New미야자키 하야오 감독의〈On Your Mark〉를 AI로 다시 만들었습니다 — 제작 과정과 프롬프트 전부 공개
UGC·숏폼Higgsfield · 시댄스 2.5
Prompt10,476자
Create a 30-second first-person POV video: a Shibuya intersection in Tokyo during rush hour, amidst heavy rain. Through the eyes of a young woman: tested by a gust of wind and whizzing traffic on the crowded curb; surging into the intersection with the flow of people; dodging a cyclist crossing the road; brushing past a small dog being held by a stranger; sprinting across the last few meters with the last wave of people as the traffic light begins to flash; turning back to look at the glowing intersection after reaching the other side. Hyper-realistic, photorealistic, atmospheric, and calmly observable documentary style—absolutely no humor, absolutely no staged shots. Scene Setting: Dusk in the blues, torrential rain, gusts of wind sweeping across the intersection. The omnidirectional intersection is vibrant: hundreds of pedestrians hold transparent plastic umbrellas, giant video billboards and building LED screens splash ever-changing lights onto the wet asphalt, taxis and buses idle at the stop line with wipers swishing back and forth, steam rises from street drain grates, and the reflections of every light source cast long streaks of light in the puddles. All screen content consists of abstract motion graphics and unrecognizable generic signs; no brands, no readable text, no identifiable IP. The rain is physically real: dense, straight rain streaks are swept diagonally by gusts of wind in front of the bright screens, raindrops explode on umbrellas, water droplets bounce off shoulders, and water flows continuously in the drains. POV locked, one-shot, no editing, eye level at the natural eye level of a 162cm tall woman. Only her hands, wrists, and forearms appear in the shot; her face and body are never shown, not even in reflections in shop windows, puddles, or car body panels. Light to medium skin tone, slender and realistically shaped hands, short, unpainted nails, a thin silver ring on the middle finger of the right hand, and a dark green raincoat with beige inner cuffs. The right hand holds the curved handle of a transparent plastic umbrella; the wet edge of the umbrella and the ribs remain visible at the top of the frame throughout the entire video. All these details remain completely consistent throughout. First Frame and Route: The first frame immediately places the character in the crowded waiting crowd at the curb, a row away from the curb: the foreground is filled with the back, shoulders, and umbrella; a red pedestrian traffic light glows on the opposite side of the road; a large screen shines brightly through the rain; a bus is driving along the stop line, creating a fan-shaped wave of water. No empty shots at the beginning, no delayed entrance. Route: Curb (withstanding the wind and avoiding the splashes) – surge into the intersection with the crowd – abruptly stop to avoid a bicycle crossing – weave between two streams of umbrellas in the center of the intersection, cross a deep puddle – drift to the right past a small dog – traffic lights flash, jog with the last of the crowd – leap onto the opposite curb – stop – turn 180 degrees to face the intersection. Action 0–6 seconds – She stands in the dense crowd, breathing; raindrops tap on her umbrella a few centimeters above the camera. A bus drives past close by, the water splashed by its tires hitting the curb – she twists her entire upper body to the left to avoid it, water still splashes onto her right sleeve, the sleeve instantly becoming darker, the image following her body's twist before returning to center. A sudden gust of wind swept into the intersection: the umbrella billowed outwards, half of the ribs folded upwards and outwards—she immediately lowered the handle with her right hand, raised her left hand to grasp one of the ribs, and with a quick flick of her wrist, the umbrella snapped back to its original shape, water droplets flying off. The traffic light changed, and the intersection's warning music began. 6–14 seconds — The entire crowd moved simultaneously, and she stepped off the curb, her head movement driven by her actual heel-to-toe steps, breathing, and body momentum. She had only taken three steps when a cyclist wearing a raincoat darted across the gap in the traffic on her right—the bicycle bell rang sharply, the headlight casting a shimmering streak of light on the wet asphalt—she stopped abruptly, leaned back half a step, the umbrella suddenly slumped, the cyclist grazing the edge of the umbrella as he passed, his rear wheel kicking up a trail of water. She exhaled and rejoined the crowd. The umbrellas brushed past each other within centimeters, their surfaces making a soft, damp rustling sound against hers; she slightly raised her umbrella and slipped under a taller one. A deep puddle lay across her path in the center of the intersection—she quickened her pace, leaping across it in a short bound, the camera noticeably dropping and bouncing upon landing, someone behind her kicking straight into the water, splashing water everywhere. 