1 paper · 1 filter
Jiahao Mei, Heinrich Dinkel, Yadong Niu +7
Audio generation has long been fragmented, with speech, music, and sound effects produced by domain-specific models that fail to jointly generate coherent audio scenes from a singl…