Prompts de Seedance 2.0: 7.965+ exemplos e guia

[主役キャラ] = 黒髪の女の子 [相手/対象] = ポニーテールのお母さん [舞台] = 家の中 [異常ギミック] = テレビ画面 【動画スタイル】 15秒。日本のフルカラーアニメ。予算5億円規模の高密度2Dアニメーション。予告編のような高速カット。目、手、足、背中、反射、広角の引き、急速プッシュインを交互に使う。字幕なし、透かしなし、BGMなし。音は呼吸、環境音、カットごとの短い効果音、最後の一言。 【映像構成】 [主役キャラ]の目線だけで[舞台]と[相手/対象]の意味が変わっていく。[異常ギミック]を各ショットのつなぎに使い、最後に「見ていた側」と「見られていた側」が反転する。 【Shot 1：片目のフック】 [主役キャラ]の片目を画面いがいっぱいに映す。瞳の中に[舞台]が反射している。[異常ギミック]が瞳の奥を横切る。 【Shot 2：足元】 [舞台]の床。足音が一回。足元だけが光や影を踏む。 【Shot 3：手元】 [主役キャラ]の手が小さく動く。何かを掴む直前で止まる。 【Shot 4：対象】 [相手/対象]が一瞬だけ映る。ピントは浅く、顔や全体を見せすぎない。 【Shot 5：反射】 鏡、窓、スマホ、金属、瞳の反射に[主役キャラ]が映る。実物の位置と反射の位置が少し違う。 【Shot 6：広角の引き】 [舞台]全体を一瞬見せる。どこに[主役キャラ]がいるのか分からない構図にする。 【Shot 7：接近】 急速プッシュイン。[相手/対象]のすぐ近くに[異常ギミック]が現れる。 【Shot 8：反転】 カメラが振り返るように高速パン。そこに[主役キャラ]が静かに立っている。 【Shot 9：目線固定】 [主役キャラ]がカメラを見つめる。動かない。背景だけが少し揺れる。 【Shot 10：最後の一言】 [主役キャラ]が短く「最初から見てた」と言う。最後のフレームは片目の接写に戻って終わる。

【概要】1985年、田舎の祖父母の家で夏休みを過ごす10歳の男の子の1日。帰省中の若い叔父が家庭用ビデオカメラで、声をかけずに離れた場所から記録している。480p、16:9、15秒。7ショット（各約2秒）のハードカットで、7か所すべて別の場所。トランジションやフェードは使わない。被写体はカメラに向かって何もしない：目線を送らない、物を見せない、手を振らない、ポーズを取らない。遊びに没頭している。演技ではなく、日常の途中を切り取った断片。会話・セリフ・ナレーションは一切なし。 【被写体】美貌の定義：将来ハンサムになると誰もが言う、子役級の日本人の美少年。水に入った瞬間の子どもの顔が一番いい。顔：短い黒髪に寝癖、太い眉、切れ長できりっとした二重の目、日焼けで鼻の頭の皮がむけている（現実的な肌の質感）、白い歯、あごに小さなかさぶた。服（普段着）：白いランニングシャツ（洗いざらしで少し黄ばんでいる）、緑の半ズボン、麦わら帽子（つばがほつれている）、首に白いタオル、素足にゴム草履。仕草の癖：帽子を後ろにずらす、弟をちらっと見る。全ショットで顔・髪・服装が完全に同一。 【登場人物】弟（6歳）：まん丸の顔、坊主頭、白いランニングと赤い半ズボン。兄のやることを全部真似する。怖がり。誰もカメラを見ない。 【時代・場所・光】1985年8月の田舎。①古い日本家屋の縁側（朝の光、緑色の渦巻き型の蚊取り線香から細い煙、卵かけご飯）②田んぼのあぜ道（真夏の白い光、自転車）③川の浅瀬（きらめく水面）④川の岩場（小さな岩から浅瀬へ）⑤川原の石の上（スイカ、強い日差し）⑥神社の境内の木（木漏れ日、セミ）⑦夕方の縁側（線香花火：細いこよりの先で小さな火の玉がぱちぱちと火花を散らす手持ち花火。青い夕闇）。看板・文字は画面に入れない。 【カメラ】1985年の家庭用ビデオカメラの質感：色がにじみ、解像度が低く、水面の反射が白飛びし、輪郭が甘い。日付表示は出さない。手持ち、自然な揺れ、不完全な構図、ズームの迷い、露出のゆらぎ。被写体から3〜6mの距離。被写体はカメラを気にしない。スタビライズ、ジンバル、ドローン、スローモーション、映画的照明、商業カラーグレーディングはなし。カメラは常に撮影者が手に持っている位置（立つ・座る・しゃがむ・歩きながら・隣の席から）から撮る。人がいられない場所（空中・水中・天井・真上・車の外から走る車内・被写体の目の前数センチ等）からのアングルは使わない。撮影者は同じ空間に確かにいる一人の人として、被写体が動けば少し遅れて追い、ときどき構図が甘くなる。 【ショット】（各約2秒。各行＝場所・光／没頭していること／内側の感情と、それが漏れる小さなしぐさ／カメラ位置） 1. 縁側・朝。卵かけご飯を急いでかきこんでいる。目はもう庭の先を見ている。感情：早く川に行きたい。蚊取り線香の煙が横に流れる。カメラ：縁側の外から。 2. あぜ道・自転車。弟を荷台に乗せて立ちこぎ。口を真一文字に結んで前を見る。感情：兄貴風の頼もしさ。風で帽子が後ろにずれる。カメラ：後ろから追う、揺れる。 3. 川の浅瀬。腰まで浸かって網を構え、息を止めて魚を見つめる。感情：集中。弟が後ろでじっとしている。カメラ：岸から、水面の反射が白飛び。 4. 岩場。小さな岩から浅瀬へ飛び込み、水しぶきの中から顔を上げて大笑い。感情：解放感、爽快。カメラ：岸から少し離れて、しぶきが画面にかかる。 5. 川原の石の上・スイカ。かぶりついて、種を遠くへ飛ばし、弟をちらっと見てにやりとする。感情：競争心。カメラ：横から、強い日差し。 6. 神社の木・木漏れ日。虫取り網で捕まえたセミを弟の目の前に突き出し、仰け反る弟を見て大笑い。感情：いたずら心。カメラ：弟の後ろから。 7. 縁側・夕方。線香花火の火の玉を真剣な目で見つめ、落ちた瞬間、消えた先をしばらく見つめている。感情：一日が終わる小さな寂しさ。カメラ：縁側の端に座って横から。その途中、およそ00:14で録画は突然ブラックアウトに切り替わる。フェードアウトなし。 【小物・道具のディテール】網は竹の柄に緑の網。自転車は黒い実用車で荷台つき。麦わら帽子はつばがほつれている。スイカは縞で、皮の近くまで食べる。蚊取り線香は緑の渦巻きで陶器の豚の入れ物に入っている。線香花火はこよりの先に赤い紙。 【文字の禁止】画面内に読める文字・ロゴ・看板・ラベル・数字を一切出さない。 【物理・一貫性】現実世界の物理。余分な指、融合した手、歪んだ解剖学、浮遊物、消える物体、突然の変形はなし。水しぶき、濡れた髪と服、火花は現実的に。足は地面に接地。麦わら帽子、タオル、弟の服装は全ショットで同じ。 【音】自然な環境音のみ（ショットごとに切り替わる）：茶碗の音、自転車のチェーンと風、川の流れ、飛び込みの水音、スイカをかじる音、セミの大合唱、線香花火のぱちぱち。言葉は一切なし。撮影者のときどきの小さな笑い声・息づかいと、被写体の笑い声・息づかいのみ可。音楽なし。ナレーションなし。人工的な音響効果なし。 【雰囲気】大人が見て「こういう夏休みがあった」と思う、昭和の田舎の少年のなんでもない1日の記録。演技ではなく、遊びの途中の断片。懐かしく、まぶしく、深く人間的. カメラがたまたまそこにあったような感じを優先。

HALF A BEAT LATE — 15 seconds. Dark gothic psychological horror. 16:9. Real-time action, controlled cuts, continuous spatial and emotional progression. REFERENCES Image 1 defines adult female jester Sin and her identical evil mirror double: exact face, white makeup, red nose, dark eye makeup, painted tear marks, red twin-horn hair, gold crown with bells, ivory ruff, red-and-blue court costume, pointed shoes. Image 2 defines the room and opening frame: burgundy curtain and warm wall sconce on the left, charcoal walls, wooden floor, large antique gold mirror on the right. Preserve the room, mirror frame, and light direction. IDENTITY AND STAGING SIN starts outside the mirror: increasingly frightened, resisting capture. DOUBLE starts inside the mirror: amused, disobedient, smiling, physically assertive. After the exchange, DOUBLE is outside in the room and SIN is trapped behind intact glass. Begin on Image 2’s empty-room composition. Sin enters from lower-left, walks diagonally to the mirror, and stops about 55 cm from the glass, torso facing it, back and partial profile toward camera. Keep her slightly off the mirror center so both real and reflected faces stay visible. Camera remains on the mirror’s left-front side; the glass stays broadly visible, not edge-on. Natural perspective, readable facial detail, enough depth of field for both faces. During the exchange, Sin passes inward through the screen-left half of the mirror while Double steps outward through the screen-right half. Their paths stay separate. Afterward Double takes one short settling step toward screen-left, ending in the left foreground while trapped Sin remains visible inside the mirror on the right. PERFORMANCE LOCK Sin stops smiling once she notices the anomaly and never smiles again. Double keeps smiling and never copies Sin’s fear. Sin shakes her head, retreats, gasps, loses balance, resists. Double refuses to follow, advances, grips, pulls, plants its feet. After the exchange, Sin panics inside the mirror while Double turns toward camera, smiles wickedly, then whispers. Both faces stay alive and natural, never frozen or deformed. 0-3 SECONDS — THE SMILE REMAINS A heel clicks offscreen. Sin enters immediately, takes two or three short grounded steps to the mirror, relaxed posture, chin slightly raised, small habitual smile. Weight transfers naturally through each foot; skirt and crown bells follow with soft motion and sound. Her reflection appears normal at first. Sin briefly adjusts the side of her crown with her right hand, lowers it, exhales, and her mouth turns neutral. The reflected hand lowers about half a second late. Double keeps smiling and slowly shows its teeth. A restrained push-in keeps both faces readable: Sin neutral and unsmiling, Double still smiling. 3-5.5 SECONDS — OPEN DISOBEDIENCE Sin looks from Double’s mouth to its eyes, brow tightening, then slowly shakes her head once. Double does not shake. It tilts its head slightly and widens its amused smile. Sin catches her breath, draws her chin back, retreats about 20 cm, and raises her right hand at chest height, palm toward the mirror. Double leans closer to the glass as Sin increases her real distance. Keep both opposing movements visible in the same oblique frame. 5.5-8 SECONDS — GRIP, THEN PULL Hard cut to a same-side medium shot holding both faces, Sin’s raised right wrist, and the mirror’s left edge. A localized patch of glass turns into a dark elastic membrane. Double’s left hand reaches through into the room and grips Sin’s right wrist firmly. Double pulls inward. Sin’s eyes snap to the captured wrist; her lips part in a sharp inhale. Her left hand grabs the mirror’s left frame, rear foot braces on the wooden floor, knees bend as she pulls backward. Double keeps its cold smile and retracts with steady force. Sin loses her hold on the frame. Her right shoulder is pulled forward first, then her torso; her rear shoe begins to slide. She resists the whole time and never walks voluntarily into the mirror. 8-11.5 SECONDS — ONE IN, ONE OUT The camera withdraws slightly to reveal the mirror boundary and landing foot. Double steps outward through the screen-right half of the mirror and plants one pointed shoe firmly on the wooden floor before taking weight. Sin is pulled inward through the screen-left half. Double keeps hold until Sin crosses the surface, then releases her wrist. Their identical faces pass on separate paths: Sin terrified, trembling, ragged breath; Double composed with a small victorious smile. Red hair and moving fabric briefly occlude the crossing point. Use one concealed continuity cut inside that occlusion only. When the occlusion clears, the exchange is complete: inside the mirror Sin stumbles one step away, then turns and lunges back toward the glass; outside, Double completes its settling step toward screen-left. The membrane seals back into intact reflective glass before Sin reaches it. Her palms hit the inner surface and stop. The gold frame stays rigid. No shattering. 11.5-13 SECONDS — TURN FIRST Hard cut to a same-side medium close shot that clearly contains both faces. Double starts facing the mirror, then shoulders and torso turn toward the camera until its face is fully visible. Show the full body turn before any frontal smile. Behind it, trapped Sin faces outward from inside the mirror. She looks down at her blocked palms, urgently scans the frame edges, then looks at Double. She strikes the glass once and opens her mouth to call out, unheard. Sin remains an independent trapped person, not a synchronized reflection. 13-14 SECONDS — THEN THE EVIL SMILE Double finishes turning and stands still, shoulders relaxed, eyes on the lens. Only now does one corner of its mouth rise into a clearly wicked, satisfied smile with a little tooth visible. Keep Sin sharply readable behind it: raised inner brows, tense lower lids, downturned mouth, fast chest movement, palms pressed to the glass. Her fear continues throughout Double’s smile. 14-15 SECONDS — FINALLY, “SHH” Double raises its right index finger to its lips, slightly gathers its smile, and whispers directly to camera: “Shh.” The voice is low, intimate, steady. Sin strikes the glass again with one hand while the other stays planted, mouth open in a silent plea. She never mirrors the finger gesture, smile, or gaze. Hard cut to black at exactly 15 seconds. COLOR AND LIGHT Single consistent palette: charcoal brown #211E1B for walls and deep shadow, dark burgundy #4B171A for curtain and red costume shadow, aged brass #9C7544 for mirror frame, crown, and bells, ivory grey #D8D0C1 for makeup and ruff, deep blue-grey #29313B for blue costume fabric and cool shadow. Preserve Image 2’s warm left-side sconce at about 2700-3000K. Localized old-gold highlights, subtly cool shadows, neutral midtones, restrained saturation, dense textured blacks. Protect the white makeup from orange or blue contamination. Mirror glare never hides the reflected face. No color, exposure, or lighting reset during the exchange. AUDIO Heel clicks, fabric movement, and delicate bells establish the room. Sin’s interrupted breath builds into gasps during resistance. Add wrist-grip friction, shoe scrape, frame contact, and restrained membrane tension. After the mirror seals, Sin’s voice and breathing are inaudible from the room; her impacts become faint muffled thuds. Double’s final “Shh” is the only spoken line. No music, narration, or subtitles. CONTINUITY LOCKS Exactly one original Sin and one evil Double. Preserve separate identities, correct wrist ownership, grip contact, resistance, and inward/outward travel paths across cuts. No merged bodies, duplicate occupants, extra hands, or position jumps. Before the anomaly, the reflection behaves naturally. After the exchange, both performances remain independent. Final image contrast must stay clear: Double o

