
Use @video_1 as the STRICT reference for camera, choreography, movements, timing, environment and train transitions. Replace the original person with my character from @image_1–@image_5. All five images show the SAME person wearing five different outfits. IDENTITY LOCK: Preserve the exact same face, facial features, hairstyle, body proportions and appearance throughout the entire video. Never change the person's identity between outfit transitions. OUTFIT LOCK: @image_1 = LOOK 1 @image_2 = LOOK 2 @image_3 = LOOK 3 @image_4 = LOOK 4 @image_5 = LOOK 5 Reproduce each outfit exactly, including clothing, shoes and accessories. STRICT 1:1 REPLICATION: Copy the original camera angle, framing, background, movements, gestures, poses, transition timing, train direction and speed. Keep the camera stationary. Each passing train must completely hide the character before revealing the next outfit. Preserve the exact original sequence and rhythm. CRITICAL — CLEAN BACKGROUND: The original reference contains a large text graphic behind the character at the beginning. DO NOT COPY THIS GRAPHIC. Replace the entire area behind the character with the natural metro station background. The opening frame must show ONLY the metro platform and my character. Do not generate any letters, titles, signs or graphic panels behind the character. ABSOLUTE NO-WATERMARK RULE: NO WATERMARKS. NO ORIGINAL TEXT. NO OPENING TITLE. NO BACKGROUND GRAPHICS. NO USERNAMES. NO CREATOR HANDLES. NO LOGOS. NO SUBTITLES. NO SOCIAL MEDIA UI. NO INSTAGRAM OR TIKTOK ICONS. NO END CARDS. These restrictions apply to EVERY FRAME, including the first and last. The reference is used for physical action and camera timing, never for graphical overlays. OUTPUT: Ultra-photorealistic live-action. Vertical 9:16. Original video duration. Five accurate outfits. One consistent character identity. Exact train-wipe transitions. Only the metro environment, my character and the passing train. Completely clean footage.