Audizer
Vocals

Level the vocal before you compress it

Before any compressor, adjust the gain of each phrase so the whole take sits within about 3 to 5 dB. Raise the quiet lines, lower the loud ones, and limit any single move to around 6 dB. The compressor then takes 3 to 5 dB off every phrase evenly.

The problem with compressing first

A compressor has one threshold. A verse sung at −24 dBFS and a chorus belted at −10 dBFS can't both be right for it. Set the threshold for the chorus and the verse passes through untouched and gets lost. Set it for the verse and the chorus is crushed by 12 dB, with pumping and a dull tone.

Engineers solved this decades ago by riding the fader into the compressor. In a DAW the equivalent is clip gain: split the take into phrases and set each one's level.

How to do it by hand

  1. Split the vocal at the gaps between phrases.
  2. Find the level the take sits at most of the time. That's your reference.
  3. Go phrase by phrase. If a line is 6 dB under the reference, raise it 4. If one is 5 dB over, drop it 3 or 4.
  4. Inside a phrase, fix single words that jump out or disappear. Split around the word and adjust it 2 to 4 dB.
  5. Don't flatten everything. Moving each phrase about 70% of the way to the reference keeps the song's shape.
  6. Now set the compressor. With an even input, 3:1 with 3 to 5 dB of gain reduction will hold on every line.

Numbers to aim for

MeasureRaw take, typicalAfter levelling
Spread between quietest and loudest phrase8 to 14 dB3 to 5 dB
Largest single gain moveabout 6 dB
Compressor gain reduction afterwardsswings from 0 to 12 dBsteady 3 to 5 dB

In our own test mix, the lead vocal's phrase levels were spread across 10.8 dB as recorded. After phrase levelling the spread was 4.3 dB.

Why not use two compressors instead

Serial compression helps and is worth doing. It still can't tell a quiet, intimate line from a line that was sung off-mic. Clip gain is a decision about each phrase. Compression is a rule applied to all of them. You want the decision made first, then the rule.

Quick answers

What is clip gain?
A gain setting on the audio clip itself, applied before the track's plugins. Changing it changes what the compressor receives.
How even should a vocal be before compression?
Phrases within about 3 to 5 dB of each other is a good target. Tighter than that starts to remove the dynamics of the performance.

Audizer does this when you ask.A Mac DAW with an engineer built in. Say what you want and it makes the edit. Free download.

Get Audizer

Updated 2026-11-15. Corrections: support@intuzu.com