Mid-Autumn girl holding Cantonese-style mooncake short video AI Image Prompt

9:16 vertical short video, 15 seconds, hyper-realistic 3D rendering, cinematic soft warm light, 4K HD, clean image with no grain. An East Asian little girl over 1 year old, round chubby cheeks, big bright eyes, sweet smile, cute twin ponytails, wearing a light-colored small flower hair accessory, dressed in a light-colored new Chinese-style Hanfu, with a small rabbit embroidery pattern on the garment. The little girl holds an oversized, exquisite, clearly three-dimensional patterned Cantonese-style mooncake with both hands (with '中秋快乐' printed on the mooncake), dark blue hazy blurred night background, eye-level shot, strong Mid-Autumn Festival atmosphere, delicate character details, realistic skin texture. The little girl speaks naturally facing the camera, lip movements precisely match every syllable of each line, lips open and close slightly and naturally, expression changes gently with the lines, eyes look directly at the camera, movements are coherent and smooth, no exaggerated performance, face not distorted, does not open mouth wide, no long freeze. Gentle and clear childlike milky voice, moderate speaking pace. 0-2s: Holds the oversized mooncake steadily with both hands, body leans slightly forward, head gently sways left and right with small amplitude, arms gently lift toward the camera, line: 'The Mid-Autumn Festival is here, good fortune has arrived at the door!' 2-5s: Hugs the mooncake tightly, little hands rest lightly on the mooncake surface, head turns slightly to one side, then turns back to face the camera, line: 'Wishing uncles, aunts, grandpas, and grandmas good health and long life and abundant blessings,' 5-10s: Both hands gently offer the mooncake forward toward the screen, shoulders sway slightly, head nods twice, line: 'Also wishing friends in front of the screen a full moon and reunion, may all things go as you wish!' No subtitles, no watermark, no text of any kind, no logo, no UI buttons.