映画のような AI ビデオ生成Seedance 2.5

Create a 15-second ultra-photorealistic documentary clip in 16:9 that feels completely real, as if filmed by a wildlife photographer during an actual expedition. The footage must be indistinguishable from authentic DSLR field footage, not a movie scene, not a commercial, and not CGI. Camera: Handheld Sony FX3 with a 24–70mm lens. Natural operator movement, slight breathing, tiny framing corrections, realistic autofocus adjustments, authentic exposure shifts, no cinematic camera moves. 0–4 seconds A wildlife photographer quietly walks along a remote mountain lake at blue hour. Gentle wind ripples the water. Birds can be heard in the distance. The camera casually scans the shoreline with no dramatic music or editing. 4–8 seconds Without warning, a small section of the lake begins glowing with thousands of tiny blue lights beneath the surface. The camera hesitates in surprise, zooms in naturally by hand, and struggles to keep focus as the operator whispers, "What is that?" 8–12 seconds The glowing area rapidly expands across the lake. Millions of bioluminescent organisms swirl beneath the water, creating hypnotic patterns that look completely natural. The camera shakes slightly as the operator slowly backs up in disbelief. 12–15 seconds A single fish jumps out of the glowing water, sending brilliant blue ripples across the lake. The camera instinctively follows the splash, then holds on the surreal glowing lake for two seconds before the clip ends naturally. Visual Style: Real wildlife documentary, authentic field footage, National Geographic realism, BBC Earth style, Sony FX3, full-frame DSLR look, true optical bokeh, realistic skin tones, physically accurate water reflections, natural dynamic range, subtle sensor noise, gentle film grain, no CGI look, no AI artifacts, no oversaturated colors. Audio: Only authentic environmental sound wind, birds, water, footsteps, distant insects, subtle human breathing, and one quiet whispered reaction.

A young woman leans softly on a metal railing beside a quiet lake at golden hour. She rests her head on her folded arms and looks calmly into the camera. Her long strawberry blonde hair moves gently in the breeze. Warm sunlight glows around her hair and face. Her white lace summer dress moves slightly with the wind. The camera slowly pushes closer while keeping her eyes sharp. She blinks naturally and gives a very subtle soft smile. Green trees, distant water, and shoreline stay blurred with creamy bokeh. Natural skin texture, realistic hair and fabric, shallow depth of field, soft cinematic lighting, smooth natural motion. No text, no dialogue, no artificial HDR.

The subject is based on @ image1. Face shape, features, skin tone, damp slicked-back short black hair, black satin cami dress, layered silver necklace, and silver earrings must match @ image1 exactly — no facial distortion allowed. Ignore @ image1's white studio background, frontal standing pose, and bottom text. Even if the image quality is rough, the facial features should remain clearly recognizable. 9:16 vertical format. The entire clip should have the low-quality look of an old iPhone everyday video. Rough noise and compression artifacts persist throughout. Early summer dawn on a rooftop — the sun already harsh, the air still and humid. 0-4 seconds: Hanging wet laundry on a rooftop clothesline, morning light harsh and already bright, wiping sweat from her forehead with her wrist as she reaches up to clip a sheet. 4-8 seconds: Ducking back into the stairwell shade, leaning against the doorframe — a long exhale of relief from the shadow's coolness. 8-12 seconds: Standing selfie angle against the backdrop of damp hanging laundry, low and slightly husky voice saying "…doing laundry before it gets too hot. Yeah." (Korean: "…너무 더워지기 전에 빨래하는 거예요. 네.") 12-15 seconds: Looking out at the hazy city skyline beyond the rooftop railing, a hanging sheet rippling in the wind behind her, the frame trembling slightly, final frame freezes. Camera: 1–1.5 second loose/casual cuts. Old iPhone handheld vertical shooting, slight body sway, autofocus hunting in changing light, harsh sun causing exposure flicker between shade and direct light. Sound: No background music. Ambient city hum as base layer, fabric flapping in wind, distant traffic murmur, faint birdsong, voice close-up but slightly muffled. Style: Uncolored, low-quality phone video texture. No color grading, no film look, no cinematic look, no beauty filter, no CG texture. Preserve authentic skin texture (pores, flyaway hairs, sweat sheen). No merging shots or skipping scenes. No on-screen text.

