The big advantage of separate tracks (stems): you can listen to, mute, or isolate any instrument afterward. For that to work, the rehearsal needs to be recorded separately in the first place. Here’s how:
The best way: split at the source
A multitrack recorder records each channel separately. Every microphone and every DI box (guitar, bass, vocals, drums) goes to its own input — at the end you get a separate file per instrument. No splitting needed afterward, best quality.
- Each instrument on its own channel (mic or DI).
- Drums either as a stereo sum or via multiple mics, depending on the recorder.
- Set levels cleanly before the take — nothing should clip.
- Export one file per track after the rehearsal (WAV), convert to MP3 if needed, upload.
You can find out which devices work well for this in the hardware guide.
No multitrack setup? AI separation as a fallback
If you only have a stereo recording (e.g. from a phone), services like Moises or LALAL.ai can split it afterward into drums, bass, vocals, and “everything else”. Convenient, but never as clean as real separate channels — still often good enough for a quick breakdown.
Afterward
Upload the tracks to your rehearsal — in the player you can then mute and unmute individual instruments and practice tricky spots slowly.