Parallel compression on drums, vocals and the mix bus: a step-by-step guide

Ordinary compression has a built-in trade. Push it hard enough to make a drum kit or a vocal feel dense and loud, and you start shaving off the very thing that made it exciting: the front edge of each hit, the little swells in a voice. Parallel compression is the way out of that trade. Instead of compressing the sound itself, you compress a copy, hard, and tuck that copy underneath the untouched original.

The result is upward rather than downward in feel. The loud parts barely change, because the dry signal still carries them. The quiet parts, the room tone, the snare's tail, the breath before a phrase, come up, because the squashed copy is loudest exactly where the original was quietest. Engineers often call it New York compression, after the dense drum sounds associated with that city's studios, but the idea is the same in any DAW.

If you are new to threshold, ratio, attack and release, start with our compression basics guide. This piece assumes you know those controls and focuses on doing the parallel version well.

Two ways to set it up

Option 1: a Dry/Wet knob on one insert. Many modern compressors have a mix control. In Ableton's Compressor, "Dry/Wet adjusts the balance between the compressed and uncompressed signals". Set the compressor up aggressively, then turn the mix down until the original's movement returns. This is the quickest route and the right one for a single vocal or a single bass.

Option 2: a send to a return (aux) track. Put the compressor on a return, set it fully wet, and send the tracks you want to it. The Live 12 manual notes that returns can process audio sent to them from numerous tracks, which is the point: one compressor reacts to the whole kit at once, and you choose how much of each drum feeds it. Our bus processing and groups guide covers the routing in detail.

Which to choose? Use the return when you want different amounts of different sources (lots of snare and room, almost no kick) or want to EQ the squashed signal separately. Use the knob when one source is enough and you want fewer channels.

Pre or post? The same manual says every Live return has a Pre/Post toggle that decides whether the send is taken before or after the track's mixer stage. For parallel compression, post-fader is almost always right: when you ride the snare fader down, its share of the crushed signal goes down with it, and the balance holds.

Step by step: parallel drums

This is where the technique earns its reputation.

1. Route the kit. Create a return, add a compressor, and send kick, snare, toms and room mics to it. Leave overheads lighter if cymbals start to hiss.
2. Crush it. High ratio (8:1 or more, or a limiter-style setting), low threshold, fast attack, and a release that lets the compressor recover before the next beat. You want 10 dB or more of gain reduction. Solo the return; it should sound flat, pumping and almost ugly. That is correct.
3. Start silent. Pull the return fader all the way down, unsolo, and play the full mix.
4. Fade it up slowly. Bring the return up until the kit feels bigger and more sustained, then back off a touch. The usual sign you have gone too far is that the cymbals wash forward and the kick loses its click.
5. Shape the copy, not the kit. If the crushed signal adds mud, high-pass it. If it adds harshness, a gentle top-end cut on the return keeps the size without the fizz.
6. Check in mono and at low volume. Parallel drums should still read as "more kit" on a phone speaker, not as louder hi-hats.

Step by step: parallel vocals

Vocals want a lighter hand. The goal is intimacy and consistency, not power.

1. Compress the main vocal normally first. A gentle insert compressor controls the big peaks; the parallel layer is extra.
2. Use a medium setting on the copy. Around 4:1, medium attack so consonants still come through, and enough gain reduction that quiet words sit up.
3. Blend low. On a vocal the parallel layer often sits much lower than on drums. If you can clearly hear it as a separate effect, it is probably too loud.
4. Watch the sibilance. Heavy compression lifts "s" and "t" sounds along with everything else. A de-esser before the parallel compressor, or on the return, fixes most of it.

Our vocal chain basics guide explains where this sits relative to EQ and de-essing.

Step by step: the mix bus

Parallel compression on the whole mix is the riskiest use and the one most likely to be overdone.

1. Use the mix knob, not a return. On the stereo bus a Dry/Wet control is simpler and avoids routing the whole mix twice.
2. Go gentle. Low ratio, slow attack, a few decibels of reduction on the copy, and a mix setting that is mostly dry.
3. Level-match before you judge. Parallel compression nearly always makes things louder, and louder tends to sound better at first. Bypass and compare at matched level, or you will fool yourself.
4. Leave headroom for mastering. If the mix is going to a mastering engineer or a mastering tool, it is usually better to keep this light and let that stage decide on overall density. Our limiting and true peak guide explains what happens next.

Common mistakes

Phasey or hollow sound. If the parallel path is delayed relative to the dry path, the two copies interfere. Most DAWs compensate for plug-in latency automatically, but check that compensation is switched on, especially if you are using an external hardware compressor or a high-latency plug-in on the return.

Using it everywhere. A parallel return on drums, another on vocals, another on bass and a mix knob on the bus add up to a mix with no quiet left in it. Pick the one or two places where density helps the song.

Not soloing the copy. People often blend in a parallel signal that is barely compressed. If the soloed return does not sound obviously squashed, it is not doing the job; you are just adding a second, slightly louder copy.

Forgetting the arrangement. Parallel compression lifts quiet detail, including noise, bleed and edits. Clean those up first.

What we think

Parallel compression is usually taught as a loudness trick, and that undersells it. Its real value for an artist is that it lets you keep the performance. A singer who drops to a near-whisper on the second verse, a drummer who plays ghost notes you can barely hear, these are choices, and ordinary heavy compression flattens them into sameness. A parallel layer lets you lift the quiet parts into earshot without erasing the difference between quiet and loud.

That makes it a decision about feel, not just level, and it is worth treating as one. Before you reach for it, ask what the dynamics of the part are saying and whether you want the listener to hear more of that or less. Some of the best records are dense because the band played them dense. Others are moving because the gaps were left alone. A crushed return can fake the first; it cannot recover the second once it is gone.

Our practical rule: set up parallel compression on drums by default, try it on vocals when the quiet lines are getting lost, and be sceptical of it on the mix bus. And always make the final call with the levels matched, because the version that is merely louder will win every unfair test.