Cut the silent stretches out of a talking video with ffmpeg on Windows without losing sync:…

Guide ID: art_01M77W4AD3ECV1RDH31C3PSRRX

Editorial publication checks are separate from community and evidence review. The evidence labels retain this revision’s reported inspection, execution and review status.

Guide title

Cut the silent stretches out of a talking video with ffmpeg on Windows without losing sync: detect them, cut picture and sound together, then check the result

Coverage

For people with a recorded talk, lesson or screen recording who want the dead air gone without the picture and the sound drifting apart. You get a Windows PowerShell script that lists how many silences ffmpeg finds at several loudness limits, cuts them from picture and sound together while keeping a short pad of quiet, and reports a pass or check verdict from a second look at the result. It ran only on generated clips with steady noise, so real speech, long files and other rooms are untested.

Example

On a generated clip, ffmpeg's audio-only silence filter shortened the sound to 23.4 s while the picture stayed at 30 s; the script's cut kept picture and sound changing together to within a thirtieth of a second.

Useful starting points

• Measure your room noise first: in the test a loudness limit below the noise peak found no silences at all.
• Dry-run before writing a file and read the list of cuts and the limit ladder.
• Keep a short pad of quiet at each cut and listen afterwards, because the checks cannot tell you whether a word was clipped.