CLA Compressors Sidechain Question

Recording, Mixing and Producing
#1

I’m having a hard time understanding the sidechain knob in the CLA Compressors.

In the CLA-2A manual, it says, “HiFreq increases voltage amplifier gain in the peak reduction circuit, for frequencies above 1 kHz, leaving lower frequencies unaffected. When set to Flat, the CLA-2A will provide equal
reduction to all frequencies. The more you move away from the Flat position, the less sensitive the compressor is to lower frequencies, resulting in less compression. This control may also be used as sort of a de-esser.”

In the CLA-3A manual, it says, “The more you move away from the Flat
position, the more sensitive the compressor is to higher frequencies, resulting in heavier compression. This control may also be used as sort of a de-esser.”

Hm,m. Both describe it slightly differently, and neither are what I understand a side-chain to do. Could someone please clarify?

My understanding is that the side-chain should basically apply a high-pass filter, not to the output of the compressor itself, but to the incoming signal that will trigger the compressor. I understand that to mean that the compressor will then not be triggered so easily by the low-end stuff, which can throw things off cuz of all the energy down there. It would be triggered by the higher frequencies, cuz they are elevated and hit the compressor first. So the compressor, for example, will not be so much triggered by the kick or bass for example (so no pumping) but by, say, the highe frequency guitars, vocals, etc., because they’ll hit the compressor first.

So, my understanding is that side-chaining is about the frequency of the signal that triggers the compressor only. Meaning that after it’s triggered it’ll be applied equally to the full frequency spectrum, correct? So, I’m thinking the CLA-2A manual is incorrect when it says, “When set to Flat, the CLA-2A will provide equal reduction to all frequencies.” Well, I mean, that’s kind of correct but misleading: once triggered the compressor will always provide equal reduction to all frequencies, no? Similarly, it’s not about “leaving lower frequencies unaffected.” It’s not about what’d coming out of the compressor, but what’s going into it.

If you want to apply compression only to high frequencies (which is different than a side-chain), you’d use a multi-band compressor (or maybe a de-esser which is pretty much a special application of a MB Compressor).

I’m thinking the CLA-3A manual is closer, but still not quite correct. It’s not that it’s more sensitive to high frequencies, but that that the sidechain increases the elevation of the high frequencies over the low, so that they would trigger the compressor first.

Please tell me I’m correct. Just trying to understand this!

#2

Well, there are different uses for side-chains. What you may be looking at is whether its an internal or external side-chain. In this case, it’s an internal side-chain used to:

I think this is a pretty correct description of internal side-chain. I’d guess it’s using some sort of EQ or frequency shaper to alter the input signal to the compressor. I would say that the higher frequencies aren’t “elevated” though, just isolated, in a sense. They’re not boosted, it’s just a HPF as you said. And it’s not that they hit the compressor “first”, just that it’s only those frequencies that hit the compressor - or the full range of those frequencies hit the compressor, and only a limited range of the lower frequencies hit the compressor input.

It’s not really multi-band, but in a way it kind of is. It’s like making it a 2-band MBC in a way. Except the filter may be much more of a gradual crossover or something. It may limit the compressor sensitivity to lower frequencies, but not completely eliminate it, depending on your settings.

This may not be the most technical explanation, but I’d suggest you try playing around with it. It’s important to get the concept of course, but not always necessary to fully understand the technicalities - unless you just want to.

Well, yes, but keep in mind whether it’s an internal or external side-chain. In either case, really, the side-chain is shaping what the compressor is doing, and what its impact on the output signal will be. But for internal side-chain on a compressor it’s typically an EQ filter type situation. Whereas if you’re “ducking” a track with a compressor with a side-chain from another track (external side-chain, say ducking the bass guitar when the kick drum hits), then it’s more of a volume/level regulator.

Well, as your manual says, the internal side-chain compressor can be nearly if not the same thing as a de-esser, and even a simple 2-band compressor in some ways. A MBC is just expanding on this idea to give you more options and more bands. It’s using a (EQ type) “crossover”, or multiple ones, to split the bands. And then it’s usually giving you some more features over each band, like Threshold, Gain, Range, Attack & Release. You can actually make a “poor man’s MBC” just by splitting a signal into (say 4) bands, overlapping the freq’s with dB/octave slope, and putting a compressor with each EQ band. The compressor (for that band) will only impact the freq’s it is given.

1 Like
#3

Thanks, Stan. Please clarify for me though: whether internal or external side-chain, once the compressor is triggered it’ll be applied equally to all frequencies, yes?

So, maybe it’s a high frequency shelf that triggered it, maybe it’s another track, but once it kicks in it it compresses the entire track?

