SEP 4, 2026 · PREPRINT
SCAPES: Semantically Conditioned Autoregressive Prior for Environmental Sounds
arXiv
This is a preprint describing a novel machine learning architecture for sound synthesis with no peer review, clinical validation, or comparative efficacy data against established methods.
Reported
Model parameters36 million
Training resourcesingle consumer-grade GPU
Training duration convergence cri…approximately twice the source audio durati…