Rescue a Bad Live Recording with Ableton Stem Separation

Brian Funk turns a rough iPhone voice memo of his band into a listenable mix using Ableton Live’s Stem Separation — a practical trick for any band or songwriter.

A single iPhone voice memo recorded at a live bar gig — no soundcheck, no mic placement, no gain staging. Just a phone on a counter and a band playing. Brian Funk took that rough, unbalanced recording of his three-piece rock band Some Good Evil and used Ableton Live’s Stem Separation to pull it apart into usable individual stems, then rebalanced the whole thing into something worth listening to.

What It Does

Ableton Live’s Stem Separation analyses a stereo audio file and splits it into four isolated stems: vocals, drums, bass, and guitar. The process runs offline, so for a 29-minute recording you are looking at roughly eight minutes of processing time — enough to make a cup of tea and come back to something workable.

The real value here goes beyond the obvious use case of grabbing an a cappella or a drum loop from an existing track. For bands, songwriters, and anyone who records rough ideas on their phone, Stem Separation offers a way to retrospectively fix the balance of a performance that was never meant to be a proper recording. Buried vocals get lifted. An overpowering guitar gets tamed. The drums, even captured on a phone mic, can be shaped independently.

Artefacts and phasing are part of the deal — the tool is working with a single stereo source, not multitrack stems — but for documentation recordings and quick demos, the trade-off is absolutely worth it.

How to Use It

• Import your audio: Drag your recording into an Ableton Live session as a clip in the arrangement or session view.
• Right-click the clip: Select Separate Stems from the context menu.
• Choose your quality setting: High Quality gives the best separation results; expect longer processing times for longer files.
• Click Separate Now: Ableton processes the file and returns four new clips — Vocals, Drums, Bass, and Guitar — on separate tracks.
• Review each stem individually: Solo each track to check what has been captured and note any bleed or artefacts before you start mixing.
• Apply compression per stem: The Glue Compressor works well here — a gentle squeeze on the vocals brings them forward without sounding over-processed.
• Rebalance the faders: With each element on its own track, adjusting the overall mix is straightforward. Bring up the buried vocals, pull back anything that dominates, and nudge individual elements until the balance feels right.
• Compare against the original: Toggle between your separated mix and the raw recording to hear the difference and confirm you are moving in the right direction.

Why It Matters

Stem Separation is often discussed in the context of remixing or sampling, but its most practical application for working musicians is exactly this: recovering something that was recorded casually and turning it into a useful reference or shareable document. If you are in a band, rehearse regularly, or sketch ideas on your phone, this workflow means no informal recording is ever truly lost. A rough capture becomes a starting point rather than a dead end.