Stem Separation in Ableton Live FAQ
Stem Separation, introduced in Live 12.3 Suite, uses machine learning to extract individual elements from audio clips. This FAQ answers common questions about system requirements, modes, performance, and optimization.
System requirements
Performance
Processing modes
Options and output
Troubleshooting
System requirements
Which editions of Live offer Stem Separation?
Stem Separation is available in Live 12 Suite, starting with Live 12.3. It is not available in Live Standard, Intro, or Lite.
What are the minimum system requirements for Stem Separation?
Stem Separation works on all systems that support Live 12.3. For detailed information visit Live Minimum System Requirements.
Note: macOS 26 Tahoe is not yet supported. Please see the Mac Compatibility Guide for Ableton Live for more information on macOS compatibility.
Does Stem Separation use GPU acceleration?
Macs with Apple silicon use the GPU (Graphics Processing Unit) for accelerated stem separation processing. Intel Macs and Windows computers process stem separation using the CPU (Central Processing Unit), which can take longer. Learn more about GPU acceleration in the external article What is GPU Acceleration?
Does Stem Separation require extra disk space to install or use?
On Macs with Apple silicon, Stem Separation models increase the installed size of the Live by ~2GB, compared with the previous version of Live 12. We recommend maintaining around 10 - 20% free disk space on your system (internal) drive in order for Ableton Live to perform as expected. Learn more about Computer specifications for running Ableton Live.
Performance
How much RAM do I need for Stem Separation?
While 8 GB RAM is the minimum requirement, systems with only 8GB RAM will experience slower processing of stems, especially on Intel Macs and Windows. To optimize performance, close other applications during Stem Separation. For frequent High Quality mode use, 16 GB RAM is recommended.
Why is Stem Separation slow? Why does it seem to run faster on a different computer or in a tutorial video?
Stem Separation performance depends primarily on your computer's processor architecture, and the amount of installed RAM. Macs with Apple silicon using the GPU for Stem Separation are fastest, while Intel Macs and Windows PCs use slower CPU processing.
Processor types
| Computer or device | Processing unit used for Stem Separation | Expected speed |
|---|---|---|
| Macs with Apple silicon macOS 11 Big Sur through 15 Sequoia | GPU | Fastest processing of stems |
| Note: due to an incompatibility, computers running macOS 26.1 Tahoe use the CPU instead of GPU, which results in slower stem processing. | ||
| Intel Macs | CPU | Slower stem processing as compared to Macs with Apple silicon |
| Windows PC | CPU | Slower processing times as compared with Macs with Apple silicon |
| Push 3 (standalone) | Push processor | Only High Speed mode is available |
Additional factors affecting Stem Separation speed:
- RAM: 8 GB of installed RAM is adequate for Stem Separation but performance is likely to be slow; 16 GB RAM is recommended to use High Quality mode.
- Processing mode: High Speed is faster; High Quality takes longer but produces better results.
- Clip length and complexity: Longer clips require more processing time.
- System resources: Close other RAM-intensive programs on your computer to optimize processing time.
Why don't the separated stems sound how I expect?
In addition to your computer's processor architecture and the amount of installed RAM, results may also depend on:
- The original audio material: If the source differs significantly from any of the separation models used to train the software, clean separation may not be possible.
- Processing mode: if you’ve selected High Speed mode and are not satisfied with the results, try High Quality mode.
- The quality of the source file: low bit-rate MP3s or audio files with a low sample rate may result in poorer sounding stems.
If you’ve tested numerous high-quality sources using both modes and still experience unsatisfactory results with ALL stems, please contact Ableton Support and include the following information with your request:
- A Status Report or Crash Report
- An example of an affected Live Set including the original sample and the created stems that didn’t sound as you expect. Use Collect All and Save to copy audio files into the Project folder and compress the Project folder before sending
- A clear description of the issue including the steps needed to reproduce it
What does "Waiting for decoding to finish" mean?
Stem separation requires Live to first analyze samples for tempo and waveform data. The numbers shown (e.g., 1/2) reflect completed vs. remaining processing tasks, which may include:
- Decoding (if needed)
- 4-Stem separation in High Speed mode
- Individual separation passes for each type (Vocals, Drums, Bass) in High Quality mode
- Processing Others stem in High Quality mode
Processing modes
What's the difference between High Speed and High Quality separation?
High Speed mode processes stems in a single-pass, offering good quality results as quickly as possible. Select mode for:
- Separating stems on Push (standalone)
- Lower available RAM (less than 16 GB)
- Simpler audio sources
- Previewing stems
High Quality mode uses a multi-step process, yielding superior stem isolation with longer processing times. Select this mode (not available on Push) for:
- Lower signal-to-distortion ratio
- Higher available RAM (16 GB recommended)
- Projects requiring maximum stem quality
How can I speed up High Quality mode processing?
For faster processing, select only the stems you need, and exclude Others. To create the Others stem, Live extracts all other stems first, which increases processing time.
Does Stem Separation work on Push (standalone)?
Stem Separation is available on Push (standalone) in High Speed mode only. High Quality mode is not supported on Push.
Options and output
What stem types are available?
You can separate audio into Vocals, Drums, Bass, and Others, which includes all audio not identified as vocals, drums, or bass.
What are the properties of audio files created by Stem Separation?
All separated stems are rendered as 44.1kHz, 16-bit WAV files, regardless of Live's preference settings.
Troubleshooting
Live crashed during Stem Separation on a Mac with Apple silicon. What should I do?
- Go to System Settings → General → Software Update to install the latest security updates for your macOS version.
- Make sure you have at least 2 GB free disk space on your system (internal) drive.
- Try using High Speed mode instead of High Quality mode.
- Process shorter audio clips.
- Test Stem Separation in a simple Set with no third-party plug-ins, and follow the linked steps for Troubleshooting a crash.
- Contact Ableton Support if crashes persist after following the above steps.