30-sec cinematic Korean folk-horror ritual. Keep the Word character’s face and outfit consistent throughout. 1. 0–4s: Word character kneels pale and weak in a dim traditional Korean house; worried family, candles, incense, shaman preparing the ritual. 2. 4–8s: Salt circle and talismans surround her; candlelight flickers as her eyes twitch and fingers convulse. 3. 8–12s: Her head snaps unnaturally; guttural inhuman voice, dark veins spreading, body twisting as the ritual intensifies. 4. 12–16s: Eyes turn black; shaman chants frantically and throws rice/salt; her body convulses violently. 5. 16–20s: She rises unnaturally; shaman presses a burning talisman to her forehead as wind tears through the closed room. 6. 20–24s: Talisman burns, flames surge, she collapses and becomes completely still. 7. 24–27s: Her eyes reopen human; family rushes to her in relief while the exhausted shaman watches. 8. 27–30s: Wide aftermath: scattered salt, burnt talismans, melted candles. One intact talisman emits faint dark smoke, leaving an unsettling unresolved ending. Style: cinematic Korean folk horror, dark wooden interior, warm candlelight, deep shadows, tense atmosphere, realistic, unsettling, dramatic camera movement.

Duration: 20秒 Aspect ratio: 9:16 Overall style: POV男友第一人称短剧，真实手机手持轻微呼吸晃动，室内自然光通透，手机滑动+点击瞬间换装，节奏轻快甜宠，竖屏 【场景】明亮客厅/卧室，落地窗自然光，背景简洁温馨，地面干净无杂物 【角色】女主（@图片1，开场穿角色原本服装）；男主为POV视角，全程不露脸，仅画面下方出现持手机的男性手 【道具】一部现代全面屏手机，竖持。屏幕是简洁时尚穿搭App：上方大预览图，下方横向可滑动的服装卡片。每次滑动有跟手惯性，点击卡片时屏幕轻微缩放+短震反馈 SHOT 1（00:00-00:03.5）开场展示 画面：POV，女主站画面中央，开心地原地转半圈展示穿搭，双手轻提衣摆，眼睛亮亮看镜头。 台词：女主:"老公，看我今天穿得怎么样？" 音效：室内底噪，衣料轻响。 Constraints: 女主全身入画，开场服装=角色原本服装。 SHOT 2（00:03.5-00:06.5）掏出手机+第一次滑动 画面：画面下方男主手举起手机入镜，屏幕亮起，显示穿搭App。拇指从右向左滑过卡片，停在@衣服1的卡片上，预览图放大。 台词：男主（画外音，慵懒）:"还行。换这套看看。" 音效：屏幕滑动轻摩擦声。 Constraints: 手机屏幕清晰可读，卡片上能看出是@衣服1；男主只出手，不露脸。 SHOT 3（00:06.5-00:09.5）第一次点击瞬变 画面：拇指点在@衣服1卡片上，屏幕短震。点击落下的同一帧，女主服装瞬间变成@衣服1，站位、姿势、发型完全连续；她低头看衣服，瞪大眼张嘴，再抬头看镜头。 台词：女主（惊讶）:"诶？！" 音效：清脆"滴"+换装"叮"，衣料展开微响。 Constraints: 换装发生在点击同一帧，单帧瞬变；可加一道从上到下扫过身体的柔和微光；禁止叠化、黑闪、人物跳位。 SHOT 4（00:09.5-00:13）第二次滑动+点击瞬变 画面：拇指再次左滑，卡片停在@衣服2，点一下。同一帧女主换成@衣服2；这次她反应过来，脸颊微红，手指绞在身前，害羞小幅扭身，眼神躲闪又忍不住看镜头。 台词：男主（画外音）:"这套也行。" 女主（小声）:"你……乱点。" 音效：滑动声+"滴"+"叮"，轻快音乐渐起。 Constraints: 滑动必须完整看见从@衣服1卡片滑到@衣服2；换装仍是点击同一帧硬切。 SHOT 5（00:13-00:16.5）第三次滑动+点击瞬变 画面：再滑到@衣服3，点一下。同一帧换成@衣服3；女主已经习惯，咬一下下唇，抬手整理新衣服领口/裙摆，眼神带点得意看镜头。 台词：女主:"这套……还挺好看的嘛。" 音效："滴"+"叮"，音乐节奏加快半拍。 Constraints: 三次反应递进：惊讶→害羞→得意；每次都有小动作，禁止呆立。 SHOT 6（00:16.5-00:20）第四次滑动+点击+收尾 画面：最后一滑停在@衣服4，拇指点下。同一帧女主瞬变为@衣服4；她直接笑出来，上前一步凑近镜头，狡黠眯眼。手机仍举在画面下方，屏幕停在@衣服4预览。 台词：男主（画外音，满意）:"这套定了。" 女主（笑着凑近，嗔怪）:"你好坏啊——我好喜欢！" 音效："滴"+"叮"，音乐在女主台词落点收在甜甜重拍。 Constraints: 结尾女主凑近但不过分贴脸；手机始终在画面下方可见；服装顺序锁定：开场=原本 → @衣服1 → @衣服2 → @衣服3 → @衣服4，不可乱。 【导演铁律】 1. 换装铁律：每次换装发生在“点击”同一帧，单帧瞬变。女主站位、姿势、发型、脸型跨越换装帧完全连续，只有服装改变。 2. POV铁律：镜头=男主眼睛，全程轻微手持呼吸晃动；男主只出手+画外音，禁止露脸。 3. 手机铁律：每次必须先滑动再点击，滑动跟手、卡片切换清楚；禁止凭空换装、禁止不看屏幕直接变。 4. 服装引用：开场=@图片1原本衣服；第1次=@衣服1；第2次=@衣服2；第3次=@衣服3；第4次=@衣服4。 5. 台词口型清晰，与点击节拍对齐。 Negative: 叠化换装，黑闪，白闪，烟雾，旋转遮挡，人物瞬移，姿势断裂，脸型变化，发型变化，鞋子乱变（除非衣服图自带），露男主脸，横屏，字幕，水印，Logo，背景突然更换，多人入画，宽松居家服除非衣服图就是，镜头剧烈抖动，手机屏幕模糊不可读。

【风格】 真人实拍质感的拟人猫咪搞笑短视频。猫咪保持真实的毛发、猫脸、胡须、耳朵和肉垫，身体能够拟人坐立、抱臂和操作麻将，但不能变成人脸、人手或卡通身体。室内柔和顶光，清晰自然的毛发细节、皮革纹理和麻将塑料反光。笑点依靠动作、停顿和两只猫的反应形成，不依靠夸张卡通特效。 【时长】 10秒，9:16竖屏。全程单一连续镜头，正常速度，不切特写、不慢放。 【场景】 现代居家麻将房，米白色墙面、简洁装饰画、右后方浅灰色落地窗帘，天花板有方形灯槽和嵌入式照明。前景是一张绿色绒面自动麻将桌，深色桌框，左下方露出一部分圆形控制盘。 麻将为蓝色牌背、白色牌面，牌面有红绿黑色图案。桌面中间散放少量打出的牌，近处与左右两侧排列着其他玩家的牌墙。西部牛仔猫面前有一排横向整齐排列的麻将，牌的蓝色背面朝向摄影机，看不见里面的花纹。画面下方和右侧偶尔出现其他玩家的真人手与前臂，完成摸牌、放牌和整理牌墙的小动作，不露出其他人的脸。 【角色一：坐着打麻将的西部牛仔猫】 使用参考图一中的灰褐色猫形象。 它坐在麻将桌对面的深色椅子上，位于画面中央偏右，面对摄影机。桌面高度约在它腹部，两只前爪能自然伸到桌面。帽檐略微上翘，始终能看清眼睛。气质像自以为牌技高超的冷面高手，动作越花哨，表情越淡定。嘴里还夹着一根烟，非常酷拽的表情。 只借用参考图的猫咪外貌和牛仔服装，手中只操作麻将；不加入参考图中的枪械、滑板、设定卡文字或其他装备展示。 【 角色二：站着围观的工牌猫】 使用参考图二中的银灰色虎斑猫：颈部佩戴挂绳，胸前悬挂透明卡套，卡套内有猫咪证件照，工牌大小适中且始终可辨认。 工牌猫直立站在牛仔猫身后偏左的位置，头部高于坐着的牛仔猫，身体露出上半身。前半段两只前爪交叉抱在胸前，工牌从交叠前臂下方露出。它始终面无表情地盯着牛仔猫的动作，眼神带审视和嫌弃，身体基本不动，形成和牛仔猫花式动作的反差。 【镜头与构图】 摄影机位于麻将桌另一侧，接近坐着的玩家视线高度，略微向下拍摄桌面。绿色麻将桌占画面下半部，两只猫占中上部。牛仔猫位于前方中央偏右，工牌猫位于后方偏左，始终保持这组前后关系。 轻微手机广角，固定位置手持微晃。两只猫的脸、牛仔猫的双爪、面前整排麻将和工牌猫的工牌都尽量留在画面内。镜头不绕行、不突然推近，结尾依靠动作制造冲击。 [00:00–00:01.40] 动作段1： 开场牛仔猫已经坐在桌前，挺着胸，微微抬起下巴，眼睛半眯，摆出胸有成竹的神态。 它先用右前爪沿桌面向画面左侧轻快伸出，拨动一张麻将，再把爪子收回。随后两只前爪分别按住面前全是背面的横排麻将的左右两端，向中间轻轻归拢，将牌列理齐。动作干净利落，肩膀几乎不动，像对桌上的局面完全有把握。 工牌猫在身后抱臂站立，只用眼睛跟随它的爪子移动，没有笑容，也不帮忙。右侧一只真人手把一张麻将放入桌面中间，随即收回。 音效：清脆的麻将碰撞声，轻微室内环境声。 [00:01.40–00:03.20] 动作段2： 牛仔猫从面前牌列中拿起一张蓝背麻将，左前爪在胸腹前摊开，肉垫朝上，形成承接姿态；右前爪将麻将压在左爪上 亲亲抚摸，摆出一副不用看牌就知道的表情。 工牌猫抱臂不动，目光向下，盯着它搓牌的双爪。 音效：几下节奏紧凑的“嗒、嗒、啪”麻将声。 [00:03.20–00:05.80] 动作段3： 牛仔猫右爪再次夹持一张麻将，从桌面抬起，向摊开的左爪做一次短距离递送。麻将贴近两爪之间翻转，随后被右爪带回桌面，轻轻磕在面前牌列旁。 它略微前倾，双爪贴着绿色绒面滑回自己的牌列两侧，再将横排麻将理成一条直线。完成后停顿半拍，抬起眼睛看向摄影机方向，下巴微抬，保持故作高深非常自信的的冷静表情。 桌面右侧和下方的真人手继续正常摸牌、整理牌墙，衬托牛仔猫刻意耍帅的表现。工牌猫的头部略微低下，眼神更加严肃，但仍然抱臂不动。 音效：麻将擦过绒面的轻响，零星碰牌声；停顿时减少音效，突出它装腔作势的神态。 [00:05.80–00:07.70] 动作段4： 牛仔猫上身向前探，右前爪从自己面前伸向画面左下方的蓝背牌墙，爪垫压住最外侧一张麻将，把它从牌墙端部拖出，再沿桌面带回自己面前。 右爪收回时，它的身体重新坐直，头部却故意向上抬，目光看向上方偏右，仿佛不用看牌也知道下一步。左爪搭在自己的牌列旁，右爪把新摸的牌放在牌列附近。 。 音效：摸牌“咔”的一声，随后是一小段尴尬的安静。 [00:07.70–00:08.65] 动作段5： 牛仔猫胸有成竹的表情，做出一副要麻将胡了的动作：转两只前爪分别伸到自己面前整排麻将的两端，向内夹住牌列，试图把整排麻将同时提起并向上翻动。 它双肘弯曲，双爪快速上抬，但中间的麻将没有得到支撑，牌列立刻从中间断开。靠近双爪的几张牌先被带起，中间多张麻将接连滑落；蓝色牌背和白色牌面在空中交替翻转，落回绿色桌面。 失手只发生在桌面上方很短的高度，麻将以不同角度撞击、弹跳、翻倒，向桌面中央和左右散开，碰乱附近打出的牌。整排麻将从有序直线变成一桌凌乱，桌子本身不翻倒。 牛仔猫双爪仍停在半空，短暂保持原本准备展示的姿势，随后发现爪中已经空了。牛仔猫冷面高手的气势瞬间消失，表情略显尴尬。 音效：从一声失手的“咔”迅速变成密集的“哗啦啦、噼里啪啦”，最后几张麻将零星弹跳收尾。 [00:08.65–00:09.25] 动作段6： 牛仔猫双爪尴尬地悬在胸前，先低头看散乱的麻将，再慢慢转头望向身后的工牌猫。耳朵略向两侧压低，眼睛睁圆，表情从自信变成心虚。 工牌猫低头看一眼桌面，再看牛仔猫，眼睛略微眯起。它终于松开交叉的前爪，其中一只爪子从胸前抬起，肘部弯曲，爪垫朝向牛仔猫帽顶。胸前工牌随着解除抱臂的动作轻轻摆动。 保留短暂的对视停顿，让观众先看到牛仔猫意识到自己闯祸。 音效：最后一张麻将落稳的轻响。 [00:09.25–00:10.00] 动作段7 工牌猫抬起的前爪从上方向前下方快速挥出，用柔软爪垫在牛仔猫的帽顶和后脑方向狠狠拍了一下，动作短促明确，像对炫技失败的嫌弃提醒。 牛仔帽受到触碰后直接被扇飞。牛仔猫立即缩脖子、低头，身体向画面右侧躲开，双爪本能抬到脸和帽檐旁挡住，肩膀耸起。它仍坐在椅子上，不摔倒、不离开画面。 工牌猫的前爪跟随动作伸直，再开始收回，脸上始终保持无语的严肃表情；胸前工牌因挥爪轻轻晃动。桌面上留着刚才散落的麻将，其他玩家的真人手短暂停在桌边。 音效：一下轻巧的“啪”，。 【动作与一致性要求】 坐着打麻将的始终是西部牛仔猫，站在身后围观并在结尾抬爪的始终是工牌猫，不交换位置、服装或身份。牛仔帽、围巾、皮衣、虎斑花纹和工牌保持连续一致。 前半段突出“冷脸、稳坐、花式搓牌、抬头装高手”，后半段突出“整排提牌失手、麻将散落、空爪僵住、心虚回头、被轻拍帽顶”。必须先失手撒牌，再有工牌猫的反应和拍帽动作。 麻将是有厚度的硬质长方体，蓝背白面，接触双爪时有明确支撑，掉落时有重力、碰撞与反弹。麻将不凭空复制、不变成纸片、不长时间悬空。工牌

