A pianist presses a key softly, producing a gentle quiet note. Then strikes the same key hard, producing a much louder note.
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 piano
- a pianist
Event: a pianist presses a piano key
Audio presence
- a piano
Sound: a piano note
Key standards
Visual physical commonsense (V-PC)
- The same single piano key is pressed twice, first gently and then with clearly more force.
Audio physical commonsense (A-PC)
- The second note is noticeably louder than the first note.
- Each note has the characteristic timbre of a struck piano key.
Cross-modal physical commonsense (AV-PC)
- Each note onset aligns in time with a visible keypress.
- The louder note corresponds to the forceful keystrike.
- The quieter note corresponds to the gentle keypress.