MMAudio

Comparisons with state-of-the-art methods in VGGSound

Example 1: Wolf howling. Back to index

Ground-truth

Ours

V2A-Mapper

FoleyCrafter

Frieren

VATT

V-AURA

Seeing and Hearing

Comparisons with state-of-the-art methods in VGGSound

Example 2: Striking a golf ball. Back to index

Ground-truth

Ours

V2A-Mapper

FoleyCrafter

Frieren

VATT

V-AURA

Seeing and Hearing

Comparisons with state-of-the-art methods in VGGSound

Example 3: Hitting a drum. Back to index

Ground-truth

Ours

V2A-Mapper

FoleyCrafter

Frieren

VATT

V-AURA

Seeing and Hearing

Comparisons with state-of-the-art methods in VGGSound

Example 4: Dog barking. Back to index

Ground-truth

Ours

V2A-Mapper

FoleyCrafter

Frieren

VATT

V-AURA

Seeing and Hearing

Comparisons with state-of-the-art methods in VGGSound

Example 5: Playing a string instrument. Back to index

Ground-truth

Ours

V2A-Mapper

FoleyCrafter

Frieren

VATT

V-AURA

Seeing and Hearing

Comparisons with state-of-the-art methods in VGGSound

Example 6: A group of people playing tambourines. Back to index

Ground-truth

Ours

V2A-Mapper

FoleyCrafter

Frieren

VATT

V-AURA

Seeing and Hearing

Comparisons with state-of-the-art methods in VGGSound

Moving train

Water splashing

Skateboarding

Synchronized clapping

Failure cases

Human speech

Unfamiliar vision input