Pro Audio Files

2 Effective Ways to Use Parallel Compression

➥ Learn how to control shape, tone and dynamics with compression

Parallel compression refers to the technique of duplicating a signal, compressing the copied signal, and then blending it back in with the uncompressed signal. Some compressors also have a ‘mix’ knob where you can blend the compressed and uncompressed signals. The benefits of parallel compression are manyfold, though there are a few pitfalls.

Setting up a parallel compression chain can be daunting. Compression on its own can be a bit complex for those who are just getting the swing of things. Parallel compression complicates this by coupling all the versatility of regular compression with treating a signal in extremes.

The parallel signal isn’t necessarily going to sound good on it’s own, so how do you know when you’ve compressed it correctly?

Well — it isn’t so tough if you can find your center — which in all cases of mixing is: what are you trying to accomplish?

If you know what you want to hear, you’ll know whether or not you’ve set up the chain correctly.

In my book, there are two reasons to use parallel compression:

  • To get a very “forward” sounding compression.
  • To highlight and bring out a specific tone in the signal.

The “In Your Face” Compression Sound

One way to set up an effective parallel compression chain is to use very heavy settings.

Turn the threshold down almost all the way (or turn the ‘input’ up a lot). The idea is to find that spot where the very edge of the signal’s sustain is where the compressor starts to release. Set the attack as fast as possible.

Think of the ratio as texture control, with higher ratios creating more of a crunchy/distorted sound. A 2:1 setting will be a bit more “natural” but still extreme at such a low threshold. Any higher ratios will start pushing in some heavy pump and eventually some chompy distortion (which could be a good or bad thing).

For the in-your-face clean sound, 2:1 is a good place to start. Even lighter ratios will still be dramatically effective. I find the release works best when it’s pretty slow. The amount of reduction should be most minimal at the very edge of the sustain of your signal, but the release needs to be set slow enough so that you don’t bring up the room tone or create a wooshing noise. Unless you want to highlight the “air” — but I’ll get to that next.

When you blend the compressed signal back in, you’ll get a natural sounding compression curve that seems to beef the sustain of the signal a little better than regular compression, resulting in a more “forward” sound.

Targeting a Specific Sound

This is probably the most useful way to utilize parallel compression — and unsurprisingly the one I hear about the least. You can use parallel compression like a phase-free EQ!

If you use a compressor with an external sidechain, or adjustable sidechain, you can feed the sidechain with an EQ’d version of the source sound. If you EQ out the part of the signal you want to enhance, the compressor will only be acting on the opposite frequencies.

In other words, if you want to bring out the low end of a kick drum, put a high-pass filter on a mult of the kick feeding the sidechain. The compressor will dodge the low end of the signal. When blended back in, it’s as if you EQ’d up the low end, but without the artifacts.

In addition, there are no rules against EQ’ing the parallel signal post-compression to bring out a certain tone, or doing anything else to it for that matter.

For a deeper look at compression, check out Mixing with Compression.

Here’s a video you might like on using parallel compression to bring out the attack or sustain of a snare drum:

Here’s a playlist of more compression tutorials:

Missing our best stuff?

Sign up to be the first to learn about new tutorials, sales, giveaways and more.

We will never spam you. Unsubscribe at any time. Powered by ConvertKit
Matthew Weiss

Matthew Weiss

Matthew Weiss is a Grammy nominated and Spellemann Award winning audio engineer from Philadelphia. Matthew has mixed songs for Snoop, Sonny Digital, Gorilla Zoe, Uri Caine, Dizzee Rascal, Arrested Development, 9th Wonder, !llmind & more. Get in touch: Weiss-Sound.com.

Free Video on Mixing Low End

Download a FREE 40-minute tutorial from Matthew Weiss on mixing low end.

Powered by ConvertKit
  • Frank Nitsch

    Hi Matthew,

    thanx for the article. I like to learn about advanced topics a lot. Compression and all its variants like multiband or parallel compression definitely leave plenty of room for improving knowledge and experience. 😉
    I’m a bit confused after reading the above statement about using a side chained parallel compressor for EQing. The filtered side chain signal would trigger the compressor only based on the frequencies passing the EQ or cut filter. But when the side chain triggers, the compressor would still tame the whole frequency spectrum, wouldn’t it? Wouldn’t this rather be a good use case for a multiband compressor instead? It would limit the compression to the desired frequency bands and leave the low frequency range in your example untouched.

    Take care

    Frank

    • There’s pros and cons to both. Using parallel compression to target a frequency range may not be as specific, but it also doesn’t have any of the phase artifacts that come from equalizing effects. In certain situations where you specifically want to target isolated incidents of frequency bumps multi-band might be better. For example, low frequency bumps from a vocalist swerving into the microphone. Other times, you want to emphasize frequencies without smearing them – in vocals, using wide band de-essing is an example of this. Likewise, if a vocalist has a tendency to shoot excited notes through her/his nose you may want to use a parallel compression chain to compress those notes harder – but not actually change the tonality.

      Think of it as rearranging the dynamic structure of a sound to emphasize areas where certain frequencies don’t exist. You’re shifting the focus of the sound, rather than actually changing it’s tonality through phase manipulation. If that makes sense.

      I’ll post up another comment if I can figure a better way to explain it, but until then – I suggest setting up a parallel chain that reacts more toward certain frequencies, and setting up a multi-band compression chain and try to produce the same results. You’ll hear that it functions quite differently.

  • Great article and quite well written.

