A person taps an empty glass bottle with a mallet and it rings at a bright high pitch. Then taps a half-filled bottle, and it rings noticeably lower.
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 glass bottle
- a mallet
Event: a glass bottle is tapped with a mallet
Audio presence
- a glass bottle
Sound: a bottle tap ring
Key standards
Visual physical commonsense (V-PC)
- An empty glass bottle and a water-filled glass bottle are each visibly tapped by the mallet in sequence, with each tap visibly contacting the bottle.
Audio physical commonsense (A-PC)
- The tap ring from the water-filled bottle is noticeably lower in pitch than the tap ring from the empty bottle.
Cross-modal physical commonsense (AV-PC)
- Each ring sound aligns in time with a visible tap of the mallet against a bottle.
- The lower-pitched ring corresponds to the water-filled bottle.
- The higher-pitched ring corresponds to the empty bottle.