Keeps only what's shared identically between the left and right channels — typically the centered lead vocal — and discards the rest. Same honest caveat as the Vocal Remover: this is phase-cancellation math, not AI separation, so other centered elements (bass, kick) come through too.
The Vocal Extractor isolates sound that is common to the left and right channels (the “centre”) within the vocal frequency range — a quick way to pull out centre-panned lead vocals, dialogue or narration from a stereo mix.
It is the opposite of the Vocal Remover: instead of subtracting the centre, it keeps only the centre.
The Mid signal (L+R)/2 is band-passed between your low and high cuts with 4th-order Butterworth filters, gained, and passed through a true-peak limiter. Output is mono. Instruments panned to the centre (bass, kick, snare) leak through at low levels, which is why the default low cut is 150 Hz.
Pull dialogue from a stereo film mix.
Hear the vocal line alone to learn it.
Remove stereo ambience from a mid-centred voice.
Check how much of a mix is centre-panned.
| Method | Band-limited mid signal |
|---|---|
| Output | Mono, limited to −1 dBTP |
No — only centre-panned content can be isolated this way.
Mono has no stereo difference, so the result is the same as the input band-passed.
You'll get whatever was mixed dead-center — usually the lead vocal, but often with some bass, kick drum or reverb bleeding through too, since this is phase-cancellation math, not trained AI separation.
Quick reference tracks, checking lyrics, or rough sampling — not a substitute for a proper isolated vocal stem.
This tool runs entirely in your browser using the Web Audio API. Nothing is uploaded, stored or shared — there is no server to send your file to, so it works offline once the page has loaded.
Be the first to share your experience with this tool — your feedback helps others.