Compression basics for dialogue: smoother voice levels in any editor

Dialogue is naturally dynamic. Some words are loud, some are quiet, and the distance between them can be too big for a video mix. Compression reduces the distance between the loud and soft parts. When used correctly, it makes dialogue easier to listen to without sounding crushed.

What compression does

A compressor lowers the level of any signal that goes above a set threshold. The ratio controls how much it lowers. A 3:1 ratio means that for every 3 dB the signal goes above the threshold, only 1 dB passes through. This makes the loud parts quieter, which lets you turn the whole clip up and brings the quiet parts forward.

Set the threshold around the average level

The threshold is the level where compression starts. For dialogue, set it near the average level of the speech, not the peaks. If the threshold is too high, nothing happens. If it is too low, the voice sounds squashed. The goal is to see gain reduction on loud words but not constant activity on every sound.

Use a gentle ratio for dialogue

Start with a ratio between 2:1 and 4:1. This is gentle enough to keep the voice natural while still controlling peaks. Higher ratios like 8:1 or more are closer to limiting and can sound aggressive unless you need a very controlled broadcast sound.

Attack and release shape the response

Attack is how fast the compressor reacts. A medium attack of 5 to 15 milliseconds lets the start of consonants through before compressing. Release is how fast it stops compressing. A release of 40 to 100 milliseconds works well for speech. If the release is too fast, you hear pumping. If it is too slow, the voice stays flat.

Add make-up gain last

After compression, the overall level is lower. Use make-up gain to bring the average level back to where it needs to be. A typical target is around -12 dBFS average. Always listen to the compressed signal against the rest of the mix.