A highly specific prompt for Seedance 2.0 that replicates K-pop dance movements using depth video references and image inputs for character and background…
Korean K-pop MV style, the background is replaced with a neon street in Seoul at night. The appearance and clothing of the female lead strictly refer to @image1, and the appearance and clothing of the male lead strictly refer to @image2. All movements, positions, body postures, interaction processes, and rhythms of the male and female leads strictly refer to @depthvideo, completely replicated without modification. At the beginning of the video, the female lead appears first and naturally speaks: "I will say to him in Korean: 오빠, 정신 차려!" The first half "I will say to him in Korean" uses Chinese, and the second half "오빠, 정신 차려" uses natural Korean pronunciation, with a slightly hurried and reminding tone, and the lip shape is accurately synchronized with both languages. Subsequent character appearance orders and all movements continue to strictly follow @depthvideo. Keep the appearance of the two characters consistent throughout, with stable facial features and natural, smooth movements. 9:16 vertical screen, one-shot till the end, no subtitles, no text, no watermarks, no character misalignment, role swapping, or limb deformation.