14 May 2026 | 7 min read
Audio compression explained: dynamics, not file size
Two things are called compression in audio and they have nothing to do with each other. This one is about controlling loud and quiet, not shrinking files.
File compression makes a WAV into a FLAC or an MP3. Dynamic range compression makes a whispered line and a shouted one sit in the same mix. Same word, different job. This article is about the second one.
The five controls
- Threshold: the level above which the compressor starts working.
- Ratio: how firmly it pushes down what crosses the threshold. 3:1 means three dB in, one dB out.
- Attack: how quickly it reacts. Fast catches transients, slow lets them through.
- Release: how quickly it lets go. Too fast breathes, too slow chokes the next word.
- Makeup gain: what you add back afterwards, because compression only ever turns things down.
A starting point for dialogue
Threshold set so the meter shows three to five dB of reduction on normal speech peaks, ratio around 2.5:1 or 3:1, attack near 10 milliseconds so consonants keep their edge, release around 100 to 150 milliseconds so it recovers between phrases. Add makeup gain until the compressed version matches the bypassed one in loudness, then judge.
The mistake almost everyone makes
Comparing compressed and uncompressed at different levels. Louder always sounds better for the first ten seconds. Match the loudness before you decide whether the compressor helped, or you will keep choosing the setting that is simply louder.
What to fix before reaching for a compressor
- Bad gain staging: fix the levels at source first.
- Distance changes: ride the fader or ask for a better boom operator position.
- Clipping: no compressor undoes a truncated waveform.
- Noise: compression raises the noise floor along with everything else.
Clean levels before the mix make all of this easier. Trim each track numerically in the tools here, then send balanced stems into your DAW.