A person stands outside a closed glass window and speaks directly toward the camera; the window remains visibly shut the entire time, yet the voice sounds fully clear, bright, and direct, as if no glass barrier were present.
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 glass window
Event: a person speaks toward the camera through a closed glass window
Audio presence
- a person
Sound: a person speaking
Key standards
Visual physical commonsense (V-PC)
- A person stands outside a glass window and speaks directly toward the camera with continuous visible mouth motion.
- The glass window remains visibly closed and intact throughout the clip with no openings or gaps between the person and the camera.
Audio physical commonsense (A-PC)
- The voice sounds fully clear, bright, and direct with the unobstructed timbre of close-range speech in open air, without any low-pass muffling or attenuation.
Cross-modal physical commonsense (AV-PC)
- The fully clear voice character is present in the audio while the person is visibly speaking from outside the closed glass window.