No processor gets used on more of my channels than compression, and no processor gets abused more often by everyone else. The problem isn’t difficulty. It’s that most explanations stay hopelessly vague. Being told a compressor “glues the mix together” gets you no closer to knowing where the attack knob goes.
This is the practical version instead. What the box really does, what every knob controls, starting points for vocals, drums, bass and 808s, plus the three techniques (sidechain, parallel, multiband) that handle 90% of what you’ll ever need.
Before we start: fancy hardware is not required for any of this. The stock compressor in your DAW does the job, and Serum 2 users already have a decent compressor sitting on the FX buses; details in my Serum 2 effects guide.
What compression actually does
A compressor automatically pulls down whatever gets loud. Nothing more. The gap between a track’s softest and loudest moments is called dynamic range, and the compressor narrows that gap so the level holds steadier.
Two controls set the basic behavior:
- Threshold: the level at which compression kicks in. Signal sitting below it goes through unchanged.
- Ratio: how firmly the level gets reduced once the threshold is crossed. With 2:1, a signal poking 4 dB over the threshold emerges just 2 dB over. With 4:1, that same 4 dB overshoot exits at 1 dB. Push past 10:1 and you’ve essentially built a limiter.
Here’s a habit worth building early: read the gain reduction meter instead of staring at knob positions. Gentle leveling means 2–4 dB of reduction on the biggest hits. A meter stuck at 10 dB around the clock means the threshold sits too low, and the track is going to come out squashed.
Also, level-match with makeup gain every single time you A/B. Louder-and-compressed always wins the comparison, but that’s your ears being fooled, not the compressor doing you favors.
Attack and release: where the sound actually changes
How much compression happens is set by threshold and ratio. What it sounds like is set by attack and release. All the tone shaping happens right here.
Attack controls how quickly the compressor clamps once the signal passes the threshold.
- Slow attack (10–100 ms) lets the first transient sneak through before compression grabs the body. You get punch, that “in your face” quality. Perfect for snares, plucked sounds, anything where the click is the point.
- Fast attack (0–5 ms) grabs the transient along with everything else. Sounds get rounder, smaller, pushed back in the picture. That’s what you want for pads, backing vocals, background parts, and for calming sharp clicks and nasty peaks.
Release controls how quickly the compressor backs off once the signal falls under the threshold again. You’re usually working between 40 ms and a few seconds.
- Fast release gets the compressor breathing along with the material. Sustain and room come up, energy increases. Take it too far, though, and pumping appears (sometimes exactly what club music wants).
- Slow release smooths the gain reduction into invisibility. The right call for vocals and buses where the compressor should never announce itself.
On rhythmic material, try this: dial the release so the gain reduction meter lands back at zero just before the next hit arrives. Now the compressor rides the groove instead of wrestling with it.
Compressor types in plain terms
Open any plugin folder and you’ll find emulations named after circuit designs. What those labels mean for your ears:
- VCA (think SSL bus compressor or dbx 160): quick, clean, surgical. The do-everything option. It’s my go-to for drum buses and mix bus glue.
- Optical (think LA-2A): inherently slow and smooth, since the gain element is literally controlled by light. Hard to make sound bad on vocals, bass and guitars.
- FET (think 1176): blisteringly fast, aggressive, with a hint of grit baked in. The classic pick for snares, room mics and vocals that need to be right up front.
- Vari-mu (think Fairchild): tube circuitry, gentle behavior, less a peak catcher and more a “thickener”. You’ll find it on mix buses and mastering chains.
Straight talk: early on, which type you pick matters far less than where your attack and release sit. A plain digital compressor set correctly wins against a vintage emulation set wrong, every single time.
Starting settings by instrument
Treat these as launch points, not law. Dial them in, then tweak with your eyes on the gain reduction meter.
| Source | Ratio | Attack | Release | Gain reduction |
|---|---|---|---|---|
| Lead vocal | 3:1 | 10–30 ms | 100–300 ms | 3–6 dB |
| Kick / snare | 4:1 | 20–40 ms | 50–150 ms | 3–5 dB |
| Drum bus | 2:1–4:1 | 10–30 ms | fast / auto | 2–4 dB |
| Bass / 808 | 4:1 | 10–30 ms | 100–200 ms | 3–6 dB |
| Pads / keys | 2:1 | 1–5 ms | 200–500 ms | 2–3 dB |
| Mix bus | 1.5:1–2:1 | 30 ms | auto / 100–300 ms | 1–2 dB |
Vocal tip: a single compressor pulling 8 dB announces itself immediately. Two in series, each doing 3–4 dB, sound like nothing happened. Oldest move in vocal mixing, still undefeated.
