Silence Remover
This silence remover lets you load an audio file and automatically cut out silent gaps — tune the threshold, minimum length, and how much padding to leave — then export a lossless WAV, all in your browser.
ℹ This detects silence by level (loudness), not by understanding speech — it’s an energy gate, not voice detection. Tune the threshold to your recording: too high and quiet words get cut, too low and gaps stay. Output is WAV (browsers can’t write MP3/AAC). Everything runs locally — your file is never uploaded.
How It Works
Your file is decoded to raw samples in the browser using the Web Audio API’s AudioContext.decodeAudioData. The tool measures a short-window loudness envelope (≈20 ms RMS) across the whole file and marks every stretch that stays below your threshold as silence. A silent stretch is only removed if it lasts at least your minimum length, and a slice of padding is left at each edge of the cut so transitions don’t click. The kept segments are joined back together and exported as a lossless 16-bit WAV — an audio to WAV conversion that preserves every sample of the kept audio. The waveform shades the regions that will be removed so you can dial in the settings before rendering. Nothing is uploaded. If you need to check the level of your recording’s noise floor before choosing a threshold, the noise floor analyzer can give you a per-octave-band reading from a live mic.
This is level-based gating (sometimes called a downward expander or noise gate in audio software), not speech recognition: it can’t tell a quiet word from background hiss, so set the threshold for your recording. A common starting point for podcast editing and other voice recordings is −45 dB with a 500 ms minimum gap; for music or acoustic sessions you may need to go lower (−55 dB or below). Use “Leading/trailing only” to trim silence from the start and end of a clip, or “All silences” to auto-cut every quiet gap throughout. For MP3 output, re-encode the WAV in a desktop app such as Audacity, Adobe Audition, or FFmpeg. You can also trim a specific section of the output if you only need part of the processed file.