A hiker stands at the edge of a mountain valley and shouts loudly across it. A quieter, slightly duller echo of the shout returns from the far cliff.
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 hiker
- a mountain valley
Event: a hiker shouts across a mountain valley
Audio presence
- a hiker
Sound: a human shout
Key standards
Visual physical commonsense (V-PC)
- A single hiker is shown shouting across the open space.
- The setting is a mountain valley or canyon with the opposite cliff or ridge clearly visible.
Audio physical commonsense (A-PC)
- A quieter and duller copy of the shout returns in the audio after a clearly perceptible delay.
- The returning sound is a discrete copy of the original shout's content with reduced amplitude and dulled high frequencies.
Cross-modal physical commonsense (AV-PC)
- The original shout onset aligns in time with the visible shouting action of the hiker.
- A discrete delayed echo of the shout follows the original shout in the audio track.