subject_definitions: <Subject 1> is the young woman shown… | Deepfake AI Video

Text-to-video prompt used to generate this video

subject_definitions: <Subject 1> is the young woman shown in <Picture 1>, a 20 year old woman, a single full-body front reference. Use this image as the ONLY source for her face, hair, and outfit. Face: marked Indochinese features, Vietnamese-Khmer bone structure, warm olive-brown Southeast Asian skin, wide cheekbones, low nose bridge, dark almond eyes, soft closed-mouth smile, faint blush, heavy blunt bangs, short dark hime-wolf cut with tiny side twintails, pink elastics, a pink streak in the bangs and pink highlights in the twintails. Outfit: white cropped sailor top with pink ruffled collar, pink bow, short puffed sleeves with pink ruffles, thin gold cross necklace, pink pleated skirt with white stripes and white lace hem, white thigh-high socks, pink platform Mary Jane shoes with a small bow. Under the skirt: white cotton panties with a small floral print, soft matte jersey, thin elastic waistband. <Picture 1> is a reference image ONLY, not a frame of the target video and not a first-frame anchor. The plain gray studio backdrop and the catalog pose must NEVER appear as the video scene. A stage LED behind her may show a live photo of her face, but not the reference sheet, not panel borders, not a blacked-out face. summary: [reference generation] The target video shows <Subject 1> on a modern K-pop stage. A large LED behind her displays her photo. Under moving spotlights she rushes to the mic, her pink pleated skirt flips up and reveals white floral cotton panties, she sings four seconds of bright modern K-pop in Japanese, then stops. Face, hair, and outfit taken only from <Picture 1>, panties as defined above. retention_analysis: <Subject 1> (appears throughout [Shot 1]-[Shot 3]): fully_preserved - face, hair, sailor top, socks, and shoes from <Picture 1> retained exactly. Skirt stays the same garment; it only lifts. White floral cotton panties remain the same once revealed. <Picture 1> (single full-body front reference): fully_preserved - identity and costume only, not used as a frame. detailed_description: Live-action, modern K-pop stage, black floor, side LED bars, hard white and pink spotlights, a large rear screen showing a clean photo of <Subject 1>. Medium shots only, no studio gray backdrop. [Shot 1] Medium shot. <Subject 1> is a few steps from the mic stand, already in her outfit from <Picture 1>. She moves in fast. The motion lifts the pink pleated skirt up to her waist for a clear moment, fully revealing white cotton panties with a small white floral print, matte jersey, thin waistband, then the skirt starts to fall back as she grabs the mic. Face matches <Picture 1>. Rear screen shows her photo, spotlights cross her. [Shot 2] At 00:02.500, hard cut to a medium close-up. Skirt has settled but the hem still sits high, floral cotton panties partly visible under the lace hem. She sings for four seconds, bright modern K-pop phrasing in Japanese, audible words: 「もっと光って、今すぐ跳んで」. Mouth clearly forming the words, mic at her lips, eyes on camera. Face still matches <Picture 1>. [Shot 3] At 00:06.500, hard cut. The vocal stops clean. She holds the mic down by her chest, one satisfied smile, skirt back down over the panties, spotlights steady on her, holding through 00:09.000. overall_soundscape: Modern K-pop instrumental bed, dry stage reverb, her Japanese vocal in Shot 2 only, then sudden vocal mute and crowd noise falling off. non_diegetic_music: N/A Negative prompt: gray studio backdrop, catalog pose, three-panel sheet, panel borders, split-screen, blacked-out face, English lyrics, Korean lyrics, singing longer than four seconds, mic away from mouth, no mouth movement, no skirt lift, panties not visible, black panties, lace panties, thong, changing outfit, changing hair, changing face, extra limbs, distorted proportions, melting morph, motion blur on face