Inside a moving car in daylight, a person speaks while traffic noise is heard outside; the side window is clearly shown fully closed, then rolled all the way down while the car keeps moving, yet the outside sound becomes more muffled and distant after the window is opened.
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 car
- a side window
Event: a person speaks inside a moving car as a closed side window is rolled down
Audio presence
- traffic noise
- a person
Sound: ['traffic noise', 'a person speaking']
Key standards
Visual physical commonsense (V-PC)
- The clip shows the inside of a moving car in daylight with a person speaking and the side window initially fully closed.
- The side window is then visibly rolled all the way down during the clip while the car continues moving, with the window remaining fully open for the rest of the clip.
Audio physical commonsense (A-PC)
- Outside traffic noise is audible at the start of the clip with the loudness and brightness expected through a closed side window.
- After the window is fully open, the outside traffic noise becomes more muffled and distant in the audio rather than louder and clearer.
Cross-modal physical commonsense (AV-PC)
- The shift to more muffled and distant traffic noise aligns in time with the moment the side window is visibly rolled all the way down.