What is Sidechain Compression?

guinsu

Member
Ok, this probably sounds dumb, but I couldn't turn up much on google and I see this phrase get tossed around a bit. What is it and where would a ME use it?
 
guinsu said:
Ok, this probably sounds dumb, but I couldn't turn up much on google and I see this phrase get tossed around a bit. What is it and where would a ME use it?

Basically a side-chain compressor uses another signal other than the main input to control the amount of compression. Often times it is used in mastering to compress the audio at a particular frequency range by EQing the original signal in a separate path and feeding it into the side chain to control level.

Very useful for things like de-essing, removing plosives and so forth. Very similar to multi-band compression, but only works on one band at a time.
 
Also for things like ducking. You could have a music bed going through the compressor and a voice over send going to the side chain. When the voice comes up, the music bed will "duck".
 
masteringhouse said:
Basically a side-chain compressor uses another signal other than the main input to control the amount of compression. Often times it is used in mastering to compress the audio at a particular frequency range by EQing the original signal in a separate path and feeding it into the side chain to control level.

Very useful for things like de-essing, removing plosives and so forth. Very similar to multi-band compression, but only works on one band at a time.

I've always been a little sketchy on how de-essing works. So you take a vocal track, EQ it to EMPHASIZE the "ess" and then feed that into the compressor to control the amount of compression on the un-EQ'ed track? Is that right?
 
reshp1 said:
I've always been a little sketchy on how de-essing works. So you take a vocal track, EQ it to EMPHASIZE the "ess" and then feed that into the compressor to control the amount of compression on the un-EQ'ed track? Is that right?

Yes, you feed the EQ'ed track into the sidechain, and the compressor channel is inserted on the main track as usual. Compressors are stupid, they react to level, not frequency. Boosting the problem freqs makes them much higher in level than the rest of the signal, and so makes the compressor react only when the selected freqs go above the threshold.
 
easychair said:
Yes, you feed the EQ'ed track into the sidechain, and the compressor channel is inserted on the main track as usual. Compressors are stupid, they react to level, not frequency. Boosting the problem freqs makes them much higher in level than the rest of the signal, and so makes the compressor react only when the selected freqs go above the threshold.

Okay thanks, that's what I figured but was never quite sure.
 
easychair said:
Compressors are stupid, they react to level, not frequency. Boosting the problem freqs makes them much higher in level than the rest of the signal, and so makes the compressor react only when the selected freqs go above the threshold.


Are you kidding??? Compressors are probably one of the best tools you can use for mixing. And you can get compressors that act on certain bands of frequencies. Of course with some over lap since a 10th order filter might tear your system apart (or maybe my system since it isn't strong enough).

Why would you want to boost the the 'problem' area anyways???

Timmy J
 
I think easychair just meant that single-band compressors don't care about frequency. The side-chain of the compressor can be tapped into so that it causes the compressor to react (compress) depending on the eq shape provided to it which simply hypes a part of the spectrum whenever that component is present in the program material. When the single-band compressor kicks in however it compresses everything at that point in time - across the entire freq spectrum. Unless the compressor itself is hyped due to some extra feature - T-Racks compressor is hyped a bit like this but not due to compression.

As far as multi-bands go I've set up 16th order compressors (96db/octave) in the DAW and it doesn't hurt - just depends on the type of cross-over character you like!
 
kylen said:
As far as multi-bands go I've set up 16th order compressors (96db/octave) in the DAW and it doesn't hurt - just depends on the type of cross-over character you like!

kylen,

What are you using for your DAW? I'm in the midst of looking for a new station and would like as much input as possible. I've read a bit that AMD opeteron (I think that's how you spell it) are fairly good, but I can't remember what chipset to use it with.

Thanks,
Timmy J
 
tpreager said:
kylen,

What are you using for your DAW? I'm in the midst of looking for a new station and would like as much input as possible. I've read a bit that AMD opeteron (I think that's how you spell it) are fairly good, but I can't remember what chipset to use it with.

Thanks,
Timmy J
I'm on a P4 2.6GHz with an 845 chipset - in the midst of upgrading my chipset to get onboard firewire and usb2 as well as spdif i/o. I haven't finished yet so it's not a good time for me to recommend it...I can say I'm trying a ABit IC7-G motherboard for the upgrade, antec neopower 480w and either 1GB 2700DDR or 3200DDR I'm not sure yet. I thought I'd wait till they straighten out 64bit a little more before I consider the cpu upgrade...
 
Once in a life time

When I first read about what sidechaining is, there was an example of letting the kick determine how much compression to apply to the bass guitar, the result being (after hours of tweaking attack/release times, I guess) the kick attack getting through right before the note from the bass guitar hits you. I was wondering if anyone knows if this is the effect on Talking Head's Once in a Lifetime? It sounds like the bass just swells up from the kick attack, and it's really freaking cool, even if it is more obvious than you'd probably want on a straight-ahead rock record.
 
I accidentally came upon all this stuff about side chain compression. There is huge confusion going on here. Side Chain compression is not Side Chained compression. Notice the lack of the "ed" at the end of the word. The effect noted here is called Ducking and is achieved, as you say by hitting the threshold of the side chain input of a compressor with a foreign signal.
Side Chain compression is where a signal itself has been sidechained (parallelled) to a different set of channels on your mixer and had the socks compressed off it. This side chained ,heavily compressed signal is then mixed with the original uncompressed form of the signal. It is a very effective way of keeping your vocal on the front of the speakers without it having to be loud. Neither one of the two signals sounds like it's going to do any good but mix them together and magic happens.
This form of compression is also known as Parallel Compression or New York Compression (since they reach for it as a cure-all in New York apparently)
I hope this is enough to stop all the articles I've seen on the net getting it wrong. It probably isn't. Just because they've all got it wrong doesn't make them right though. They are all describing "Ducking" not Side Chain compression.
Gregg Jackman.
 
just to make it even clearer

I see that I'm being doubted here. It is not a matter of opinion but fact. Mike Brevelle first wrote about side chain compression in 1977. It has nothing to do with feeding a compressor's side chain with another signal. It is what Bob Katz decided to rename Parallel Compression. He feared people would make the mistake you all are making. There is nothing wrong with the technique you are all desdribing, except that it isn't called Side Chain compression. It is called "Ducking".

Gregg Jackman
grammy award winning engineer of 41 yrs experience.
 
Yes, Brevelle did write about parallel compression and did call it side chain compression; That is a fact.

But everybody read it and wondered why he called it side chain compression when side chain compression (using a key input) already existed.
 
Last edited:
my last word

Steenamaroos last comment makes the assumption that “everybody” thought Brevelle was wrong.

Some of us understood it.

I can't help it if loads of you didn't.

Mike Brevelle was accurate in his description of side chain compression.

Let’s imagine we’re using a DJs voice routed to the mix bus so we can hear him loud and clear, and creating a second path of it (side chain) to hit the key input of a compressor. The compressor is strapped across a music mix of some sort. When the DJs voice rises above the threshold set on the compressor it moves into gain reduction and the music gets ducked.

Did the side chain get compressed?? ( second path of the DJ )
Of course not. Therefore this is not side chain compression.

Side chain compression is where the side chain (second path) of a signal gets compressed. It’s really very simple. The clue’s in the name.

I’ll leave this discussion now and go back to being a recluse. I’m not comfortable with the possibility of being seen as a heretic. Galileo was more comfortable about that.
Mind you he had to do time for it.

Gregg Jackman
 
Back
Top