14–22 seconds — Two streams of people crossed diagonally; she slowed her pace, stepping aside to let a man carrying a briefcase hurry past. Then, from the opposite direction: a young woman in a yellow raincoat, a small, brown puppy tucked into the zipper of her shirt, only its wet head visible, its ears flattened by the drizzle. The POV focused on the puppy; her pace slowed; as the two passed side-by-side, the puppy turned its head to meet the camera's gaze, its eyes following the lens for a full frame before the flow of people dispersed. Her left hand briefly rises into the frame, making an unconscious little wave. 22–30 seconds — The pedestrian traffic light across the street begins to flash, indicating a quickening of the music. The flow of umbrellas around her accelerates simultaneously, and she transitions from a brisk walk to a jog—the camera's vertical movement becomes larger and faster, her breathing becomes heavier, the umbrellas above her head bounce with her steps, the rhythm of raindrops hitting the umbrellas is disrupted by her running, the umbrellas and shoulders of runners on either side quickly sweep past the edge of the frame, and streaks of light on the wet asphalt flow backward beneath her feet. With the last three steps, she leaps onto the steps, the noticeable impact causing the image to lurch forward before stabilizing, stopping at the edge of the awning, the rise and fall of her chest clearly felt. She turns around: the entire intersection—a slowly flowing river of glowing umbrellas in the rain, illuminated by the lights of the four screens—fills the final few frames, the traffic light turns red, the last few runners splash past, and the intersection gradually empties behind them. Optics and Camera: A fixed 47° diagonal field of view throughout, natural human perspective, straight architectural lines, and clearly readable depth of field. The camera's performance is like her head: weighted inertia, restrained shoulder sway, natural head turns, stable delay after each step, subtle up-and-down movements while walking; the image suddenly tilts forward and rebounds when stopping abruptly, quickly flicks past and corrects itself when twisting to avoid waves, and the amplitude and frequency of undulations during jogging are realistically amplified—all the vigorous movement comes from the body, not camera effects. There is no sense of floating, no gimbal-like smoothness, no fisheye effect, no lens changes, and no zoom. There is a brief autofocus breath between the nearby umbrella, the dog, and the distant large screen, which quickly recovers. Occasionally, a single raindrop hits the lens and slides off; when the gaze shifts from the dimly lit crowd to the dazzling large screen, there is a slight fluctuation in exposure. Physics: Hyper-realistic anatomy, materials, and contact. The umbrella's folding is physically accurate: in a gust of wind, half the ribs flip upwards and outwards, the plastic canopy stretches into a reverse arc, and when pulled back by a flick of the wrist, the ribs spring back one by one, water droplets are flung out tangentially—the umbrella never tears, never slips from your hand, and never deforms or distorts. The water waves on the bus have realistic mass and parabolic trajectory; when they hit your sleeves, the fabric darkens, and water droplets roll off. The bicycle has a realistic sense of speed and tilt, with the rear wheel's water trail following a centrifugal direction. Hundreds of pedestrians use realistic walking cycles and avoid each other; no one's body or umbrella weaves into each other. The cushioning of the knees upon landing from a jump is transmitted to the screen. Rain follows gravity and wind direction; water splashes have mass, puddles ripple correctly, and wet asphalt maintains physically accurate mirror reflections, breaking with each footstep. The puppy moves with the realistic weight of a small animal—blinking in the drizzle, its head turning with natural inertia. Her fingers flex and extend on the umbrella handle, the tendons clearly visible. Visual Lock-On: Hyper-realistic Digital Cinematic Quality: Clean, modern sensor visuals, natural colors, no VHS, no stylized color grading, no neon oversaturation—the screens are bright because they are naturally bright, not due to filters. Soft, blue-toned ambient light combined with the mixed screen light from wet surfaces. Stable real-time 24fps, each frame independent and continuous; no slow motion, no speed changes, no stuttering, no ghosting, no time trailing. Audio: Only naturally synchronized sounds: the dense sound of rain on the plastic umbrella overhead, the layered sound of rain on the street, gusts of wind whistling through the gaps between buildings, the low rumble of a bus passing by and the splashing of water, the rapid ringing of bicycle bells and chains, the Shibuya intersection's traffic cue music (the tempo quickens at the end), hundreds of wet footsteps, the increased breathing and footsteps during a jog, the sound of umbrella ribs folding and springing back, the rustling of umbrella fabric against each other, the hiss of tires slicing through water, the distant noise of traffic and crowds, and the soft whimper of a dog passing by. Her vocal responses are limited to two instances: a small exhale as she dodges the bicycle, and a soft, amused exhale upon seeing the puppy. There is no narration, dialogue, music, subtitles, logo, or watermark.
