing, where the pre/post signal is handled inside the the plugin rather than via busses. Does anyone know if such a plugin exists? If not, can someone make it please?

Thanks,

JG
Wavesfactory's Trackspacer.
 
The new Sonible Smart Compressor can do spectral side-chaining which definitely sounds different than straight ducking. So it will compress the frequencies given by the side-chain as opposed to just the level.

It can be used in a similar way to Wavesfactory Trackspacer, in that one could have the bass ducking the kick or vise versa to create a "cleaner" low end. Or do Kick drum side chaining on the entire mix, but it mostly will be ducking the low end and if it is ducking the high end, it will be quicker than the low end ducking. This type of compression isn't without artifact, but I have had some great or unique results.

My favorite use has been side-chaining the music mix to the VO (and SFX). It sounds different than straight compression ducking, especially in places where music needs to keep impact, yet have the VO cut through. I feel that I'm able to get more music in the mix without getting in the way of the VO (or SFX) this way. Anyways, it's interesting to experiment with...
 
Top Bottom