Same face, three cuts, one prompt.
Seedance cuts like an editor when asked. This is the consistency test that proves it’s no fluke — one character, three locations, identical identity.
A cinematic three-shot sequence with clean editorial cuts, following the same [dancer in a flowing crimson dress with a sleek dark bun].
Shot one: [she rises onto pointe on a city rooftop at golden hour, wind catching the fabric, skyline glowing behind her].
Shot two: [in a vast mirrored dance studio she spins in slow motion, the crimson dress blooming outward, reflections multiplying around her].
Shot three: [on a rain-soaked street at night she leaps through a puddle under warm streetlights, water droplets frozen mid-air around the arc of her jump].
The same [dancer, same dress, same hairstyle] in every shot. Audio: [elegant cinematic strings building with a deep percussive hit on each cut, rain ambience in the final shot].
Customize: the key is repetition — name the identical features again after the shots ("same dress, same hairstyle"). A description at the top plus a recap at the end locks in the character. Photorealistic human face close-ups belong in reference mode, not here.