GoScript
Back to Blog
GoScript Journal

Voiceover Compression: Level Control Without the Pump

·7 min read·GoScript Team /Design

Compression lowers the loudest moments so the quiet ones can be raised later. On a voiceover it keeps words at an even level, which lets the listener relax instead of chasing the volume. Set it well and nobody hears it. Set it badly and the track breathes in a way that is impossible to unhear.

The two knobs people get wrong are attack and release. They control how fast the compressor grabs and lets go, and together they decide whether the voice sounds natural or seasick.

Ratio and threshold decide the feel

A ratio near 2 to 1 or 3 to 1 is plenty for spoken word. The threshold sets where compression begins, and for a voice you want it low enough to catch the loud syllables without clamping down on every breath.

Watch the gain reduction meter and aim for peaks of three to five decibels on the hottest lines. If it never moves, the compressor is asleep. If it never rests, you are crushing the life out of the read.

Compression should make the voice easier to follow, not louder. The moment you hear it working, you have probably gone too far.

Attack fast, release slower

A quick attack catches the start of a loud word so it does not spike, while a slower release keeps the compressor from pumping between syllables. Fast attack plus fast release is the recipe for that gasping, unnatural sound.

I set release long enough that the compressor holds through a phrase rather than loosening between every word. If the level rises and falls in a visible rhythm, ease the release until the movement smooths out.

Stack two gentle stages

One heavy compressor does damage. Two light ones, each doing a couple of decibels, stay transparent while still controlling the level. The first handles broad movement, and the second catches the stray peaks that slip past.

After compression, apply makeup gain so the track returns to the level you started with before you compress. Keep the gain reduction modest and the whole chain sums to a voice that sounds even and untouched.

Key takeaways

  • Use a low ratio and let the threshold do the heavy lifting.
  • Set a fast attack with a slower release to avoid pumping between words.
  • Split control across two gentle stages instead of one aggressive one.
  • Add makeup gain so the compressed level matches the original, then compare.

Turn this into your next script.

Analyze a channel, extract its structural DNA, and generate scripts engineered from patterns that actually work.

Analyze a Channel