Added 2026.08.24
Prompt
Create a 30-second first-person POV video: a Shibuya intersection in Tokyo during rush hour, amidst heavy rain. Through the eyes of a young woman: tested by a gust of wind and whizzing traffic on the crowded curb; surging into the intersection with the flow of people; dodging a cyclist crossing the road; brushing past a small dog being held by a stranger; sprinting across the last few meters with the last wave of people as the traffic light begins to flash; turning back to look at the glowing intersection after reaching the other side. Hyper-realistic, photorealistic, atmospheric, and calmly observable documentary style—absolutely no humor, absolutely no staged shots. Scene Setting: Dusk in the blues, torrential rain, gusts of wind sweeping across the intersection. The omnidirectional intersection is vibrant: hundreds of pedestrians hold transparent plastic umbrellas, giant video billboards and building LED screens splash ever-changing lights onto the wet asphalt, taxis and buses idle at the stop line with wipers swishing back and forth, steam rises from street drain grates, and the reflections of every light source cast long streaks of light in the puddles. All screen content consists of abstract motion graphics and unrecognizable generic signs; no brands, no readable text, no identifiable IP. The rain is physically real: dense, straight rain streaks are swept diagonally by gusts of wind in front of the bright screens, raindrops explode on umbrellas, water droplets bounce off shoulders, and water flows continuously in the drains. POV locked, one-shot, no editing, eye level at the natural eye level of a 162cm tall woman. Only her hands, wrists, and forearms appear in the shot; her face and body are never shown, not even in reflections in shop windows, puddles, or car body panels. Light to medium skin tone, slender and realistically shaped hands, short, unpainted nails, a thin silver ring on the middle finger of the right hand, and a dark green raincoat with beige inner cuffs. The right hand holds the curved handle of a transparent plastic umbrella; the wet edge of the umbrella and the ribs remain visible at the top of the frame throughout the entire video. All these details remain completely consistent throughout. First Frame and Route: The first frame immediately places the character in the crowded waiting crowd at the curb, a row away from the curb: the foreground is filled with the back, shoulders, and umbrella; a red pedestrian traffic light glows on the opposite side of the road; a large screen shines brightly through the rain; a bus is driving along the stop line, creating a fan-shaped wave of water. No empty shots at the beginning, no delayed entrance. Route: Curb (withstanding the wind and avoiding the splashes) – surge into the intersection with the crowd – abruptly stop to avoid a bicycle crossing – weave between two streams of umbrellas in the center of the intersection, cross a deep puddle – drift to the right past a small dog – traffic lights flash, jog with the last of the crowd – leap onto the opposite curb – stop – turn 180 degrees to face the intersection. Action 0–6 seconds – She stands in the dense crowd, breathing; raindrops tap on her umbrella a few centimeters above the camera. A bus drives past close by, the water splashed by its tires hitting the curb – she twists her entire upper body to the left to avoid it, water still splashes onto her right sleeve, the sleeve instantly becoming darker, the image following her body's twist before returning to center. A sudden gust of wind swept into the intersection: the umbrella billowed outwards, half of the ribs folded upwards and outwards—she immediately lowered the handle with her right hand, raised her left hand to grasp one of the ribs, and with a quick flick of her wrist, the umbrella snapped back to its original shape, water droplets flying off. The traffic light changed, and the intersection's warning music began. 6–14 seconds — The entire crowd moved simultaneously, and she stepped off the curb, her head movement driven by her actual heel-to-toe steps, breathing, and body momentum. She had only taken three steps when a cyclist wearing a raincoat darted across the gap in the traffic on her right—the bicycle bell rang sharply, the headlight casting a shimmering streak of light on the wet asphalt—she stopped abruptly, leaned back half a step, the umbrella suddenly slumped, the cyclist grazing the edge of the umbrella as he passed, his rear wheel kicking up a trail of water. She exhaled and rejoined the crowd. The umbrellas brushed past each other within centimeters, their surfaces making a soft, damp rustling sound against hers; she slightly raised her umbrella and slipped under a taller one. A deep puddle lay across her path in the center of the intersection—she quickened her pace, leaping across it in a short bound, the camera noticeably dropping and bouncing upon landing, someone behind her kicking straight into the water, splashing water everywhere. 