Split Audio by Silence

How it works

1. Choose an audio file.
2. Set the silence threshold and the shortest gap that should split.
3. Review the detected split points and adjust them.
4. Download the tracks as a ZIP.

How the split points are found

The page measures the level of the audio in 10 millisecond steps. A stretch that stays at or below the silence threshold for at least the shortest silence you set counts as silence, and the file is cut in the middle of it. Silence at the very start or end of the file is ignored, so you never get an empty first or last part.

The detected silences are highlighted on the waveform and the cut points appear in the list, where you can move them, join parts or add your own before exporting.

Settings that matter

Threshold: raise it (closer to 0 dB) if the gaps contain hiss, hum or room noise; lower it if quiet passages are being cut. Shortest silence: for tracks separated by a clear pause, 1 to 2 seconds works; for speech with short breaths use a longer value so only real breaks count. Shortest part: a silence that would leave a part shorter than this is skipped, which stops a pause inside a track from splitting it.

Measuring by Peak reacts to any short click; Average (RMS) ignores isolated clicks and suits recordings from vinyl or tape.

Limits

Splitting at silence works when the tracks really are separated by quiet gaps. Live recordings with applause, crossfaded DJ mixes and tracks that run into each other have no silence to find, so use markers for those. Silence detection is a level measurement, not an understanding of the music, so always check the list of parts.

Questions