← all videos

Static camera in a quiet living room. A dog barks on the right side of the frame. Then a man speaks from the off-screen left side: "Easy, buddy."

C1-2-07 Steady State · source_anchoring principle: Multi-Source Spatial Separation (Wave)
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 dog
  • a living room

Event: a dog barks in a living room

Audio presence

  • a dog
  • a man

Sound: ['dog bark', "man's voice"]

Key standards

Visual physical commonsense (V-PC)

  • A visible dog stays at a stable position on the right side of the frame and visibly barks during the bark event.
  • The speaking man is not visible in the frame.

Audio physical commonsense (A-PC)

  • The dog bark carries the timbre of a real canine bark.
  • A male voice clearly says the phrase "easy buddy" or a close variant.

Cross-modal physical commonsense (AV-PC)

  • The dog bark is localized to the right side of the audio field, matching the dog's visible right-side position.
  • The man's voice is localized to the left side of the audio field, matching his off-screen left position.