♪ Music — Stem separation
Demucs
An open-source music source-separation model from Meta research that splits songs into vocals, drums, bass and other stems.
Tool open-source
Music / Workflow
Hum → stem split → arrange in the DAW
Workflow
Record a hummed or played idea, split a reference into stems for study, and rebuild the arrangement with your own instruments.
Why it's hybrid
Ideas begin with a human performance and finish with human playing and mixing.
Steps
- Record a voice memo or rough take
- Split reference audio with Demucs or Logic Pro's Stem Splitter for study
- Transcribe parts and re-record with real instruments
- Use assistive EQ or mastering only as a starting point
| Medium | Music |
|---|---|
| AI role | assist |
| Training data | unknown |
| Craft | arranging, recording |
| Last verified |
Tools in this workflow
♪ Music — Mastering
iZotope Ozone
A mastering plugin suite whose Master Assistant analyzes a mix and proposes a starting chain you can then adjust module by module.
Plugin paid
♪ Music — DAW
Logic Pro
Apple's DAW, which includes AI-powered Session Players and Stem Splitter alongside a full traditional recording and editing environment.
Tool paid
Sources: [1]
Related
♪ Music — Vocal synthesis
ACE Studio
An AI singing voice workstation that turns MIDI and lyrics into vocal performances.
Tool
♪ Music — Stem separation
Demucs
An open-source music source-separation model from Meta research that splits songs into vocals, drums, bass and other stems.
Tool open-source
▲ Games · ♪ Music — Voice & SFX
ElevenLabs
An AI audio platform for text-to-speech voices, voice changing and sound-effect generation.
Tool freemium