A person speaks next to a powerful waterfall. Their voice is almost completely drowned out by the waterfall's continuous roar.
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 waterfall
Event: a person speaks next to a waterfall
Audio presence
- a person
- a waterfall
Sound: ['human speech', 'waterfall roar']
Key standards
Visual physical commonsense (V-PC)
- The waterfall shows continuous flowing water, and the person visibly moves their mouth while speaking next to it.
Audio physical commonsense (A-PC)
- The waterfall roar is substantially louder than the voice throughout, and the speech is largely unintelligible.
Cross-modal physical commonsense (AV-PC)
- The visible mouth movements of the person coincide in time with faint voice segments underneath the dominant waterfall roar.