A mixed song is a single waveform: every instrument is added together. Separating it back into parts — vocals, drums, bass and the rest — is called stem splitting. There are two main ways to do it.
Phase cancellation: the classic trick
In most songs the lead vocal sits in the centre of the stereo image, while many instruments are spread left and right. Subtract one channel from the other and everything in the centre cancels out, leaving an instrumental. This is what SoundBooster's vocal remover does. It's instant and runs on your device, but it can only split centre from sides — other centred instruments go with the vocal, and reverb on the vocal can stay behind.
AI separation
AI models trained on thousands of songs with their original stems learn what each instrument sounds like, so they can separate drums, bass, guitar and piano as well. They're much cleaner, but need a powerful computer or a paid online service.
What people use vocal removal for
- Instrumentals for karaoke, or a cappellas for remixes.
- Practising an instrument by muting it and playing along.
- DJ edits and mashups.
Where it struggles
- Centred instruments — bass, kick and snare often sit in the middle with the vocal.
- Stereo effects — wide reverbs and doubled vocals don't cancel completely.
- Mono recordings — with no stereo difference there's nothing to separate.
- Low-quality sources — heavily compressed MP3s leave more artefacts.
Getting cleaner results
- Start from the best file you have: WAV or a high-bitrate file beats a low-quality stream rip.
- Pick songs with a centred vocal and a wide mix — most modern pop and rock.
- Use the parts in context: small artefacts that stand out in solo usually disappear in a mix.
Try it in the vocal remover, then rebalance the parts or send them to the studio. Only separate music you have the right to use.
