A presenter stands at a brightly lit podium and addresses an audience through a microphone connected to two large loudspeakers mounted on the left and right walls. The voice sounds loud and full with slight compression. The power suddenly cuts and the lights and loudspeakers lose power, but the presenter keeps talking. The voice now sounds soft, clean, and natural, with the room's natural reverb.
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 presenter
- a podium
- a microphone
- two wall-mounted loudspeakers
- stage lights
- an audience
Event: a presenter speaks at a lit podium and the venue power cuts mid-presentation
Audio presence
- wall-mounted loudspeakers
- a presenter
Sound: amplified speech transitioning to unamplified speech
Key standards
Visual physical commonsense (V-PC)
- The presenter, podium, microphone, and two wall-mounted loudspeakers are visibly present throughout the clip.
- The stage lights are visibly on before the power-cut moment and visibly dark after.
- The presenter visibly continues mouthing speech across the power-cut transition.
Audio physical commonsense (A-PC)
- Before the power-cut, the voice carries the band-limited compressed timbre of an amplified loudspeaker.
- After the power-cut, the voice carries the soft unamplified human-voice timbre with the room's natural reverb.
Cross-modal physical commonsense (AV-PC)
- The voice's timbre shift coincides in time with the visible power-cut moment.
- The amplified-voice phase corresponds to the period when the stage lights and loudspeakers are visibly powered.
- The unamplified-voice phase corresponds to the period when the lights and loudspeakers have visibly lost power.