Free Online Audio Silence Remover
Detect silent gaps and remove them from audio recordings.
Windows quieter than this level are treated as silence.
Gaps shorter than this are kept, so natural pauses are not chopped out.
A little padding avoids cutting off the start of a word.
Detection uses short-term RMS energy. A quiet room tone below the threshold still counts as silence, so raise the threshold little by little until the waveform marks the right gaps.
Audio without silence
ReadyAbout the Audio Silence Remover
Long pauses are natural in interviews, lectures, podcasts, and voice memos, but they make a recording feel slower than it needs to be. This tool scans an audio file for quiet stretches, lets you decide how quiet and how long a gap must be before it counts as silence, and stitches the remaining audio back together. It is useful for tightening a rough voice take, cutting dead air between sentences, or preparing a recording for a time limit. Decoding and editing happen in the browser, so the file stays on your computer.
How detection works
The audio is divided into twenty-millisecond windows. For each window the tool measures RMS energy across the channels and converts it to decibels. A window is marked silent when its level falls below your threshold. Consecutive silent windows are grouped, and only groups longer than the minimum length are removed. Short gaps inside a sentence are therefore preserved, which keeps speech sounding natural rather than clipped.
Choosing a threshold
The default of −40 dB works for many clean recordings, but no single value fits every file. A noisy room can sit around −35 dB, so a −40 dB threshold would never detect silence; raise it toward −30 dB instead. A very clean studio recording may have a noise floor near −60 dB, where −45 or −50 dB gives more precise cuts. Watch the waveform: removed regions are shaded red, so you can see immediately whether the marks match the visible gaps.
Padding and joins
Removing silence exactly at the detected edge can clip the first consonant of the next word. The padding control keeps a small amount of sound at both ends of every cut. Longer padding sounds more conservative, while zero padding produces the tightest result. The tool joins segments directly, without crossfades, so a sudden cut in continuous background noise may produce a small click. Keeping a little padding usually avoids that.
Output and limits
MP3 is compact and convenient for podcasts and messaging. WAV keeps 16-bit PCM and is better if you plan to edit further. The export always preserves the original sample rate and up to two channels. Silence removal changes the timing of the recording, so any captions or transcripts that refer to specific timestamps will need to be retimed after editing.
Frequently asked questions
Can I remove only the silence at the beginning and end?
This tool removes qualifying gaps everywhere in the file. For trimming only the start and end, use an audio trimmer instead.
Why did it remove a quiet word?
The threshold is too high or the minimum length too short. Lower the threshold, increase the minimum length, or add padding so quieter speech is treated as content.
Does the result keep the original quality?
The samples are copied without resampling, but exporting to MP3 adds one more lossy encoding step. Choose WAV if you want to avoid that additional generation loss.