Create an 11-second cinematic dark-fantasy scene featuring a beautiful blonde female warrior queen wearing intricate medieval silver-and-black metallic armor and a delicate dark crown. SCENE 1 — 0–3 seconds: Begin with a dramatic top-down aerial shot. The warrior is flying/falling through thick stormy clouds with her arms stretched outward like she is diving through the sky. Her long blonde hair and pieces of her clothing move naturally in the powerful wind. Far below, a vast fantasy landscape is barely visible through the mist. The camera rapidly moves downward toward her while maintaining a powerful cinematic perspective. SCENE 2 — 3–5 seconds: Transition smoothly into a frontal tracking shot as she moves through the clouds. Her armored shoulders and arms remain extended. Her expression is serious, fearless and emotionally intense. Soft sunlight breaks through the clouds behind her, creating a dramatic halo around her silhouette. Tiny particles and mist pass rapidly across the camera lens. SCENE 3 — 5–9 seconds: Cut into an extreme cinematic close-up of her face. A huge majestic black raven/eagle suddenly flies beside her face, its dark feathers contrasting against her pale skin and blonde hair. The bird moves extremely close to her, almost touching her cheek. Feathers are released into the air and float dramatically around them. She looks directly ahead with an intense, mysterious expression. SCENE 4 — 9–11 seconds: The bird slowly moves away while she turns her head toward it. Her expression changes subtly from determination to curiosity and emotion. Wind continues moving her hair and armor naturally. The background remains filled with bright clouds and atmospheric mist. VISUAL STYLE: Epic high-budget fantasy film, photorealistic CGI, dark medieval fantasy, dramatic natural lighting, realistic metallic armor, highly detailed skin and hair, volumetric clouds, atmospheric fog, floating feathers, realistic bird movement, cinematic depth of field, subtle film grain, anamorphic lens characteristics, rich shadows, dramatic contrast, sophisticated color grading. CAMERA: Smooth aerial tracking, fast cinematic push-in, gradual transition to extreme close-up, shallow depth of field, realistic handheld micro-movement during the close-up, carefully controlled focus pulls, cinematic composition. MOTION: Natural wind physics, realistic hair movement, physically accurate feathers, believable bird flight, flowing fabric, subtle facial expressions and eye movement. No exaggerated or cartoonish motion. MOOD: Mysterious, powerful, emotional, majestic, supernatural and epic — like a scene from a premium dark-fantasy feature film. Maintain consistent facial identity, anatomy and armor throughout the entire video. No face distortion, no extra limbs, no duplicated bird, no unnatural body movement, no text, no s

Create a 15-second, 16:9, 24fps ultra-photorealistic skincare UGC selfie video. REFERENCE PRIORITY Image 1 = exact character identity. Image 2 = exact product appearance. Additional storyboard images = composition, timing and motion only. Keep the woman’s exact face, freckles, skin tone, blue-green eyes, long light-brown caramel hair, blue hoop earrings, pearl necklace, white linen shirt, body proportions and identity consistent throughout. Use Image 2 as the exact Crème de la Mer Moisturizing Cream reference. Preserve the same creamy-white jar, lid, shape, proportions, material, label layout and packaging colors throughout. Never turn it blue, translucent, gel-like, or replace it with another product. STYLE Authentic end-of-day skincare vlog filmed by the woman herself on a smartphone. Relaxed, intimate, slightly tired, low-energy and unperformed. Not a polished beauty ad. No presenter behavior, exaggerated reactions, product-selling gestures or glossy TVC lighting. LOCATION Mediterranean-style villa bedroom at golden hour. White stucco walls, rumpled white bedding, open doorway, soft bougainvillea and olive leaves in the background. Warm natural sunlight enters from frame-left, creating soft golden highlights on her hair, cheek and jaw. No ring light or studio lighting. CAMERA Horizontal 16:9 selfie composition. Keep her face visually dominant, usually 40–60% of frame width. Use asymmetric framing with the product, applying hand or environment on the opposite side. Mostly head-and-shoulders or tight upper-chest framing. Authentic smartphone footage: arm’s-length handheld camera, small tremor, subtle horizontal drift, minor framing corrections, occasional autofocus shifts, mild wide-angle distortion, natural motion blur, slight rolling shutter, subtle sensor noise. Never feel gimbal-stabilized or professionally operated. SHOT 1 — 0:00–0:04 Horizontal selfie close-up. Her face is slightly right-of-center. She pushes loose hair away with the back of her wrist and briefly checks herself on the phone screen. The OPEN white La Mer jar casually enters the lower-left beside her cheek. Natural autofocus shift between face and jar. She exhales softly. Dialogue: “Okay. Last thing, then bed.” SHOT 2 — 0:04–0:07 Horizontal detail shot. Jar on the left, softly blurred part of her face on the right. She gently presses one fingertip into the thick white cream. Show realistic deformation: dense, silky, opaque cream folds under the finger and forms a soft peak when lifted. Briefly show the front label clearly: “CRÈME DE LA MER™” “LA MER” “moisturizing cream” Then she rubs a small amount between two fingertips. Natural rack focus: label → cream → fingertips. Dialogue: “It’s so rich.” SHOT 3 — 0:07–0:11 Return to horizontal face close-up. Her face slightly left-of-center, applying hand on the right. She warms the cream and gently presses small amounts onto both cheeks, forehead and chin. Do not create a thick white mask. The cream appears briefly, then gradually absorbs. Freckles, pores, under-eye shadows and fine skin texture remain visible. Only a subtle hydrated sheen develops. Dialogue: “I just warm it up and press it in.” SHOT 4 — 0:11–0:15 Closest three-quarter selfie framing. She presses two fingertips into her cheek, lifts them, then slowly turns into the warm sunlight. A subtle hydrated sheen catches the cheekbone while freckles and pores remain visible. She looks mostly at her phone screen, gives a tiny delayed half-smile and small nod. The white La Mer jar appears naturally in soft focus near the lower frame. She reaches toward the phone to stop recording, causing a slight natural sideways tilt. Dialogue: “Morning me is gonna be so happy.” PERFORMANCE End-of-day energy: relaxed, slightly tired, casual and natural. Include subtle micro-behaviors: slow blinking, small exhale, natural swallowing, tiny head tilt, eyes checking the screen, loose hairs falling across her cheek, small posture shifts, delayed half-smile. VOICE Soft female voice, slightly tired and flat. Close smartphone microphone sound. Natural breathing. No studio-clean voice-over. SKIN REALISM Pore-level realism. Visible freckles, pores, fine forehead texture, faint under-eye shadows, tiny imperfections, vellus facial hair, moist eyes, natural catchlights, real lip texture. The moisturizer must behave like a real thick cream before absorbing. No plastic skin, beauty-filter smoothing, artificial glow or waxy face. STRICT NEGATIVE No face drift. No character replacement. No hairstyle or outfit change. No packaging change. No blue jar, blue lid, blue cream or translucent gel. No Laneige or other skincare product. No product duplication. No fake label. No commercial presenter pose. No ring light. No glossy ad lighting. No extreme beauty retouching. No plastic skin. No watermark, subtitles, UI or random text.

Create a 30-second ultra-realistic cinematic educational video set inside a modern university physics laboratory and lecture classroom, featuring a confident male physics professor entering the room with a coffee cup and teaching materials while students prepare for class. Show the professor explaining physics concepts on a large digital board, with mathematical formulas and a clear inclined-plane diagram visible behind him as students watch attentively. Transition to a detailed close-up of a small metal ball positioned on a transparent inclined track, emphasizing the experiment and the principles of motion and gravity. Show the professor and students gathering closely around the apparatus as he demonstrates how the ball moves along the track, using natural hand gestures and realistic physical interactions. Capture students reacting with curiosity and discussing the experiment while the professor carefully guides them through the observation. Include cinematic close-ups of the rolling ball, the transparent track, the professor’s hands, and the students’ focused expressions. Maintain realistic gravity, friction, momentum, object weight, contact, and natural movement throughout the experiment, with no physically impossible motion. Use smooth camera movements, shallow depth of field, realistic classroom lighting, detailed facial expressions, and authentic university laboratory surroundings. End with the professor standing beside the completed experiment as the students observe the result, creating a professional, inspiring, photorealistic educational-science atmosphere.

