sidechain compression

When you normally use a compressor, you have signal in and signal out. The compression is determined by the level coming into the compressor. As the signal passes the threshold, compression begins.

When you use sidechain, you use another signal to control compression. If that signal does not exceed the threshold, then the compressor does nothing, the output is the input. As the sidechain signal exceeds the threshold, the signal coming into the compressor is attenuated, no matter what its level.
 
A typical use is "ducking" one signal under another. Sidechaining the kick drum to the bass guitar's compressor so that the bass is attenuated when the kick drum is hit is fairly common. The side chain usually has a filter on it so you can be selective about frequency ranges will trigger the compressor.
 
I've only used it once to reduce the guitar level whenever the singer sang. When the singer stopped it went back to normal guitar volume. The amount we reduced it was very slight otherwise it would be too noticeable. So the threshold depended on the singer's voice and the ratio we set really low like 1:2 or something.
 
A compressor has a "detection" circuit in it, which an audio signal is passed through in order to trigger the compressor. Typically, you want to compress your track based on the detector reading that same signal.

Sometimes, however, you want to compress a track while the detector is responding to a completely different audio signal. For example, if the dynamics of the track are already under control, but you want it to "hide" under another track that is clashing with it.

Compressing the bass guitar to the kick was mentioned and is very common, especially in Dance music where the bass is usually a synth with constant sine wave characteristics. Sidechain compression is often used on pads and other persistent tracks to "duck" their volume down out of the way of the rhythmic elements like the kick (especially in a 4 on the floor pulse). It can also be used on drum overheads or room mics, sidechained to the kick or/and snare, so they don't become out of control. You can use expanders to do the opposite and elevate sounds at the timing of another track. You can sidechain a gate so it turns a different track on on/off in time with another etc. etc.
 
Another common use is to side chain the kick drums to the pad synths in dance music.

This creates the effect of the pad 'breathing' with the track as it fades in on every off beat (assuming the kick drum is on the beat)
 
Side chain compression can also mean the same thing as parallel compression to some, mostly older, engineers.

De-essing is probably the most common use of side chain processing. Use an eq ahead of the detector to emphasize the sibilance so the compressor reacts most strongly to that frequency range.

Just about the only thing I ever do in the detector circuit is to high pass it on some vocals.
 
Side chain compression can also mean the same thing as parallel compression to some, mostly older, engineers.

I thought that parallel compression was mixing a heavily compressed signal with the original signal (and that it's also called New York compression). Did the meaning of side-chain compression change?
 
Another use is to use the sidechain to EQ the signal going into the detection circuit.

In that case, the signal going to the compressor is split in two. One going to the compressor and one going to the EQ

The output of the EQ is sent to the sidechain input on the compressor

You set the EQ to make the compressor react (or not react) to certain frequency ranges. This is how to make a De-esser, simply pull out everything below 4k or so on the EQ. This way, only a very loud signal above that frequency will trigger the compression.
 
I thought that parallel compression was mixing a heavily compressed signal with the original signal (and that it's also called New York compression). Did the meaning of side-chain compression change?

That is what "parallel compression" means. Some older engineers call that side chain compression while most younger engineers consider side chain compression to refer to processing the signal to the detector circuit. So, yes, the meaning did shift.

A side chain is just a parallel signal path. A compressor's detector circuit is one kind of parallel signal path, a bus with duplicate signal is another kind of parallel path.
 
A side chain is just a parallel signal path. A compressor's detector circuit is one kind of parallel signal path, a bus with duplicate signal is another kind of parallel path.

I don't see a side chain as a parallel signal path. It is an alternate input to the detector. Parallel compression to my mind is one or more compressors in parallel, including, in some cases, a null compressor in the original track to compensate for propagation delays through the compressed paths
 
I don't see a side chain as a parallel signal path. It is an alternate input to the detector. Parallel compression to my mind is one or more compressors in parallel, including, in some cases, a null compressor in the original track to compensate for propagation delays through the compressed paths

The detector path is the side chain, an internal parallel signal path. The access to it is typically called the side chain insert which allows you to insert a processor on the side chain.
 