    I’d just like to include that Parallel Compression is great for an additional reason:

    You can compress and EQ the living daylights out of a summed collection of signals, such as drums, and still retain a good amount of its original dynamic range as you have full control over the blend between the two. It’s so easy to destroy your dynamics through compression – parallel compression allows us to really bring our tracks to life in a musical, and LOUD, way while still having the advantage of ‘natural’ sounding transients. Compressing and limiting the bejesus out of your drum tracks while keeping it’s original character is like having your audio cake and eating it to, in 24 bits.

    word,

    Jon

  • Ok – I have an addendum that may make things more clear.

    Parallel compression can be used at a tool to isolate a unique part of a signal. For example, you can set the attack a bit longer and the release a bit longer. With an extreme threshold setting, this will leave a primarily attack based signal remaining. This is GREAT if you are trying to enhance the attack of the source sound. Similarly, you have a hyper fast attack and a quick release with not as extreme of a threshold (but still heavy). This will squelch your attack and leave primarily the sustain behind.

    If the source sound is more complex, like a vocal, you can eq a side chain to primarily pull down certain tones. For example, some singers will tend to project more into their noses on louder notes. That means that not only will the note be louder, it will also be midrangy-er. Using a sidechain, you can focus the compressor to react specifically to that midrange. When you blend the compressed signal back in, you will have a more dynamically AND tonally even vocal – without phase artifacts.

    Lastly, you can pair parallel compression with other processors. You can eq the parallel signal to really emphasize or de-emphasize certain frequency. You can pair the compressor with a gate. There’s a lot of possibilities. You can compress the parallel and uncompressed signal together – so the effects of the parallel compression become more prevalent on quieter parts of the signal. Possibilities are endless.

  • Mike Schoonmaker

    Parallel compression really is a great trick, I use it on almost every project i master (Gigantic Masteirng).

    I actually find myself setting up “manual” parallel compression (stemming out the audio and comp/eqing before summing again) much more often than doing it solely through a plugin like fabfilter’s pro-c… Its just much more flexible and gives you the opportunity to eq any negative side effects of heavy compression or mix in only the compressed low end, etc.

  • DontWorryImAPilot

    In almost every rock song I’ve worked on (I say “almost” because it’s possible I just don’t remember the times I didn’t use it) I use parallel compression on drums.

    One thing I think people forget: you don’t have to send the entire set of drums to the parallel compression buss.

    Send (post fader) the kicks, the snares, and the toms at unity gain to a super fast attack med release compressor and you get big aggressive hits. Send a mono room mic at -6db and you’ll get some more depth and sustain in with the attack of the close mics. Send overheads in at -9db and you’ll start to get some more cymbals but, keeping the overheads lower in the parallel buss, you’re not going to get blasted as much with washy frequencies. I almost never send cymbal spot mics to the parallel buss but sometimes a touch of hihat is helpful.

    I find that sometimes the kick’s send will need to be lowered. Sometimes the snare needs a bit more. Sometimes the toms are perfect at unity. Etc. Don’t be afraid to balance that stuff out in the parallel buss just like you would in the main drum buss.

  • Hi I’m a self-taught mixing noob who is always wondering if I’m doing it right…

    I frequently use heavy compression on the sent signal from volume-automated vocals or horns. doing it could sorts up the tendency of them being over-automated.
    Also using heavy ambience on the sent signal from hard comp/limiter channel can prevent ambience being too harsh + sometimes it fixes phasing issue… as I guess.

    But I’m not really sure if It is a good choice. any opinion from anyone will be much appreciated. do you guys think it’s a good method?

  • Blue Eagle

    Hey Matt !

    Just thanks for your insights, and your always clear way to express the foundations of mixing.
    Many times helpful

    ” The idea is to find that spot where the very edge of the signal’s
    sustain is where the compressor starts to release.”

    Do you mean, the spot where Needles / Gain Reduction, just recover from the Transients ? To zero GR

    Or should we compress some of the sustain, and then see the Gain Reduction not always recovering ?

  • Blue Eagle

    Globaly, how many GR do you apply exclusively in the compressed channel ?

  • Blue Eagle

    Hey Matt !

    Just thanks for your insights, and your always clear way to express the foundations of mixing.
    Many times helpful

    ” The idea is to find that spot where the very edge of the signal’s
    sustain is where the compressor starts to release.”

    Treshold wise,

    Do you mean, the spot where the Needle / Gain Reduction, recover fully from the Transients ?

    Or should we compress some of the sustain, and then see the Gain Reduction not recovering ?

    ~*~

  • Jay Niesler

    Best compression explanation! Thank you.

  • Adam Lovatt

    I realise this article is pretty ancient by internet standards, but I wanted to point out a pretty common mistake that it falls prey to.

    “If you EQ out the part of the signal you want to enhance, the compressor will only be acting on the opposite frequencies.

    In other words, if you want to bring out the low end of a kick drum, put a high-pass filter on a mult of the kick feeding the sidechain. The compressor will dodge the low end of the signal. When blended back in, it’s as if you EQ’d up the low end, but without the artifacts.”

    This is flat-out wrong. This is not how a compressor works at all. Yes, high-passing the detector will keep low-end content from *triggering* the compressor, but whatever low-end is happening at the same time will be compressed just as much as the highs.

    If it works for you, then obviously do it, but the explanation relies on one of the most enduring (and annoying) compressor myths.

    Cheers, and thanks for what’s otherwise a great post.

Recommended Courses