30秒｜16:9｜8镜｜多镜头序列｜Hard Cuts 多镜头序列，镜头之间使用明确硬切。全片保持同一个知夏、同一只三花猫@a2b2a146-e6b3-4238-9a2b-f7130db0c416“花生”、同一颗五彩毛茸茸球、统一环境逻辑与视觉风格。无音乐、无配乐，仅真实自然同期声。

Create a 15-second ultra-realistic early-2000s Sony MiniDV home video. Subject: Young Korean woman , naturally attractive, realistic skin, minimal makeup, long dark hair loosely tied back. Oversized faded lavender T-shirt, loose gray lounge pants, white socks, thin silver necklace. Preserve identical identity, face, hair, proportions and appearance throughout. Location: Small old Seoul apartment balcony on a sunny summer morning—concrete walls, metal railing, clothesline, colorful laundry, plastic basket, potted plants, neighboring buildings and distant rooftops. Bright, slightly hazy sunlight. Style: Authentic personal early-2000s MiniDV footage with gentle handheld sway, imperfect framing, subtle autofocus/exposure adjustments, faded colors, soft contrast, DV compression and mild digital noise. Smooth continuous motion throughout: no stutter, judder, frame skipping, duplicated frames, stop-motion, excessive motion blur, speed changes, low-frame-rate look, stabilization or cinematic movement. 00–03s: She brings the laundry basket onto the balcony, sets it down and clips the first shirt onto the clothesline. 03–06s: She calmly hangs several more clothes while the camera naturally follows her hands. 06–09s: She unfolds and hangs a white bedsheet; it briefly fills the frame before moving gently in the breeze. 09–12s: She steps back and straightens an uneven sleeve. Neighboring apartments and blue summer sky are visible behind her. 12–15s: She sits on a small plastic chair, drinks water and looks toward the rooftops with a tiny relaxed smile. Camera drifts toward the laundry moving in the breeze and cuts. Audio: Natural ambience only—birds, distant traffic, faint apartment voices, fabric rustling, clothespins, basket sounds, AC hum and gentle breeze. No music, narration or added effects. Goal: An ordinary forgotten early-2000s family recording—warm, mundane, intimate and unstaged. Nostalgia comes from everyday Seoul life and authentic MiniDV image quality, not reduced frame rate.

Create a 30-second, 1080p ultra-realistic personal home video in 16:9 landscape, showing an ordinary early morning in the life of one adult Korean woman commuting through Seoul. Use an entirely original fictional character created from the written description below. No reference image is supplied or required. The finished video should resemble genuine footage casually recorded by a close friend accompanying her during a sleepy morning commute. MAIN CHARACTER: One 23-year-old Korean woman with a petite oval face, naturally fair skin with visible texture, soft brown eyes, straight eyebrows, a small natural nose, and muted coral lips. Her under-eyes look slightly puffy from waking early. She wears bare-minimal makeup with no artificial skin smoothing. Her long black hair is arranged in a loose, messy half-up bun with several strands falling naturally around her face and neck. She wears the same oversized medium-blue denim jacket over a cream knitted sweater, charcoal wide-leg trousers, black leather loafers, and small silver stud earrings throughout. A slim dark-green canvas tote remains on her left shoulder, and she carries the same black smartphone in her right hand. Preserve her exact face, hairstyle, clothing, body proportions, tote, phone, and sleepy personality across every scene. CAMERA APPROACH: The camera is operated horizontally by an unseen friend walking and commuting beside her. She never records herself and never takes the camera. Use authentic handheld smartphone footage with gentle hand tremor, subtle walking bounce, occasional imperfect framing, mild low-light grain, small exposure changes, realistic motion blur, and autofocus that briefly searches in the dim subway lighting. The friend usually records from chest or eye height at a natural conversational distance. Avoid polished compositions and repeated direct eye contact with the lens. Make the footage feel personal, spontaneous, and slightly unplanned. OPENING — QUIET PLATFORM WAIT Begin immediately on a mostly empty underground Seoul subway platform before sunrise. Pale fluorescent ceiling lights reflect softly across the clean floor. Light condensation covers sections of the platform-screen doors, and vague red and white signal lights glow along the distant tracks. A few pigeons can be seen safely on an unused exterior section beyond the platform barriers. The first commuters of the morning wait quietly at widely separated positions. The camera approaches the woman from slightly behind and to her left. She stands near a marked waiting area with her tote on her left shoulder and her phone in her right hand. She reads something on the phone, blinks slowly, and gives a small involuntary yawn that she covers with the back of her left hand. She places that hand back on her tote strap and gently shifts her weight from one loafer to the other. The friend filming steps closer, causing the phone camera to briefly lose focus before finding her face. She notices the camera for a moment but is still too sleepy to react. A loose strand of black hair rests across her cheek while the remaining commuters continue behaving naturally in the background. TRAIN ARRIVAL — MORNING AIR AND MOVEMENT A distant train rumble grows louder. She locks her phone and lowers it beside her right thigh while keeping a secure grip on it. The camera turns briefly toward the tunnel as the approaching headlights brighten behind the glass barriers. The train enters the station with a strong rush of air. Her loose hair strands and oversized denim jacket move naturally in the draft. She narrows her eyes, turns her face away from the wind, and holds the tote strap more securely with her left hand. Reflections of passing train windows rapidly travel over her face and clothing. When the platform and train doors align and open, she waits for two adult passengers to exit before stepping aboard. The friend follows directly behind her, and the frame is briefly obstructed by the shoulder of another commuter. The camera corrects its position naturally rather than producing a perfect transition. Show one quick interior moment as she stands beside the train door, holding a metal support pole with her left hand while the tote stays trapped securely between her arm and body. She briefly closes her eyes as if she could fall asleep while standing. Keep the train interior ordinary, moderately lit, and only lightly occupied. STREET EXIT — FIRST DAYLIGHT Transition with a simple natural cut to the woman climbing the final stairs out of a subway exit. The unseen friend follows two steps behind. The outside light is pale blue and slightly brighter than the station, causing the smartphone exposure to adjust gradually. She reaches the quiet street, squints at the dawn light, and takes a deeper breath of cold air. She transfers her phone briefly into her left hand, stretches her right arm above her head, and rolls one shoulder beneath the oversized jacket. After stretching, she returns the phone to her right hand and settles the tote back into its original position on her left shoulder. The Seoul neighborhood is only beginning to wake. Show closed storefront shutters, a delivery bicycle leaning against a wall, traffic signals changing over an almost empty intersection, a city bus passing in the distance, and the first office workers walking toward the station. Keep signs and branding soft, distant, and unreadable. ODENG CART — A WARM MORNING STOP As she continues walking, she notices a small street-food cart preparing to open near the subway exit. Steam rises from a metal broth container into the cold blue morning air. An adult vendor arranges fish-cake skewers while warm light from the cart contrasts naturally with the cool street. She walks over, politely nods to the vendor, and stops beside the cart. The friend films from her right side as she chooses one folded fish-cake skewer. The vendor places the skewer in a small paper cup containing a little hot broth and hands it to her. She accepts the cup carefully with her left hand while continuing to hold her phone in her right. She lifts the cup close to her lips, blows gently over the broth, and attempts a cautious sip. The broth is hotter than expected. She immediately pulls it away, raises her eyebrows, presses her lips together, and releases a quiet amused breath through her nose. Keep the reaction small, natural, and spontaneous. She then takes a modest bite from the fish cake, chews naturally, and begins walking again while carrying the cup securely. Steam should continue rising from the broth, the skewer must remain inside the same cup, and its size must reduce naturally after the bite. FINAL MOMENT — SLEEPY COMMENT The friend walks backward a few steps while recording her from the front. She glances toward the camera with tired eyes and notices that her ordinary breakfast is still being filmed. Without stopping, she says naturally in English, “You’re always filming the boring parts.” Her voice is soft, sleepy, and playfully annoyed. After speaking, she gives the friend a small genuine smile, looks back toward the sidewalk, and continues walking. The camera allows her to move slightly farther ahead before following from behind. End with her oversized denim jacket, messy half-up bun, dark-green tote, and steaming breakfast cup visible as she walks into the pale morning street. Finish with a simple natural fade to black. AUDIO: Use synchronized original location sound only: distant train rumble, soft platform announcement murmurs, train-door chimes, compressed air, footsteps on station tiles, jacket fabric movement, quiet early-morning traffic, the vendor arranging metal containers, broth being poured, light street wind, and her natural dialogue. Background conversations remain indistinct. No music and no narration. REALISM AND CONTINUITY:

