Shorten or cut out the pauses in a recording automatically, and trim silent ends
Processed in your browser. Nothing is uploaded. Checked against sources on 16 Oct 2026.Remove Silence from Audio finds the pauses in a recording in your browser and shortens them to a quarter of a second, or cuts them out, and trims silent ends. By default a pause is anything below -45 dBFS for over 0.5 seconds. Nothing is uploaded and the result downloads as a WAV.
Press Choose an audio file or drop the recording on the page; it is decoded and measured on your device.
Read how many pauses were found and how much shorter the recording becomes.
Adjust Silence is quieter than, Pauses longer than and Leave of each pause if needed.
Press Remove silence, compare the before and after players, then press Download.
Lectures, interviews, voice notes and podcast takes are often a third silence: thinking time, page turns, a pause before every answer. Cutting each one by hand in an editor takes longer than the recording.
How pauses are found. The audio is mixed to mono and measured in 10 millisecond windows. A window is silent when its RMS level is below the threshold, given in dBFS, decibels below digital full scale. A run of silent windows longer than the minimum is a pause; anything shorter, such as the gap between two words, is never touched.
| Setting | Default | What it does |
|---|---|---|
| Silence is quieter than | -45 dBFS | the level that counts as silence |
| Pauses longer than | 0.5 s | shorter gaps are left alone |
| Leave of each pause | 0.25 s | half kept on each side of the cut |
| Trim start and end | on | silence before the first sound and after the last goes |
A quiet room usually measures around -60 dBFS and speech between -30 and -10, so -45 sits safely between them. The page states how many pauses it found and the new length before anything is cut, and warns when the threshold would remove almost the whole file.
Clean joins. Each cut falls inside silence and is faded over 3 milliseconds, so nothing clicks. Sound in only one stereo channel still counts as sound.
Tested on synthetic recordings of tones and pauses with known lengths and noise levels, checking the new length and that the sound itself is kept. The result is a 16-bit WAV at the original sample rate, and nothing is uploaded.
Drop the recording on Remove Silence from Audio. It measures the file in 10 ms steps, finds every pause longer than half a second below -45 dBFS, and says how much shorter the result will be. Press Remove silence, check the before and after players, and download the WAV.
Start at -45 dBFS, which suits most voice recordings. If a noisy room or a hum stops the pauses being found, move the slider right to about -35. If quiet words at the ends of sentences are being cut, move it left to -55. The page says when almost everything would be removed.
No. Only pauses longer than the Pauses longer than setting, half a second by default, are touched, and each keeps a quarter of a second so speech still breathes. The short gaps inside sentences, usually under 0.3 seconds, stay exactly as they were.
Yes. Choose Nothing under Leave of each pause and every pause is cut out, leaving the sounds back to back; the Remove Pauses page opens that way. Each join is faded over 3 milliseconds so it never clicks, and the silence at the start and end is trimmed too.
Browsers can decode MP3 and M4A but cannot encode MP3, so the edited audio is saved as uncompressed 16-bit WAV at the original sample rate, which plays everywhere. A minute of stereo is about 10 MB. Compress it to MP3 or M4A afterwards if a smaller file is needed.
No. The audio is decoded, measured and edited by your own browser, and the WAV is written on your device. Nothing is sent to a server, so interviews, lectures and voice notes stay private. Files up to 300 MB and three hours long work.
Rates and rules change; check the source before you rely on a figure.
Still stuck? Ask about Remove Silence from Audio and anybody who knows can answer. No account needed.
Same job, other tools: Frequency Generator – Generate a steady test frequency for speakers, headphones or instruments · Sweep Tone Generator – Sweep smoothly between two frequencies over a time you choose · Audio Sweep Generator – Play a slow logarithmic sweep and hear every rattle in the room.
On the related tools: Mic Test – See your voice on a live waveform and hear it played straight back.