#4

It depends. I believe the most common use of internal side-chain is an EQ filter, though there are probably other uses. So what’s happening is some frequencies are being bypassed or minimized and will not trigger the compressor in those frequencies. Since low end has a lot energy, typically, and can over-activate the compressor, some of those frequencies may get rolled off and less affected by the compressor. So in that sense, some frequencies are “flowing through” or bypassing the compressor, while others are being compressed or more compressed than those others.

A typical use for external side-chain is ‘ducking’ where you want one track to get out of the way of another. The classic example is ducking the bass guitar when the kick drum hits. So you’d put a compressor on the bass track, external side chain to the kick track, and adjust the compressor so that when the kick drum hits it compresses the bass track just enough to make room. In this case, typically, all frequencies may be compressed (though you could probably combine internal and external side chains to control this very specifically). Mostly all you care about there is reducing the bass track volume briefly, so it’s kind of a volume knob to manage that, and would be easier than trying to draw in automation.

#5

I found this very interesting, and maybe somewhat relatable your OP in terms of what happens with a compressor. The difference is that I think a compressor plugin with an internal EQ sidechain will perhaps act in a “parallel” type mode where the EQ sidechain affects the compressor input but allows the original signal through to some degree. (?) Like a HP filter on the internal sidechain would affect the compressor input, but wouldn’t HP the whole signal flowing through. I’d actually like to know if that’s the case now that I think about it.

Anyway, in this example Warren Huart shows you how using an EQ plugin before a compressor can impact the compressor response. In this case, Warren is actually doing HP with the EQ on the acoustic guitar track, so that low end signal is being eliminated from the track and the compression (compressor input and output). So it’s different than your question, but still a concrete example of how low end can affect a compressor, thereby squashing the higher frequencies more than necessary.

Watch from the time-stamp for like 4 minutes to get the gist of it.

#6

Ugh. I did something wrong when I clicked the link - maybe I double-clicked or something - and it started at the beginning. Took me forever to find your timestamp at around 44:00.

:frowning:

My fault, not yours. Thanks for posting. Yes, this is very relevant to what I’m talking about! He describes beautifully how the low end of the guitar triggers the compressor, which then influences the full audio spectrum on the guitar, even the high end, which wasn’t loud enough anyway (see 45:15, also 47:20: “…the compression of the low end … was effecting the way you hear the high end … it was giving us a decibal-and-a-half of compression … which was squashing the highs which aren’t high enough to trigger the compressor themselves”).

So that’s my understanding of what a basic sidechain does, “kind of.” Like his video, the internal EQ/sidechain takes out the low end, so only the high end triggers the compressor, but when triggered, it is triggering the compressor of the whole track. The part that’s different in the way he’s doing it is that his EQ will also highpass the low end of the guitar, in addition to triggering the compressor which effects all remaining frequencies. In an internal sidechain, “my understanding” is that it keeps the low end from triggering the compressor, but without EQing out the low end. Trigger only.

So, I’ve been trying to read up now on the split-mode in the C1 and Ren compressors. I guess the idea there is that in addition to the sidechain (the red line) which controls what frequencies trigger the compressor, the main signal is then also split into a complementary passive band (blue line) which blocks those frequencies from being processed by the compressor in any manner whatseover. So it’s kind of a “sidechain-plus” kind of effect, which is basically now a simple dynamic EQ (or multiband? There’s another thread somewhere that Jonathan started about the difference between those two; but I don’t wanna go off-topic).

So…

I “think” (please verify!) that a basic EQ internal sidechain (like in the Omni maybe?) would do the same thing as creating a pre-fader send to another track (and routing it so it doesn’t send it’s signal to the master buss), slapping a highpass EQ on it, and then using that as a sidechain to trigger the compressor on the original track. That would be a goofy way of doing it, but in trying to understand what’s happening under the hood, I think that’s what it’s doing basically. So the idea is it filters out the low end so it doesn’t trigger the compressor but does not EQ the sound of the original signal.

As I’m reading over your notes from earlier, if I am following correctly, I believe that’s what you were trying to help me understand when you were talking about how important it is to remember the difference between an internal and external sidechain.

#7

That may be the gist of it. I guess it’s hard to explain all this clearly, and I’m trying to describe my understanding of it without having the DSP engineer grasp of all the details. I think you’re getting there with the ‘split’ function, and multiband is probably a good analogy. Maybe I’d say “pass-through” rather than blocking anything. In other words, an internal side-chain could impact part of the spectrum, and let other parts pass-through relatively unaffected (or at least modified to a degree).