A person punches a soft pillow with a muffled dull thud. Then punches a hard wall with a louder sharper impact and a slight ring.
Seedance 2.0ByteDance · proprietary
Kling 3.0 OmniKuaishou · proprietary
Veo 3.1Google · proprietary
LTX-2.3Lightricks · open
Ovi 1.1character.ai · open
JavisDiT++JavisDiT · open
MagiHumanGAIR-NLP · open
Per-prompt rubric
Basic standards
Visual presence
- a person
- a soft pillow
- a hard wall
Event: a person punches a pillow and a wall
Audio presence
- a pillow
- a wall
Sound: a punch impact
Key standards
Visual physical commonsense (V-PC)
- A person punches a soft pillow first, then punches a separate hard wall.
- The fist visibly contacts each surface during its corresponding punch.
Audio physical commonsense (A-PC)
- The pillow punch sound is dull and muffled with no audible reverb tail.
- The wall punch sound is louder and sharper than the pillow punch, with at most a slight short ring.
- Each impact sound has the natural timbre of a fist striking its respective surface.
Cross-modal physical commonsense (AV-PC)
- Each impact sound onset aligns in time with the visible fist contact on its surface.