生成一段约11秒、9:16竖屏真人电影感短视频。人物变身前的五官、发型、身材比例、服装造型严格参考 @变身前角色图。人物变身后的五官、发型、身材比例、服装造型严格参考 @变身后角色图。人物前后必须明确为同一个人，只是处于完全不同的两个造型状态。保持核心五官、脸型、年龄和身份一致。变身后使用的剑、剑柄、剑鞘、剑身长度、材质、纹样、结构严格参考 @剑参考图。剑在整个后半段保持完全一致，不允许长度改变、剑柄变化、剑鞘消失、剑身弯曲或重新生成不同造型。不要生成背景音乐。所有环境声、动作声和兵器声必须与人物动作严格同步。整体导演要求视频采用：生活化人物动作→ 扫帚舞动→ 动作逐渐收束→ 姿势匹配→ 瞬间变身→ 拔剑→ 英雄式人物展示的连续结构。前半段是真实、随意、略带玩心的生活化手机视频质感。后半段变成具有电影感、武侠感和人物压迫感的角色展示。整个变化的高潮来自：同一人物动作姿态连续 + 瞬间换装 + 拔剑声音爆点。变身采用明确的动作匹配式硬切 Match Cut。不要使用身体融化、服装生长、粒子包裹、闪电、魔法光效、渐变换装或长时间Morph。0.0–0.7秒画面镜头从很低的位置开始。最开始不要立刻完整展示人物。首先看到：地面、人物小腿、鞋子、长柄扫帚。扫帚刷头位于画面较近的位置，占据明显前景。人物一只手已经自然握住扫帚杆。人物严格保持 @变身前角色图 的完整造型。动作人物开始将扫帚从地面自然提起来。扫帚杆原本斜向地面。随着手臂抬起，杆身逐渐向上旋转。身体开始准备向前移动。动作有真实重量。手掌必须真实抓握杆身。运镜低机位。摄影机缓慢：tilt up 向上抬镜 + slight pull back 轻微后退。像摄影师拿手机站在人物正前方拍摄。带非常轻微真实手持晃动。不要稳定器式机械移动。声音自然户外环境底噪。轻微风声。远处鸟鸣。树叶或环境摩擦声。扫帚刷毛擦过碎石或地面：沙沙——扫帚离开地面时加入轻微刷毛拖动声以及木杆晃动声。不要BGM。0.7–1.4秒动作镜头逐渐展示人物全身。人物向前迈出一小步。主手控制扫帚杆。另一只手顺势接住杆身中段。人物把扫帚从身体侧面向前带起。扫帚刷头从画面下部快速划出一条弧线。动作感觉像：普通人突然兴起，把长柄扫帚当作长棍随意耍一下。不是专业武术套路。运镜镜头继续略微后退。逐渐稳定在略低于腰部的观察高度。人物完整全身保持在画面内。不变焦。声音扫帚第一次快速划过空气：whoosh——同时保留脚步踩地声。鞋底落地声音需要对应步伐。1.4–2.1秒动作人物开始第一次明显旋转扫帚。双手自然交替换位。扫帚从身体一侧：抬向胸前→ 从身体前方经过→ 绕向另一侧。刷头快速经过人物头部附近。人物视线自然跟随扫帚运动。肩膀转动。腰部跟随。身体重心随扫帚方向自然转移。禁止只有手臂机械转动而躯干完全不动。运镜摄影师继续非常缓慢后退。轻微左右调整镜头，以确保扫帚和人物始终完整处于构图中。声音每一次明显挥动同步出现空气风切：whoosh。第二次略强：fwoosh。扫帚刷头靠近镜头快速经过时：WHOOOSH——声音比之前更厚、更近。2.1–2.9秒动作进入前半段最明显的一次扫帚运动。人物将扫帚杆抬高。扫帚刷头经过肩部以上甚至头顶附近。杆身在画面中形成明显斜线。人物双手沿杆身自然调整握位。身体随惯性旋转约30–60度。扫帚从高处快速绕向身体另一侧。随后逐渐向下落。动作过程中保持真实人体惯性。扫帚不能：穿过人物身体、穿过手掌、突然缩短、突然变长、刷头朝向错误。运镜摄影机做很轻的横向跟随。整体仍然位于人物前方。不要360度绕拍。不要突然推近。声音扫帚上扬：whoosh——。快速从头部上方经过：FWOOSH——。扫帚方向快速改变时增加一次短促风切。声音空间位置跟随扫帚移动。2.9–3.7秒动作人物开始结束舞扫帚动作。扫帚运动幅度明显降低。人物逐渐把扫帚带到身体侧面。同时抬起一条腿。动作顺序必须明确：支撑腿站稳→ 另一条腿抬膝→ 身体轻微倾斜保持平衡。不能突然把腿瞬移到高处。扫帚逐渐成为身体构图的一部分。声音风切声明显减少。增加：衣料摩擦、鞋底调整、双手在扫帚杆上滑动的细小摩擦。环境声逐渐变得更清楚。3.7–4.7秒动作人物抬起的腿由屈膝逐渐向身体侧方展开。动作：抬膝→ 小腿打开→ 脚向侧面伸展。不是武术高踢。更像自然、有一点顽皮感的侧抬腿姿态。人物上身略微向另一方向倾斜维持平衡。头部稍微转向侧面。扫帚保持在身体一侧。杆身接近竖直或略微倾斜。运镜镜头运动明显减少。保持人物全身。只有轻微真实手持漂移。声音衣物自然摩擦声。扫帚杆被重新握稳的轻微：咯 / creak。脚步调整声。没有明显风切。4.7–5.6秒动作人物完成最终变身前的姿态。一条腿支撑。另一条腿打开。身体约3/4侧向镜头。一只手控制扫帚。另一只手自然形成平衡动作。人物姿态逐渐稳定。重点建立一个非常清晰的人体轮廓。这个姿态将成为后面变身镜头的Match Cut基础。运镜接近固定镜头。只有非常轻微手持呼吸感。声音环境风声。衣料声。人物非常轻的呼吸。扫帚已经基本停止产生风切。5.6–6.30秒动作人物进一步稳定。扫帚停止高速运动。人物保持侧腿打开的姿势。身体仍然存在极小的自然惯性。头部转向侧面或略微看向镜头方向。绝对不要提前出现任何变身迹象。此时必须仍然100%保持：@变身前角色图 的服装、发型、身体造型。声音声音空间开始明显收紧。环境声逐渐降低。风声减弱。最后约0.15–0.2秒进入近似短暂静音。只保留极弱环境底噪。形成爆点前的声音留白。6.30–6.45秒变身瞬间人物刚刚形成最稳定、最清晰姿态的一刻：执行一次明确的：HARD CUT / MATCH CUT。下一帧直接切换为：@变身后角色图 的完整角色状态。不要渐变。不要叠化。不要中间形态。不要一半原服装一半新服装。不要头发先变、衣服后变。不要五官Morph。整个角色造型必须在一个明确剪辑点瞬间完成切换。场景变化与此同时，环境允许同步从普通生活户外场景变成更具有电影感的森林 / 山林 / 古道 / 浓密绿色植物环境。但人物在画面中的：位置、身体大小、重心、身体朝向、腿部姿态与上一帧保持视觉对应。使观众首先感受到：同一个动作没有断，但人已经完全变了。声音变身瞬间加入一次短促电影级低频冲击：THUMP / BOOM。不是爆炸。更像非常短的低频电影冲击。同时叠加：WHOOMP 短促空气冲击声。以及服装瞬间展开：FLAP。如果角色存在金属饰品：加入极轻的：CLINK。低频冲击只出现一次。不要持续轰鸣。6.45–6.65秒动作变身已经完全结束。人物此时严格是 @变身后角色图。剑和剑鞘严格参考：@剑参考图。人物顺着变身前的身体惯性自然调整重心。抬起的腿开始回落。身体逐渐进入更稳定、更有力量的站姿。人物的一只手自然接近剑柄。另一只手稳定剑鞘或自然控制身体平衡。不能突然瞬移手掌到剑柄。必须看到手掌自然完成：靠近→ 接触→ 握住三个动作阶段。6.65–7.35秒拔剑核心动作这是整条视频最重要的动作和声音爆点。人物动作人物手指先抓紧剑柄。另一只手稳定剑鞘。剑的造型严格参考：@剑参考图随后连续完成一次真实拔剑：握紧剑柄→ 手腕发力→ 剑刃开始离开剑鞘→ 手臂向外拉开→ 剑刃沿剑鞘真实轴线滑出→ 剑完全离鞘。必须是一把真实剑从一个真实剑鞘里被拔出来。剑身必须从剑鞘内部连续出现。禁止：剑直接凭空出现在手里、剑鞘自动消失、剑瞬间变长、拔出两把剑、剑刃穿过剑鞘、剑柄与剑身方向错误、剑身从剑鞘侧面穿出。身体动作拔剑过程中肩部、肘部、手腕和腰部共同参与。身体略微转向拔剑方向。重心下沉一点。拔剑具有重量和阻力。不要像抽出一根没有重量的塑料棒。运镜变身完成后镜头重新稳定。拔剑时允许摄影机产生一次极其轻微的向人物靠近。不要切特写。仍然保持人物大部分身体可见。拔剑声音手指抓紧剑柄：非常轻微的皮革或缠柄摩擦：creak。剑刃开始移动：shrrrk——。声音由偏厚的金属摩擦开始。随着剑身越来越多离开剑鞘，声音逐渐变得更明亮、更锋利：SHHHHHK——。剑刃完全脱离剑鞘的那一帧：SHING！清脆、短促、高频明确的真人电影金属鸣响。声音结构：creak → shrrrk—— → SHHHHK—— → SHING！整个拔剑声音与实际动作逐帧同步。剑身完全离开剑鞘的瞬间必须正好对应：SHING！禁止拔剑完成后才响。禁止提前响。禁止使用动漫夸张刀剑声。禁止雷电声。禁止魔法音。7.35–7.60秒声音留白剑完全出鞘后不要立刻加入其他大声音。保留：SHING—— 约0.15–0.25秒非常短的自然金属余韵。环境声在这一瞬间保持克制。让拔剑成为整条视频最清晰的听觉记忆点。7.6–8.3秒动作人物把拔出的剑带向最终展示位置。剑刃在空气中划过一个简洁、利落的弧线。不是复杂剑法。不是连续挥砍。只需要一次非常自然的：拔剑后调整剑势。人物身体完全稳定下来。一条腿略向前。另一条腿支撑。肩部打开。人物抬头。视线从侧方逐渐转向镜头。如果衣服存在：披风、长袖、长裙、飘带随着刚才动作产生一次自然惯性摆动。声音剑刃快速划过空气：SWISH——。注意这里的剑风和前半段扫帚完全不同。扫帚声音厚、钝。剑风声音：薄、锐利、快速、高频更多。衣摆同时产生轻微：flap / rustle。森林环境声音逐渐恢复。8.3–9.3秒动作进入人物展示阶段。人物不再做大型动作。保持充满力量的站姿。剑稳定握在手中。剑和剑鞘继续严格参考：@剑参考图。人物肩膀做极小调整。下巴略微抬起。眼神逐渐变冷。呼吸自然。头发和衣摆受到非常轻微自然风影响。运镜摄影机开始：very slow push-in。缓慢向人物推进。从完整全身逐渐接近七分身。推进非常克制。不要突然变焦。不要突然切脸部特写。声音森林自然环境：轻风、树叶摩擦、远处鸟鸣。极轻人物呼吸。偶尔存在极弱金属结构细响。不要为了强调剑不断加入金属声音。9.3–10.2秒动作人物进入最终英雄姿态。眼神：冷静、坚定、具有压迫感。人物可以直视镜头，也可以略微越过镜头看向远方。剑保持稳定。不要继续舞剑。不要走很多步。不要改变握剑手。不要把剑突然换到另一侧。运镜继续very slow push-in。人物逐渐成为画面绝对视觉中心。保持真实手机/电影摄影的轻微手持呼吸感。10.2–10.93秒最终镜头人物形成最终定格姿态。剑保持稳定。衣摆、发丝、附近植物受到极轻微自然风影响。人物存在真实呼吸。镜头仍有极微弱向前推进，但速度逐渐降低。最后0.3秒接近稳定。直接结束画面。不要淡出。不要黑场。不要做结束Logo。声音风声略微增加一点。树叶轻微摇动。可保留几乎不可察觉的金属余震质感。随后自然结束。不要添加结尾音效。运镜总纲0–1.4秒低机位近景起手→ tilt up→ slight pull back→ 从扫帚和腿部逐渐展示人物全身。1.4–3秒轻手持全身跟拍。摄影师非常轻微后退和横向修正。主要让人物动作推动画面。镜头本身不要抢戏。3–6.3秒摄影机运动逐渐减弱。基本稳定在全身构图。人物动作从快速变慢。为Match Cut建立稳定pose。6.3秒瞬间Match Cut完成变身。不使用复杂转场。6.4–7.4秒保持中远景/全身构图完成拔剑。可以有极其轻微push-in。绝对不要突然切到手部特写。7.5–10.93秒very slow push-in。从全身逐渐接近七分身。最终形成英雄人物展示镜头。人物与动作一致性强约束0–6.3秒只允许出现 @变身前角色图 的造型。6.4秒以后只允许出现 @变身后角色图 的造型。两个造型之间只允许存在一个明确的切换点。前后必须是同一个人物。禁止变身过程中出现第三张脸。禁止年龄改变。禁止人物身高明显变化。禁止体型突然改变。禁止肢体增生。扫帚物理连续性扫帚必须由人物双手真实握持。每一次旋转都必须遵循真实杆状物体运动。上一帧与下一帧之间：手部位置、杆身角度、肩部方向、身体重心必须连续。禁止扫帚瞬移。禁止刷头翻转。禁止杆身穿过身体。剑物理连续性变身后的剑严格参考 @剑参考图。剑、剑柄和剑鞘必须属于同一套兵器。拔剑前，剑必须真实存在于剑鞘中。拔剑过程中剑刃连续从剑鞘中出现。剑拔出后剑鞘仍然存在。剑刃的长度、宽度、造型和材质全程一致。剑不能在镜头中不断自行变化。声音总原则不要背景音乐。声音全部来自真实场景和人物动作。前半段声音生活化：扫帚擦地、扫帚风切、脚步、衣服摩擦、自然环境。6.3秒之前逐渐降低声音复杂度。变身前制造约0.2秒短暂声音留白。变身瞬间：THUMP + WHOOMP + FLAP + very subtle CLINK。拔剑：creak → shrrrk—— → SHHHHK—— → SHING！拔剑声音必须成为全片最大的声音记忆点。拔剑后短暂保留金属余韵。随后恢复自然森林环境声。全程声音必须与实际动作严格同步。不使用：BGM、电子音乐、玄幻魔法音、雷电、爆炸、动漫式夸张刀剑音。整体保持真实真人武侠电影的声音质感。

[STYLE + CAMERA + ATMOSPHERE] Real found-footage phone vlog on an alien planet. She is filming as she walks, the way someone records a trip for a friend: uneven clips, hard cuts, the phone too close, then swung out at whatever just moved. Slight shake, bad exposure, soft focus, lens dirt, wind slamming the mic, then creatures and plants filling the frame. When she speaks to the lens, she says the exact short lines written below, clear and plain, in a normal speaking voice. Wind can sit under the words. It does not replace them. No transitions, no slow-mo, no grade, no text, no stickers. It feels raw because the planet is doing too much for a phone. A giant moon hangs in the sky the whole day, pale and huge, even with the suns up. [SUBJECT] A woman about 27 in alien years, off-world traveler filming her own vlog. Cool lilac-grey complexion, faint mint freckles, a short glassy hair-crest messy from wind, gold eyes. Simple field clothes the whole take: dusty linen wrap, travel shirt, boots, a small pack. Fully clothed in every clip. She talks to the lens only at the beats marked below, then forgets the camera and points it at the planet. Same face in every cut. Calm, curious, a little worn. Not performing. [PLACE] An alien wild coast, lived-in and too big for the phone. Chalk shelves, forests of tall ribbed plants that lean and creak, ground-cover that pulls back when a boot lands, a gel-river the color of bottle glass, herds crossing open amber flats, a cliff path under the giant moon. Two small suns. Wind, grit, insect-choir, distant animal calls. No postcard town. The extraordinary stuff is just outside the lane. [ON-CAMERA LINES] She only speaks these lines, in this order, in a clear everyday voice: 1. “Day two. That moon still hasn’t moved.” 2. “It’s fine. It’s just eating.” 3. “They cross here every day. Look at the legs.” 4. “Yeah. I’m not going in.” 5. “Two suns. Moon’s still up.” 6. “Stopping here. I’ll send this later.” [TIMELINE — PHONE VLOG, HARD CUTS] 0-4s: HOOK — she is already walking a chalk shelf, phone in her hand, giant moon filling the upper frame behind her. She turns the lens on herself and says, clear: “Day two. That moon still hasn’t moved.” Then she flips the phone out: a stand of ribbed plants as tall as towers, swaying, seed-pods knocking. Raw, immediate. 4-9s: She pushes through a lane of vegetation. Leaves the size of doors fold aside and settle behind her. Cut to the phone dropped low: ground-cover peeling back from her boots, pale roots clicking. Cut to her stopping, holding the phone steady, filming a slow six-legged grazer the size of a cart chewing a pod. It looks at the lens. She does not get closer. She says, clear, off the side of the frame: “It’s fine. It’s just eating.” Wind under the line, words intact. 9-15s: She crests a rise. The giant moon sits on the horizon, enormous, craters readable. She films it, then the phone dips and catches a herd of long-necked grazers crossing the amber flat, legs too many, moving like a slow traffic jam. Cut to a close of a plant opening as the herd passes, then shutting. Cut to her face for a second, squinting. She says, clear: “They cross here every day. Look at the legs.” Then the phone goes back to the animals. Overexposed, dusty, real. 15-21s: Gel-river. She stands on the bank and films the surface beading and rolling upstream. Something large moves under it, a dark shape, then a fin of shell breaks and sinks. She steps back, laughs once, and says, clear: “Yeah. I’m not going in.” Cut to her crouched, filming a cluster of cube-bodied river birds pecking the mud. One hops onto a rock and stares. She stays quiet. Cut to the moon reflected, broken, in the gel. Phone almost slips. She catches it. 21-26s: Afternoon wind. She follows a dirt track between breathing shell-boulders. A door-sized beetle crosses ahead and she waits, filming its legs, then its eye, then the plants it brushes. She stays quiet. Cut to her wiping the lens with her sleeve. Cut to a high shelf: two suns, the giant moon still up, a valley of moving forest below. She turns the lens on herself and says, clear: “Two suns. Moon’s still up.” Then she holds the wide shot. The audio is wind, the far call, and that line. 26-30s: The light drops. The moon brightens until it is the main light. She films a night-opening grove, pods glowing faint mint, small flyers stitching between them. Cut to her face, tired, hair-crest flat from the wind, the moon over her shoulder. She says, clear: “Stopping here. I’ll send this later.” Last shot: phone pointed out from the shelf at the giant moon and the dark gel-river, a herd still moving as small shapes, her breathing on the mic, then the recording ends mid-step, like a real vlog file. [REALISM LOCK] Found footage alien vlog only. Handheld phone, hard cuts, dirty lens, wind, bad exposure. Same woman, fully clothed, same face. She speaks only the six written lines, in order, intelligible, everyday English. Giant moon in the sky all day. Creatures and vegetation carry the clips: ribbed forest, ground-cover that pulls back, a six-legged grazer, a many-legged herd, cube-bodied river birds, a shape under the gel, a door-sized beetle. Raw, specific, extraordinary, and ordinary to her.

