Why is Parallel Compression Such a Big Deal??
Parallel Compression - Maybe You've Heard of It?
Parallel Compression, or New York Compression as it's referred to sometimes, gets a lot of hubub in the audio community nowadays. Youtube is littered with tutorials like "Make Your Drums Slam with Parallel Compression" and "Add Punch and Depth with Parallel Compression!" But what's the big deal? What's the difference between parallel compression and good ole regular compression?
Onward and Upward
When most of us talk about compression, we're talking about downward compression. A downward compressor is a device that pushes down any signal above a set threshold -- hence downward. For example if the threshold was 0dB, any signal coming through above 0dB would be attenuated to a quieter level. How quickly, and how much the signal is attenuated is dictated by the type of compressor and its parameters -- topics for another post. The goal of a compressor, any compressor, is to make the dynamic range (the difference between the quietest part and loudest part) smaller.
On the other hand, parallel compression is closer to what's known as upward compression. Rather than pushing signal down, upward compression takes the quieter parts of a signal and brings the level up. For example, if a recorded guitar had a dynamic range from -5dB to +5dB, upward compression would take the quietest signal and raise it, shrinking the dynamic range to something like -2dB to +5dB.
The simple premise of downward and upward compression is similar -- take something with a wide dynamic range, and shrink it. However, upward compression can be more desirable in certain situations because it usually (when done correctly) doesn't affect the loud transients (the quick, initial attack) of a signal as much. Downward compression by its very nature alters these transients, no matter how "transparent" the compressor. To avoid altering transients with regular compression you have to use a slow attack, but a lot of times you end up losing the body and ambience of the signal in exchange for that crisp attack.
Take a snare hit for example: the initial attack of the snare is a quick loud signal that appears only for a few milliseconds. The attack then decays quickly without much sustain. Downward compression shrinks the dynamic range by squashing the initial attack, but upward compression shrinks it by bringing up the quieter body and sustain of the hit after the attack, leaving the initial "crack" unaltered and natural sounding.
Notice how both parallel compression examples have a nice big initial transient, whereas the fast attack regular compression barely has any at all. The slow attack compression has a great initial transient, but it decays a little more quickly than the parallel examples. Of course, you can alter your regular compression to have a fast release to try and make up for this, but with a slow attack and a fast release you risk losing the natural sound of your source due to pumping. "Pumping" is a term used to describe the sensation of a compressor that attenuates a signal and then noticeably lets go. We generally hear these level changes as sounding unnatural and try to avoid them. However, don't throw out slow attack and fast release settings quite yet -- we're going to use them to create our parallel compression.
So How Do We Go Up?
Okay, so upward compressors sounds cool, where can I get one? The problem is that upward compressors are not at all commonplace in most studios, or plugins for that matter. The way most of us achieve upward compression is by faking it with downward compressors. We take a dry uncompressed signal, with all transients intact, and add in a super duper compressed signal, with thick, lush body and decay. What we get is all the crisp clarity of the dry signal with the added body and sustain of the squashed signal.
For example, in the analog world we would simply run one channel dry, then bus or duplicate that channel and compress the crap out of it. Then we mix in the compressed channel to taste. In the digital realm, there are a few more options: 1. You can literally duplicate the track in your DAW. Leave one track dry, compress the other, and mix. 2. Bus the dry track to an aux, compress, then mix dry and compressed aux -- though this only works if your DAW has delay compensation, otherwise phasing could be an issue. 3. The easiest way is to take advantage of a compressor plugin that has a "mix" knob. Simply increase the amount of compression by turning the knob from left (completely dry) to right (completely wet.) Usually a 12 o'clock position will achieve 50% dry, 50% compressed.
So What Does it Actually Sound Like?
Honestly, I find it hard to find any source that parallel compression doesn't sound good on. A better question than "what should I use it on?" is "when should I use it?" Although parallel compression can sound arbitrarily good on any source, it doesn't make sense for all styles of music. Parallel is great on individual drums, drum busses, bass, guitars, vocals, and master busses -- but only if you want those things to have the full body and punch of parallel compression. Parallel compression might be great for rock, blues, hip hop, pop, and alternative genres; but maybe ill-suited for blue-grass, classical, or any source that wider dynamic range is preferred. Simply put, just like with regular compression, parallel compression can be over-used, and in trying to add fullness and depth, one can easily squeeze the life out of a beautiful performance. As always, discretion is the name of the game.
Those of you who are saying "okay, okay, enough of the opinions and stuff, let's hear some real-world examples!" -- the following will provide a few bars of drums to compare. The differences in these mixes is pretty subtle, so I can pretty much guarantee that you won't hear a difference unless you're listening on nice monitors or headphones. If you don't hear it at first, don't worry -- you're not crazy. You won't hear much of a difference (especially between the compressed mixes.)
The attacks of the kick and snare are pleasantly in tact on all three due to the slow attack on the regularly compressed mix (you can hear a difference in attack between the recordings, but none of them sound as if the attack has been completely lost or compromised.) The overall tone doesn't change much. The thing you're looking for is in the background. Pretend like you're trying to listen to the room that the drums are playing in. Listen to the dry mix, then listen to the parallel mix -- you'll hear the ambience of the room really pop out. Once you can hear that, then listen back and forth between the different types of compression. This ladies and gentlemen, is the level of critical listening that will take your mixes from "pretty darn good" to "yeah, I don't know what you did, but please take my money!" Well, maybe people won't be throwing money at you, but you'll notice a difference in your mixes for sure!
I've included a fourth track with the parallel compressor settings, but mixed 100%. This should give you a good idea of what exactly it is that is blending in with the dry track. I should also note that these compression settings are dialed in for subtlety -- a higher ratio and lower threshold could easily provide a more extreme compression that could add even more body and ambience to the track. Also, if you're interested, the compressor I'm using is the Slate FG-Grey compressor pictured earlier in the post.
After reading through this post I'm realizing it could possibly come off as an ode to parallel compression and a criticism of regular compression. Please don't hear what I'm not saying: Regular compression sounds really really good and I probably use it more often than parallel compression in most of my mixes. This article is merely attempting to point out sonic differences between the techniques, and the way that altering compressor parameters produces different effects in parallel compression vs regular compression. It is my hope that the readers will find this helpful in experimenting with their mixes and finding what works best for them.
That being said, what do you think? Do you prefer regular or parallel compression? Do you have a favorite type of compressor? (ie Optical, Variable-Mu, FET, VCA.) When is it appropriate to compress vs. leaving the source dynamic breath? I'd love to hear your thoughts!