Hotel Lobby AIプロンプト：2人用の完全版

GOAL
Use @Video 1 as the reference for movement, positions, gestures and camera timing, and generate a realistic two-person rap performance, landscape 16:9, 10 seconds. Two performers stand on either side of one hanging microphone in front of an orange wall and trade bars.

REFERENCE ROLES
@Image 1 alone defines Character 1, the performer left of the mic: face, eyes, hair (or fur), skin tone, body and clothes. Do not take the photo's background or pose.
@Image 2 alone defines Character 2, the performer right of the mic: face, eyes, hair (or fur), skin tone, body and clothes. Do not take the photo's background or pose.
@Video 1 is used only for the orange wall, the centered hanging mic, the fixed frontal mid shot, the turn-taking rhythm, the hand gestures and the closing wide shot. Never copy, keep or blend the original performers' faces, sunglasses, jackets or identities.

CHARACTERS AND ANTI-SWAP RULE
Character 1 always stays left of the mic, Character 2 always right. Each keeps the features of their own photo. If a photo shows a pet, it performs standing upright and keeps its own fur and face.
The two are separate people for the whole clip: never swap sides, faces, hair or clothes, and never merge them, not during gestures, glances or the wide shot.

BEATS
0s-3s: mid shot at the orange wall. Character 1 starts the verse into the mic and gestures forward; Character 2 nods to the beat and points.
3s-6s: Character 2 steps in and answers, pushing a hand out; Character 1 keeps the rhythm with both hands; they trade looks.
6s-8s: the camera moves in slightly; both swing their hands on the beat and point at each other.
8s-10s: the camera pulls back to a wide shot; both move to the beat on the orange stage.
Lip movement follows the original audio line by line; gestures land on the beat.

KEEP
Keep the orange wall, the mic position and the lighting stable. Only Character 1 and Character 2 in frame: no third person, no background clones. Natural hands and limbs, no extra fingers, no clipping. No text, no subtitles, no watermark.

Use @Video 1 for movement, positions and timing only. @Image 1 is Pet 1 (left of the hanging mic) and @Image 2 is Pet 2 (right of the mic); each keeps its own breed, fur colour, markings and face. Both stand upright like performers in front of the orange wall and trade bars at the mic: Pet 1 opens, Pet 2 answers, both gesture with their front paws on the beat, then a wide shot of both moving to the music. Never swap or merge the two pets. landscape 16:9, 10 seconds, synced to the original audio. Only these two pets, no people, no text, no watermark.