Silence Cut Tool
Find the silent parts of a video or audio file automatically and cut them all at once, to tighten meeting recordings, voice memos and commentary.
Equivalent FFmpeg command
CLI command for your current settings. Updates live as you change controls.
ffmpeg -i input.mp4 -af silencedetect=noise=-40dB:d=0.5 -f null - What to do next
How to use
- Drop the fileDrag & drop a video or audio file, up to 500 MB.
- Pick threshold and minimum lengthDefault is -40 dB / 0.5 sec. Noisy recordings: -30 dB. Quiet studio takes: -50 dB.
- Set the padding0.15 sec (the default) keeps word endings from being clipped. Padding is used for video only.
- Run and downloadClick Cut the silence. When it finishes, a preview and the reduction appear.
Recommended settings
When it goes wrong
"No silent intervals were detected" message
Cause: BGM, room tone, or hiss sits above the threshold (e.g. -40 dB).
Fix: Raise the threshold to -30 dB. If still nothing, the recording has no silent parts.
Word endings get clipped
Cause: With zero padding, the start and fade-out of each word are cut too.
Fix: Set padding to 0.15–0.3 sec, or increase the minimum silence length.
"Too many silent intervals" message
Cause: The minimum silence length is so short that breath-length pauses are found too, more than 80 of them.
Fix: Raise the minimum silence length to 1 sec or more, or lower the threshold (for example -40 dB → -50 dB).
Video takes a long time to process
Cause: Video runs twice: once to find the silence, once to re-encode.
Fix: For long videos, cut out the part you need first, or use desktop FFmpeg.
Questions
Does video quality drop?
A little. Video mode rebuilds the video without the silent parts, so it re-encodes the picture with H.264 (CRF 23) and the sound with AAC at 128 kbps. Audio files come out as M4A (AAC 192 kbps).
How do I choose the threshold?
Start with the default -40 dB. If nothing is detected, go to -30 dB; if speech gets cut, go to -50 dB. Aim a little above the background noise in the silent parts.
Can I use audio-only files?
Yes. MP3, WAV, M4A and other audio files are processed in one pass with the silenceremove filter and saved as M4A (AAC 192 kbps). No padding is added: every silence longer than the minimum length is cut in full, at the start, in the middle and at the end.
The result feels choppy / over-cut
Increase padding to 0.3 sec and the minimum silence length to 1 sec or more. Lectures and interviews sound more natural with longer settings. If much of the recording is unwanted, you can also rough-cut it first with the Trim tool.
Everything gets flagged as silence because the recording is quiet
Normalize loudness first with the Volume tool, then run this tool; silence detection becomes more stable.
More detail (settings, limits, FFmpeg commands)
What each setting means
Browsers and limits
- Maximum file size: 500 MB
- Processing stops if more than 80 silent intervals are detected (raise the minimum silence length)
- Video comes out as MP4 (H.264 / AAC) and audio as M4A (AAC); neither is lossless
- Padding applies in video mode only
The same job with the FFmpeg command line
ffmpeg -i input.mp4 -af silencedetect=noise=-40dB:d=0.5 -f null - ffmpeg -i input.mp4 -vf "select='not(between(t,5.2,8.1)+between(t,20.3,24.0))',setpts=N/FRAME_RATE/TB,scale=trunc(iw/2)*2:trunc(ih/2)*2" -af "aselect='not(between(t,5.2,8.1)+between(t,20.3,24.0))',asetpts=N/SR/TB" -c:v libx264 -preset fast -crf 23 -pix_fmt yuv420p -c:a aac -b:a 128k -movflags +faststart output.mp4 ffmpeg -i input.mp3 -vn -af silenceremove=stop_periods=-1:stop_duration=0.5:stop_threshold=-40dB -c:a aac -b:a 192k output.m4a