If you want to learn how to compress a vocal, the short version is this: compression narrows the gap between the quietest and loudest words so the voice stays steady and present instead of jumping around the mix. You do it by setting a threshold that catches only the louder phrases, a moderate ratio, and an attack and release timed to the singer’s delivery — aiming for a few dB of gain reduction, then matching the level back so you can compare fairly. Everything below breaks those controls down, with starting points you should always fine-tune by ear.
Violet Recording is reader-supported — we may earn a commission from links on this page, at no extra cost to you.
What a compressor actually does to a vocal
A raw vocal has a wide dynamic range: whispered verse lines sit far below belted chorus phrases, and even within a single sentence the level swings word to word. That variation is natural when someone is in the room with you, but through speakers it means half the words disappear and the loud ones stab. A compressor turns down the parts that cross a level you set, and leaves everything below it alone. The result is a smaller, steadier range that sits reliably in the track.
Every compressor gives you the same core controls — threshold, ratio, attack, release and make-up gain — so once you understand them, the plugin brand barely matters. Your DAW’s stock compressor is more than capable of a professional vocal. Get the controls right before you go shopping for anything fancier.
Set the threshold and ratio to control the swings
The threshold decides where compression kicks in. Lower it and more of the vocal gets caught; raise it and only the loudest peaks are touched. A good way to dial it in is to pull the threshold down while the loudest chorus phrase plays until the meter shows the compressor working on those peaks, then back off slightly so the quiet verse words pass through untouched.
The ratio sets how firmly the level is reduced once it crosses the threshold. For most vocals, a ratio between 2:1 and 4:1 is a sensible starting point — gentle enough to sound natural, firm enough to matter. A 2:1 ratio halves anything over the threshold; 4:1 clamps harder for an aggressive, in-your-face pop or rap vocal. Save higher ratios like 8:1 and above for genuine peak-catching rather than everyday levelling. Watch the gain-reduction meter and aim for roughly 3–6 dB of reduction on the loudest phrases. If it is pinning at 10 dB or more, your threshold is too low or you are asking one compressor to do too much.
Dial in attack and release for a natural sound
Attack and release are where a vocal compressor sounds either invisible or obvious, so this is the part worth slowing down on.
- Attack is how quickly the compressor clamps down once the signal crosses the threshold. A fast attack grabs the initial consonant and can dull the vocal; a slower attack — think tens of milliseconds — lets the transient of each word snap through before the level is reined in, which keeps the vocal crisp and articulate. Start medium and adjust: too soft and lispy means slow the attack down.
- Release is how fast the compressor lets go afterwards. You want it to recover in the gaps between words rather than clamping through them. Too fast and it pumps or distorts; too slow and it never resets, so the vocal sounds squashed and lifeless. Many engineers set release by ear until the level movement breathes in time with the phrasing. An auto-release setting is a fine safety net if you are unsure.
The interplay between these two is the whole game. If a compressed vocal feels dull, lengthen the attack. If it feels jittery or pumps, lengthen the release. Small changes make an audible difference.
Putting it together: how to compress a vocal in a clean chain
Order matters, and a reliable chain looks like this: subtractive EQ first, then compression, then any tone-shaping EQ, de-essing and effects. Cutting low-end rumble and boxy mud before the compressor means the compressor reacts to the voice itself rather than to boomy energy it shouldn’t be chasing. Roll off the rumble below roughly 80–100 Hz with a high-pass filter and clean up any 200–500 Hz mud, then feed that tidier signal into the compressor.
One of the most useful pro techniques is serial compression — two compressors in series, each doing a gentle few dB of gain reduction, instead of one plugin doing 8 dB on its own. Two soft passes sound far more natural than one hard squeeze, and they give you a steadier, more controlled vocal without the artefacts. Set up a modest compressor to catch the broad swings, then a second one after it to tuck in whatever peaks remain. If EQ and compression order is new to you, our guide to EQ and compression fundamentals explains the reasoning behind this chain, and the how to mix vocals walkthrough puts it in context of the full vocal workflow.
Match make-up gain and A/B honestly
Compression lowers the loud parts, so the overall vocal ends up quieter — which tricks your ears into thinking the compressed version sounds worse when really it is just softer. Use the make-up gain (or output) control to bring the compressed vocal back up to roughly the same loudness as the uncompressed signal, then bypass and engage the plugin at matched volume. Only then can you judge whether the compression is actually helping.
Louder almost always sounds “better” for a moment, so this level-matched comparison is the single most important habit in all of compression. If the engaged version sounds steadier and more present at the same loudness, keep it. If it sounds duller or more squashed, back off the ratio or raise the threshold. For checking how your vocal level fits the finished track’s loudness, our LUFS explainer covers streaming targets, and the free tools include a quick LUFS and decibel reference.
Common mistakes and when a plugin upgrade is worth it
The most common error is over-compression: chasing so much gain reduction that the vocal loses all its dynamics and sounds flat, breathy and fatiguing. If you can hear the compressor working, you have probably gone too far. A close second is relying on the compressor to fix an unstable level that automation should handle — ride the fader for big section-to-section changes first, then let the compressor smooth the fine detail. A third is a too-fast attack that swallows consonants and leaves the vocal soft and indistinct.
Also mind sibilance: heavy compression and a presence boost can push “s” sounds into harsh territory, so a de-esser after the compressor often earns its place. Your DAW’s stock de-esser and compressor handle the vast majority of vocals with no help. That said, some third-party compressors — vintage-modelled optical or FET emulations, for instance — add a character and glue to vocals that stock plugins don’t quite reach, and that is one place the spend can be justified. Treat it as a flavour upgrade, not a fix; the technique above matters far more than the brand. For the surrounding decisions, see our notes on using reverb and delay and what mastering does once the mix is done, plus the full mixing and mastering hub.
Frequently asked questions
What compressor settings should I start with on a vocal?
A solid starting point for most vocals is a ratio around 3:1, a threshold set so you see roughly 3–6 dB of gain reduction on the loudest phrases, a medium attack that lets consonants through, and a release that recovers between words (or auto-release if you are unsure). Then make up the lost gain and compare against bypass at matched loudness. These are starting points, not rules — adjust every setting by ear to the singer and the song.
Should I compress a vocal before or after EQ?
A reliable order is a subtractive EQ first to remove rumble and low-mid mud, then compression, then any tone-shaping EQ or de-essing after. Cleaning up boomy low-end before the compressor means it reacts to the voice rather than to energy you were going to cut anyway. It is not an absolute law — if a bright tone-shaping boost is feeding harshness into the compressor, move that boost after it instead. Trust what you hear.
How much gain reduction is too much on a vocal?
There is no hard limit, but if you are consistently pushing more than about 6 dB of gain reduction with a single compressor, the vocal will usually start to sound squashed and lifeless. A better approach for heavy control is serial compression — two compressors each doing a gentle few dB — which sounds more natural than one plugin doing all the work. If the vocal loses its energy and dynamics, ease off; steady and alive beats flat and loud every time.