Ultra-realistic cinematic scene of a young woman standing inside a crowded subway train, quietly looking at her smartphone while passengers sit and stand around her. She wears a dark green and burgundy hoodie over a white shirt, with a backpack over one shoulder. Warm fluorescent train lighting, realistic urban atmosphere, shallow depth of field, cinematic composition, natural skin texture, soft background bokeh, subtle film grain, documentary-style photography, 35mm lens, f/1.8, high detail, photorealistic, vertical.

【概要】1986年、団地に住む8歳の女の子の1日。母が父のビデオカメラを借りて、声をかけずに離れた場所から記録している。480p、16:9、15秒。7ショット（各約2秒）のハードカットで、7か所すべて別の場所。トランジションやフェードは使わない。被写体はカメラに向かって何もしない：目線を送らない、物を見せない、手を振らない、ポーズを取らない。遊びや用事に没頭している。演技ではなく、日常の途中を切り取った断片。会話・セリフ・ナレーションは一切なし。 【被写体】美貌の定義：クラスで一番かわいいと言われる、子役級の日本人の女の子。真剣な顔と、できたときに自分でにこっとする顔が一番かわいい。顔：まっすぐ切りそろえたおかっぱ（前髪も一直線）、丸い顔、黒目がちの大きな目、上を向いた小さな鼻、赤い頬（現実的な肌の質感）、前歯が1本抜けている。服（普段着）：白い丸襟のブラウス（襟が少しよれている）、赤いジャンパースカート、白いハイソックス（片方がずり落ちる）、白いズック靴、髪の左側に赤いピン。仕草の癖：できたものを自分の目の前で眺める、ソックスを引き上げる。全ショットで顔・髪・服装が完全に同一。 【登場人物】友達2人：ポニーテールで黄色いTシャツの子（そばかす）、三つ編みで水色のワンピースの子（眼鏡）。誰もカメラを見ない。 【時代・場所・光】1986年夏の団地。①ベランダ（洗濯物の間、朝の光）②団地の中庭のゴム跳び（2人が長い平たいゴムひもを足首の高さで両側から張り、その間で跳ぶ遊び。午前の光）③給水塔の下の日陰④集会所前の縁石（あやとり：輪にした1本のひもを両手の指にかけて形を作る遊び。木漏れ日）⑤公園の砂場（泥だんご：泥を丸めて乾いた砂をまぶし、手で磨いてつやつやにした球。白い光）⑥団地の郵便受けの前（夕方の橙）⑦団地の階段（夕方）。看板・文字は画面に入れない。手紙の宛名面は映さない。 【カメラ】1986年の家庭用ビデオカメラの質感：色がにじみ、解像度が低く、明るい空が白飛びし、輪郭が甘い。日付表示は出さない。手持ち、自然な揺れ、不完全な構図、ズームの迷い、露出のゆらぎ。被写体から2〜5mの距離。被写体はカメラを気にしない。スタビライズ、ジンバル、ドローン、スローモーション、映画的照明、商業カラーグレーディングはなし。カメラは常に撮影者が手に持っている位置（立つ・座る・しゃがむ・歩きながら・隣の席から）から撮る。人がいられない場所（空中・水中・天井・真上・車の外から走る車内・被写体の目の前数センチ等）からのアングルは使わない。撮影者は同じ空間に確かにいる一人の人として、被写体が動けば少し遅れて追い、ときどき構図が甘くなる。 【ショット】（各約2秒。各行＝場所・光／没頭していること／内側の感情と、それが漏れる小さなしぐさ／カメラ位置） 1. ベランダ・朝。干した洗濯物の間を縫って歩きながら、赤いピンを付け直している。感情：はりきっている。口を結んでピンを押し込む。カメラ：室内側から、逆光でシーツが白く光る。 2. 中庭・ゴム跳び。張られたゴムひもを跳び、足を引っかけて転びかけ、自分でおかしくて笑う。感情：夢中で楽しい。カメラ：正面、腰の高さ、少し離れて。 3. 給水塔の日陰。3人で並んで座り、水筒のコップの麦茶を飲む。コップの縁に口をつけたまま遠くの空を見て、それから次の子に渡す。感情：暑くてほっとする。カメラ：横から。 4. 集会所前の縁石・木漏れ日。あやとりで「はしご」ができて、自分の手を目の前で眺め、にこっとする。感情：できた満足。カメラ：斜め前から、手にピント。 5. 砂場。磨いた泥だんごを両手で包んで、光にかざして角度を変えている。感情：誇らしい。目がきらきらしている。カメラ：しゃがんで泥だんごに寄る。 6. 郵便受けの前・夕方。背伸びして郵便受けを覗き、封筒（無地、宛名面は映さない）を取り出して胸に抱く。感情：わくわく。カメラ：横から。 7. 団地の階段・夕方。封筒を抱えて駆け上がる後ろ姿。踊り場を曲がって見えなくなる。感情：早く伝えたい。カメラ：階段の下から見上げたまま。その途中、およそ00:14で録画は突然ブラックアウトに切り替わる。フェードアウトなし。 【小物・道具のディテール】ゴムひもは白くて平たい幅広のゴム。水筒は赤いプラスチックでふたがコップになる。泥だんごはつやつやの黒褐色。あやとりのひもは赤い毛糸の輪。封筒は白い無地。赤いピンは丸い頭のプラスチック。 【文字の禁止】画面内に読める文字・ロゴ・看板・ラベル・数字を一切出さない。封筒は無地。 【物理・一貫性】現実世界の物理。余分な指、融合した手、歪んだ解剖学、浮遊物、消える物体、突然の変形はなし。足は地面に接地。赤いピン、ジャンパースカート、友達の服装は全ショットで同じ。 【音】自然な環境音のみ（ショットごとに切り替わる）：洗濯物のはためき、ゴム跳びの足音と笑い声、水筒の音、木漏れ日の鳥、砂の音、郵便受けの金属音、階段の足音。言葉は一切なし。撮影者のときどきの小さな笑い声・息づかいと、被写体の笑い声・息づかいのみ可。音楽なし。ナレーションなし。人工的な音響効果なし。 【雰囲気】大人が見て「私もやった」と思う、昭和の団地の女の子のなんでもない1日の記録。演技ではなく、遊びと用事の途中の断片。懐かしく、いとおしく、深く人間的。カメラがたまたまそこにあったような感じを優先。

Multi-shot sequence with hard cuts between shots. Keep the same characters, cat, environments and visual style throughout. No music or score — only natural diegetic sound. SHOT 1 A dirty, exhausted 18-year-old girl with dark auburn-brown hair in a messy low ponytail, warm brown eyes, faded olive-green military jacket, backpack and gloves, holds a bolt-action rifle inside an abandoned ornate mansion. Sweat, dirt, a cheek scrape and tears cover her face. Her hands tremble. She closes her eyes as a tear rolls down her cheek, then hears a faint sound and slowly lowers the rifle, visibly heartbroken. SHOT 2 Over her shoulder, an infected golden-orange tabby cat sits weakly on the debris-covered parquet floor. Its fur is dirty and matted, eyes cloudy and pale, with small wounds and dark infected veins visible beneath its fur. The cat trembles and lets out a soft, frightened meow. SHOT 3 She slowly crouches beside the cat, carefully places the rifle on the floor, wipes her tears and gives it a small broken smile. She quietly speaks to the cat in a gentle voice, reaches into her jacket and pulls out a small worn red toy mouse. She gently places it on the floor and nudges it toward the cat. SHOT 4 Extreme low-angle close-up of the small red toy mouse sliding slowly through dust across the cracked parquet floor. It stops directly in front of the cat's paws. The cat looks down at it. Hold on the still toy and the cat's trembling paws. HARD CUT / TRANSFORMATION SHOT 5 Same low camera angle and exact toy position, but now on a clean sunny residential basketball court. A healthy golden-orange tabby cat suddenly reaches forward and grabs the toy mouse. Its eyes are clear and warm, fur clean and fluffy, movements energetic and playful. The cat bats the toy around excitedly. SHOT 6 Wide shot of the basketball court. The woman gently tosses the toy mouse across the court. The happy cat runs after it, pounces on it, plays with it, then runs back toward her and drops the toy at her feet, ready to play again. SHOT 7 Close-up of the clean, unharmed woman laughing warmly while affectionately scratching the cat behind its ears and gently rubbing its neck. The cat purrs, leans against her hand and happily flicks its tail. SHOT 8 Wide peaceful shot. The woman and cat sit together beside a pristine dark-green vintage convertible parked beside the basketball court. She sits on the ground beside the car, opens a small tin of cat food and gently feeds the cat. The cat eats peacefully, then rubs its head against her hand. She smiles and gently strokes its back. CONSISTENCY Same woman throughout: dark auburn-brown hair, loose low ponytail, warm brown eyes, olive-green military jacket, khaki cargo pants and brown boots. Same cat throughout: golden-orange tabby coat, distinctive facial markings, same eye shape, same body size and proportions. Infected, injured and frightened in the mansion; completely healthy, clean, pla

15秒、3:4縦、i2v。 @image1をファーストフレームとして使用。 動画テーマ： 3×3の9コマ写真コラージュが、シャッター音と各コマごとのフラッシュをきっかけに生きた動画として動き出す、不思議で可愛い15秒動画。 9コマすべて同じ一人の成人日本人女性。

A hyper-realistic, photorealistic 8K CGI advertisement sequence for a futuristic luxury electric concept sports car, styled like a high-end commercial with a sleek, powerful, and sophisticated mood. Setting & EnvironmentSet in an infinite, minimalist virtual studio with a deep black void background and a perfectly polished, highly reflective black obsidian floor that crystal-clearly mirrors the scene. The atmosphere is crystal clear with subtle light bloom. Lighting Key Light: A cool-white (6000K) softbox panel light descends from above, growing larger to cast clean defining highlights over the final car. Emissive Accent Light: Silvery-white and vibrant electric-blue light glows directly from the fluid liquid metal, projecting dynamic reflections onto the floor. Subject Transformation Initial Form: A single, pristine sphere of mercury-like liquid metal hovering above the floor, rippling with high reflectivity and surface tension. Final Form: A low-profile, aerodynamic futuristic sports car with smooth, single-piece molded body panels in deep gloss black pearlescent paint that subtly shifts to dark blue, featuring razor-sharp LED light bars. Camera ChoreographyAn uninterrupted, dynamic camera move starting as an extreme close-up macro (85mm, f/1.8 shallow depth of field) focused on the hovering liquid drop. The camera smoothly pulls back and accelerates into a sweeping clockwise orbit, transitioning to a wide shot (35mm, f/5.6 deep focus) to reveal the complete vehicle sharp and in full view. Action & Animation Sequence 0–2s: The liquid metal sphere slowly pulses and ripples to a low-frequency bass hum. 2–5s: The sphere explodes outward in zero-gravity as liquid tendrils stretch to form the car's wireframe, traced by glowing electric-blue energy lines, accompanied by high-tech chimes and metallic crackles. 5–7s: Liquid metal flows over the energy frame, solidifying into smooth body panels and forming wheels to a heavy bass thrum. 7–8s: The car gently settles onto the reflective obsidian floor, its razor-sharp LED headlights and taillights flash on with a powerful pulse, fading into the subtle hum of a high-tech electric motor.

时长：20秒 画幅：9:16竖屏 主题：同一句韩语，韩剧里与现实中的反差。 形式：单人双状态，五组A/B硬切对照。 【人物绑定】 ​ 是唯一人物角色卡。全片保持同一人物身份，面部、发型和服装沿用参考图，不重新设计长相。 A与B是同一个人的两种表演，不是两个人对话，不出现双人同框或分屏。 【状态A｜韩式爱情电影】 韩式爱情电影中的人物近景，精致、温柔、含蓄，带轻微浪漫感。 光线： 窗侧大面积柔光照亮面部，明暗过渡细腻；侧后方一束柔暖逆光轻轻勾出发丝和肩部轮廓。眼睛有自然眼神光，面部立体，不使用均匀磨平五官的美颜光。 调色： 奶油白高光、低饱和暖肤色、轻微青灰冷调阴影。整体清透，黑位略柔，亮部平滑过渡。背景偏米白、淡灰与浅暖色，避免整张画面泛黄、泛粉。 摄影质感： 中长焦人像镜头观感，浅景深，眼睛与嘴部清晰，窗帘和远处小灯柔化为细腻散景。高光带极轻微柔晕，保留皮肤纹理与细微胶片颗粒。 每个A镜头只有极轻、几乎察觉不到的缓慢推近。 表演： 目光柔软，呼吸自然，动作幅度小。开口前有短暂情绪停顿，声音轻柔，语尾轻落，像韩剧中对亲近的人说话。 【状态B｜普通手机直拍】 现实室内手机前置直拍，简单灰墙背景，较直接的正面照明。白平衡中性，反差和锐度比A更直接，背景也更清楚。 取消A的逆光轮廓、浅景深散景、柔晕与电影调色，但仍保持正常曝光和真实肤质。 机位固定，构图比A略紧，保留轻微手机广角近距离感，不夸张拉伸面部。 表演： 直视镜头，眉眼明显发力，身体稍向前探，语气短促、直接、不耐烦。喜剧来自突然变脸和说话方式，不演成失控尖叫。 【构图与剪辑】 统一头顶至腰部的竖屏中近景，眼睛位置保持在画面相近高度。 A说完后直接硬切B，画面质感、神态与声音同时改变。 每次切换都保持同一张脸可辨认，不用变形、闪白、黑帧或变装转场。 SHOT 1｜00–04秒｜别看 Subject： 同一人物，温柔含蓄与直接斥责形成反差。 Action / Performance： 00–02秒，状态A： 人物先看向侧下方，轻轻转眼看向镜头，带一点含蓄笑意，柔声说“别看”。说完略收回目光，保持自然羞涩。 镜头极缓慢推近，柔暖逆光照亮少量发丝。 02–04秒，硬切状态B： 人物直接盯住镜头，眉头收紧，头稍向前探，用明显更重的语气重复“别看”，随即补上粗口。 最后一个字落下，目光仍停在镜头上，留一点不悦的静止。 Dialogue： A：“보지 마.” B：“보지 마, 이 개새끼야.” Environment / Camera / Style： A为柔光窗帘与奶油色散景，韩式爱情电影质感。 B为灰墙、直接光与固定手机近景。 Constraints： 两句均由画面内人物本人说出，先说完A，再切B，口型准确。 SHOT 2｜04–08秒｜别这样 Subject： 同一人物，亲昵劝阻与不耐烦训斥。 Action / Performance： 04–06秒，状态A： 人物轻轻歪头，肩膀放松，目光带一点柔软的责怪，低声说“别这样”。语尾很轻，像面对亲近的人。 面部清楚，背景柔化，镜头保持细微推近。 06–08秒，硬切状态B： 头摆正，目光直直看向镜头，短促地说“别这样”，紧接着补上粗口。 下巴只随重音向前点一下，说完停止动作。 Dialogue： A：“하지 마.” B：“하지 마, 이 개새끼야.” Environment / Camera / Style： 沿用A的精致电影光线与B的普通手机直拍，硬切瞬间清楚显出质感差异。 Constraints： 不用夸张摇头或甩头。表情变化集中在眼神、眉间、下颌与语气。 SHOT 3｜08–12秒｜别走 Subject： 同一人物，真诚挽留与强硬命令。 Action / Performance： 08–10秒，状态A： 人物目光稍微抬起，像在看即将离开的人，嘴唇轻启，用带一点不舍的声音说“别走”。 说完保持期待的眼神，轻微吸气，不落泪、不追出画面。 柔光留住眼睛里的细小反光，镜头缓慢靠近一点。 10–12秒，硬切状态B： 人物突然睁大眼睛，头略向前探，把同一句话说成强硬制止，紧接着补上粗口。 表情像在命令对方站住，保持眼神压迫感。 Dialogue： A：“가지 마.” B：“가지 마, 이 개새끼야.” Environment / Camera / Style： A浅景深、柔和高光、冷暖细腻分离；B中性颜色、清晰背景、固定机位。 Constraints： A的情绪含蓄，B的语气强硬。不要两段都演成哭泣或大喊。 SHOT 4｜12–16秒｜别听 Subject： 同一人物，温柔安慰与暴躁纠正。 Action / Performance： 12–14秒，状态A： 人物轻轻叹气，短暂垂眼，再抬眼看镜头，小幅摆一下手。 先自然说出感叹词，再温柔地说“别听”，像在劝对方不要把难听的话放在心上。 表情有理解与安慰，动作细小。 14–16秒，硬切状态B： 人物眼睛睁大，眉毛抬起，带着明显不耐烦说“别听”，随后快速补上粗口。 头部只做一次短促前倾，咬字清楚、重音明确。 Dialogue： A：“아이고… 듣지 마.” B：“듣지 마, 이 개새끼야.” Environment / Camera / Style： A为韩剧中安慰人的电影近景；B像手机镜头前直接训人。 A的轻柔叹气与B的干脆发声形成声音反差。 Constraints： A的感叹词与后半句自然连贯。两段声音不重叠，不增加其他人的回应。 SHOT 5｜16–20秒｜别哭 Subject： 同一人物，心疼安抚与最后的不耐烦吐槽。 Action / Performance： 16–17秒，状态A： 人物微微向前倾，双手在腰胸之间轻抬，做一个很小的安抚动作，眼神心疼，柔声说“别哭”。 保持电影柔光与清楚的嘴部细节。 17–20秒，硬切状态B： 人物立刻以烦躁口吻说“别哭”，补上粗口。 接着视线移到旁边，肩膀松下来，压低声音快速嘟囔，像已经受够对方总哭。 最后以未说完的“씨…”收尾，闭嘴，留下一个无语的表情。不要突然笑场。 Dialogue： A：“울지 마.” B：“울지 마, 이 개새끼야. 맨날 쳐 울고 앉아있어, 씨…” Environment / Camera / Style： B状态持续到片尾，不再切回A，不加通关式音效或结束标题。 Constraints： 末段先强硬制止，再低声抱怨，音量有变化。尾句语速加快 but 发音清楚，不扩写新台词。 【声音与口型】 所有韩语都是画面内人物的现场对白，绝不是旁白。 人物说话时，嘴唇、下颌、呼吸与韩语发音同步。 同一基础声线： A轻柔、亲近、带自然气息，不能用气声糊掉发音； B更直接、更响、重音清楚，带不耐烦，不能全程破音尖叫。 保留轻微呼吸、叹气与室内底噪。 A段可有极低音量、稀疏的浪漫钢琴底音；切到B时钢琴立即停止，只剩干净直接的人声与房间底噪。下一段A再轻轻恢复同一音乐动机。 音乐始终让位与台词，不加入罐头笑声。

一、任务目标 生成一段完整连续的 15 秒中国仙侠电影短片。 核心喜剧主线只有一条，并且要让观众 第一遍就立刻看懂： 一名敌方剑客故意挑拨师姐与小师妹，篡改师姐原本说过的话，让小师妹误以为师姐认为她“三招必败”。 真正反转是： 她们从来没有讨论“小师妹能不能打赢”，而是在讨论这个敌人究竟能撑几招。 二、整体风格 整体必须同时具备以下气质： 电影级写实质感 纯古风中国仙侠美学 真正史诗级高手对峙的镜头语法 一本正经的冷面喜剧 默片式反应节奏 港式动作喜剧的利落效率 清楚的“三拍递进” 明确的铺垫与回收 Arri Alexa 电影机质感 稳定清晰的面部微细节 细腻胶片颗粒 自然体积光 喜剧必须克制，不夸张，不胡闹，不靠人物变蠢。 三、人物身份锁定 角色 ID A｜@图片 1｜剑仙师姐 始终保持同一个人物： 25–30 岁东亚女性 身材高挑纤细 椭圆脸 深色杏眼 黑色长发半挽 白玉簪固定 同一套白色刺绣真丝汉服 银色腰封 玉佩 白色布靴 唯一一柄银色长剑 角色 ID B｜@图片 2｜小师妹 始终保持同一个人物： 20–25 岁东亚女性 身形娇小 圆润灵动的脸 黑发编辫 同一套青绿色亚麻汉服 深色腰带 木簪 黑色布鞋 唯一一柄深色钢剑 其他角色 敌方剑客 一名敌方剑客 负责挑拨、误导和最终出手 不可抢走喜剧情绪中心 年长师父 同一个年长师父 只在关键时刻补最后一刀 必须保持安静、克制、一本正经 四、环境 DNA 与空间原则 本轮新上传的全部背景与地点参考图，共同决定同一套环境 DNA。 正式构图前，先无声整合并统一以下内容： 真实地形 建筑语言 材质 植被 水体 天气 山雾 主光方向 倒影 空气纵深 在此基础上，重新规划出一个： 本轮独有、完整统一、可信的新空间。 环境规则 背景中的以下元素始终自然鲜活： 风 水体 植被 云层 远处普通弟子 空间环境声 但必须满足： 背景绝对中立 不制造包袱 不制造反转 不主动推动剧情 不替人物表达情绪 五、三镜头结构 镜头 1｜0–5s 景别 全景或远景 画面内容 同一个剑仙师姐 同一个小师妹 一名敌方剑客 一名年长师父 四人位于统一空间中，地理关系清楚稳定。 敌方剑客站在两人正前方，年长师父只在数步之外安静存在。 剧情与对白 敌人故意看着小师妹挑拨道： “你师姐亲口说，你与我交手，三招必败。” 同一个小师妹非常缓慢地转头看向师姐，问： “三招？” 同一个白衣剑仙没有任何紧张，只平静纠正： “我说的是他。” 镜头重点 第一拍要先让观众以为矛盾是“师姐看不起小师妹” 小师妹转头这一瞬必须清楚 师姐的纠正必须平静、确定、不带解释欲 这一句要直接翻第一层误会 镜头 2｜5–10s 景别 中景或牛仔景 连续性要求 保持以下内容全程稳定： 同样两名女子 同一个敌人 同一个师父 同样服装 同样长剑 完全一致的地理空间 剧情与对白 小师妹立刻重新看向敌人。 这一次她不是生气，而是真的有些惊讶，说： “三招？师姐，你也太给面子了。” 敌人刚才得意的表情当场僵住。 浅景深后方，同一个师父从头到尾没有插话，只安静地抬起一根手指。 敌人发现后，恼火问： “你什么意思？” 镜头焦点转移到师父。 师父保持绝对认真，只回答： “一招。” 完整停顿半拍。 敌人终于彻底破防，大喊： “欺人太甚！” 随后主动拔剑，向师姐正面冲去。 镜头重点 第二拍完成喜剧递进 “三招”升级成“一招” 师父这一刀必须极干 停顿要留足，让观众和敌人同时意识到自己成了被预测的对象 敌人破防后立刻进入动作段 镜头 3｜10–15s 景别 特写或极近特写 连续性要求 保持以下内容稳定： 同一个师姐 同一个小师妹 同一个敌人 同一个师父 服装、长剑、站位、空间关系全程一致 动作设计 同一个师姐始终保持放松，甚至不把自己的银色长剑完全拔出。 她只完成一个： 非常清楚 物理合理 极短促 极精准 的动作组合： 侧身闪避敌人正面冲击 用仍然带鞘的同一柄银色长剑 短促敲击敌人持剑手腕 顺势借敌人的前冲力量改变其重心 结果： 敌人的长剑旋转飞脱 安全插入旁边地面 敌人自己单膝落地 全程不得出现血腥伤害 动作后的对白 全场安静。 同一个小师妹慢慢举起一根手指，认真说道： “师父猜得真准。” 同一个剑仙师姐低头观察败下阵来的敌人，平静回答： “其实半招。” 镜头焦点再次转向师父。 师父一本正经说道： “我已经给他留面子了。” 结尾画面 极近特写中： 敌人依次看看师父、师姐、小师妹 发现自己一句都反驳不了 小师妹用力咬住嘴唇憋笑 师姐仍然非常正式地朝他行一个武者承让礼 精准切在： 敌人彻底无语的表情上。 六、表演原则 师姐 始终平静 不炫耀 不解释过度 强者松弛感要稳定 喜剧成立依靠认真，而不是玩笑腔 小师妹 反应要快 情绪不夸张 从“误会”到“听懂”再到“补刀”，节奏要清楚 憋笑感要真实克制 敌人 前半段必须真心自信 中段逐渐僵住 后段破防但不能演成小丑 笑点来自后知后觉，不来自愚蠢 师父 台词极少 每句都像盖棺定论 必须一本正经 越认真越好笑 七、动作与镜头要求 动作身体力学清晰自然 敌人冲击要真有气势 师姐应对要极省、极准、极短 最终胜负必须一眼看懂 不要复杂炫技连招 不要拖泥带水 不要光污染特效 不要随机剑气 摄影要求 16:9 横屏 严格三个连续清晰镜头 运镜干净克制 前景、中景、远景保持自然视差 镜头服务人物关系与包袱回收 不做花哨炫技运动 八、声音与同步要求 原生同步普通话对白 口型精准 喜剧停顿明确 视线关系准确 拔剑、碰撞、长剑落地、归位等声音清楚可辨 环境声持续自然存在 不生成字幕 重点利用 Seedance 2.0 的能力： 多模态参考连续性 多人物复杂运动稳定性 导演级运镜控制 15 秒多镜头同步音视频输出能力 九、严格避免 blurry bad quality low quality low resolution noisy jpeg artifacts watermark text error deformed mutated bad anatomy poorly drawn hands bad composition out of frame disfigured inconsistent character changing clothes face morphing background shift glitching cuts disappearing props 额外避免： 字幕 过度夸张表演 无意义搞笑表情 复杂难懂的反转 动作软绵绵 花哨慢镜 随机法术 环境制造笑点 配角抢戏 敌人突然降智 十、最终效果目标 整段短片的观感应该是： 先制造“师姐似乎看低小师妹”的表面误解 很快翻转成“原来被评价的是敌人” 最后再用动作和师父补刀，把包袱彻底坐实 最终关键词： 仙侠、冷面喜剧、高手对峙、三拍

一段15秒、9:16竖屏的智能手机视频，极其逼真的实拍质感。白天，从临街公寓楼上窗口向下斜拍，一位冷艳的成年女子在人行道上遭到人形机器人抢包，机器人逃跑时自己摔倒，女子从容取回包，并用一种出人意料的方式教训它。 【拍摄视角】 摄影者站在公寓窗口，手机镜头朝下斜俯拍楼下近处的人行道与相邻的沥青道路。画面左边缘露出很窄的一条深色窗框，浅灰色混凝土人行道位于左侧，深灰色道路位于右侧，路缘在画面中呈斜线延伸。没有天空、远处天际线或大范围城市全景。 俯拍角度能同时看见人物头顶、肩部、身体正面的一部分和完整脚步。人物足够大，能够辨认抢包、跌倒、捡包及手臂动作，避免高空航拍般的人物过小。 手机基本保持原来的方向，仅有轻微自然手持晃动，突发动作时有幅度很小、稍慢半拍的构图修正。始终在同一个窗口拍摄，连续一镜到底，没有电影式运镜、切镜、特写插入或场景过渡。 【角色与道具】 女主使用成年预设女主角#1 ，保持预设脸部、发型、身材与身份，完整穿着预设服装#2 ，鞋履和配饰沿用预设，不自行改款。 她左手提着一个中等大小的印花手提包，包口闭合，提带没有缠绕手腕。全片只有一个包，不使用拐杖。 她以自然克制的海妖风步态行走：肩部放松下沉，颈部舒展，下巴略抬，身体随着重心转移形成柔和流动的S形线条；步幅适中，双脚沿较窄的步行轨迹自然交替，双腿前后错位但不刻意交叉。神态冷淡，有距离感，不夸张扭腰，不像舞台走秀。转身、弯腰和用力时保持真实的人体平衡。 一个黑白相间的未来主义人形机器人 ceYCnpdUfKl6 跟在她身后几步远。机器人拥有流畅、可信的机械身躯，白色装甲板、黑色关节、深色无特征面罩，接近成年人的身高。关节活动有机械重量感，面罩不显示表情、文字或图案。 【0—3秒】 女主沿着人行道缓缓朝画面下方走近，左手的印花包轻轻摆动。机器人在她身后正常行走，随后加快两步，从她靠道路的一侧接近。 接近段不提前搞怪，让机器人看起来确实正在寻找抢包机会。女主没有察觉，也不回头看它。 【3—6秒】 机器人伸出右手，抓住印花包的提带，突然将包从女主左手中夺走，随即跨下路缘，转向道路逃跑。 女主左臂被带动一下，立即松手站稳，转头看向机器人，没有跌倒或被拖行。 机器人刚跑出两步便回头看她，转身与落脚没有协调好，一只脚尖碰到自己的另一只脚踝。它连续小碎步试图找回平衡，身体却越晃越大，最终侧髋撞地，翻成仰躺，两条腿短暂抬起后落回沥青路面。 摔倒突然、笨拙而沉重，有明确的脚部失误和补救过程，不做后空翻，不悬空旋转。右手在撞击时松开，包落在人行道边、机器人右肩旁。包口保持闭合。 【6—9秒】 女主停顿半秒，低头看着四脚摊开的机器人，然后从容走到它身边。 她双脚站稳，微屈膝俯身，用左手捡回自己的包。机器人仰躺着，右肘弯曲，刚才抢包的右手缓慢抬起，手掌半张。 女主拿回包后，视线落到那只机械手上。她没有立即离开，而是停住片刻，伸出空着的右手抓住它的右腕。 【9—12秒：核心喜剧动作】 女主左手提包，右手握着机器人的右腕，带动它松弛的前臂，让它用自己的手掌朝自己的额部装甲拍一下。 “啪。” 她把它的手掌拉开十几厘米，短暂停顿，再拍一下。 “啪。” 两下分明、干脆，接触过程完整可见：机器人刚才用这只手抢包，现在女主直接拿这只手教训它自己。 机器人右肘保持自然弯曲，肩关节没有反折。女主只移动它的一条松弛前臂，不抬起整台机器人。每次拍到额部时，机器人头部仅轻轻晃动，不飞起、不碎裂、不爆火花。 女主神态始终冷淡，动作利落，没有大喊大叫，也没有夸张挥臂。拍完后松开它的手腕，那条机械前臂顺着重力落回胸甲，发出短短一下碰撞声。 【12—15秒】 女主直起身，整理一下左手的包带，转身回到原来的行进方向，自然恢复冷艳从容的海妖风步态。 她平静地继续走，像刚处理完一件小麻烦。机器人仍仰躺在道路上，右手搭在自己胸前，双腿狼狈地摊开。 结尾保留女主离开的全身和地上的机器人，让从容的人与失败的机械小偷同时出现在画面里。不增加新的反转动作。 【视觉真实性】 自然日光和真实建筑阴影，混凝土接缝、路缘磨损、沥青颗粒与细小污渍清晰可信。服装褶皱随身体动作变化，包随惯性摆动，机器人装甲反射周围真实环境，人物与机器人共享一致的光线和接触阴影。 整体具有电影级的材质与光影可信度，但仍像手机偶然拍下的街头事件。保留轻微手抖、适量手机压缩、自然运动模糊和细微自动曝光变化。不使用美颜、过度锐化、夸张景深虚化、广告式布光或重度电影滤镜。 【声音】 公寓窗口距离下的自然街道收音：轻微风声、远处城市底噪、较远的脚步声、机器人撞地的沉闷机械声，以及两次短促的手掌碰撞装甲声。声音保留空间距离，不像贴近人物录制。 无对白、无旁白、无配乐、无罐头笑声、无卡通音效。 【连续性要求】 女主外貌与服装、机器人外形、手提包、街道结构、光线方向全程一致。 手提包严格遵循：女主左手携带—机器人右手抢走—随机器人落地—女主左手捡回—女主带走。 打击动作严格保持女主右手握住机器人右腕，机器人用自己的右手掌拍自己的额部。手腕、手肘和肩膀结构连续，不增加肢体、不左右互换、不穿模。 无剪辑、无场景过渡、无慢动作、无文字、无字幕、无水印、无过度CGI特效。

TITLE: THE SURVIVAL RUN Create a single-page, premium Hollywood disaster-action storyboard in 16:9 widescreen format with 12 cinematic panels covering one continuous 30-second survival sequence. REFERENCE IMAGE 1 (Image1): Use as the EXACT main character reference. Maintain the same facial identity, hairstyle, beard, body proportions, black shirt, black pants and overall appearance consistently in every panel. No character, face, hairstyle or clothing changes. Dust and debris may gradually appear as the disaster progresses. REFERENCE IMAGE 2 (Image2): Use as the EXACT storyboard design and layout reference. Match its premium cinematic panel arrangement, numbering, timecodes, black caption strips, typography hierarchy, borders and professional Hollywood storyboard presentation. STYLE: Photorealistic live-action Hollywood disaster film, ARRI Alexa 35, anamorphic cinematic lens, realistic Kuala Lumpur, Malaysia, practical stunt realism, believable destruction physics, handheld tracking, cinematic lighting, motion blur, dust, shattered glass, sparks and realistic debris. CORE RULE: ONE CONTINUOUS SURVIVAL SHOT. Every movement must logically connect. No teleportation or unexplained location changes. The same man constantly runs, stumbles, recovers, jumps, slides and changes direction while the collapsing city becomes the antagonist. STORYBOARD PANELS: 1. 00:00–00:03 — EARTHQUAKE Low-angle behind his feet as he runs through Kuala Lumpur. The ground shakes, buildings tremble, glass explodes and concrete crashes behind him. He looks back and accelerates. 2. 00:03–00:06 — COLLAPSE Front tracking shot. A building façade collapses behind him. He changes direction and jumps over a metal barrier as debris destroys his original path. 3. 00:06–00:09 — CAR IMPACT At an intersection, vehicles slide toward him. He dives at the last second, a car passes overhead, he rolls, recovers and keeps running. 4. 00:09–00:12 — ROAD CRACK A huge fissure tears across the road. He runs onto a damaged car as a springboard and leaps across collapsing pavement. 5. 00:12–00:15 — FALLING GLASS Side tracking shot. High-rise windows explode above him. Glass rains down as he shields his face and ducks beneath a falling sign. 6. 00:15–00:18 — COLLAPSE A building collapses ahead. He spots a narrow gap, drops low and slides beneath falling debris, then emerges dusty and runs. 7. 00:18–00:21 — VEHICLE CHAOS A bus crashes into cars. A spinning vehicle approaches. He jumps onto its hood, springs over the roof and lands on the opposite sidewalk. 8. 00:21–00:24 — ELEVATED ROAD FAILURE An elevated road collapses ahead. He performs a desperate full-body gap jump, lands hard, rolls, recovers and runs.

Created a 30-second cinematic luxury jewelry advertisement video in a warm and elegant style. The video opens with a beautifully decorated bedroom, soft morning sunlight, and a premium jewelry gift box placed on a wooden table. A woman gently picks up the elegant box, creating a delicate and emotional unboxing moment. The camera moves into cinematic close-up shots of her face and hands as she reveals the jewelry. A sparkling blue gemstone pendant is shown in detailed macro shots with soft golden lighting. She wears the necklace gracefully, highlighting its elegant design and luxurious shine. The scene continues with a peaceful morning moment near the window, holding a cup in warm sunlight. Close-up shots capture the pendant sparkling naturally against her skin with realistic shadows and reflections. The video ends with a premium jewelry box displaying the complete matching jewelry set. Soft cinematic lighting, warm golden tones, luxury commercial aesthetic, smooth camera movements, shallow depth of field, photorealistic details, 9:16 vertical format.