The old way didn't work
For decades, "vocal removers" used phase inversion — flipping one stereo channel and summing to cancel anything center-panned. It worked on some songs, terribly on most. Modern reverb, stereo width, and production tricks broke the method completely. You'd end up with hollow, metallic audio and vocal bleed everywhere it was supposed to be gone.
The Stemulator way
Stemulator's separation engine has learned the acoustic signature of a human voice across a vast corpus of multitrack recordings. The engine recognizes what a voice sounds like — not where it's panned — so it can remove vocals from any mix regardless of production style. You get a clean instrumental that sounds like the original minus one layer, not a gutted approximation of it.
Who uses a vocal remover?
- Karaoke hosts & singers — strip vocals to sing over the real instrumental
- Producers — pull acapellas for remixes and mashups
- Students — learn a guitar or piano part without the vocal distracting
- Content creators — use instrumentals as background music in video
- Translators & dubbers — replace original vocals with a new language
- Worship & cover bands — rehearse with the exact instrumental of the real song
Remove or isolate — both in one pass
Every vocal-remover job on Stemulator produces both outputs: the instrumental (no vocals) and the isolated vocal acapella. One upload, two files. Lead and backing vocals can also be split into separate stems.
Supported formats
Upload: MP3, WAV, FLAC, M4A, AAC, OGG, WMA. Download: lossless WAV, AIFF, or MP3.