Find the dead air in talking-head videos, podcasts and screen recordings. Tune the threshold, see every silent gap on a waveform and export a cut list.
Decoding audio...
| # | Start | End | Length |
|---|
Runs entirely in your browser. Nothing is uploaded to any server.
Removing long pauses makes tutorials, vlogs and podcasts feel faster and more engaging.
The waveform shows exactly which parts count as silent before you touch your editor.
Use the timestamps in any editor, or run the ffmpeg command for an automatic rough cut.
The audio is decoded with the Web Audio API and split into 20-millisecond windows. For each window the tool calculates the RMS level across all channels and converts it to decibels relative to full scale. Any run of windows quieter than your threshold, lasting at least the minimum duration, is marked as silence.
Padding trims each silent region inward so a little natural breathing room stays around speech, which avoids choppy, clipped words. A threshold around -40 dB works for most voice recordings; raise it toward -30 dB for noisy rooms or lower it toward -50 dB for very clean studio audio.
Everything quieter than this level counts as silence.
Short gaps between words are ignored so speech is not chopped up.
The command trims and concatenates every non-silent part in a single pass.
No. The file is read and decoded entirely in your browser using the Web Audio API.
Your browser must support the audio codec inside the file. AAC in MP4 and Opus or Vorbis in WebM work in most browsers; some MOV or MKV files may not.
Start at -40 dB. If background noise is being kept, raise the threshold; if quiet speech is being marked as silence, lower it.
Yes. Trimming at exact times requires re-encoding, which ffmpeg does automatically with its default codecs for the output extension.
There is no fixed limit, but the whole audio track is decoded into memory, so very long files may be slow or fail on low-memory devices.
More free tools from My Panda Toolbox.