Silence Removal / Compaction Preview
Preview how much threshold-qualified silence would be shortened, then optionally render a new PCM WAV while retaining a bounded amount of each pause.
Choose a browser-decodable audio file.
What this tool does
It first shows a non-destructive preview of each sustained below-threshold pause, how much would be retained, and the estimated output duration. Rendering removes only the excess portion of those detected pauses.
How to use it safely
Start with a conservative threshold such as −45 to −55 dBFS and keep 100–300 ms of each pause. Audition the result before publishing. Aggressive settings can make speech unnatural or remove quiet wanted audio.
Processing method
Build 120 uses hard decoded-sample splices without automatic crossfades. It does not claim breath-aware editing, speech understanding, podcast mastering, room-tone synthesis, or artifact-free silence removal.
Privacy
The selected file is decoded and processed locally in the browser by this Build 120 tool; it is not intentionally uploaded to MesraTools.
Limitations
Silence and segment boundaries are threshold-derived editing evidence, not semantic chapters or speaker turns. Browser memory and codec support vary by device. Generated WAV/TAR/text files do not preserve every source metadata field.
Related tools
Use Silence & Active Segment Detector to inspect boundaries, Detected Segment Exporter for one range, Audio Pause Statistics for descriptive pause evidence, and Audio Loop & Cue Builder for manual looping.