Main Subject: Young Korean woman, early 20s. Natural appearance. Black wavy hair loosely tied into a slightly messy ponytail with wispy bangs. Light grey zip hoodie worn open over a faded white tank top. Relaxed blue jeans. Black canvas sneakers. Natural skin texture, minimal makeup, ordinary everyday look. Consistent appearance throughout the video. Location: Authentic Korean residential neighborhood during a light summer rain shower. Narrow concrete alleys, low-rise houses, small gardens, potted plants, bicycles, utility poles, overhead cables, wet pavement, puddles, rain-darkened walls, leafy trees. Camera Style Early-2000s consumer DV camcorder. Friend casually filming. Heavy handheld shake. Imperfect framing. Subject frequently drifts toward frame edges. Late reframing. Autofocus hunting. Lens breathing. Exposure pumping. Motion blur. Rolling shutter. Digital compression artifacts. Slight sensor noise. Soft detail. Faded colors. Low contrast. Occasional blown highlights. No stabilization. No cinematic movements. No modern color grading. The camera operator behaves like a real friend with no filmmaking experience. 00:00–00:02 Standing beneath a small apartment awning while rain falls across the street. She leans forward and extends her hand into the rain to test how heavy it is. A passing vehicle somewhere off-frame splashes water nearby. She immediately laughs and pulls her hand back. The camera misses part of the reaction because the operator reframes too slowly. Rain droplets briefly appear on the lens. Autofocus hunts between her face and the rainy street. 00:02–00:04 She steps into the drizzle and starts walking down the narrow alley. The camera follows unevenly from behind. Her sneakers splash through shallow puddles. Wet strands of hair begin sticking to her cheek and neck. Wind moves her hoodie slightly. The operator almost loses her from frame while turning a corner. 00:04–00:06 Passing a row of potted plants outside a house. She casually brushes rainwater from a large leaf. The camera unexpectedly focuses on the leaf instead of her face. Exposure shifts as the scene moves from open sky into tree shade. She continues walking before camera catches up. 00:06–00:08 A neighborhood cat shelters beneath a parked bicycle. She crouches down briefly and says a few friendly words. The cat sniffs her hand. Camera focus repeatedly jumps between the bicycle wheel, the cat, and her face. One moment is partially blocked by the operator's movement. 00:08–00:10 Walking again. Rainwater drips from overhead branches onto her shoulder. She reacts naturally and laughs to herself. The camera is slightly too close and awkwardly reframes. A gust of wind moves nearby leaves and loose hair strands. 00:10–00:12 Near a small intersection. She pauses while waiting for a scooter to pass in the distance. The camera drifts toward the wet road before finding her again. She wipes rain from her forehead using her sleeve. Her clothing shows subtle dampness from the weather. 00:12–00:15 Walking away down the rain-darkened street. The camera trails several steps behind. She suddenly realizes she is still being filmed. She glances back over her shoulder. A small genuine smile appears. She looks away again and keeps walking. The operator lowers the camcorder slightly. Recording abruptly cuts to black mid-motion as if the camera was accidentally switched off. Audio (Natural only) Light rainfall. Water dripping from roofs and leaves. Footsteps on wet concrete. Distant neighborhood traffic. Occasional scooter passing far away. Birds sheltering in trees. Rustling leaves. Faint residential chatter. Cat sounds. Subtle wind. No music. No narration. No sound design. No added effects. Generation Priority Authenticity > Beauty Believability > Cinematics Natural Imperfections > Technical Perfection The final result should look indistinguishable from a real early-2000s camcorder scene.

