1 paper · 1 filter
Koichi Saito, Julian Tanke, Christian Simon +7
Prevailing Video-to-Audio (V2A) generation models operate offline, assuming an entire video sequence or chunks of frames are available beforehand. This critically limits their use…