14–22 seconds — Two streams of people crossed diagonally; she slowed her pace, stepping aside to let a man carrying a briefcase hurry past. Then, from the opposite direction: a young woman in a yellow raincoat, a small, brown puppy tucked into the zipper of her shirt, only its wet head visible, its ears flattened by the drizzle. The POV focused on the puppy; her pace slowed; as the two passed side-by-side, the puppy turned its head to meet the camera's gaze, its eyes following the lens for a full frame before the flow of people dispersed. Her left hand briefly rises into the frame, making an unconscious little wave. 22–30 seconds — The pedestrian traffic light across the street begins to flash, indicating a quickening of the music. The flow of umbrellas around her accelerates simultaneously, and she transitions from a brisk walk to a jog—the camera's vertical movement becomes larger and faster, her breathing becomes heavier, the umbrellas above her head bounce with her steps, the rhythm of raindrops hitting the umbrellas is disrupted by her running, the umbrellas and shoulders of runners on either side quickly sweep past the edge of the frame, and streaks of light on the wet asphalt flow backward beneath her feet. With the last three steps, she leaps onto the steps, the noticeable impact causing the image to lurch forward before stabilizing, stopping at the edge of the awning, the rise and fall of her chest clearly felt. She turns around: the entire intersection—a slowly flowing river of glowing umbrellas in the rain, illuminated by the lights of the four screens—fills the final few frames, the traffic light turns red, the last few runners splash past, and the intersection gradually empties behind them. Optics and Camera: A fixed 47° diagonal field of view throughout, natural human perspective, straight architectural lines, and clearly readable depth of field. The camera's performance is like her head: weighted inertia, restrained shoulder sway, natural head turns, stable delay after each step, subtle up-and-down movements while walking; the image suddenly tilts forward and rebounds when stopping abruptly, quickly flicks past and corrects itself when twisting to avoid waves, and the amplitude and frequency of undulations during jogging are realistically amplified—all the vigorous movement comes from the body, not camera effects. There is no sense of floating, no gimbal-like smoothness, no fisheye effect, no lens changes, and no zoom. There is a brief autofocus breath between the nearby umbrella, the dog, and the distant large screen, which quickly recovers. Occasionally, a single raindrop hits the lens and slides off; when the gaze shifts from the dimly lit crowd to the dazzling large screen, there is a slight fluctuation in exposure. Physics: Hyper-realistic anatomy, materials, and contact. The umbrella's folding is physically accurate: in a gust of wind, half the ribs flip upwards and outwards, the plastic canopy stretches into a reverse arc, and when pulled back by a flick of the wrist, the ribs spring back one by one, water droplets are flung out tangentially—the umbrella never tears, never slips from your hand, and never deforms or distorts. The water waves on the bus have realistic mass and parabolic trajectory; when they hit your sleeves, the fabric darkens, and water droplets roll off. The bicycle has a realistic sense of speed and tilt, with the rear wheel's water trail following a centrifugal direction. Hundreds of pedestrians use realistic walking cycles and avoid each other; no one's body or umbrella weaves into each other. The cushioning of the knees upon landing from a jump is transmitted to the screen. Rain follows gravity and wind direction; water splashes have mass, puddles ripple correctly, and wet asphalt maintains physically accurate mirror reflections, breaking with each footstep. The puppy moves with the realistic weight of a small animal—blinking in the drizzle, its head turning with natural inertia. Her fingers flex and extend on the umbrella handle, the tendons clearly visible. Visual Lock-On: Hyper-realistic Digital Cinematic Quality: Clean, modern sensor visuals, natural colors, no VHS, no stylized color grading, no neon oversaturation—the screens are bright because they are naturally bright, not due to filters. Soft, blue-toned ambient light combined with the mixed screen light from wet surfaces. Stable real-time 24fps, each frame independent and continuous; no slow motion, no speed changes, no stuttering, no ghosting, no time trailing. Audio: Only naturally synchronized sounds: the dense sound of rain on the plastic umbrella overhead, the layered sound of rain on the street, gusts of wind whistling through the gaps between buildings, the low rumble of a bus passing by and the splashing of water, the rapid ringing of bicycle bells and chains, the Shibuya intersection's traffic cue music (the tempo quickens at the end), hundreds of wet footsteps, the increased breathing and footsteps during a jog, the sound of umbrella ribs folding and springing back, the rustling of umbrella fabric against each other, the hiss of tires slicing through water, the distant noise of traffic and crowds, and the soft whimper of a dog passing by. Her vocal responses are limited to two instances: a small exhale as she dodges the bicycle, and a soft, amused exhale upon seeing the puppy. There is no narration, dialogue, music, subtitles, logo, or watermark.
구글 계정으로 3초, 따로 작성할 양식은 없습니다.
처음이시면 이 버튼으로 가입까지 함께 진행됩니다.