Individual audio track comparisons with baseline methods

Example 4 Back to index

Track 1: "The crowd murmurs excitedly in the background" (All videos are identical)

MMAudio MMAudio + Negative prompting MMAudio + NAG (ours)

Track 2: "Hooves clatter noisily against the metal ramp" Back to index

MMAudio MMAudio + Negative prompting MMAudio + NAG (ours)

Track 3: "Cows moo loudly as they move down the ramp" Back to index

MMAudio MMAudio + Negative prompting MMAudio + NAG (ours)

Track 4: "Dust particles rustle softly as they settle in the air" Back to index

MMAudio MMAudio + Negative prompting MMAudio + NAG (ours)

Track 5: "The gate slams shut with a loud metallic clang" Back to index

MMAudio MMAudio + Negative prompting MMAudio + NAG (ours)

Composite audio Back to index

MMAudio MMAudio + Negative prompting MMAudio + NAG (ours)