Scene & Mood: A monochrome intimate studio beauty portrait with noir shadow and restrained movement. The mood is quiet, sculptural, and breath led: a close view of a bare shoulder, neck, jawline, nose, and softly parted lips, with the upper face held in darkness. Keep the image elegant, realistic, and tactile, not glossy or artificial. Frame Map: Extreme close up portrait composition. The lower half of the face sits just right of center, lips near the middle of frame, nose catching a narrow strip of light, eyes fully hidden under a deep shadow band. One bare shoulder fills the lower left foreground softly out of focus. Raised arms create dark curved shapes at the top and sides of frame. Background remains black gray negative space with shallow depth. Subject stays close to camera, intimate portrait distance, with no full face reveal. Subject Lock @ image1: @ image1 keeps the same face, hair, wardrobe, body shape, and silhouette throughout. Preserve the shadowed upper face, visible nose, lips, jawline, neck, bare shoulder, raised arm pose, soft skin texture, natural lips, and quiet editorial stillness. Maintain the same black and white tonal palette, hard directional light, fine grain, and sculpted contrast. No change in identity, no new wardrobe, no added jewelry, no makeup transformation, no glamour shine. Cross Frame Rules: Keep the subject locked in the same extreme close up framing and depth. Do not reveal the eyes. Do not swap pose, screen side, facial structure, skin tone, shoulder position, or raised arm silhouette. The face remains partially hidden by shadow, lips remain the emotional focus, and the body stays close to camera. No sudden camera jump, no morphing, no smoothing into plastic skin, no added text or graphic elements. Movement: 10 seconds total. The subject makes one slow inhale and exhale. Lips part slightly on the inhale, then soften closed without speaking. The chin lowers by a few millimeters, then returns almost still. Tiny micro motions: subtle breath in the throat, faint pulse at the lips, barely visible shoulder rise, soft skin movement, slight tension release in the raised arms. A narrow shadow edge drifts gently across the nose and upper lip as if from a slow flag or hand-shaped shadow moving off camera. Camera performs a very slow push-in, nearly imperceptible, holding the same composition and shadow design. Last Frame: At exactly 10 seconds, hold on the extreme close up. Lips are softly parted, nose rim lit, upper face still hidden in darkness, shoulder blurred in the lower left foreground, arms framing the head as dark shapes. Camera is nearly locked, ending in quiet tension with the subject motionless except for a final faint breath. No on-screen text, no captions, no signage typography, no rendered text in the frame. World Plate: Minimal black and white studio space with no visible set edges. Hard directional key light from upper front right, deep shadow from above, soft falloff into a dark gray black background. Subtle air depth, mild haze, natural shadow texture, and fine photographic grain. Palette stays silver, charcoal, and soft skin gray tones. Background remains empty and still, with no props or distractions. Sound Bed: Quiet diegetic studio room tone only: faint breath, slight skin and fabric movement from raised arms, soft body shift, barely audible air in the room. No music, no score, no lyrics, no stylized sound design. Capture Realism: Natural skin texture with visible pores, matte skin, soft highlight rolloff on nose and lips, lifted shadow detail, realistic grain, no waxy smoothing, no over sharp edges, no clipped whites, no crushed blacks. Skin should look human and softly lit, not mirror-glossy. Keep the monochrome contrast rich but breathable. Camera Capture: 10 seconds, 75mm intimate portrait lens, 24fps, 180 degree shutter, wide latitude cinema capture, very slow controlled push in, shallow depth of field, fine 35mm grain, soft highlight rolloff, lifted shadow detail, black and white editorial grade, realistic studio portrait capture.

Ultra-cinematic Pixar-style 3D animation, adorable chibi version of Erling Haaland with long blond ponytail, fair skin, rosy cheeks, expressive blue eyes, wearing a red long-sleeve shirt, navy apron, pink rubber gloves, and a white hairnet, traditional Korean hanok courtyard, warm autumn afternoon sunlight, authentic kimchi-making station, ultra-detailed vegetables, realistic food textures, cozy atmosphere, shallow depth of field, HDR, 8K, one continuous sequence. The scene opens with the adorable chibi football star standing behind a large wooden bowl filled with freshly seasoned napa cabbage. Around him are baskets of fresh napa cabbage, green onions, garlic, ginger, red chili flakes, and traditional Korean earthenware jars. He enthusiastically mixes the kimchi seasoning by hand, folding every cabbage leaf carefully with exaggerated cute movements while smiling. The camera slowly circles around him as he continues preparing the kimchi with genuine curiosity. His ponytail gently sways in the breeze while sunlight creates soft highlights across the courtyard. Every ingredient appears vibrant and highly detailed with realistic textures. Curious about the flavor, he picks up a freshly seasoned piece of kimchi and carefully brings it toward his mouth. The camera cuts to an intimate close-up as he takes his first bite. His eyes widen in surprise from the spicy kick, cheeks puff slightly, eyebrows lift, and he freezes for a brief moment while processing the intense flavor. A second later, his expression transforms into pure delight. He smiles broadly, his eyes sparkle with excitement, and he happily nods in approval. He laughs warmly while pointing at the kimchi as if telling everyone how delicious it is. The camera pulls back to reveal the beautiful traditional courtyard filled with earthenware jars and colorful ingredients. He proudly places both hands on his hips, then raises one gloved hand with a cheerful thumbs-up toward the camera, beaming with satisfaction. Style: Pixar-quality 3D animation, ultra-realistic food rendering, cinematic storytelling, wholesome, cozy Korean cultural atmosphere, expressive facial animation, physically accurate lighting, smooth camera movement, premium animated commercial, masterpiece, vibrant colors, ultra-detailed textures, 8K HDR.

模式：R2V，15秒，高冲击竖屏动作序列。 【参考】 @Image 1 只锁定红色近战战士身份：黑红角状重甲、熔岩红光骷髅脸、巨大沉重体型、 brutal 近身搏杀风格。 @Image 2 只锁定幽灵赛博女巫身份：苍白兜帽、漂浮身体、青白色电路光、烟雾灵体、全息法术面板、远程施法者。 @Image 3 只锁定环境：崩塌中的竖向赛博朋克巨城、暴雨、破碎平台、空中桥梁、垂落电缆、绿色蒸汽、火光、深不见底的垂直空间。不要复制 @Image1 或 @Image2 的背景。 【动作原则】 全片只有一个动作目标：红色战士不断向上追近女巫。两人必须频繁同框，距离越来越近。红色火尾追逐青白电弧，作为贯穿每个切点的连续轨迹。 【时间轴】 0.0-1.0s 竖向远景：女巫在上方断桥倒飞，红色战士在下方平台蓄力，二者上下距离清楚。 1.0-2.0s 低角度仰拍：红色战士爆跳起身，平台被踩裂，红色火尾向上拉长。 2.0-3.0s 女巫过肩镜头：她回身甩出一道青白电弧，电弧贴着雨幕向下切。 3.0-4.0s 横向跟拍：红色战士用前臂重甲硬挡电弧，火星和蓝白碎光朝镜头飞散。 4.0-5.0s 高角度俯拍：女巫穿过垂落电缆上升，红色战士踩墙追上，靴底刮出金属火花。 5.0-6.0s 中景同框：女巫贴墙急退，红色战士从侧面扑近，距离缩短到一臂之遥。 6.0-7.0s 女巫双手张开，生成半圆青白护盾；红色战士一拳撞上护盾，护盾边缘裂开。 7.0-8.0s 快速甩镜：冲击波把雨水震成白色水雾，碎片从两人之间向下坠落。 8.0-9.0s 贴身追拍：女巫绕断桥柱子飞行，红色战士用手爪抓住柱面转向，继续追。 9.0-10.0s 侧面宽景：两人沿建筑外墙垂直上升，红色战士连续踩两块小平台，每块平台都崩裂。 10.0-11.0s 中近景：女巫释放第二道更窄的电弧，红色战士低头冲过，电弧擦过肩甲爆出火花。 11.0-12.0s 近景动作：红色战士伸手抓向女巫斗篷，指尖擦到布料边缘，斗篷被拉扯但她挣脱。 12.0-13.0s 女巫瞬移半米，留下青白故障残影；红色战士的重拳砸中她刚才的位置，断桥塌陷。 13.0-14.0s 最终冲撞：女巫在最高平台前架起破裂护盾，红色战士带着红色火尾正面撞上。 14.0-15.0s 定格对峙：护盾布满裂纹，红色战士离女巫非常近，几乎抓住她；红光和青白光在中心僵持。 【镜头】 快速硬切，每个切点都接同一条上升追击线。镜头可以低角度、过肩、侧面跟拍、高角度俯拍、贴身手持，但不要随机绕圈。每个镜头只表现一个动作和一个结果。 【VFX / 声音】 主特效只用两种：红色火尾与青白电弧。碰撞产生火星、碎光、雨水白雾、下坠金属碎片。原创赛博工业战斗音乐，重鼓点卡住每次起跳、格挡、撞盾、近身擦过。音效：暴雨、装甲伺服、金属断裂、电流爆响、火焰轰鸣。无对白。 【锁定】 红色战士只近战追击，女巫只飞行远程防守。两人不要分开太远，不要变成单人摆姿势，不要新增角色，不要大法阵远距离收尾，不要用浓烟遮住人物，不要血腥画面，不要新增字幕或水印。

Create a cinematic luxury fashion advertisement video using the female model image as reference. The camera performs a smooth 360-degree horizontal orbit around the woman sitting on the metal folding chair and the suspended matcha splash. The camera moves slowly with a premium commercial gimbal movement, creating realistic depth, parallax, and a dramatic bullet-time effect. The woman remains completely frozen in the exact same pose, maintaining the same facial expression, hairstyle, outfit, and body position. She looks like a high-fashion editorial model captured in a perfect frozen moment. The matcha drink splash remains suspended in mid-air like a sculptural liquid artwork. The flowing matcha creates elegant curves, floating droplets, and realistic reflections while appearing frozen in time. As the camera moves around, subtle details become visible: liquid texture, glass reflections, fabric details, denim texture, skin highlights, and studio shadows. The background remains minimal and unchanged. Only the camera movement creates the cinematic effect. Style: luxury beverage commercial, Vogue fashion film aesthetic, Apple-style cinematic product video, ultra-realistic, 8K, high-speed photography, shallow depth of field, realistic lighting, smooth motion, premium advertising quality.

Use the uploaded image as the exact first frame. Preserve the woman’s identity, curly hair, black outfit, seated pose, cigarette position, smoke, textured white wall, chair, natural outdoor lighting and composition. Create a realistic cinematic drama scene. The woman sits quietly against the wall, holding the cigarette near her face. Soft afternoon sunlight touches her hair and skin. Thin smoke moves slowly across the frame. The background stays slightly blurred with subtle street ambience. Camera Movement: Slow cinematic push in from a medium shot to a closer portrait. Very gentle handheld motion, shallow depth of field, natural film grain, soft contrast, warm highlights, realistic skin texture. Action & Emotion: She looks directly into the camera with calm pain in her eyes. Her expression is controlled but emotional, like she is finally saying something she has held inside for a long time. She pauses, breathes softly, smoke drifting between her and the camera. in a low, steady, intimate voice, she says: “You asked why I left… I didn’t leave you. I left the version of me that kept begging to stay.” Timing: 0–2s: Smoke curls slowly. She holds eye contact, silent. 2–7s: She speaks the first line with natural lip-sync and subtle facial movement. 7–9s: She exhales gently, eyes softening. 9–10s: Final emotional beat, camera holds on her face. Audio: Soft city ambience, distant traffic, faint wind, cigarette ember crackle, minimal cinematic piano pad underneath. No text on screen. No dramatic overacting. Style: Hyper realistic cinematic film, emotional indie drama, natural lighting, shallow depth of field, 35mm lens look, warm color grading, subtle film grain, realistic smoke physics. Negative Prompt: No extra people, no change in face, no distorted hands, no warped cigarette, no exaggerated smoke, no cartoon style, no text, no logo, no unnatural mouth movement, no sudden camera shake.

