Composite audio comparisons with state-of-the-art methods

Example 9 Back to index

Captions for each audio tracks (all sounds should be contained)

  1. Engine roars powerfully
  2. Wheels grind heavily on pavement
  3. Horn blasts loudly
  4. People chatter excitedly
  5. Bicycles click softly as they move
MMAudio + NAG (ours) MMAudio MMAudio + Negative prompting
Seeing-and-Hearing FoleyCrafter FoleyCrafter + Negative prompting