Meta has launched a new open-source audio AI model called SAM Audio to isolate sounds from complex audio mixes.
The model can separate specific sound sources—like voices, instruments, and effects—from mixed audio tracks.
SAM Audio is open-source, letting creators and developers integrate it into apps and tools for editing or sound analysis.
Use cases include podcast editing, music remixes, noise reduction, and improved audio accessibility.
Meta says the model is optimized for efficient performance, allowing higher-quality sound separation without heavy hardware.
By open-sourcing the tool, Meta encourages community experimentation and innovation in audio AI.