How to set a compressor: ratio, attack, release and makeup gain

0

A compressor does one thing: it reduces the gap between the loud and the quiet parts of a signal. Everything else follows from that. Yet it is the tool that locks up the most mixes, because its controls interact and we tend to set them staring at a meter instead of listening. Let’s take each parameter in turn, then see how to combine them by source.

What a compressor actually does

Compressing means controlling dynamics. The compressor watches the incoming level and, the moment it crosses a threshold, turns it down by an amount you choose. We reach for it for three distinct reasons that are worth keeping separate: taming peaks that jump out (an uneven bass, a vocal that spikes on certain syllables), making a source denser so it holds its place, or gluing a group of elements together on a bus. Those three jobs do not call for the same settings.

Before you even touch the compressor, make sure your levels are sane: good gain staging upstream keeps you from fixing a simple level problem with compression.

The controls, one by one

Threshold

This is the level at which compression kicks in. Above it, the signal is turned down; below it, it passes untouched. Lowering the threshold compresses more of the signal. Set it by watching how much gain reduction you get, not by aiming at an absolute number.

Ratio

The ratio says how hard the signal is turned down once past the threshold. At 2:1, two decibels over the threshold come out as one; at 4:1, four become one; past 10:1 you are limiting. A gentle ratio (1.5:1 to 3:1) adds density; a high ratio reins in stubborn peaks.

Attack

Attack sets how quickly the compressor reacts once the threshold is crossed. A fast attack grabs the transient (the initial crack of a snare, the pluck of a string) and rounds it off; a slow attack lets that transient through before acting, which preserves punch. It is the control that changes a sound’s character the most.

Release

Release sets how long the compressor takes to stop acting once the signal drops back below the threshold. Too short and it breathes between notes and can pump; too long and it stays clamped and smothers the dynamics. It is often timed to the tempo and phrasing of the track.

Knee

The knee describes how compression sets in around the threshold. A hard knee applies the ratio abruptly the instant the threshold is crossed: clean, audible, useful for limiting. A soft knee eases compression in on either side of the threshold: more discreet, ideal on a voice.

Makeup gain

Compressing lowers the overall level. Makeup gain brings it back up to the starting volume. Beware the classic ear trap: louder always seems better. Set makeup so the compressed and uncompressed sounds match in perceived loudness, otherwise you are judging a volume boost, not the quality of the compression.

A Teletronix LA-2A optical compressor mounted in a studio rack
Credit: Teletronix

Setting it by source

There is no universal preset, but there are reliable starting points:

  • Vocals: 2:1 to 4:1, soft knee, medium attack (10-30 ms) to keep the consonants, medium release. Aim for 3 to 6 dB of reduction on peaks. If sibilance jumps out, that is not the compressor’s job: reach for a de-esser.
  • Bass: 3:1 to 6:1, slower attack to let the finger attack through, short release for density. Bass is often the source that gains the most from being held.
  • Drums and buses: slow attack to preserve the crack, release timed to the tempo so it breathes. If you want punch and density, think parallel compression rather than crushing the track.
  • Mix bus: gentle ratio (1.5:1 to 2:1), 1 to 2 dB of reduction, soft knee. You are after cohesion, not an effect.

If it is a sound’s attack you really want to shape, without touching the volume, a transient shaper will often do a more surgical job than a compressor.

Common traps

Three mistakes come up again and again. First: too much ratio, too soon, which flattens the life out of a sound. Second: an attack so fast it eats the transients and leaves everything limp. Third: trusting the meter over your ears. Set it, then A/B at matched perceived loudness: if the compressed sound loses its depth, back off. And remember that compression never replaces a good balance: often a touch of EQ fixes what looked like a dynamics problem.

My take

A compressor rewards restraint. Two or three well-placed decibels do more for a mix than twelve that flatten everything. Set attack and release by ear, listening to what happens between the notes rather than watching the needle. A good setting shows itself most when you bypass it: if the sound loses its life, you went too far. Start gentle, push until it serves the track, then back off one notch. It is almost always the last notch you added that was the one too many.

Share.

About Author

After 20+ years in professional audio: live sound engineering, studio technical direction (Deep Forest, Pierre Jacquot), head of digital marketing at Playback.fr. A first-hand witness to the analog-to-digital shift, I track the whole audio landscape and break it down here — no fluff.