Rough cutFree · no account · nothing is uploaded

Silence detector (preview)

Drop a file and see where the gaps are and how many minutes you could get back. The file is decoded by your browser, on your machine: nothing is uploaded anywhere.

Drop an audio or video file here, or click to choose one (300 MB maximum)

Silences
0
Recoverable time
—
File duration
—

The threshold is a level: −40 dB catches real gaps, −30 dB also catches breaths, −55 dB only catches absolute silence. The padding leaves a breath before and after each cut.

How the detection works

Your browser decodes the file (Web Audio), then it's analysed in 20-millisecond windows: for each one we compute the average level (the RMS value) and compare it to your threshold. Consecutive windows below the threshold form a silence; only those longer than the minimum are kept, and the padding is trimmed from both ends so you don't clip the end of a word.

Only the left channel is read: enough to spot gaps, and it keeps memory use sane on a long file. Past 300 MB the tool refuses — export the audio track as MP3 and drop that.

What this preview does not do

It doesn't cut anything. It gives you a list of timecodes and one number: the time you'd save. Cutting for real means being in the editing software, on the timeline, with the transitions and the audio following along. That's the plugin's job.

Frequently asked questions

Does my file go to a server?

No, and you can check: cut your connection after the page has loaded, the analysis still runs. Decoding happens through your browser's Web Audio API, on your machine.

Why a 300 MB limit?

Because a full audio decode takes several times the file size in RAM, and a tab that goes over gets killed by the browser. For a long interview, export the audio track as MP3 (tens of megabytes for an hour) and drop that.

My ProRes MOV doesn't work.

Browsers don't decode ProRes, MXF, or most MKV files. H.264 MP4, M4A, MP3, WAV and AAC do work. A quick MP3 export of the audio track solves it every time.

Which threshold should I use?

Start at −40 dB. If it misses obvious gaps, move up towards −35 or −30 (it gets more aggressive and starts catching breaths). If it cuts mid-sentence, move down towards −50.

What is the padding for?

To avoid cutting flush against the word. 60 ms on each side keeps the cuts breathable; at 0 the edit feels choppy.

Can I import these markers into Premiere?

The export is a tab-separated text file (start, end, name, length) plus a CSV, readable by a spreadsheet or an import script. Premiere doesn't import marker lists natively — which is exactly what the plugin does instead, placing the cuts straight on the timeline.