Silence Trimmer
Strip dead air from both ends of a recording and fade the edges cleanly
100% In-Browser · Private & Instant · Free Forever
Your file is decoded, processed, and re-encoded in this browser tab. Nothing is uploaded to an OfflineTTS processing API.
Interactive Workspace
Silence Trimmer
Recordings usually start with a second of room tone and end with another. This tool finds the first and last moment the audio rises above a threshold you set, keeps a small amount of padding around them, and applies fades to the new edges.
Why padding matters: trimming exactly at the threshold produces a file that starts abruptly the instant sound begins. Keeping a tenth or two of a second of the original silence preserves the natural attack of the first word or note. The default is 0.15 s.
Choosing a threshold: the right value depends on the noise floor of the recording. A quiet studio take might have a floor at -60 dBFS, so -45 dBFS is a safe threshold. A phone recording in a café might sit at -40 dBFS, where a -45 dBFS threshold would never trigger — the tool would find no silence to remove. The measurement is deterministic, so if nothing is trimmed, raise the threshold above the noise floor.
Fades: every trim introduces a discontinuity at the new boundary. A short fade — 0.05 to 0.2 seconds — removes the click without being audible as a fade. The default 0.1 s is a reasonable compromise.
Good uses: tightening voice notes before transcription, preparing samples for a sampler, cleaning up a batch of recordings before publishing, and removing the dead air at the end of a track that fades to silence.
Files are decoded, measured, and re-encoded inside this browser tab. Nothing is uploaded and no result is retained after you close the page.
How It Works
Load a recording
Add the file with unwanted silence at the start or end.
Set the threshold
Choose a level above the recording's noise floor.
Set padding and fades
Keep a little silence, and fade the new edges.
Trim and export
The tool reports the new duration; download WAV or MP3.
Features
Automatic Detection
Finds the first and last sample above your threshold.
Adjustable Padding
Keeps a configurable amount of the original silence.
Click-free Fades
Applies matching fade in and fade out at the new edges.
Repeatable
Deterministic thresholds mean the same result every run.
Thresholds, Noise Floors, and Real Recordings
Every recording has a noise floor: the level of the quietest content in the file, usually room tone, preamp hiss, or both. A threshold only works if it sits above that floor. Set it below and the tool sees constant sound and trims nothing. Set it far above and the tool will cut into the quiet beginning of the actual performance.
The noise floor is not usually a fixed number. A recording made with automatic gain control, common on phones, has a floor that moves as the gain adjusts. In that case no single threshold is ideal for the whole file, and the practical approach is to pick a value that handles the section you care about and check the result by listening rather than relying on the duration readout alone.
If the recording is very noisy, clean it first with the noise remover. Lowering the noise floor makes the silence detection much more reliable, because the difference between room tone and speech becomes larger and the threshold has more room to sit between them.
What Happens at the New Edges
A digital audio signal is a sequence of samples. Trimming at a point where the waveform is at, say, 0.4 of full scale means the file begins with a sample at 0.4 and the previous context is gone. Playback starts with a step from silence to 0.4, and that step is a broadband click. A fade removes the step by ramping the level from zero over a few milliseconds.
The fade length is a trade-off. Very short fades of one or two milliseconds remove the click but can still colour the attack of a percussive sound. Longer fades of 300 ms or more change the character of the beginning. Between 50 and 200 milliseconds is the range where most material sounds untouched.
Files are decoded, measured, and re-encoded inside this browser tab. Nothing is uploaded and no result is retained after you close the page.
One consequence of trimming is worth noting: the trimmed file is shorter, so any later processing is faster and uses less memory. For a batch of long recordings, trimming first and processing afterwards is measurably quicker than the other way round.
Related Tools
Audio Trimmer
Drag a selection on the waveform, add fades, and export just that part
Audio Joiner
Merge files in order, with optional normalization, crossfades, and gaps
Noise Remover
Learn the noise floor of a recording, then gate it out band by band
Audio Converter
Change sample rate and channel count, normalize, and re-export
Silence Trimmer — FAQ
Why did nothing get trimmed?
The threshold is probably below the recording's noise floor, so every sample counts as sound. Raise the threshold — a café recording might need -35 dBFS while a studio take works at -50 dBFS. The tool reports the resulting duration, so you can see immediately whether anything changed.
Does it remove silence in the middle of a file?
No. Only the leading and trailing silence is trimmed. Removing pauses between sentences requires a different pass and changes the rhythm of the speech, so it is deliberately not attempted here.
Why do I need fades if the edges were already silent?
Because the trim point is where the level crosses the threshold, which is not the same as zero. Cutting there leaves the waveform at a non-zero value, and the jump to silence is a click. A short fade removes it.
Can I trim several files the same way?
Yes — process them one at a time with the same settings. Because detection is deterministic, files with similar noise floors get consistent treatment.
Does trimming reduce quality?
No. Trimming removes samples; it does not re-encode them. WAV export is sample-exact. Only the MP3 export is a lossy step.
More Local Audio Tools
Every audio utility here runs in the browser tab. Browse the full set of editing, conversion, and measurement tools.