1. Tune Threshold to Your Noise Floor
Watch the waveform overlay as you drag the Silence Threshold slider: green bands are kept speech and red bars are cut pauses. Stop right when room hum turns red while quiet breaths stay green.
THE TOOLBOX / MEDIA & AUDIO
Automatically detect and remove awkward pauses from podcasts, voiceovers, lectures, and voice memos—or split recordings into separate chapter WAV files—100% privately in your browser via Web Audio API.
Upload a podcast, voiceover, or lecture—or click “Load Sample Audio with Pauses”.
Silence regions highlight automatically so you can tune threshold and padding before exporting.
AUDIO EDITING & THRESHOLD GUIDE
Watch the waveform overlay as you drag the Silence Threshold slider: green bands are kept speech and red bars are cut pauses. Stop right when room hum turns red while quiet breaths stay green.
Setting Remaining Pause Between Cuts to 0.15s–0.25s makes podcasts sound tight and energetic without the robotic “breathless” feel of zero-gap jump cuts.
Recording vocabulary clips, sound effects, or sample packs? Leave a 1-second pause between takes, then click Export Chapters (ZIP) to slice every take into its own numbered WAV file automatically.
Make precise manual cuts in the Audio Trimmer, stitch multiple clips in the Audio Joiner, capture voice tracks in the Online Voice Recorder, or extract audio with MP4 to MP3 Converter.
| Recording Type | Threshold | Min Silence | Padding / Kept Gap |
|---|---|---|---|
| Studio Podcast / Voiceover | −38 to −42 dBFS | 0.50 – 0.80 s | 0.12 s pad / 0.20 s gap |
| YouTube Talking Head | −34 to −38 dBFS | 0.35 – 0.50 s | 0.08 s pad / 0.10 s gap |
| Lecture / Phone Voice Memo | −26 to −32 dBFS | 0.80 – 1.20 s | 0.15 s pad / 0.25 s gap |
It scans the audio waveform in 20-millisecond windows and measures root-mean-square (RMS) loudness in decibels full scale (dBFS). Any stretch quieter than your chosen threshold for longer than the minimum silence duration is flagged for trimming or splitting.
For clean studio microphones, −40 dB to −35 dB catches pauses while keeping soft consonants intact. For noisier room recordings or phone voice memos, try −30 dB to −26 dB and add 0.10–0.15s of speech padding.
Speech padding preserves natural word endings (like “s” and “t” sounds) so edits never sound clipped, while an 8-millisecond fade-in/fade-out prevents audible clicks or pops at splice points.
Yes. Every non-silent region is listed as a chapter with exact timestamps. You can download individual chapter WAVs, export all chapters in one ZIP archive, or copy YouTube-ready chapter timestamps.
No. Decoding, silence detection, waveform rendering, and WAV/ZIP encoding run 100% locally in your browser using the Web Audio API.