2022 has been an exciting year for machine learning (ML) and realtime neural audio synthesis. We share how we achieved real-time differentiable DSP (DDSP), in an expressive and transcultural instrument, operating with low latency in DAWs and mobile devices worldwide. Our talk will cover:
- building an intuition for DDSP and the capabilities of an audio machine learning system,
- what problems we faced and how we achieved real-time DDSP in both pro-audio and consumer applications,
- how we facilitated creative musical expression in our ML system,
- how we approached testing, and
- how we see realtime ML audio processing in future.
IF YOU ARE ATTENDING ONLINE, ALL TALK SESSIONS CAN BE ACCESSED FROM THE MAIN LOBBY:
https://conference.audio.dev
