How-to guide
How to remove background music from a voice memo
Voice memos recorded in public spaces — cafes, gigs, transit — usually pick up music along with the speech you actually want. VocalSplit treats spoken voice the same way it treats sung vocals and pulls it out of the background bed.
The separator is trained on sung vocals but works remarkably well on speech. Male speech with music bleed tends to clean up more reliably than whispered or very soft voice memos.
Step-by-step
- Export the voice memo from your phone. On iPhone: Share → Save to Files. On Android: most voice recorders save as M4A or MP3 directly.
- Open VocalSplit.io. Any modern browser. No install, no account.
- Upload the voice memo. M4A, MP3, WAV and AAC all work. Voice memo files are usually small, well under 100 MB.
- Pay for the split. $0.99 gets you your first separation. Credits never expire if you save the rest for another time.
- Download the speech and music stems. The speech shows up in the vocals stem and the background ambience goes into the instrumental.
Tips for better results
- For a clean speech track, discard the instrumental stem and keep only the vocals stem.
- Whisper or very soft speech is harder to separate cleanly than normal conversational volume.
- Mono recordings work just as well as stereo — you do not need to convert first.
Try VocalSplit free
Upload a song and get clean vocals and instrumental stems in under 15 seconds. First split is $0.99.
Split a song