AudioShake turns messy recordings into editable sound

Effortless Audio Cleanup for Broadcasts

Source separation is moving from studio cleanup into live broadcasts and accessible production tools, while raising unresolved questions about music rights and licensing.

3 key takeaways

1AudioShake separates mixed recordings into editable components for broadcast, film restoration, music, and karaoke workflows.
2Real-time performance depends on spectrograms, training data, and deployment choices spanning cloud systems and local workstations.
3The technology’s expanding reach brings practical access alongside difficult questions about generative AI, ownership, and licensing.

Don't miss

Jessica Powell traces AudioShake’s origin to Tokyo karaoke, where missing songs prompted the idea of separating vocals from instrumentals.

The brief

Jessica Powell, AudioShake’s CEO, explains how source separation makes mixed recordings editable by isolating voices, instruments, and unwanted background sounds.

The technology serves sports broadcasts, film restoration, music, and karaoke, with cloud, workstation, or local deployment chosen according to the model and workflow.

Real-time separation in noisy stadiums depends on spectrograms, training data, and models that learn sound characteristics without isolated recordings of every speaker.

AudioShake began with a practical Tokyo karaoke problem: Jessica Powell and co-founder Luke wanted to perform songs missing from karaoke catalogs.

The episode closes on the harder boundary around the technology: making AI music tools useful while fitting them into rights, licensing, and established industry structures.

Weekly email

State of the Pod, every Monday

The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.