IMAGE-TO-VIDEO (use the provided night Madrid fashion photo as the first frame / identity, outfit, and lighting reference) Create an ultra-realistic vertical 4:5 video, 6–8 seconds. Keep the same night street location in Madrid, the same lighting (warm street lamps + soft bokeh), and the same outfit: short elegant evening dress, long open coat, stiletto heels, rings visible, perfect manicure. Preserve the same framing and camera angle as the reference. Camera: - Looks like a professional handheld or lightly stabilized camera at night (smooth but still natural). - Subtle forward tracking as she walks, no sudden pans, no zoom jumps, no reframing. Action & performance: - Violeta walks down the street with strong, stylish runway-like confidence (but still natural for real life). - She starts with a serious, composed expression. - While walking, she looks around and glances up at the buildings and facades beside her (curious, admiring mood). - Around the middle of the clip, her expression softens and she transitions into a gentle, natural smile (not exaggerated), then keeps walking. - Coat moves elegantly with each step; heels land naturally; hair moves subtly. Realism & quality: - Natural gait in heels (no sliding feet), correct body mechanics. - Hands and fingers remain correct; rings and manicure stay consistent. - No face morphing, no flicker, no warping, no “melting” background. - Keep night lighting consistent and reflections stable if pavement is slightly wet.