Parallel Compression

Certain mix techniques have an aura of mystery about them. The first time someone explained side chain compression to me, my eyes crossed a bit and I got lost in my studio for a week trying ideas out. Parallel compression, sometimes called New York compression, can be similarly mystifying, but it is a delightfully useful technique once you understand how to apply it correctly. 

Background

Effects processors fall into two broad categories. Some are intended to create something that is blended with the original signal. A delay is the classic example of this. Think of how useless a delay effect would be if you didn’t also have the original signal to compare it to. Your sound would just be…late. Almost all time based effects, like delay and reverb, fall into this category. The processing just doesn’t make sense to our ears without the original signal present as well.

Things like tonal and dynamics processors fall into a different category. They are generally not geared to add something to a signal, but rather to fundamentally change the signal they are processing. This is not really the kind of control that makes sense on an equalizer or a noise reduction processor. Why would you want to mix back in some of the audio you corrected with this processor? Blend some of the muddy or noisy signal back into the what you had just improved? I don’t think so!

A good cue for which type of processing effects fall into is the presence of a blend or “wet/dry” knob. At its core, this is purely a balance control that allows you to mix between how much of the processed and unprocessed signal you are listening to. 

So (spoiler alert) parallel compression is really all about putting a blend control back into a compressor, which is generally considered one of those processors where this approach doesn’t really make sense. However, it can yield some effective results and it can be applied in the analog domain, or within your DAW.

The basic idea is to split the signal you are processing, either with a mult in a patchbay in the analog world or a bus in a DAW. In a DAW, you would instantiate the compression plug in on the bus channel or aux input rather than on the main channel so that you are compressing a separate version of the signal.

While the compression settings will vary a bit based on the song and application, you are going to compress much more aggressively than if you were putting the compressor inline. A ratio of 10:1 with a fairly low threshold ( -26 dBFS) is a good starting spot but it is not unusual to get very aggressive with the compressor/limiter settings in these applications.

After that, you can balance the level of the compressed and uncompressed signals by adjusting the fader that the compressor is set up on. A good approach to this is to set up the compressor and then pull the fader for this channel all the way down. Slowly bring it up under the uncompressed signal until it is at a level where you begin to hear its impact. 

A couple of things will happen immediately. You will hear a certain thickness or warmth that was not there before. Part of the reason is that you are getting two different tonal characteristics from the compressed and uncompressed signals. The uncompressed channel allows the transients to come through but the compressed channel adds some character. Now try muting the channel the compressor is on and you should feel like something is missing.

Since you are combining two tracks that are basically identical, you are increasing the level of whatever you are processing which is below the threshold of the compressor. If you set both the compressed channel fader and the uncompressed fader to unity everything below the compressor threshold would be 6 dB louder. As the signal crosses the threshold in the compressed channel, the compressor will keep this channel from getting any louder so the uncompressed audio will be what you mainly here. This is a bootstrapping approach that can be less damaging to your audio in general but helps bring up the volume and impact of quieter passages.

Note that when you are doing this technique you are combining essentially identical signals so having delay compensation enabled on your DAW is critical or you will hear lots of comb filtering. Also, some compressor plug-ins with look ahead features induce a lot of latency and could outpace the capabilities of some systems to account for this much delay.

Application

Classic places to apply parallel compression are full mixes, drums and vocals but each have areas where you need to pay attention.

On drums and full mixes one of things to be cautious about is the fact that drums usually have the loudest transients in a mix. If you have short attack and release times, this can cause the compressor to “breathe” in time with the beat. This can be a cool effect or it can be distracting. To avoid this, you could create a bus in your DAW that has everything routed to it EXCEPT the drums and then apply compression to this channel. If you set your bus sends to be prefader any changes you make in your main mix will not impact what is going to the compressor. Blending this back in to your output mix avoids most of those pumping issues but can add a sense of fullness to the mix and help emphasize quieter passages. 

The technique applied to vocals can be extremely effective. I have spent hours applying clip gain to quiet passages of a lead vocal trying to get it to punch through a dense mix. Applying this to a vocal allows you to help with that deficiency but also allows you to get some cool color by using a compressor that has a lot of character to add small bits of distortion to a vocal. A saturation plug-in like the Waves Abbey Road processor can add subtle life, or you can go over the top with something like the Soundtoys Decapitator for some screamo type effects without destroying your intelligibility.

Of course, if you look for a plugin that already has a “Wet/Dry” mix built into it you don’t have to jump through all the routing hoops. Softube’s Grand Channel already features a wet/dry mix that is helpfully labeled “parallel inject”. If you place this compressor plug-in on your stereo out and play with the wet/dry balance you are doing the same thing without all the internal DAW wiring. 

Softube labels this as parallel inject but you will also see it as a "blend" control