The detector path is the side chain, an internal parallel signal path. The access to it is typically called the side chain insert which allows you to insert a processor on the side chain.

Its not an internal parallel signal path. The detector controls the gain reduction of the compressor. Without the sidechain, the detector gets its input from the signal being compressed. In sidechain mode, the detector gets its signal from another source. There is nothing parallel about it. The detector is used either way, in the same way. Its just the source that changes
 
Its not an internal parallel signal path. The detector controls the gain reduction of the compressor. Without the sidechain, the detector gets its input from the signal being compressed. In sidechain mode, the detector gets its signal from another source. There is nothing parallel about it. The detector is used either way, in the same way. Its just the source that changes
The detector is not in the signal path that makes it to the output. After the input, the signal is split and sent to both the detector path and the amplifier that does the gain reduction.

If there is a sidechain button, it interrupts the signal feeding the detector and activates the sidechain input.

If there is a sidechain insert, it works like a channel insert on a mixer. If you use an insert cable, you can insert an EQ and EQ the signal going into the compressor. If you just use a straight cable, you can feed the detector any signal you want.

Internally, there are two parallel paths.
 
Okay, while we're on compressors, I have a C1-L1 plug, and although I get Attack/Release/Ratio/Threshold/Makeup. I even kind of understand the mid scoop knob and kind of love having a variable Knee knob. Even has a nice variable filter knob with it's own on-off switch, which I suppose would be great for funk bass and some other things I can think of...leave the bottom end alone and jut compress from X Hz and up. I've even done some experimentation with Alpha ans Sigma Attack modes, and although I don't understand why, I can hear the difference and use them accordingly...

But, can someone define (in real terms) what the difference is between Stereo mode, Mono L Mode, Mono Mode, Mid/Side Mode, Mid Mode, Side mode? There's another one that just says R-S All a bit overwhelming when you look at the old soviet units and their plug in equivalents, and see all those modes and have no idea because I can't seem to get them to sound different no matter the signal...:confused: Maybe it's just my ears.

Well, they sound different, but it's hard to tell what to use where...is there a guide to using this thing that I've missed?
 
Its not an internal parallel signal path. The detector controls the gain reduction of the compressor. Without the sidechain, the detector gets its input from the signal being compressed. In sidechain mode, the detector gets its signal from another source. There is nothing parallel about it. The detector is used either way, in the same way. Its just the source that changes

Nope, the side chain is the path with the detector as distinguished from the main signal path through the gain control section. The side chain insert or input is the access point to the side chain. The side chain signal can be tapped from before or after the gain controller. In the diagram below the side chain is shown with red arrow and the side chain insert would be between the splitter and rectifier blocks.

Comp-Block-Diagram.gif
 
Mid/side mode normally has one channel compressing the mono information (mid) and the other channel compressing the stereo information. Think of it like a mid/side mic technique where one channel is compressing the signal from the cardioid mid mic and the other channel is compressing the signal from the figure 8 mic that is picking up the left and right.

There is an explanation here

https://www.soundonsound.com/sos/apr12/articles/reaper-0412.htm
 
Internally, there are two parallel paths.

Yes, but they end at the VCA, not at the output, which is what I would define as parallel paths. If the signal came into the device and was split two ways and converged at the output, then that would be a parallel signal path. However, one fork of the split doesn't make it to the output, it only controls the attenuation of the main signal.
 
Yes, but they end at the VCA, not at the output, which is what I would define as parallel paths. If the signal came into the device and was split two ways and converged at the output, then that would be a parallel signal path. However, one fork of the split doesn't make it to the output, it only controls the attenuation of the main signal.

But even by your definition, they are parallel until the VCA. It is still two separate paths internally. The way you described it above made it sound like you thought the detector was in the signal path before the vca, which isn't true.
 
Back
Top