提示词 / 授权来源
Prompt 153 · @john_my07
@john_my07 · 文本生成视频
Create a 30-second ultra-realistic live-action-style video featuring the same clearly adult Korean woman throughout. The concept, locations, actions, and visual language must be completely different from the previous residential home-video concept. MAIN SUBJECT: Young Korean woman in her early 20s, natural everyday appearance, warm and approachable personality. Black wavy hair with wispy bangs, naturally styled. She wears a fitted dark navy knit top, relaxed beige wide-leg trousers, white minimalist sneakers, and a small black shoulder bag. Minimal makeup, realistic skin texture, subtle facial asymmetry and natural imperfections. Maintain EXACTLY the same face, identity, body proportions, hairstyle, clothing, accessories, and appearance in every shot. CONCEPT: A spontaneous modern-day afternoon in Seoul — she is meeting a friend, navigating a busy contemporary neighborhood, grabbing a drink, and catching a brief sunset moment. It should feel like genuine footage captured casually by a friend on a modern smartphone, NOT a commercial, music video, fashion shoot, or AI-generated sequence. VISUAL STYLE: Extremely photorealistic modern smartphone footage. Natural 2026 smartphone-camera appearance with realistic computational photography, accurate skin tones, subtle HDR, natural dynamic range, realistic exposure, fine sensor detail, mild digital sharpening, authentic smartphone motion blur, and occasional autofocus adjustment. Use ordinary handheld smartphone recording with imperfect but believable framing. Slight hand movement and natural walking shake. Occasional tiny focus adjustments. Realistic rolling shutter during quick movement. No artificial cinematic camera moves, no impossible transitions, no excessive bokeh, no dramatic lighting, no film grain, no stylized color grading, no slow motion. ENVIRONMENT: Modern Seoul neighborhood with contemporary apartment buildings, glass storefronts, pedestrian crossings, buses, bicycles, small cafés, convenience stores, Korean signage, parked cars, people naturally going about their day, street trees, traffic lights, and realistic urban clutter. Everything should have believable scale, reflections, shadows, perspective, and physical interaction. 00:00–00:05 The video starts naturally with the phone already recording. She exits a modern subway station into bright afternoon sunlight. She squints slightly and adjusts her hair while checking her phone. The camera operator follows a few steps behind. A bus passes in the background, briefly changing the lighting and creating realistic motion blur. 00:05–00:10 She crosses a busy but ordinary pedestrian intersection. The camera remains at normal human height. She waits for the signal, glances at the approaching traffic, then notices her friend filming and gives a quick amused smile. A cyclist passes behind her without interacting with her. 00:10–00:15 They enter a small modern café. The camera follows through the doorway. She orders an iced drink at the counter, briefly points at something on the menu, pays with her phone, and steps aside. Reflections appear naturally on the glass door and counter. Background customers remain ordinary and slightly out of focus because of distance, not artificial blur. 00:15–00:20 Outside again. She takes the first sip while walking and immediately makes a subtle satisfied expression. She checks a message on her phone while walking, then nearly walks past her friend because she is distracted. She looks up, laughs quietly, and continues. 00:20–00:25 They arrive beside a modern riverside pedestrian path. She rests her drink on a railing and looks across the water. The camera moves naturally to her side. Wind gently moves her hair and clothing. Distant pedestrians, cyclists, bridges, buildings, and water remain physically consistent in the background. 00:25–00:30 The sun is beginning to set. She turns back toward the camera and says casually, “Wait, look at that.” She points toward the skyline. The camera instinctively turns away from her toward the sunset, briefly overexposing before the smartphone automatically corrects exposure. The final frame settles on the realistic skyline and water while her voice and footsteps remain audible. AUDIO: Authentic location audio only. No music and no cinematic sound effects. Include realistic Seoul street ambience, traffic, pedestrian crossing sounds, footsteps, distant conversations, subway-station ambience, café sounds, refrigerator/air-conditioning hum, cup and ice sounds, bicycle bells, wind, and distant city noise. Dialogue should sound naturally captured by a smartphone microphone, with realistic changes in volume depending on distance. STRICT REALISM / ANTI-AI REQUIREMENTS: Every human must have anatomically correct hands and exactly five fingers per hand. NEVER generate extra fingers, duplicated hands, fused fingers, missing limbs, duplicated people, distorted faces, malformed ears, warped teeth, floating objects, impossible reflections, or physically impossible body positions. Hands must interact correctly with phones, cups, doors, clothing, and other objects. If she holds the drink, only ONE believable hand should occupy each grip position. Her phone must remain a normal rectangular smartphone with consistent dimensions. Maintain strict object continuity: the same drink, phone, bag, clothing, accessories, and hairstyle must remain consistent throughout. Objects cannot spontaneously appear, disappear, duplicate, change color, or change shape. Human motion must obey real-world physics. Walking, sitting, turning, drinking, touching objects, hair movement, clothing movement, reflections, shadows, glass surfaces, water, traffic, and background pedestrians must behave naturally. Faces must remain stable and recognizable from frame to frame. No identity drift, facial morphing, plastic skin, beauty-filter effect, unnatural eye movement, frozen expressions, or AI-perfect symmetry. Do not make every background person look identical. Background pedestrians should have natural variation in clothing, body shape, movement, and appearance. Do NOT make the video look like an AI commercial. No perfect cinematic compositions. No excessive shallow depth of field. No impossible camera movement. No hyper-saturated colors. No glossy skin. No artificial lens flares. The final result should be indistinguishable from authentic footage recorded by a real person on a high-end modern smartphone during an ordinary day in Seoul.