arXiv:2606.03803v1 Announce Type: cross Abstract: We present LiveBand, a real-time system that generates high-fidelity music accompaniments to live audio input, respecting strict causal constraints. Our method trains a causal transformer generator in the continuous latent space of a pre-trained causal audio autoencoder, using adversarial sequence-level supervision from a discriminator. At each timestep, the generator receives only the causally available mix context and Gaussian noise, and predicts accompaniment latents without access to future mix frames or ground-truth target latents. Trainin
Advances in causal transformer models and audio autoencoders are enabling real-time generative AI for complex creative tasks.
This development pushes the boundaries of real-time AI generation in the creative sector, potentially transforming live musical performance and production workflows.
AI can now generate high-fidelity, causally constrained musical accompaniments in real-time, moving from static pre-trained models to dynamic, interactive systems.
- · Musicians
- · Music producers
- · AI music platform developers
- · Entertainment industry
- · Traditional session musicians
Real-time AI accompaniment tools become accessible to a wider range of musicians and content creators.
The definition of live music performance expands to include human-AI collaborative acts across various genres.
New forms of intellectual property and compensation models emerge for AI-generated creative works.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI