Prompts Seedance 2.5 : 1 083+ exemples et guide

【全局设定】 30秒，16:9，顶级3D国漫CG，电影级立体光影、精细角色建模、真实布料动力学与金属质感。不是真人摄影，不是二维动画。整体凌厉、高速、招招见力，近战斩击与群体技能递进，最终全灭收剑。 【资产引用与锁定】 @女主资产：使用「3D女主.png」，锁定同一成年女性的脸型、五官、棕金眼睛、黑发高束与额间蓝色三道印记，保持参考图的深蓝黑战袍、旧金属护肩与装饰；参考图三个视角属于同一个人，视频不出现三联画。右手仅持一把参考长剑，左腰从开场起固定配套剑鞘，直至结尾不消失。 @魔兽群资产：仅对应普通四足魔兽群，锁定体型、黑色甲壳、暗红裂纹、角与四肢结构；数量有限，死亡后不刷新、不复活，不变成巨型魔兽。 @巨型魔兽资产：对应全文所有“重兽／重型魔兽”，全片仅一只，严格参考资产图的巨大体型、厚重肩甲、双角、粗壮前肢与胸甲下暗红核心；体型明显大于普通魔兽，比例始终一致。从t=1s起就在兽群后排可见，t=14s开始冲撞，t=25—26s被大招击杀，此后不复活、不再新增。 @草原场景资产：锁定起伏草原、低矮石脊、中央孤岩和浅沟的位置关系；草高至小腿，交战脚步清晰，远方双峰作为方向参照；破坏痕迹持续保留，不自动恢复。 以上资产仅在本区引用，逐秒分镜使用“女主、普通魔兽、重兽”等简称，不重复添加@。所有资产统一为3D国漫CG风格。 【战场与战斗逻辑】 阴云透出斜向日光，狂风压低草浪。开场全景交代前、中、后三层魔兽，数量有限，不刷新、不复活、不从画面外无限补充。普通魔兽低身扑咬、侧向包抄；重型魔兽沉肩冲撞、前肢砸地，攻击先有发力动作再接触。 前段0—9秒：快速小技能接光速感位移，反复穿过兽群。 中段9—21秒：移动释放群体技能，清除中心兽群，外围残敌和重型魔兽仍可见。 后段21—28秒：压缩残敌位置，移动中完成大招，杀尽所有魔兽。 最后28—30秒：明确战场无活兽，再完整收剑入鞘，结束战斗。 【位移、技能与特效】 女主位移必须表现为“整个人与手中剑收束成一道细长金色光线，外缘带青色流光”，不是人物慢跑拖着光带；光线实际穿越草原、绕过孤岩、掠过浅沟，到达攻击位置立即恢复清晰人形完成斩击。起点、路径、落点连贯，不生成分身。 小技能【掠光斩】：蹬地化光穿阵，落点挥剑，释放贴着剑路的短促青金刃弧。 群体技能【回澜剑潮】：移动挥剑连续带出三层贴地弧形剑潮，逐层推过中心兽群，敌人先被命中，再失衡、崩解。 终极技能【一剑平川】：高速折返过程中将流光压入剑锋，落脚转髋横斩，低矮而宽阔的青金剑潮覆盖剩余敌阵；不站桩结印，不召唤新武器。 特效配比：能量体50%＋水墨25%＋流体25%，红、青、金三色为主；青金属于女主，暗红属于魔兽。能量紧贴剑路，水墨沿受力方向迸裂，流体表现为附着剑锋的黏稠能量与翻卷浪沿。零血液、零断肢，魔兽死亡化作水墨粒子与失去光泽的甲壳碎屑，碎屑受重力落地。 【运镜与节奏铁律】 连续高速跟拍、一镜到底，十组三秒镜头仅表示连续调度段落，不代表硬切。每秒写明运镜、景别、动作与结果；组间借人物追击、草叶和碎屑掠镜连续换位，不跳轴、不突然瞬移机位。 低机位急推、侧向追踪、交错点环绕、快速升降和攻击方向甩镜交替；命中后短促定向震动，随即追踪敌人飞退。主体接触时清楚，背景沿运动方向拉伸，禁止全屏模糊掩盖动作。 清怪期间持续冲刺、侧蹬、滑步与折返，短蓄力融入移动；正常时间下的极高速动作，无慢动作、子弹时间、冻结帧. 敌人全部死亡后才允许自然减速收剑。 无台词、无字幕；声音为剑鸣、密集破空、兽吼、冲撞、甲壳崩裂与草浪声，最后以清晰入鞘声收束。 【镜头1（0—3秒）】 t=1s运镜贴草向前急推，FS全景低机位侧视女主从左侧石脊蹬地化成细长金光穿入前排兽群，落点恢复人形横斩扑来的魔兽，青金刃弧命中后炸开黑墨碎粒，后排重型魔兽清楚可见。 t=2s运镜沿横斩方向高速侧移，MS中景斜侧视女主顺势压身滑过第二只魔兽的扑击，反手斜挑击中其侧甲，魔兽翻向外侧压倒草丛，剑锋流体沿斩线甩成细弧。 t=3s运镜绕交错点快速弧移并拉宽，FS全景俯斜视女主侧蹬地面化光折返，从两兽夹击的空隙贯过并在末端回身一斩，两侧魔兽接触刃弧后崩解，草原留下清楚折线压痕。 【镜头2（3—6秒）】 t=4s运镜沿金光贴地追进，FS全景低角度侧视女主以【掠光斩】穿过横扑兽腹下方的空隙，现身后旋腰扫中侧甲，魔兽受力横飞，墨粒与草屑沿同一方向喷开。 t=5s运镜快速横甩至反击方向，MS中景侧视一只魔兽从孤岩旁扑向女主持剑侧，女主短促架开前爪并借接触旋身绕后，碰撞处先出金属火星，再带动其前肢偏转。 t=6s运镜贴孤岩外缘疾掠，FS全景侧视女主脚掌蹬岩化作一道金线绕过岩背，在魔兽后侧恢复人形落剑，其甲壳沿命中处炸裂，孤岩留下蹬踏裂痕与飞散石粉。 【镜头3（6—9秒）】 t=7s运镜降至浅沟上方高速跟拍，FS全景斜俯视女主贴沟化光穿过追扑间隙，在沟沿落脚转身劈出短刃弧，前兽被斩灭，后兽踩入沟底失衡，泥块沿沟壁滚落。 t=8s运镜沿沟沿侧移并收近，MS中景侧视女主向前压步连续斜斩两次，分别击碎两只扑近魔兽的肩甲与能量核心，兽体先折退再散成黑墨，镜头在第二次重击后短震。 t=9s运镜随女主跃过沟沿迅速抬高，FS全景俯视她化光绕向中心兽群外侧，落地拖剑划出第一道青金弧线，魔兽转向追咬，远处外围残敌仍在向战区奔来。 【镜头4（9—12秒）】 t=10s运镜贴女主身侧高速平移，MS中景斜视她在连续侧滑中转腕收剑，剑锋流体迅速增厚并接一记低位横斩，【回澜剑潮】第一层浪沿从剑路推出，撞翻中心前排魔兽。 t=11s运镜沿第一层剑潮斜向急推，FS全景腰高侧视女主紧随浪沿蹬地折向，再斩出第二层更宽刃浪，刚跃起避浪的魔兽被第二层截住，黑墨与甲片沿后退方向炸开。 t=12s运镜升至孤岩侧上方拉宽，FS全景俯视女主绕中心兽群奔行完成第三次斩击，三层剑潮依次推过夹击区域，中心魔兽接连崩解，外围兽群与重型魔兽仍留在远端。 【镜头5（12—15秒）】 t=13s运镜沿女主冲出剑潮的路线急降，FS全景低机位侧视她化光穿过正在落下的墨粒，落点挑斩一只冲出浪沿的魔兽，兽体向后折翻，草浪被切出狭长通道。 t=14s运镜迅速转向重型魔兽并后拉，FS全景斜侧视重型魔兽低头沉肩猛冲，女主侧蹬化光让开正面撞击，其肩甲撞碎低矮石脊，碎石先接触崩裂再沿冲势滚飞。 t=15s运镜贴撞碎的石脊高速侧跟，MS中景侧视女主在重兽外侧现身斩中前肩甲，重兽踉跄侧摆却未倒下并抬爪反扑，青金刃弧与暗红护甲光短促对撞，镜头随接触震动。 【镜头6（15—18秒）】 t=16s运镜压低跟随横扫前爪，FS全景低角度侧视女主低身滑过爪下并划开其侧甲，爪尖扫中身后碎岩而非空气，石屑向外扇开，女主出滑步立即继续前冲。 t=17s运镜贴草追踪女主折返路线，FS全景俯斜视她化作金线绕重兽外侧切入外围兽群，落点连续两记短斩逼开左右扑兽，青金弧线与翻飞黑墨标出清楚攻击落点。 t=18s运镜随斩击横向快移并拉宽，FS全景侧视女主顺势推出一道低位扇形剑潮扫过聚集的普通魔兽，前排被命中后层层倒向后排，重兽从浪沿后方重新踏稳追来。 【镜头7（18—21秒）】 t=19s运镜沿追击

Create a 30-second, 1080p, 16:9 ultra-realistic early-2000s MiniDV home video in Seoul using the provided character sheet as the ONLY character reference. IDENTITY LOCK: Keep exactly the same Korean girl, face, natural skin, dark hair, body proportions, navy hoodie, white top, beige cargo shorts, gray sneakers, burgundy crossbody bag and silver watch throughout. No beautification, face changes or outfit changes. STYLE: A close friend casually filming her on an old consumer DV camcorder during a warm summer evening in a quiet older Seoul neighborhood. Narrow streets, apartments, trees, bicycles, utility wires and a local bus stop. Strong natural handheld shake, imperfect framing, autofocus hunting, exposure shifts, motion blur, faded DV colors, mild noise and accidental camera movements. Absolutely no cinematic/gimbal look. 00:00–00:05 — She notices her bus approaching, points and says “Wait!” then suddenly runs toward it. Cameraman reacts late and runs behind her; chaotic realistic shake briefly loses framing. 00:05–00:10 — She arrives seconds too late as the bus leaves. Out of breath, she watches it go, turns toward camera and says “Seriously?” Her friend laughs and she starts laughing too. 00:10–00:15 — Walking home, she pulls a small paper fan from her bag. A breeze blows hair across her face; she struggles to move it away and laughs naturally. 00:15–00:20 — At an old drinking fountain she wets her hands and splashes her cheeks. The cold surprises her. She mischievously flicks a little water toward the camera; cameraman instinctively jerks backward while she laughs. 00:20–00:25 — She balances casually along a low curb. A cyclist approaches and rings the bell; she quickly steps down, slightly loses balance, catches herself and gives the camera an embarrassed smile. 00:25–00:30 — She walks ahead up a quiet street. Turns back, gestures toward her friend and naturally says “Come on.” She continues uphill, glances back with a small smile, camera follows imperfectly, then abrupt MiniDV cut to black. AUDIO: Natural human voices and real neighborhood sound only—footsteps, breathing, insects, distant traffic, bicycle bell, water, breeze, laughter and camera-handling noise. No music, narration, subtitles, logos or watermarks. REALISM: Natural acting, blinking, breathing, expressions and body physics. Correct anatomy and hands. No duplicated people, face morphing, broken fingers, teleportation, disappearing objects, frozen backgrounds or AI-looking movement. Keep actions physically continuous and naturally paced.

Main Subject: Young Korean woman, 24, naturally attractive, realistic skin, minimal makeup, long dark hair loosely down. Wearing a simple oversized casual T-shirt and loose lounge pants. Preserve her exact identity, facial features, hairstyle, body proportions and appearance throughout. Location: Small old Seoul apartment bedroom during a dark rainy early morning. A simple wooden dresser, small mirror, scattered hair accessories, folded clothes and an old bedside lamp. A rain-covered window shows blurred neighboring apartment buildings and wet rooftops outside. Lighting & Atmosphere: Cozy, dim blue-hour ambience. Cold blue-gray light from the rainy window mixes with a weak warm bedside lamp. Dark cloudy sky, muted colors, soft shadows, low exposure and a quiet sleepy atmosphere. Steady rain outside. Style: Ultra-realistic early-2000s Sony MiniDV home video. Completely candid and unstaged. Natural handheld movement, imperfect framing, subtle camera shake, occasional autofocus hunting, mild exposure shifts, faded colors, soft contrast, authentic DV compression, subtle low-light digital noise and microphone noise. Keep movement smooth and continuous. No stuttering, judder, frame skipping, duplicated frames, stop-motion appearance, excessive motion blur or low-frame-rate look. No stabilization or modern cinematic movement. 00:00–00:04: She stands beside the dresser getting ready, suddenly realizes she cannot find her favorite small hair clip. She looks around the dresser with a mildly confused expression. 00:04–00:08: She searches the tabletop, moves a small brush aside, checks beside the mirror and looks down around the floor. She becomes increasingly puzzled but remains calm. 00:08–00:11: She walks toward the mirror and looks closely at her reflection, still trying to figure out where the clip could be. She touches her hair absentmindedly. 00:11–00:15: Her fingers suddenly feel the hair clip already tucked into her hair. She freezes, realizes what happened, then looks directly at the camcorder with an embarrassed little smile and quietly laughs at herself. Audio: Natural location sound only—steady rain against the window, soft footsteps, faint fabric movement, small objects shifting on the dresser, distant traffic and quiet apartment ambience. No music, narration or added sound effects. Goal: Cute, spontaneous and intimate, like a tiny everyday mistake accidentally captured on an old home video. The humor should feel completely natural and understated. Her realization should be the main payoff, followed by a genuine embarrassed smile toward the camera. Motion quality: Keep all movements smooth, continuous and realistic. Her searching gestures should be slow and simple rather than frantic. The vintage DV aesthetic must come from image quality, autofocus, exposure, color and handheld operation—not from reduced frame rate or choppy motion.

Create a 30-second, 1080p ultra-realistic emotional live-action scene about a young woman discovering her boyfriend kissing another woman inside a quiet café. Focus entirely on authentic human emotion, natural crying, subtle facial expressions and believable body language. No melodrama or exaggerated acting. MAIN CHARACTER Young Korean woman in her early 20s, naturally beautiful, realistic skin texture, minimal makeup, long black hair loosely styled, wearing a fitted pastel-pink top, casual denim shorts and simple sneakers. Maintain the exact same face, hairstyle, outfit, body proportions and appearance throughout the entire scene. LOCATION A cozy modern café on a rainy evening. Warm interior lighting, large windows covered with rain droplets, wooden tables, soft background chatter, cups and plates, a few customers quietly talking. SCENE She enters the café carrying a small handbag, casually looking around for her boyfriend. She suddenly notices him sitting at a table near the window. He is kissing another young woman. She immediately stops walking. Her expression changes from confusion to disbelief. She slowly lowers her handbag. She stares at them silently for a few seconds, struggling to process what she is seeing. Her eyes begin to fill with tears. Show realistic tears naturally forming along her lower eyelids, then slowly rolling down her cheeks. No exaggerated crying. She takes a shaky breath. She walks toward their table. Her boyfriend notices her. His expression immediately changes to shock. The other woman turns around and becomes uncomfortable. The girlfriend looks directly at him with trembling eyes and quietly says: “How could you do this to me?” Her voice cracks naturally. She wipes one tear from her cheek, but more tears continue falling. Her boyfriend stands up and tries to explain. She shakes her head and takes a small step backward. She says through tears: “I trusted you.” She looks at him for one final moment. Her expression shows heartbreak, disappointment and disbelief rather than anger. She turns around and walks toward the café door while quietly crying. Her boyfriend takes one step after her but stops. She reaches the door, pauses briefly, wipes her face and leaves. The camera remains inside the café as the door closes behind her. CAMERA / VISUAL STYLE Handheld cinematic camera with subtle natural movement. Start with a medium shot following her into the café, then slowly move closer when she notices the couple. Use realistic close-ups of her eyes, trembling lips, tears and subtle facial reactions. Natural shallow depth of field, realistic skin pores, authentic reflections in the café windows, soft rainy evening light, natural motion blur, 24fps, photorealistic live-action. Do not make her crying look beautiful or glamorous. It should look spontaneous, un

30-second, 16:9, 24fps short film dedicated to raw, visceral post-apocalyptic hand-to-hand and improvised weapon combat. High-budget gritty theatrical live-action quality. No subtitles, no UI, no logos, no voiceover, no BGM—retain only gravel crunching, heavy rusted iron impacts, bone breaks, guttural grunts, wind tearing through wreckage, torn cloth flapping, and ragged panting.Characters Protagonist: Leather-clad drifter using a reinforced steel pipe and spiked gauntlet, hyper-aggressive instinctual fighting style. Scavengers (4 Raiding Mutants): Distinct designs—scrappy armor, bone masks, wielding machetes, heavy chains, and spiked clubs. Non-stop feral encirclement. Core & Action Savage, continuous momentum with zero breathing room. Entire fight takes place in a rusted vehicle graveyard. Features brutal duck-and-weaves, vaulting off car hoods, using one enemy as a human shield against a chain attack, low sweeping leg-takes, and improvised counter-strikes utilizing car doors and debris. One-Time Speed Ramp A single 0.15-second micro-slowdown at the 10-second mark: A swung spiked club brushes past the protagonist's face, dislodging a drop of blood and dirt from their cheek before the protagonist grabs the attacker's wrist and drives a knee into their elbow joint. Camera & Environment Gritty handheld camera, close-proximity tracking, violent whip pans re-acquiring targets mid-strike, mud and rust spraying onto the camera lens. Set in a desolate wasteland of stacked rusted car chassis, flying gravel, kicking up dust clouds under high winds, with broken glass showering down on heavy impacts. Ending Sequence Protagonist redirects a chain attack around a rusted pillar, uses the tension to vault forward, delivers a flying knee to Scavenger A, spins mid-air to crack Scavenger B with the steel pipe, and finishes the final attacker by driving them headfirst through a cracked windshield. Final frame captures the protagonist gasping for air amidst the wreckage.

Use @[Image] as storyboard reference for cinematic sequence. use as first shot 01 A 10-second mesmerizing cosmic meditation journey transitioning into a magical fractal portal. Visual style: abstract sacred geometry, photorealistic 8K, ethereal glowing runes, swirling liquid metal ribbons of gold and deep purple, starry deep space background, cinematic volumetric lighting, 3D render aesthetic. CRITICAL: Absolutely NO frames, borders, film strips, numbers, UI, or text overlays. Render as seamless full-screen cinematic shots only. Ignore grid layout completely. Sequence Description: The video begins with the full view of the cosmic spiral. The ribbons inscribed with ancient arcane runes begin to rotate slowly in a hypnotic, clockwise motion, creating a mesmerizing flow of magical energy. The textures of the ribbons shift like liquid gold and violet silk, catching the light as they move. (0-4.0s): The central black singularity "breathes," pulsing with bursts of prismatic starlight. Sharp rays of neon blue, magenta, and gold radiate outward from the center in rhythmic waves, syncing with a slow, calming tempo. The background stars twinkle softly. (4.0-7.0s): The camera begins a slow, steady dolly zoom (push in) towards the center. The sensation is one of deep focus and spiritual awakening. The runes on the ribbons shimmer and glow brighter, flowing along the curves like streams of pure mana. The edges of the frame slowly blur (vignette effect), drawing all attention to the glowing eye of the universe. (7.0-10.0s): As the camera moves closer to the event horizon, the light intensifies. The spiral rotation accelerates slightly but remains smooth and fluid, never jerky. The gold and purple colors blend into a warm, radiant aura. The video ends with the viewer feeling pulled gently into the magical void, leaving a sense of peace and infinite possibility. ZERO TEXT. Style reinforcement: continuous hypnotic rotation, smooth easing camera movement, magical particle effects, glowing rune textures, meditative trance atmosphere, high contrast between deep shadows and radiant light, NO text, NO grid, NO numbers, ultra-high-end abstract visual quality.

Cinematic hyper-realistic 3D CGI collectible-figurine render, like a live-action blockbuster VFX sequence: physically-based rendering (PBR), octane-render-quality lighting and materials, fabric with visible weave and natural drape, metal with realistic wear

Create a realistic luxury perfume review video featuring a stylish young woman in an elegant modern setting. She holds the Bleu de Chanel perfume bottle toward the camera, showing the bottle clearly. She then sprays a small amount on her wrist and neck, gently smells the fragrance, smiles naturally, and gives a genuine personal reaction. She looks directly into the camera and speaks naturally in a warm, confident female English voice with perfect lip-sync: “I’ve been trying Bleu de Chanel, and honestly, I really love it. It smells fresh, clean, sophisticated, and very elegant. It’s perfect for everyday wear, but it also feels special enough for an evening out. Definitely a fragrance I’d recommend.” Use natural facial expressions, realistic hand gestures, accurate lip movement, and authentic influencer-style delivery. Keep her voice clear, smooth, friendly, and conversational—not robotic or overly commercial. Use soft flattering lighting, realistic skin texture, shallow depth of field, cinematic camera movement, natural perfume spray mist, realistic reflections, and a premium lifestyle background. Keep the perfume bottle consistent and recognizable throughout. Photorealistic, luxury fragrance commercial, authentic product review, realistic physics, smooth motion, 4K, vertical . No distorted hands, no warped bottle, no changing face, no flickering, no incorrect text.

Create a polished 30-second cinematic comedy video with the pacing and visual quality of a premium commercial mixed with a high-budget spy movie. MAIN CHARACTER: One adult woman throughout the entire video. Keep her facial identity, hairstyle, skin tone, body proportions, and overall appearance consistent in every shot. Natural realistic acting and expressions. VISUAL STYLE: Ultra-realistic live-action, cinematic commercial, high-end spy-thriller aesthetic, realistic lighting, natural human movement, professional camera work, realistic environments, subtle comedy, dramatic pacing, realistic physics, premium production quality. 00:00–00:04 — THE MISSION Late at night, the woman suddenly remembers that her favorite dessert is inside the refrigerator. She slowly sits up in bed with a serious expression. Treat this moment like the beginning of an extremely important spy mission. Cut to a dramatic close-up of her eyes. 00:04–00:08 — STEALTH MODE She quietly walks through a dark hallway toward the kitchen. Use dramatic low-angle shots, subtle shadows, slow camera movement, and cinematic suspense. She moves carefully as if she is infiltrating a highly secure location. 00:08–00:12 — THE FRIDGE She reaches the refrigerator. She looks around suspiciously. She slowly opens the refrigerator door. Bright refrigerator light illuminates her face dramatically. She looks inside. Her expression suddenly changes from confidence to complete shock. The dessert is missing. 00:12–00:17 — THE INVESTIGATION She immediately switches into detective mode. Show quick cinematic shots of her searching the kitchen: - Looking under the kitchen table - Checking drawers - Examining the countertop - Looking at suspicious footprints - Inspecting an empty dessert container Make everything visually dramatic and completely serious, even though the situation is ridiculous. 00:17–00:22 — THE DISCOVERY She notices a tiny clue leading toward another part of the kitchen. She follows it carefully. The camera tracks behind her like a spy thriller. She suddenly discovers the dessert box. She looks relieved. Dramatic music builds. She slowly opens the box. 00:22–00:26 — THE TWIST The box is completely empty. Only a single spoon remains inside. She freezes. Hold on her shocked expression for comedic timing. Then slowly reveal another adult person sitting casually in the background, eating the last bite of the dessert. The person looks completely innocent. 00:26–00:30 — FINAL SHOT The woman slowly turns toward the camera with an exhausted, speechless expression. Cut to the person casually taking another bite. End with a clean cinematic freeze-like moment. Display: “Case closed. 🍰” Then subtly display: “Made With SuperCool” Keep the branding elegant and minimal. IMPORTANT: - Keep the same woman throughout. - No face morphing or identity changes. - No duplicate main characters. - Realistic human movement. - Realistic kitchen environment. - No cartoon or CGI appearance. - No exaggerated facial distortion. - No random text or logos. - No watermark. - Keep the comedy subtle and natural. - The first 20 seconds should feel genuinely like a serious spy movie. - The humor should come mainly from the contrast between the extremely serious cinematic treatment and the ridiculous dessert investigation. - Make the final reveal unexpected but easy to understand. - Premium cinematic quality suitable for a professional social media advertisement. - Vertical 9:16 composition.

アーチの下に立つ2人。Aがゆっくり顔をカメラに向けて微笑む。Bが笑って半歩前に出る。手持ちのスマホ、わずかな手ブレ、逆光とレンズフレア、浅い被写界深度。カットを割らない。

Cinematic anime clip, 15 seconds. Lush tropical jungle island, bright sunshine piercing through dense green canopy, vibrant green everywhere, dramatic rock formations, muddy slopes, hanging vines, rocky ravines, golden light throughout. One location

Vertical 9:16 cinematic video, hyper-realistic, photorealistic, ARRI Alexa LF, Master Prime lenses, 1080p HD. Dusk close-up of a stylish individual, showing realistic skin pores, peach fuzz, fabric texture, and orange neon reflected in the eye. Fast cut to a busy London street at twilight, with dynamic tracking, wet pavement, realistic traffic, and cinematic reflections. The subject walks briskly past a traditional red phone booth glowing with warm orange ambient light. Macro close-up of hands adjusting a vintage analog watch with an orange dial, rugged nylon strap, and brushed metal details. Rapid transition to a figure running up a dark industrial staircase with dramatic warm orange and deep blue lighting. Capture realistic muscle movement, flowing motion, natural shadows, cinematic grain, and authentic depth of field. Final stable close-up of the subject staring confidently into the camera, with soft orange light illuminating half the face. Kodak Portra 400-inspired film look, ARRI color science, natural camera noise, high contrast, and immersive textures. Professional cinematography, realistic motion blur, shallow depth of field, warm orange and moody blue cinematic color grading.

Create a 30-second ultra-photorealistic American ranch lifestyle vlog following the same young ranch woman through one authentic working day, from sunrise to sunset. Begin with her opening the barn and feeding horses, then repairing a wooden fence before saddling a horse and riding across open Western fields. Show a genuine working ranch with weathered barns, horse paddocks, hay, dirt paths, wooden fences, pickup trucks, green fields and distant hills. Keep the character, horse, clothing and ranch environment perfectly consistent. Capture realistic horse behavior, riding biomechanics, hoof impacts, dust, grass movement, leather tack, feed, wood, hammer strikes and natural sunlight. Use handheld 2026 documentary-vlog cinematography with close-ups, wide ranch views, realistic autofocus, subtle camera shake and a smooth golden-hour ride. Natural diegetic audio only: horses, birds, wind, barn sounds, hoofbeats, tools, leather and insects. Short synchronized American English dialogue. No music, narration, subtitles, CGI, artificial movement, impossible physics, logos or watermark.

A visually striking, cinematic close-up shot capturing a woman from the chest up against a plain black background. She wears a vibrant orange sleeveless top that contrasts sharply with her skin and the dark backdrop. Her long, dark brown, curly hair falls naturally over her shoulders, and she wears simple silver hoop earrings. The lighting is soft and directional, illuminating her face with a natural glow. The video showcases a seamless, realistic transformation of her facial expression across three distinct emotional beats: 1Neutral & Subtle (0–10s): The video begins with the woman looking directly into the camera with a calm, neutral expression. She blinks naturally and offers a faint, polite smile that barely moves her lips, maintaining steady eye contact.
 2Surprise & Awe (10–20s): Her expression gradually shifts into genuine astonishment. Her eyes widen, her mouth opens slightly in an "O" shape, and her eyebrows lift upward as if reacting to an incredible sight off-camera.
 3Warm Delight (20–30s): Her surprise softens into pure joy. She breaks into a wide, radiant smile showing her teeth, her cheeks lift, and her eyes narrow slightly with genuine happiness as she continues looking at the viewer.
 Technical & Visual Style: •Framing: Medium close-up shot (bust level), centered composition.
 •Lighting: Studio studio-style soft key lighting on her face, deep black background with zero distractions.
 •Camera Movement: Static camera with no pan, tilt, or zoom; smooth, continuous frame rate capturing subtle micro-expressions.
 •Color Palette: High-contrast warm orange top against a deep black void, realistic skin tones, natural lighting highlights.

Use the uploaded reference image as the exact character reference. Keep the same face, hair, outfit and appearance throughout. The woman stands beside the giant white ice dragon, then rides on its back as it flies over snowy mountains and a frozen river. They reach an ancient stone shelter where she explores while the dragon protects her. Cinematic camera movement, epic wide shots, realistic motion, glowing blue eyes, snowfall, blue atmosphere with warm firelight, photorealistic fantasy movie quality, 16:9, no background music.

画面中只有一名明确成年的男性龙骑守望者与一条唯一的远古巨龙。男性守望者身材精瘦而强健，苍白皮肤，原创锐利面孔，铂金色长发越过肩部；全程穿同一套磨损炭黑龙骑护甲：分层旧皮革、深色锁子甲、雕塑感肩甲、金属扣件、重型长靴以及一件长黑披风，不得出现王冠、家族纹章、可识别影视游戏标志。他右手始终只握一支相同燃烧火把，左手始终空着；髋部可固定一把入鞘长剑，但从不拔剑、不碰剑。人物脸型、发长、服装轮廓、火把尺寸从第1帧至第480帧完全一致。 巨龙只有一条，而且必须始终明确是真正的四足双翼巨型爬行动物：四条强壮腿、两扇巨大折叠翼、长而肌肉发达的颈部、宽阔爬行类头骨、角质眉骨、厚重上下颚、层叠青铜黑鳞片与深琥珀色眼睛。龙头单独就大于守望者完整身体。巨大躯干与折叠双翼大部分藏在黑暗中，只允许火光逐渐揭示头、颈、前肩、翼缘以及部分庞大躯干。绝不形成站立人形轮廓，不得变成巨人、恶魔、巨魔或模糊类人怪物。巨龙始终停留在洞穴上方偏左并延伸至中央黑暗的高岩架上，不飞行、不换位置。 风格：高级预渲染AAA暗黑幻想动作游戏电影，风格化照片级CGI，明确保持计算机生成质感；高多边形数字雕刻人物与巨龙，原创游戏电影级面部，PBR旧皮革、冷金属、锁甲和龙鳞，精细头发与披风模拟，体积火光、烟雾、灰烬、火星、热空气折射、克制Bloom、强景深分离、轻微大气颗粒、方向性运动模糊、深而可读的阴影、下一代主机Cutscene品质。绝不真人实拍、真人演员穿戏服、Cosplay、普通游戏画面、动漫、插画或赛璐璐。 环境是一座由黑色玄武岩贯穿形成的巨大地下火山龙坑。顶部消失在烟雾里，锯齿状岩壁、古老巨型石支撑、高低错落的天然岩架包围深层中央地面；空气中始终漂浮煤灰、烟尘和少量余烬。地理关系严格锁定：守望者位于画面下方中央、小型石台附近；巨龙在他后上方的左侧高处黑暗中；巨龙喷火时头朝画面左侧，火焰沿上层洞穴和顶壁横扫，绝不直接击中守望者。喷火前火把是唯一清晰可见光源。 镜头规则：原生横屏16:9，无永久黑边；精确20.00秒，24fps，共480帧；一个连续镜头，无真正剪切、无跳切。虚拟ARRI Alexa 35，35mm Cooke anamorphic。用户要求的“四个镜头”转化为同一物理摄影机连续完成的四个构图阶段：紧密三分之四侧脸→极远景龙坑→同一极远景完整火焰呼吸→最终低角度守望者近景。阶段转换必须利用黑暗岩柱、烟雾、火光过曝边缘和摄影机平滑位移完成自然遮挡，不使用真正编辑剪切。摄影机整体慢而受控，仅有极轻真实设备微振，无随机手持乱晃、无快速环绕、无炫技推拉。自然速度，无慢动作、无速度坡度。 0.00-3.40秒：立即从守望者紧密三分之四侧脸开始，暖色火把照亮一侧面孔，另一侧沉入深阴影。头略低，眼睛最初看向洞穴地面，嘴闭合，表情克制而紧绷。胸廓完成一次缓慢呼吸，火把热流推动几缕铂金长发和披风边缘。摄影机只做极度克制的缓慢靠近。2.30秒后，眼睛首先向上移动；随后下巴只抬高一点，像感知到上方巨大质量发生移动。绝不能提前看见巨龙。3.40秒附近，摄影机顺着一根近距离黑色石柱掠过，前景短暂遮满画面并自然拉远进入下一构图，但仍是同一连续运动。 3.40-6.40秒：遮挡移开后形成巨大极远景，守望者缩小为下方中央极小人物，右手低垂握火把，有限琥珀光池只照亮脚边。上方巨大黑暗占据绝大部分画面，只显示一段折叠翼外缘、隆起肩部轮廓以及两点遥远琥珀眼睛；整体剪影必须明确表现四足带翼巨龙，而非站立人形。摄影机几乎静止，仅极慢向前漂移。守望者只慢慢抬脸，不走动、不奔跑、不举武器。巨龙胸腔完成一次巨大吸气，折翼皮膜出现细微拉伸，鳞片摩擦岩石，岩架下方灰尘和细碎石屑因质量移动而落下。 6.40-7.20秒：仍保持同一极远景。巨龙先由长颈肌肉收紧并向前伸展，随后巨大头颅从黑暗里慢慢推进；角眉和宽阔爬行类吻部被火把微光与眼睛反光短暂勾勒。头部转向画面左侧。动作顺序严格：下颚先下降→上颚随头部略抬→胸腔和喉部继续扩张→喉咙深处出现极暗橙色光→光从深喉向口腔前移→细烟先从牙缝漏出。火焰绝不能提前出现。 7.20-9.30秒：喉部橙光逐步增强，巨龙胸腔开始收缩，第一批细小火星从张开的口腔喷出，随后才形成一条与龙嘴物理连续的细而集中的橙白火柱，向画面左侧略向上射过洞穴顶部。喷发产生反作用力，龙头先向后退极小距离，随后颈部肌群收紧把位置稳定住。火不能凭空从黑暗出现，必须始终从张开的龙嘴内部产生。守望者仍留在下方中央完整可见。 9.30-12.80秒：同一次呼气持续，不停顿、不重新点火。细火柱逐渐变成巨大连续火流：核心接近白色，外围黄橙色，最外层为深红湍流边缘；离开龙嘴后才逐步膨胀。龙颚持续打开，颈部锁紧，胸腔缓慢收缩。强烈火光逐步揭示龙脸、长颈、前肩鳞片和折叠翼的一段翼缘，强化其真实巨龙身份。热空气压力波到达下方后，守望者铂金长发和黑披风被向后推，他只轻微屈膝扩大支撑、保持双脚扎地；右手火把仍牢牢握住，火苗被压力吹斜但不熄灭。灰尘、煤灰与余烬沿热流方向被推离。 12.80-15.20秒：同一持续火焰发展到最高规模，上半幅画面逐渐充满翻滚火焰、厚烟和热折射。亮度必须连续上升，绝不是突然白闪。高速移动火光沿玄武岩壁和石支撑向外扫动，短暂揭示巨龙庞大头部、整段长颈、前肩、折翼和一部分巨大鳞甲躯干，但其四足结构和高岩架位置不改变。摄影机曝光受到强火影响，出现受控琥珀白Bloom与局部炫光，仍能隐约辨认张开的龙颚和远处小小守望者剪影。不得复制第二条龙，不得让火焰变成生物或独立发光物。 15.20-16.50秒：火暴达到最亮状态，火焰、烟雾、热折射覆盖大部分图像，但不得变成纯白无信息画面。洞穴因连续压力出现低频震动，远处岩架掉落碎石，守望者披风猛烈抽打，头发顺热风方向拉开；龙颈持续承受反冲，头部有极小真实颤动但不改变位置。火焰保持自然速度湍流，不使用慢动作。摄影机被热流造成轻微空气折射和细小物理振动。 16.50-18.40秒：巨龙胸腔到达一次呼气末端。火流按明确顺序自然衰减：巨大火暴先收缩→宽火柱变窄→中央细火柱继续数帧→最后数颗火星离开口腔→只有在呼气真正结束后火焰才与龙嘴断开。随后龙头略微降低，先开始闭合下颚，再完全收口；两侧鼻孔喷出密集灰黑热

Create a 30-second ultra-realistic live-action short film set in a cozy Korean apartment late at night. A young woman is working alone on a creative project. The footage should feel naturally filmed, with believable acting, practical lighting, subtle handheld movement, realistic room ambience, and authentic creator energy. 0–6s — IDEA She sits tiredly at her desk, looking at an unfinished project on her laptop. She nearly closes it, then suddenly notices something that sparks an idea. Her expression changes from exhausted to quietly excited. She grabs a notebook and quickly writes down the thought. 6–13s — STARTING She opens her laptop fully, clears some space, and begins working. Show natural close-ups of her sketching, typing, scrolling, and checking the screen. Her movements become faster as she gets more invested in the idea. 13–21s — DEEP WORK Her attention is completely locked in. A flowing montage shows rapid typing, switching between creative elements, adjusting details, reviewing images, taking a quick sip of her drink, and immediately returning to the screen. Her desk becomes increasingly messy while the room remains quiet and atmospheric. Use energetic but believable editing, occasional handheld shifts, subtle autofocus changes, and realistic screen reflections. Avoid overly flashy transitions. 21–24s — FAILURE She finally finishes and clicks the button to generate the video. The progress bar approaches completion, then suddenly stops. A plain fictional message appears: “VIDEO GAGAL DIBUAT” She becomes completely still. 24–30s — REACTION She leans closer to the monitor, hoping she read it incorrectly. After realizing the attempt has failed, her excitement fades into tired disappointment. She removes her headphones, exhales slowly, and pushes the laptop back slightly. Quietly: “Not tonight…” Hold on her exhausted face as the cool monitor light falls across her eyes, then end naturally. STYLE: photorealistic live action, cozy Korean interior, natural skin texture, realistic hands and movement, warm practical lights mixed with monitor glow, subtle film grain, shallow depth of field, authentic late-night ambience. NEGATIVE: no anime, cartoon look, CGI appearance, exaggerated acting, comedy reaction, dramatic music, excessive VFX, unrealistic hands, fake interfaces, recognizable platform branding, or overly polished commercial visuals.

Premium law-firm phone-durability commercial, deadpan editorial prestige, one continuous 30-second take, 16:9. SUBJECT: a composed woman lawyer — charcoal wool suit, dark hair pulled back cleanly, seated at the near end of the table; an opposing lawyer — tailored navy suit, short dark hair, clean-shaven, seated at the far end. A modern flagship smartphone with a burnt-orange and deep midnight-blue unibody, a raised horizontal dual-lens camera bar, and seamless edge-to-edge glass front, placed screen-down and slid across the table. A clear glass water pitcher, two-thirds full, mid-table. SETTING: a vast high-end law-firm conference room. A single immensely long polished oak table runs the full length of the room, grain warm and visible, stretching into deep perspective toward a far wall lost in soft shadow. Tall windows along one side throw flat overcast daylight across the wood. Legal folders stacked in neat piles at intervals. Sparse dust motes hang almost still inside the diagonal light beams; the air elsewhere is clean, no haze, no fog. STYLE: cinematic realism, color-negative rendition with fine 35mm grain, cool overcast key warmed by oak bounce light, blacks holding suit and wood detail, a thin bright specular line along the phone's glass and metal edge. No CGI, no plastic sheen, no digital cleanliness. CAMERA: one continuous take, no cuts, no transitions. Smooth, mechanically stabilized dolly and Steadicam movement throughout — no handheld shake anywhere except a brief real impact vibration at the slam. Captured at 24fps with a true 180-degree shutter, motion fluid and continuous, never choppy, never stepping. 0-3s [MEDIUM TWO-SHOT] Fixed eye-level framing holds both lawyers across the table's full depth, completely still, no drift. She speaks calmly, hands still. <room hush> Dialogue language: English, calm dry delivery. {My client's been more than fair.} 3-6s [CLOSE INSERT] Slow macro push to the tabletop. Her hand sets the phone screen-down on the oak; one small confident push sends it gliding forward. <a soft click of glass meeting wood> 6-18s [LOW LATERAL TRACKING] Extremely low, tabletop-height smooth dolly glides alongside the sliding phone at matching speed for the full length of the table, never overtaking, never settling. The phone clips the base of the glass pitcher; water surges audibly without spilling, and the phone continues undisturbed, slowing as it nears the far end. <long glass-on-wood whoosh> <glass clink, water slosh> 18-21s [MEDIUM ON THE OPPOSING LAWYER] Fixed eye-level medium shot, holds completely still. He looks down at the phone in front of him, face controlled, deadpan. {Not a chance.} 21-23s [CLOSE INSERT] Static close insert at tabletop height, holding firm through the impact. His hand drives the phone down screen-first in one sharp strike; nearby folders and the pitcher vibrate briefly, then stillness. <a sharp slam, a short wooden vibration tail>

Day 1 — “Rainy Convenience Store” Create a 30-second, 16:9, 1080p ultra-photorealistic personal home-video capturing an ordinary rainy evening in the life of a young Japanese woman. The footage should feel spontaneous and genuinely recorded by a friend, not like a commercial or music video. MAIN SUBJECT Young Japanese woman in her early 20s, naturally attractive, realistic skin texture, minimal makeup, relaxed and slightly playful personality. Shoulder-length dark-brown hair worn loosely with natural flyaways and slightly damp strands from the rain. Wearing a muted sage-green lightweight hoodie, relaxed dark denim jeans, white canvas sneakers and a thin silver necklace. Maintain the same face, hairstyle, clothing, body proportions and overall appearance throughout the entire video. SETTING A quiet older residential neighborhood in Osaka during a warm summer evening after rainfall. Narrow wet streets, small residential buildings, vending machines, bicycles under awnings, utility poles, tangled overhead cables, parked scooters, rainwater running beside the pavement, small plants outside homes and a modest neighborhood convenience store. The environment should feel lived-in and ordinary. Avoid famous landmarks, recognizable brands, advertisements or tourist locations. CAMERA / VISUAL AESTHETIC Raw personal footage casually recorded by a friend on a small early-2000s consumer digital camcorder. Handheld movement, imperfect framing, occasional autofocus hunting, slight exposure pumping, mild motion blur, soft digital detail, faded colors, subtle sensor noise, accidental zoom adjustments and realistic camera-handling movements. The camera operator should sometimes react slightly late to her movement. No gimbal, no stabilization, no drone shots, no cinematic lens effects, no polished commercial cinematography. 00–05s — LEAVING HOME She opens the front door of a small apartment building and steps underneath the awning. Rain is falling steadily outside. She checks the sky, pulls her hood loosely over her head and looks toward the camera with a small amused smile. She says: “I thought it stopped raining.” She steps carefully onto the wet pavement. 05–10s — WALKING THROUGH THE RAIN The camera follows beside her as she walks down the narrow residential street. Her sneakers make small splashes through shallow puddles. She passes parked bicycles covered with rain droplets. A scooter moves through the background, briefly spraying water from the road. She adjusts her hoodie sleeve while walking. The camera briefly loses focus on her and focuses on a nearby rain-covered bicycle before recovering. 10–15s — CONVENIENCE STORE STOP She notices a small convenience store ahead. She walks inside while the camera remains just outside the entrance for a moment. Warm interior lighting contrasts naturally with the cool rainy street. She comes back out holding a small transparent plastic bag containing a bottled drink and a

30 seconds | 16:9 | photorealistic live-action | ONE CONTINUOUS TAKE | 2 FACE REFERENCES FICTIONAL RECREATION — Hyper-realistic unscripted water-park obstacle challenge filmed as if captured by a real production camera. Two provided face references are used strictly for identity consistency. Preserve facial identity, facial proportions, hairstyle, skin texture, age, and recognizable features throughout the entire shot. No face swapping, no identity drift. CORE CONCEPT A seemingly normal water-park obstacle challenge progressively turns into a spectacular chain reaction of physical mistakes. The entire sequence happens in ONE UNBROKEN TAKE with no cuts, no transitions, no time jumps, and no impossible camera teleportation. The humor comes entirely from believable physics, bad timing, momentum, slipping, water impact, and genuine human reactions. CHARACTERS SUBJECT A — Adult contestant based on FACE REFERENCE 1. Athletic casual appearance, realistic swimwear, wet hair and skin, natural body proportions. SUBJECT B — Adult contestant based on FACE REFERENCE 2. Standing nearby and reacting naturally throughout the sequence. Both identities remain locked. Clothing, body proportions, hairstyle, wetness, accessories, and facial characteristics remain continuous from beginning to end. SETTING Large modern outdoor water park on a bright summer afternoon. Giant inflatable obstacle course floating in a large pool, climbing platforms, floating balance beams, slippery ramps, overhead water jets, inflatable barriers, spectators behind safety rails, lifeguards, wet concrete, colorful umbrellas, realistic resort architecture. Hundreds of subtle background actions create authentic environmental life without distracting from the main action. CAMERA / ONE-TAKE CHOREOGRAPHY 00:00–05:00 — Camera begins approximately three meters behind Subject A at pool level, handheld but professionally operated. Subject A runs toward the first floating obstacle while Subject B watches from the side. Camera follows naturally, matching walking and running speed. Small realistic handheld movements, natural autofocus, water droplets occasionally crossing the lens. 05:00–10:00 — Subject A jumps onto a wet inflatable platform. Their landing is slightly unstable. One foot slides forward unexpectedly. They recover for a moment, laugh nervously, then continue. Camera moves closer while maintaining continuous spatial relationship and realistic human-scale perspective. 10:00–15:00 — Subject A attempts to cross a narrow floating balance beam. Their foot slips. They grab a flexible overhead rope, swinging sideways with real momentum. The rope stretches and pulls their body back toward the obstacle. Subject B reacts from the adjacent platform, stepping backward to avoid the swinging contestant. 15:00–20:00 — Subject A regains balance but accidentally kicks a large inflatable obstacle. It shifts dramatically. A nearby water jet activates directly into the obstacle, creating a sudden burst of spray. Camera operator instinctively steps backward while continuing to film, briefly losing perfect framing before quickly reacquiring Subject A. 20:00–25:00 — Subject A tries to escape across a slippery ramp but loses traction completely. Their body slides uncontrollably downhill. Subject B reaches toward them but misses. Subject A collides softly with a large inflatable barrier, rebounds, spins naturally, and falls backward into the pool. 25:00–30:00 — Huge realistic splash completely covers the camera lens for a moment. Camera operator wipes the lens while remaining in the same physical location. Through droplets and distorted reflections, Subject A resurfaces laughing and trying to regain composure. Subject B is laughing in the background. Camera settles into a final medium-wide composition showing both contestants, the obstacle course, lifeguards, and surrounding spectators. PHYSICAL REALISM Every movement follows believable human biomechanics and real-world water physics. Correct body weight, inertia, friction, momentum, buoyancy, splash displacement, wet-surface slipperiness, inflatable deformation, rope tension, water spray and gravity. No floating bodies, rubbery limbs, impossible recoveries, teleportation, or exaggerated cartoon physics. CAMERA REALISM Authentic handheld operator behavior. Natural micro-jitter, slight bobbing while walking, imperfect framing during sudden action, realistic autofocus hunting, exposure adaptation between bright sunlight and shaded obstacles, lens water droplets, temporary obstruction, realistic motion blur. The camera physically moves through the environment rather than cutting between viewpoints. LIGHTING Strong natural afternoon sunlight, realistic reflections on wet skin and inflatable surfaces, physically accurate water highlights, soft shadows, subtle atmospheric haze. Sunlight remains consistent throughout the continuous take. AUDIO Pure location sound. Running footsteps, splashing water, inflatable material flexing, distant conversations, excited spectators, water jets, lifeguard whistles, Subject A laughing and reacting naturally, Subject B shouting a spontaneous reaction, heavy splash impact, water dripping directly onto the camera microphone. No music. No cinematic sound effects. VISUAL CHARACTER Extremely photorealistic live-action footage. Premium high-end sports-documentary realism combined with authentic spontaneous UGC imperfections. Natural skin pores, wet hair strands, realistic fabric saturation, detailed water droplets, physically accurate reflections, realistic crowd behavior, subtle lens imperfections, natural exposure, realistic depth of field. Absolutely no CGI appearance, no plastic skin, no artificial cinematic overgrading. CONTINUITY LOCK ONE SINGLE CONTINUOUS TAKE from 00:00 to 30:00. No cuts. No hidden cuts. No camera teleportation. Subject A and Subject B remain identical to their respective face references throughout. Preserve wardrobe, hairstyle, accessories, body proportions, wetness progression, environment, obstacle positions, sunlight direction and physical geography. Every action must logically cause the next action. FINAL QUALITY TARGET The footage must be indistinguishable from an expensive real-world water-park challenge filmed by a skilled human camera operator. The spectacle comes from authentic physical interaction and unpredictable human behavior—not CGI, fantasy, exaggerated destruction, or artificial visual effects.

【风格】 真人写实的高空奇观搞笑短视频，手持手机广角摄影质感。真实人物、直升机内饰、火山岩和橙红色熔岩，结合电影级超现实泡沫效果。自然光、真实材质、轻微飞行振动和运动模糊，突出高空落差、喷发逼近的压迫感，以及结尾人物狼狈大笑的喜剧反差。 白色泡沫具有细密气泡、湿润反光、黏附感和重量。喷涌时成团翻卷，撞击后飞溅、堆积，附着在人物和机舱表面后缓慢滑落。不是烟雾、棉花、雪或纯白光效。 【时长】 15秒，9:16竖屏。全程一个连续手持长镜头，正常速度，不切镜、不慢放、不定格。前半段展示和投掷动作清楚，中间留出等待，后半段泡沫突然爆发，结尾留足人物大笑的时间。 【场景】 一架直升机悬停在火山口上方，侧面舱门完全敞开。摄影机位于机舱内，靠近舱门。暗色门框、灰黑色防滑地板、金属门槛、深色座椅与安全带真实可见。 下方是一个巨大的近圆形熔岩池，橙红色熔融物缓慢翻滚，金黄色亮斑不断变化，四周覆盖黑色火山岩壳，分布细密发光裂缝。远处山脊笼罩在灰色薄雾中。开场人物、机舱与镜头均干净，没有泡沫。 【角色】 一名成年黑人女性探险者，棕色马尾，荧光黄色短袖上衣、灰绿色工装长裤、深色靴子，佩戴黑色安全背带与带麦克风的航空耳机。安全系绳连接机舱固定点，人物始终留在舱内。 开场兴奋地向摄影机展示道具；推动巨物时有真实负重感；泡沫冲来时本能缩头后退；最后满身泡沫，看向摄影机哈哈大笑，笑声与嘴型、呼吸和肩膀颤动同步。 【道具】 一颗巨型白色曼妥思薄荷糖，直径约80厘米、厚约30厘米，圆润厚实的扁圆造型，正反面略微鼓起，边缘圆滑，表面是哑光白色糖衣，完全不透明。 糖果竖立在人物胸腹前，上缘接近胸口，下缘接近膝盖，宽度超过躯干。必须双臂环抱，双手扣住左右两侧，前臂夹紧，屈膝承重。不能单手拿起，不能变成小糖粒、薄纸片、球体或透明冰块。全片只出现这一颗糖。 【镜头】 开场近距离手持中景，人物位于左侧，巨型白色糖果占据中央，敞开的舱门位于右侧。摄影机同时拍到人物表情、双臂、糖果完整形状与门槛。 糖果离手后，摄影机向右下方俯转，追踪下坠，但始终留在机舱内。随后持续俯瞰熔岩落点，等待反应。泡沫冲来后，摄影机迅速后撤并转回人物，最后稳定在满身泡沫的人物中近景。 [00:00–00:01.30] 动作段1：抱着巨型曼妥思，兴奋展示 人物侧身站在敞开的舱门旁，双脚前后错开，膝盖微屈。双臂抱紧巨型白色曼妥思，糖果竖在胸腹前，占据画面中央大部分空间，头部和肩膀露在上方，裤腿露在下方，明确建立巨大尺寸。 她看向摄影机，露出兴奋的笑容，双手重新扣紧糖果两侧，将它向怀里提稳一下。前臂绷紧，肩膀略微耸起，表现糖果的分量。马尾被舱门外的风吹向侧后方。 音效：直升机旋翼声、机舱振动声、高空风声。 [00:01.30–00:02.30] 动作段2：转向舱外，屈膝送到门口 人物将目光转向舱外下方，肩膀和躯干一起向门口转动。双膝进一步弯曲，身体重心降低，把糖果下缘移到金属门槛附近。 她双臂逐渐伸出，糖果从贴近胸腹的位置向外移动，保持明确厚度和重量。朝向熔岩的白色表面出现轻微暖色反光，其余部分仍是哑光白色。 摄影机轻微向右下方倾斜，重点转向双手、巨型糖果与门槛。 音效：衣物和背带摩擦声，人物短促用力呼气。 [00:02.30–00:03.10] 动作段3：推出门槛，双手松开 人物双腿逐渐伸展，双臂合力把糖果推过门槛。糖果顶部先向舱外倾斜，重心越过边缘后，左右手先后松开，糖体完全离手。 巨型糖果沿向外下方的轨迹坠落，缓慢开始翻转。人物双掌短暂停在空中，随后迅速收回，身体重心退回机舱。 摄影机立即沿糖果运动方向向下俯拍，人物脸部退出画面，左上角短暂保留前臂，门槛与防滑地板留在下方前景。 音效：风声增强，鞋底摩擦地板声，旋翼低频持续。 [00:03.10–00:04.60] 动作段4：糖果翻转下坠，落入熔岩 巨型曼妥思从近处清晰的厚扁圆形，逐渐变成远处的非常小的白色圆片。下坠时正面倾斜、侧面转向摄影机，再继续缓慢翻转。糖果实际尺寸与形状不变，只因距离增加而在画面中连续缩小。 摄影机持续向下俯拍，将圆形熔岩池移到中央。左下角留着一条斜向金属门槛，保持高空舱门视角。 糖果以倾斜姿态落入熔岩亮区，局部表面短暂凹下，周围熔融物向外翻开。白色糖体被翻涌的熔岩逐渐遮没，没有立即出现大喷发。 音效：旋翼声、风声，一声遥远低沉的落入声。 [00:04.60–00:07.00] 动作段5：等待反应，白色小泡逐渐聚集 摄影机保持俯视同一处落点，只有轻微手持振动。熔岩继续缓慢翻卷，金黄色纹理改变形状，周围黑色岩壳与裂缝保持固定位置。 先保留一小段没有明显变化的等待。随后落点出现零星白泡，细小气泡接连鼓起、破裂，逐渐聚成一小片浓密泡沫。 白色泡沫缓缓扩大，中央开始隆起，但大部分熔岩池仍然可见。画面形成“好像只是冒一点泡”的错觉。 音效：持续旋翼声，细微嘶嘶声和起泡声逐渐变密。 [00:07.00–00:09.00] 动作段6：泡沫骤然爆发，巨大白柱直冲镜头 落点中央突然隆起，白色泡沫由小团迅速变成粗大的喷涌柱。新的泡沫不断从底部向上推挤，顶部成团翻卷膨胀，整体以惊人的速度升向高空。 泡沫从中心向外覆盖熔岩池，橙红色熔岩逐渐只剩外围一圈。巨大的白色浪头向两侧展开，直接朝舱门逼近，气泡纹理随接近而越来越清楚。 摄影机先抬起一点视线追踪泡沫，随后明显抖动并开始后撤。几团先到的泡沫越过门槛，拍在地板和镜头边缘，留下黏附的白色湿点。 音效：起泡声骤然增强为浑厚喷涌声、湿润扑打声，人物从画面外短促惊呼。 [00:09.00–00:11.50] 动作段7：泡沫灌满舱门，人物从头到脚被覆盖 摄影机后撤并转向舱内，人物、座椅和内壁重新进入画面。人物抓紧把手，屈膝缩头，另一只前臂抬到脸前，向后退一步。 紧接着，白色泡沫浪越过门槛涌入机舱，先撞上她的前臂和肩膀，再铺向胸腹、双腿和靴子。另一团泡沫盖住头顶与耳机，形成歪斜的厚厚泡沫帽，马尾只露出几缕。 座椅靠背、坐垫、门框和地板同时被泡沫覆盖。泡沫碰到表面后黏住、堆积，再缓慢滑落；地面泡沫向舱内推进，包住靴面。黄色上衣逐渐只剩零星色块，黑色背带大部分被埋住。 镜头被一团泡沫短暂擦过，局部泛白，随后泡沫滑向边缘，重新看清人物。舱外喷涌开始减弱，人物稳稳留在舱内，满身泡沫，没有被冲走。 音效：连续“噗、哗、啪”的泡沫撞击声，人物惊呼开始转为忍不住的笑声。 [00:11.50–00:15.00] 动作段8：发现满身泡沫，擦开脸后哈哈大笑 摄影机稳定在人物中近景. 整个机舱和门框覆盖大片白色泡沫，脚边地板堆着厚厚泡沫。她从头到脚都被裹住，肩头和手臂还不断有泡沫缓慢滑下。 她先低头看自己的胸口，再抬起挂满泡沫的手，短暂愣住。随后用手掌抹开眼睛和嘴边的泡沫，让五官露出来；头顶、耳机、脸颊边缘和衣服仍保留大量泡沫。 她抬眼看向摄影机，先咧嘴，随即仰头发出自然畅快的“哈哈哈哈”大笑。眼睛眯起，嘴巴张开，肩膀和胸口随着笑声连续起伏。低头再看一眼自己狼狈的模样，又笑得更厉害。 她举起满是泡沫的手，向摄影机展示。镜头轻轻靠近，让笑脸清晰可见

30-second 16:9 realistic stranger video-call recording. Two imperfect webcams, dark chat interface, subtle compression, ordinary room lighting, natural background noise, no music. 0–5s: A young South Asian man and a Japanese woman connect. They exchange a quick greeting and smile. 5–10s: He asks her to name the first unusual object she can see in her room. She looks around and answers: Woman: “A red mug.” 10–15s: He laughs softly and says: Man: “Okay, remember that.” He reaches down and picks up a small empty-looking paper bag from beside his chair. 15–20s: Without any visible trick, he folds the bag closed, brings it toward the webcam, then opens it. Inside is a red mug matching what she described. Keep the reveal completely physical and unexplained—no VFX, glow, smoke, teleportation, or magical effects. 20–24s: He slowly turns the mug around so she can clearly see it, looking amused but remaining calm. Man: “This one?” 24–30s: She stares at the screen, confused. After a brief video-call delay, she leans toward the camera and says: Woman: “Wait… what?” Her expression gradually changes from confusion to disbelief. Hold on her reaction and cut. Make the conversation casual and spontaneous. Webcam imperfections, slight autofocus shifts, realistic pauses, natural facial movement, believable hands, and restrained reactions. Avoid: cinematic visuals, perfect studio lighting, artificial skin, exaggerated acting, screaming, supernatural effects, visible object before the reveal, distorted hands, extra fingers, rapid cuts, or rushed dialogue.

snow sliding off the branches and lifting away in soft sheets, leaving the dark green needles behind on branches that do not move

Create an ultra-realistic 30–40 second Korean found-footage horror sequence inside a high-rise apartment building at 2:17 AM. A young Korean woman carrying grocery bags enters an empty elevator. An elderly security guard watches the CCTV monitors downstairs.