Audio Silence Remover

How to use the audio silence remover

1. Choose your recording.
2. Set a threshold — start at −40 dB and raise it if nothing is found.
3. Set a minimum gap of at least half a second, then remove and download.

About this tool

Finds sections quieter than a threshold and removes them. The obvious use is tightening a recording — an interview with long pauses, a screen recording with dead air while something loads, or a voice memo with silence at both ends.

Three controls, and each exists because the naive version fails without it.

The threshold is in decibels rather than as a raw amplitude, because that is how quietness is actually judged. −40 dB is a reasonable starting point for a clean recording; a noisy one may need −30 dB or higher, since room tone and hiss are never truly silent and a strict threshold will find nothing at all.

The minimum gap stops the tool cutting between words. Ordinary speech contains hundreds of brief silences, and removing them produces something that sounds like a badly spliced tape. Half a second is a sensible floor.

The keep-either-side padding leaves a fraction of each silence in place. Cutting exactly at the threshold clips the natural decay at the end of a word and creates an audible click — precisely the artefact the fade tool exists to remove.

Level is measured over short windows rather than sample by sample, because every waveform crosses zero constantly and a per-sample test would report silence thousands of times a second in the middle of a loud passage.

The result is exported as WAV, because browsers can decode MP3, M4A, OGG and FLAC but can only encode WAV. That means the output is uncompressed and several times larger than a compressed input — a real trade-off, and the alternative would be bundling an encoder several times the size of this entire site.

Common uses

• Tightening an interview with long pauses.
• Cutting dead air from a screen recording.
• Trimming silence from the start and end of a voice memo.

Frequently asked questions

Nothing was found. What should I change?

Raise the threshold towards zero. A noisy recording is never truly silent, so −40 dB may be stricter than any part of it reaches.

Why is my speech being chopped up?

The minimum gap is too short. Ordinary speech contains hundreds of brief silences, so a floor of half a second or more keeps words intact.

Why keep part of each silence?

Cutting exactly at the threshold clips the decay at the end of a word and produces a click. A small pad keeps the result sounding natural.