Kaip su AI perkelti save į bet kokią aplinką

Use @video1 as the base for performance, lip movement, facial expressions, gestures, body movement and timing. Use @image1 for the new outfit and location. Use @image2 as the wide view of the same room. Use @image3 as the side view of the same room. Replace the background and outfit from @video1 completely, from the very first frame to the last frame. The whole video takes place in the room from @image1, @image2 and @image3, and she wears the outfit from @image1 in every shot. The original room and clothes from @video1 never appear. Keep the exact same person and face from @video1: identical facial features, expressions, micro-expressions, eye movement, skin texture, pores and fine lines. Lip movements must match @video1 and the original audio exactly, perfectly synced with every word. Keep the original voice and audio from @video1 unchanged. Hand and body movements must follow @video1 exactly, with the same timing. The room and outfit stay the same in every shot: same furniture, layout, sea view and lighting. Camera: 0–4.8s — starts at the same framing as @image1, smooth cinematic pull-out, the camera clearly moves back 4.8–9.4s — medium shot from the right side of the frame, side angle as in @image3, stable static camera 9.4–16.7s — front angle, starts as an extreme wide shot showing the whole room as in @image2, then a continuous slow push-in until it reaches the same framing as @image1 16.7–19.3s — close-up of the face, 3/4 side angle, camera slightly above eye level, static camera 19.3–23.3s — smooth drone-like orbit: the camera starts on the right side and circles around her to the left side, keeping her centered in frame Photorealistic and natural, like a real camera recording. Real skin texture, no smoothing, no beauty filter, no plastic AI look. Bright natural daylight from the windows, consistent in every shot. No face drift. No identity change. No flickering. No lip-sync errors. No distorted hands or extra fingers.