Remove Silence From Audio Files Online

Threshold guide (how quiet is "silent"):

• ·−50 dB — Catches only true silence (no signal at all). Safe for almost any recording. Will not accidentally cut breathing, background noise, or quiet room tone.
• ·−40 dB — Catches very quiet sections. May catch soft breaths and very quiet background noise in clean recordings. Good starting point for well-recorded podcasts.
• ·−30 dB — More aggressive. Catches breathing sounds and quiet pauses. Risk of cutting natural speech pauses. Test with preview before using.
• ·−20 dB and above — Very aggressive. Will cut significant portions including breaths and soft sounds. Use only for specific cases like removing all non-speech sections from a recording.

Minimum silence duration:

• ·0.5 seconds — Removes all pauses longer than half a second. Produces very tight, energetic-feeling audio. Natural for fast-paced content; can feel rushed for slower conversational styles.
• ·1–2 seconds — The standard podcast setting. Removes dead air (2+ second pauses) while preserving natural pacing (under 1 second pauses). Best starting point.
• ·3–4 seconds — Conservative removal. Only cuts extended pauses — the truly accidental ones.

After removal: Normalise the resulting audio to a consistent LUFS target. Silence removal can change the perceived loudness balance of the track.

Silence removal vs manual trimming: Manual trimming with the Trim tool gives you exact control over which specific sections to remove, but requires listening to find them. Silence removal automates the detection — faster for long recordings with many gaps, but requires testing the threshold settings to ensure only genuine silences are removed.

Silence removal vs noise removal: These are different operations. Silence removal deletes sections that are below a volume threshold. Noise removal reduces background noise (hiss, hum, fan) in sections where audio is present. Use silence removal to cut dead air; use noise removal to clean up what remains.

Silence removal vs dynamic compression: Dynamic range compression reduces the volume difference between loud and quiet sections — it does not remove them. Silence removal actually cuts the quiet sections out, reducing total duration. For podcast editing, use both: remove silence first to cut dead air, then compress to even out volume variation in the remaining audio.

Dead air — the extended silence between sentences, the pause where a speaker forgot their train of thought, the gap where a question was asked on a separate mic that did not get recorded — is the enemy of listener engagement. In podcast production, silences longer than 2–3 seconds signal a problem. In voice recordings and interviews, unnecessary pauses slow the pace and make content feel unpolished.

Silence removal works by scanning the audio's amplitude waveform against a configurable threshold. Any section where the signal stays below the threshold for longer than a minimum duration is flagged as silence. Those flagged sections are removed, and the surrounding audio is joined together.

The two settings that determine what gets cut are the threshold (in dB) and the minimum silence duration (in seconds). These settings need to work together: a threshold that is too high cuts breathing sounds and natural pauses; a threshold that is too low misses silences that still sound quiet. A minimum duration that is too short removes natural pacing; too long leaves too much dead air.

The right approach is to preview the detected silences before committing. This tool shows which sections will be removed so you can adjust the settings until only the genuine dead air is flagged — then apply with confidence.