Loading tool…
Dead air makes recordings drag. This tool finds every pause quieter than your threshold and longer than your minimum, shades it on the waveform, and either removes it or shortens it to a length you choose. You can listen to the tightened version before you download it.
Processing locally on your device
Overview
Silence detection works on a level envelope. Your browser decodes the file and measures the RMS level of every 10 ms window across all channels, in dBFS. Any run of windows below the threshold that lasts at least the minimum pause length counts as silence. Because the envelope is computed once, moving the sliders updates the detected pauses, the shaded regions and the time-saved figure instantly, even for long recordings.
Three settings shape the result. The threshold decides what counts as quiet: around −45 dBFS suits clean studio speech, while noisy rooms or video calls may need −35 to −30 dBFS so that background hum is treated as silence. The Auto button estimates the room’s quiet floor (the 10th percentile of all window levels) and sets the threshold about 10 dB above it. The minimum pause, typically 0.5 to 1 second, leaves breaths and natural gaps between words alone. Padding keeps a little quiet, often 100–200 ms, on each side of the sound so speech doesn’t start or end abruptly.
There are two modes. Remove deletes each pause except the padding, which gives the tightest result. Shorten reduces every pause to a fixed length instead, such as 0.4 seconds, keeping the rhythm of conversation while still cutting the dead air. Leading and trailing silence is trimmed in both modes. Every join is smoothed with a 5 ms crossfade so the edits don’t click, and the stats show how many pauses were found, how much time is saved and the new length.
The cutting is done in JavaScript on the decoded samples, so the Result preview is ready almost immediately and WAV export is instant. Other formats are encoded with FFmpeg in WebAssembly after a one-time 31 MB download. Decoded audio takes memory, about 21–23 MB per stereo minute at 44.1–48 kHz, so recordings over about 30 minutes can be slow on phones.
| Recording | Threshold | Min. pause | Padding / pause length |
|---|---|---|---|
| Studio podcast | −45 dBFS | 0.7 s | Shorten to 0.4 s |
| Zoom or Teams call | −38 dBFS | 1.0 s | Shorten to 0.5 s |
| Lecture hall | −32 dBFS | 1.5 s | Remove, 250 ms padding |
| Voice-over takes | −50 dBFS | 0.5 s | Remove, 100 ms padding |
Drop a recording (audio or video) onto the silence remover.
Set the silence threshold, or click Auto to estimate it from the quietest parts.
Set the minimum pause length so breaths aren’t counted.
Choose Remove with padding, or Shorten each pause to a fixed length.
Check the shaded regions and time saved, switch to Result to listen, then export.
Use cases
Shorten long thinking pauses to around 0.4 seconds instead of cutting them out entirely. Conversations stay natural, and an hour-long episode often loses several minutes of dead air.
Remove the gaps while a teacher changes slides or writes on the board, so students reviewing the recording get to the content faster.
Strip the silence between takes and lines from a long recording before cutting it into clips, so every line is easy to find in your editor.
Tighten rambling voice memos before sharing or transcribing them, which also reduces the file size and transcription time.
Pro tips
Answers to common questions about the silence remover.
Try Auto first. As a guide, −45 dBFS suits clean studio speech and −35 dBFS a typical video call. If pauses aren’t detected, raise the threshold toward −30. If quiet words are being cut, lower it toward −50.
Not if the minimum pause is longer than a breath. Values of 0.5 seconds or more usually leave breaths and short gaps alone while catching real dead air.
Remove deletes each pause and keeps only the padding on either side of the sound. Shorten leaves every long pause at a fixed length you choose, which keeps the rhythm of speech.
You can load a video and the tool works on its soundtrack, but it exports audio only. Cutting the audio changes its timing, so it would no longer line up with the original picture.
Each join gets a short 5 ms crossfade, and with padding the cuts fall in quiet parts, so they are normally inaudible.
No. ToolEnhancer processes files on your own device inside the browser. They are never sent to, stored on or seen by our servers.
Yes. The silence remover is free to use with no sign-up, no watermarks and no daily limits.