LiveBand generates real-time music accompaniments from audio

Arxiv CS.SD (Sound/Audio ML)

LiveBand uses a causal transformer in latent space of a pretrained model to generate accompaniments with strict causal constraints. The system produces high-fidelity audio in real time, respecting latency requirements for live performance.

liveband

Arxiv CS.SD (Sound/Audio ML)·Marco PasiniJavier NistalBen HayesMathias Rose BjareStefan LattnerGeorge Fazekas

8 days ago