Somebody That I Used To Know AI: Sing the Painted Wall Duet With a Friend

Recreate the reference video [Video 1] shot for shot: the same wall of large flat geometric shapes in dusty pink, terracotta, beige, cream and grey, the same soft studio light, the same camera angles, cuts, timing and singing. Use [Video 1] as the sole reference for the video structure, the camera, the actions, the interaction between the two singers, the environment and the timeline. The ONLY change: the cat facing the camera in the middle is now the person in [Image 1], and the cat in profile on the right, who sings toward it, is now the person in [Image 2]. Both are real adult people, shown from the chest up. [Image 1] is the sole identity reference for singer A, [Image 2] for singer B. Singer A = the person in [Image 1], in EVERY shot: copy the exact face from [Image 1] - face shape, eye color and eye shape, eyebrows, nose, mouth, skin tone, age, hair color, hair length and hairstyle, and any beard, moustache, glasses or freckles exactly as in the photo. Faces the camera and sings with feeling, the mouth opening wide on the long notes, the eyes half closed. Singer B = the person in [Image 2], in EVERY shot: copied from [Image 2] in the same way. Stands at the right in profile, turned toward singer A, and sings to them. Paint: each wears the top from their own photo, and the top, the neck and the shoulders are painted over with the same flat pink, terracotta, beige, cream and grey shapes as the wall behind them, so their bodies blend into the wall. Each face stays unpainted and clearly their own. No cat ears, no fur, no whiskers, no tails. The same real people in every frame, so their friends instantly recognize them. Never morph, never drift back to the cats' faces. Do not mix the two faces: [Image 1] is only singer A, [Image 2] is only singer B. Both change: no cat from the video stays in any shot, also at the edge of the frame in the close-up. Shot 1 [0-2.1 s]: medium two-shot: singer A in the middle facing the camera, singer B at the right in profile, turned toward A. Shot 2 [2.1-9.5 s]: hard cut, close-up: singer A's face big in the middle, singing with the mouth wide open; singer B's profile at the right edge, singing toward A. Shot 3 [9.5-14.9 s]: hard cut, medium two-shot as in shot 1: A sings to the camera, B sings toward A. Keep the left-right places, the distance between them, the framing and the length of the video. Their mouths open and close in time with the singing in the video. No dialogue, no background music. Anatomy rules: one head each, on its own shoulders. Exactly two people. No extra people, no animals, no text, no logos, no watermark.

Recreate the reference video [Video 1] shot for shot: one continuous handheld take in front of a cream diamond-tiled wall and a white door, the same light, framing, timing and moves. The ONLY change: the performer in the full-body patchwork body-paint costume is now the person in [Image 1]. Use [Video 1] as the sole reference for the camera, the moves, the singing and the timeline, and [Image 1] as the sole identity reference. Performer = the person in [Image 1], in EVERY frame: copy the exact face from [Image 1] - face shape, eye color and eye shape, eyebrows, nose, mouth, skin tone, age, hair color, hair length and hairstyle, and any beard, moustache, glasses or freckles exactly as in the photo. Their own hair from [Image 1]: NO dark curly bob wig unless [Image 1] shows it. The costume stays from the video, never the clothes from [Image 1]: a tight full-body suit painted in large flat pink, terracotta, brown, cream and grey shapes. The face stays unpainted: no painted lines on the face, no red lipstick unless [Image 1] shows it. No red wristband, no watch. The singing, matching the reference second by second: they sing straight to the camera with big, over-the-top expressions, the mouth opening very wide on the loud notes, the head tilting, the eyes closing on the long notes, standing in place while the handheld camera moves a little closer and back. The same real person in every frame, so their friends instantly recognize them. Never morph, never drift back to the original performer's face. No dialogue, no background music. Anatomy rules: exactly one person, one head, on the performer's own shoulders. No text, no logos, no watermark.

Recreate the reference video [Video 1] shot for shot: the same wall of large flat geometric shapes in dusty pink, terracotta, beige, cream and grey, the same soft studio light, the same camera angles, cuts, timing and singing. The ONLY change: the cat facing the camera in the middle is now the pet in [Image 1], and the cat in profile on the right, who sings toward it, is now the pet in [Image 2]. Both cats change: no cat from the video stays anywhere, also at the edge of the frame in the close-up. [Image 1] is the sole look reference for pet A, [Image 2] for pet B, and [Video 1] is for the moves, the cuts and the camera only. Pet A = the pet in [Image 1], in EVERY shot: the same species and breed, face shape, eyes, ears, muzzle, fur colors, fur length and markings as in [Image 1], with the natural body volume and proportions of that animal: never flattened, squashed, stretched or merged with anything. Not the coat, the ears or the face of the cat in the video. It faces the camera and sings, opening its mouth wide when the cat in the video does. Pet B = the pet in [Image 2], in EVERY shot: the same species and breed, face shape, eyes, ears, muzzle, fur colors, fur length and markings as in [Image 2], with the natural body volume and proportions of that animal. Not the coat, the ears or the face of the cat in the video. It sits at the right in profile, turned toward pet A, and sings to it. If an image shows a dog or another animal, that animal takes the same place, sitting the same way: a dog stays a dog and a cat stays a cat. Each pet looks exactly like its own photo from the first frame to the last, and the two pets look as different from each other as the two photos do. Never swap the two pets. Shot 1 [0-2.1 s]: medium two-shot: pet A in the middle facing the camera, pet B at the right in profile, turned toward A. Shot 2 [2.1-9.5 s]: hard cut, close-up: pet A's face big in the middle, singing with the mouth wide open; pet B's profile at the right edge, singing toward A. Shot 3 [9.5-14.9 s]: hard cut, medium two-shot as in shot 1. If a pet in any shot still looks like a cat from the video and not like its photo, it is wrong: redraw it as the animal in its photo. No dialogue, no background music. Exactly two animals, no people, no text, no logos, no watermark.