A person tries to talk on a phone next to a busy highway. The traffic roar drowns out their voice and they have to shout to be heard.
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 phone
- a highway
- vehicles
Event: a person talks on a phone next to a highway
Audio presence
- a person
- highway traffic
Sound: ['shouted human speech', 'traffic roar']
Key standards
Visual physical commonsense (V-PC)
- A busy highway with visibly moving vehicles is shown, and a person visibly shouts into a phone next to it during the same scene.
Audio physical commonsense (A-PC)
- The traffic roar is louder than the voice throughout the phone call and stays active during the shouting segments.
Cross-modal physical commonsense (AV-PC)
- The visible shouting mouth movements coincide in time with faint shouted-voice segments underneath the traffic roar.