Compression

The most essential dynamics tool in audio — what it does, how it works, and when to use it.

dynamicscompressionmixingmasteringvocals

Compression is the most fundamental dynamics processor in audio. It's also the most misunderstood. Half the tutorials online will have you slapping a compressor on every track with no clear reason — and the other half will tell you compression is some dark art that takes decades to master.

Neither is true. Here's what you actually need to know.

Specs

Specifications

TypeDynamics Processor

FunctionReduces dynamic range

Key ControlsThreshold, Ratio, Attack, Release, Knee, Makeup Gain

Common TypesVCA, FET, Optical, Variable-Mu, Digital

Use CasesVocals, drums, bass, mix bus, mastering

What Compression Actually Does

The simple version: compression makes the loud bits quieter. That's it. The quiet bits stay roughly the same, but the peaks get pulled down, so the overall difference between the loudest and quietest moments — the dynamic range — gets smaller. You then turn the whole signal back up with makeup gain, which has the effect of making the quiet bits louder relative to where they were.

Now the technical version.

Threshold sets the level at which the compressor starts working. Anything below the threshold passes through untouched. Anything above it gets reduced.

Ratio determines how much reduction happens. A 4:1 ratio means that for every 4 dB the signal goes above the threshold, only 1 dB comes out. At extreme ratios (10:1 and above), you're effectively limiting — the signal barely moves past the threshold at all.

Attack controls how quickly the compressor reacts once the signal crosses the threshold. A fast attack clamps down immediately; a slow attack lets the initial transient through before pulling the level down. This is where the musicality lives — get it wrong and you'll kill the punch of a snare or flatten a vocal.

Release controls how quickly the compressor lets go once the signal drops back below the threshold. Too fast and you'll hear the compressor pumping. Too slow and it never recovers before the next hit, resulting in a squashed, lifeless sound.

Knee determines how gradually the compression engages around the threshold. A hard knee switches on abruptly; a soft knee eases into compression over a wider range, which generally sounds more natural on sustained sources like vocals.

Makeup Gain compensates for the level reduction. If you're pulling peaks down by 6 dB, you add 6 dB of makeup gain to bring the overall level back up. This is where people fool themselves — louder always sounds better, so A/B your compression with level-matched bypass or you'll convince yourself everything sounds improved when it's just louder.

The Four Main Compressor Types

VCA (SSL G-Bus, dbx 160)

Voltage-Controlled Amplifier. Fast, precise, and relatively transparent. VCA compressors respond quickly and predictably, making them excellent for technical work where you need to control dynamics without colouring the sound. The SSL G-Bus compressor on a mix bus is one of the defining sounds of modern mixing — a touch of glue that holds everything together without drawing attention to itself.

FET (1176, Distressor)

Field-Effect Transistor. Fast and aggressive with character to spare. FET compressors can react faster than any other analogue type, and they add a particular harmonic distortion that sounds brilliant on drums, vocals, and electric guitars. The Universal Audio 1176 is the classic — crank the input and pull the output down for that driven, in-your-face vocal sound that's been on countless records. All-buttons-in mode on an 1176 is essentially controlled destruction, and it sounds magnificent on the right source.

Optical (LA-2A, CL 1B)

Optical compressors use a light source and a photocell to control gain reduction. The response is inherently program-dependent — it reacts differently depending on the material you feed it. This makes optical compression feel musical in a way that's hard to achieve with other types. The LA-2A is the textbook example: two knobs, virtually impossible to make sound bad on vocals. The Tube-Tech CL 1B offers more control while retaining that smooth, unhurried character.

Variable-Mu (Fairchild 670, Manley Vari-Mu)

The oldest compressor topology, using valve gain stages where the bias voltage controls the amount of compression. Variable-Mu compressors are gentle, warm, and excel at the subtle, broadstroke work — mix bus glue and mastering. They don't grab transients aggressively; they breathe with the music. The Fairchild 670 is legendary (and legendarily expensive — original units fetch six figures). The Manley Vari-Mu is the modern standard for this topology.

When to Use Compression

Vocals. Almost always. The human voice has enormous dynamic range, and a performance that sounds perfect in the room can disappear behind the instrumentation in a mix. Gentle compression (3–6 dB of gain reduction) evens things out so the vocal sits consistently in the mix without riding the fader constantly. Many engineers use two stages — a fast compressor to catch peaks, then an optical compressor for gentle levelling.

Drums. Compression on drums is about shaping the envelope as much as controlling dynamics. A slow attack lets the transient through and then pulls down the sustain, adding punch. A fast attack does the opposite — taming the crack and bringing up the body and room sound. Parallel compression (blending a heavily compressed signal with the dry original) is standard practice on drum buses.

Bass. Consistency. A bass guitar or synth bass needs to sit solidly in the low end without individual notes jumping out or disappearing. Moderate ratio, moderate attack, and you're most of the way there.

Mix bus. Subtle compression on the stereo bus — 1–2 dB of gain reduction at most — adds cohesion. It makes separate tracks feel like they belong together. The SSL G-Bus sound is the reference point here, but plenty of engineers reach for a Fairchild or Manley emulation instead.

Mastering. Even gentler than mix bus work. Mastering compression is about density and loudness without audible compression artefacts. Typically broadband with very low ratios (1.5:1 or 2:1) and careful attack and release settings matched to the tempo of the track.

Common Mistakes

Over-compressing. The most common error by far. If you can hear the compressor working on a vocal, you've probably gone too far. Aim for transparent control, not an obvious effect (unless that's specifically what you want).

Wrong attack time killing transients. A fast attack on a snare drum will obliterate the crack that makes it cut through the mix. A fast attack on a vocal will make it sound flat and lifeless. Start with a medium attack and adjust from there — let the transient through unless you have a specific reason to catch it.

Not using your ears. Staring at the gain reduction meter and targeting a specific number is a recipe for mediocrity. Close your eyes. Listen. Does it sound better or worse? If you can't tell, you probably don't need the compressor.

Compressing because you think you should. Not every track needs compression. If the dynamics are fine as they are, leave them alone. Reaching for a compressor out of habit rather than necessity is how you end up with flat, lifeless mixes where nothing breathes.

Forgetting to level-match. Louder sounds better. Always. If you add 4 dB of makeup gain after compressing, you haven't improved the sound — you've just made it louder. Match the output level to the input level before deciding whether the compression is actually helping.