A person walks on dry concrete with steady taps. Then steps into a shallow puddle, and the footsteps turn into wet splashes with sloshing and dripping.
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
- dry concrete
- a shallow puddle
Event: a person walks across a surface
Audio presence
- footsteps
Sound: footstep impacts
Key standards
Visual physical commonsense (V-PC)
- The person walks with a natural gait on both a dry concrete surface and a shallow puddle, with visible water splashing when stepping into the puddle.
Audio physical commonsense (A-PC)
- The concrete footsteps have steady dry hard taps and the puddle footsteps have distinctly wet splashes with sloshing.
Cross-modal physical commonsense (AV-PC)
- Each footstep sound coincides in time with a visible foot landing on its corresponding surface.
- The dry hard taps correspond to footsteps on the concrete surface.
- The wet splashes with sloshing correspond to footsteps in the puddle.