AudioTools

Convert

All 32 format conversions

Compress

Cut & Join

Volume

Speed & Pitch

Effects & EQ

Clean up

Channels & Format

Record & Generate

Music

Analyze

Metadata

Detect Silence

See exactly where the silence is, with times you can copy, export or use to split the file.

or drop it here

MP3, WAV, M4A, FLAC, OGG, Opus, AIFF, WMA, AMR and the sound of most videos. Files stay on your device.

How it works

  1. Choose an audio file.
  2. Set the threshold and the minimum silence length.
  3. Read the list of silent parts.
  4. Copy it, or export CSV or Audacity labels.

A list of every silent part

The table gives the start, end and length of each silent part, and the waveform highlights them. Press the play button on a row to hear it with a second of sound on either side. The list updates a moment after you change a setting.

Audio is checked in steps of 10 milliseconds. A step is silent when its level, on every channel, is at or below the threshold in dB (0 dB is the loudest possible level, so -50 dB is very quiet). A run of silent steps counts only if it lasts at least the shortest silence you set. Peak looks at the loudest sample in each step and reacts to any click; Average (RMS) measures the overall level and ignores isolated clicks.

Exporting the list

Download CSV opens in any spreadsheet and has the times in seconds and as readable timecodes. Download Audacity labels writes a text file that Audacity reads with File, Import, Labels, which puts a labelled region on a label track at each silence. Copy times puts the table on the clipboard as tab-separated text.

The table shows the first 200 parts; the downloads contain all of them, up to 50,000.

Uses

  • Find dead air and dropouts in a recording or broadcast.
  • Check where the pauses are before splitting a file or removing silence.
  • Make labels to cut a long recording in an editor.

Questions

What threshold should I use?

For clean digital recordings -60 dB finds true silence. For recordings with room noise try -40 to -50 dB. Raise it until the highlighted parts match the pauses you hear.

What is padding for here?

It shrinks each silent part by that amount where it meets sound. Leave it at 0 for exact boundaries.

How accurate are the times?

The edges are accurate to about 10 milliseconds, the size of the analysis step.

Is my file uploaded?

No. The analysis runs in your browser.