Silence Detector

About Silence Detector

This online silence detector analyzes your audio files to find and mark all silent or quiet regions. Perfect for cleaning up recordings, editing podcasts, finding pauses in speech, or preparing audio for further editing. Adjust the silence threshold and minimum duration to match your needs, then export the marker list for use in your DAW or audio editor.

What audio formats are supported?

All common audio formats are supported including MP3, WAV, OGG, AAC, M4A, FLAC, OPUS, and more.

What is silence threshold?

Silence threshold (in dB) determines what level is considered 'silent'. -60 dB is very quiet (catches almost all silence), -40 dB is moderate (standard), -20 dB is loud (only catches very quiet parts). Lower (more negative) values detect more silence.

What is minimum silence duration?

Minimum silence duration filters out very short silent moments. 0.1s catches brief pauses, 0.5s (recommended) catches normal pauses, 1s+ catches only longer silences. This prevents marking every tiny gap as silence.

How accurate is the detection?

Detection uses RMS (Root Mean Square) energy analysis with 512-sample hop size, providing accurate detection within ±10ms. The accuracy depends on your threshold settings - test different values to find what works for your audio.

What can I do with the marker list?

Export markers as a text file to: 1) Reference when manually editing, 2) Import into some DAWs as regions/markers, 3) Use as a guide for automated silence removal, 4) Document silence locations for quality control, 5) Copy to clipboard for quick access.

What does the timeline show?

The timeline visualization shows your entire audio file with green representing audio and red representing detected silence. This gives you a quick visual overview of where silence occurs and how it's distributed throughout your file.

Why would I want to detect silence?

Common uses: Cleaning up podcast recordings (find and remove long pauses), Preparing audio for editing (mark section breaks), Quality control (ensure no unwanted silence), Preparing audio for compression (identify gaps), Finding speech pauses (linguistic analysis), Detecting recording problems (dead air).

What settings should I use for podcasts?

For podcasts: Threshold -35 to -40 dB (catches normal pauses), Minimum duration 0.5-1.0s (ignores breathing pauses, catches editing gaps). For music: Threshold -45 to -50 dB (more sensitive), Minimum duration 0.2-0.5s (catches song gaps).

Can this automatically remove silence?

This tool detects and marks silence but doesn't remove it. Use the marker list with your audio editor or DAW to manually remove unwanted silence. For automatic removal, use our Silence Trimmer tool which removes silence from start/end.

What if I get too many or too few results?

Too many silence detections: Increase threshold (less negative, like -35 dB) or increase minimum duration (like 1.0s). Too few detections: Decrease threshold (more negative, like -50 dB) or decrease minimum duration (like 0.2s).

Can I use markers in my DAW?

Some DAWs support importing marker/region lists (check your DAW documentation). Otherwise, use the exported list as a reference to manually create markers at the specified timestamps in your DAW.

Which channel is measured for stereo files?

All channels are downmixed to mono before analysis: every sample is averaged across channels, then RMS is computed on the combined signal. This means silence is detected from the full mix, so audio panned hard to the left or right channel is still correctly counted as 'not silent'. Earlier single-channel analysis could mis-flag panned content as silence; the mono downmix fixes that.

What export formats and timecodes do you provide?

Three exports: (1) Export Markers gives a human-readable .txt list. (2) Export CSV gives the silence regions as CSV with columns Index, Start(s), End(s), Duration(s), StartTC, EndTC. (3) Export Cut List (Audio Segments) gives the inverse 'keep' segments between silences (including head and tail) in the same CSV columns. Numeric columns are seconds with millisecond precision (.toFixed(3)); StartTC/EndTC use MM:SS.mmm timecode. The silence CSV is your 'remove' list (feed to an ffmpeg trim/concat command), while the segments CSV is your 'keep' cut-list to import as markers/regions into Reaper, Audition, or Premiere.

What dBFS reference is used, and is it RMS or peak?

The threshold slider is in dBFS (decibels relative to full scale). Detection uses RMS (Root Mean Square) energy per 512-sample hop, compared against the linear threshold 10^(dB/20). RMS is the perceptual loudness measure editors expect for dead-air detection, so a -40 dBFS setting flags hops whose RMS sits below that level. If you need peak-strict gating, set the threshold a few dB lower than your target.

Is my audio file safe?

Absolutely! All audio analysis happens directly in your browser. Your audio file is never uploaded to any server. Everything stays private on your device.