A realistic Red Bull World Cup 2026 commercial scene with an elite freestyle footballer standing on the roof of a skyscraper overlooking a stadium city, holding a Red Bull can dripping with condensation - He opens the can, close-up on the crack and fizz; - Energy kicks in, he sprints across rooftops, juggling the ball, jumping gaps and performing impossible freestyle tricks as helicopters film him; - He volleys the ball from the rooftop toward a giant stadium screen, ending with the Red Bull can in foreground and the city celebrating below - Hyper-realistic sports commercial, golden daylight, drone tracking shots, dynamic slow motion, sweat, sparks, urban adrenaline, premium action advertising.

10-second cinematic sci-fi action sequence, 16:9, ultra-realistic, vibrant color, maximum tension, realistic destruction. Inside a futuristic reactor facility. A tiny sphere of glowing antimatter escapes containment. Everything it touches vanishes instantly. Not explodes. Disappears. The sphere bounces unpredictably through the facility. Walls vanish. Floors vanish. Entire machines disappear. A technician flees across collapsing catwalks. The sphere ricochets closer and closer. The camera stays tight. The glowing sphere enters frame repeatedly from random directions. The technician jumps a gap. The sphere touches the bridge behind him. The entire structure disappears. Cut.

Main subject: Young Japanese woman, mid-20s, natural everyday appearance, soft oversized white linen button-up shirt (sleeves rolled to elbows), high-waisted beige wide-leg trousers, simple white canvas sneakers, thin gold necklace, straight black shoulder-length hair with subtle curtain bangs and a small silver hairpin on one side. Realistic skin texture with light freckles, minimal natural makeup, calm and gentle personality. Maintain consistent identity, clothing, hairstyle, and appearance throughout the entire video. Location: Quiet traditional Tokyo side street during golden hour in early evening. Narrow residential alley with wooden machiya houses, small potted bonsai and flowers, hanging noren curtains, bicycle parked against a wall, soft lantern light starting to glow, overhead power lines, peaceful neighborhood atmosphere. No crowds or heavy traffic. Visual Style: Ultra-realistic documentary realism. Genuine candid behavior. Natural body language. Unscripted slice-of-life feeling. Strong environmental authenticity. Rich real-world details and believable human motion. Camera Style: Early-2000s consumer DV camcorder aesthetic. Friend casually recording everyday moments. Heavy handheld shake, imperfect framing, frequent autofocus hunting, lens breathing, exposure pumping when moving between sunlight and shadow, occasional motion blur, subtle rolling shutter, mild digital compression artifacts, warm faded colors, soft contrast, slight sensor noise. No stabilization. No cinematic camera moves. No modern color grading. 00:00–00:02 She sits on a small wooden bench outside a house, quietly reading a paperback book. A light breeze turns a page. She tucks hair behind her ear while smiling at the story. Camera sways gently as it focuses on her face. 00:02–00:04 She closes the book, stands up, and begins walking slowly down the alley while holding the book against her chest. The camera follows with natural drift and slight shake. 00:04–00:06 She stops at a small family-run vegetable stand, picks up a few fresh tomatoes, and chats briefly with the elderly shopkeeper. Handheld framing drifts off-center as she pays. 00:06–00:08 She continues walking, now carrying a small paper bag of vegetables in one hand and the book in the other. Golden sunlight creates moving shadows across the alley walls. 00:08–00:10 On a quiet corner, she pauses to admire blooming flowers near a wooden fence. She leans in slightly to smell them, then adjusts her shirt sleeve. Loose side-angle shot with camera breathing. 00:10–00:12 Close side profile. Someone off-camera greets her. She turns, smiles warmly, raises her hand slightly, and softly replies “Konbanwa.” The camera reacts a moment late. 00:12–00:15 She walks away down the glowing alley, looking peaceful with her book and bag. Recording cuts abruptly to black mid-motion as if the camcorder was switched off. Audio: Natural ambient sound only — evening cicadas, distant bicycle bells, soft wind, gentle footsteps on pavement, light fabric rustling, faint neighborhood voices, paper bag crinkling, page turning. No music. No sound design. No narration. Goal: Authentic Tokyo neighborhood evening captured like a forgotten home video from the early 2000s — candid, imperfect, realistic, warm, and deeply believable.