Researchers have developed a machine-learning technique that captures and models the underlying acoustics of a scene from a limited number of sound recordings. The system can accurately simulate what ...