With 808s, compression is honestly only part of the picture; saturation and a clean single-oscillator patch do more of the heavy lifting. There’s a full walkthrough in my Serum 2 808 bass tutorial, and the Synopsis 808 presets come pre-processed to park themselves at the front of a mix. Prefer samples? A properly recorded one-shot barely needs compression at all, which is the entire premise of our 808 one-shot kit.
Sidechain, parallel and multiband compression
Sidechain compression is when the compressor reacts to a signal other than the one it’s actually processing. The textbook example: compressor on the bass, kick routed into the sidechain input, so the bass dips every time the kick lands. That’s the pump running through house, EDM and plenty of hip-hop. Begin with a fast attack, release somewhere around 100–250 ms (match it to your tempo), and set the threshold for 4–8 dB of ducking. Want kick and bass to simply share space with no audible pump? 2–3 dB does it.
Parallel compression (aka New York compression) mixes a brutally squashed copy against the untouched original. Route the drums to a bus, crush that bus (fast attack, fast release, 10+ dB of reduction), then ease the fader up beneath the dry drums until they feel thick but still snap. Sustain and weight arrive, transients survive, because the original track never lost them.
Multiband compression divides the signal into frequency bands and treats each one on its own. Reach for it when the dynamics problem lives in only part of the spectrum: low end that wobbles between bass notes, or a vocal spiking harshly at 5–8 kHz (that particular job has a name: de-essing). Compress the misbehaving band, leave everything else be. When the entire track needs taming, a normal compressor is the answer; slapping multiband everywhere is the express lane to a dead-sounding mix.
Common mistakes I still hear all the time
- Compressing to fix a level problem. A part that’s simply too loud in the chorus needs a fader move or automation. Compression handles moment-to-moment dynamics, not the balance between song sections.
- Fast attack on everything. Murder every transient and you’re left with mush. Drums that lost their punch after compression almost always got hit with an attack that was too fast.
- Judging with the compressed version louder. Match the levels first, decide second.
- Stacking compressors on already-compressed sounds. Modern packs and presets mostly arrive processed and mix-ready. Loops out of something like Drums Utopia have no use for another 6 dB of squashing; a whisper of bus glue covers it.
- No gain staging. Slam the compressor input too hot and every threshold behaves nothing like the preset designer planned. Feed it around -18 dBFS average.
FAQ: audio compression
What do the attack and release knobs actually do?
Attack determines how fast the level gets turned down once the signal crosses the threshold; release determines how fast it comes back afterwards. A slow attack preserves punch, a fast one rounds the sound off. A fast release injects energy (and maybe pumping), a slow one keeps things transparent.
How much gain reduction should I aim for?
Natural-sounding territory is 2–6 dB on the loudest moments of individual tracks, 1–2 dB on the mix bus. Going harder isn’t automatically wrong — parallel drum crushing happily lives beyond 10 dB — but push a lead vocal or a bus past 6 dB and the compressor usually becomes audible.
Should I compress every track?
No. Reach for it when there’s an actual dynamics problem: a vocal jumping around, a bass with lumpy note levels, drums missing density. A steady synth pad or an already-processed loop frequently needs zero. Left alone, a track always sounds more alive than one compressed for no reason.
What’s the difference between a compressor and a limiter?
A limiter is just a compressor running a very high ratio (10:1 and beyond) with an almost instant attack. Think of it as a ceiling: nothing gets past the level you set, which is why it sits at the very end of the master chain. Compressors shape dynamics; limiters cap them.
Where to go from here
Take one song you’re currently mixing and drill a single technique per session: gentle vocal compression tonight, the sidechain pump tomorrow, parallel drums the day after. An evening spent turning knobs while watching the gain reduction meter beats a week of reading about it.
Need material to practice on? The 250+ free Serum 2 presets hand you basses, keys, pads and 808s to compress, sidechain and crush for free. Push them through the settings in this post and every knob’s job becomes obvious to your ears.
I run Monosounds.studio and make the packs myself. Ten-plus years of producing, 97+ releases, 16,000+ downloads. Every Serum 2 preset we ship passes a CPU check on a mid-range laptop and gets used in a real track before release. If it does not survive a mix, it does not go in the pack.
More about Maxim & the studio →
