A person shouts inside a long tunnel and the shout rings on. They walk out the end of the tunnel and shout again into the open air, where the voice vanishes quickly.
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 tunnel
Event: a person shouts inside a tunnel and in the open air
Audio presence
- a person
Sound: a person shouting
Key standards
Visual physical commonsense (V-PC)
- A person shouts inside a tunnel.
- The same person walks fully out of the tunnel and shouts again in clearly open air.
Audio physical commonsense (A-PC)
- The tunnel shout carries a sustained reverberation tail.
- The open-air shout decays rapidly with negligible reverberation.
Cross-modal physical commonsense (AV-PC)
- Each shout sound aligns in time with the person's visible shouting action.
- The reverberation character of each shout matches the visible space in which it occurs.