Seedance 2.5 Video Prompt Library

PROTAGONIST: Use @[Image 1](image_1) as the exact performer identity reference. Preserve her facial features, hairstyle, skin tone, and natural body proportions throughout. AUDIO / PERFORMANCE MASTER: Use @[Video 1](video_1) as the absolute audio, vocal, rhythm, and performance reference at its original playback speed. Every dance accent, body groove, facial reaction, camera acceleration, crowd response, and transition must respond naturally to the actual beat structure. Whenever vocals occur, the protagonist performs precise frame-accurate lip-sync with convincing phoneme articulation, syllable timing, rap cadence, breaths, pauses, and expression. During instrumental passages, she stops articulating lyrics naturally while continuing the choreography. ENVIRONMENT / DAYLIGHT CORE: Typical details baked in: multi-story school building, paved schoolyard (운동장), benches and trees, students in recognizable Korean uniforms (blazers/vests, skirts or slacks, some in PE kits), phones

Main Subjects: Three young adult women, best friends, wearing authentic mid-2000s casual fashion: cropped hoodies, fitted graphic tops, low-rise bootcut jeans, colorful sneakers, small shoulder bags, chunky bracelets, and playful hair accessories. Location: Ordinary suburban shopping mall in the United States, around 2005. CRT televisions in electronics stores, CD shops, food court, arcade, clothing stores, vending machines, tiled floors, and a small photo booth. No smartphones or modern LED displays. Visual Style: Ultra-realistic 2005 consumer MiniDV footage. Faded colors, soft contrast, realistic skin texture, mild digital noise, compression artifacts, blown highlights, autofocus hunting, occasional motion blur, imperfect white balance, and slight exposure pumping. Camera Style: One friend casually filming with a bulky MiniDV camcorder. Heavy handheld shake, accidental zooms, sudden pans, awkward close-ups, autofocus breathing, imperfect framing, and no stabilization. Timeline — 15 Seconds 00:00–00:03 The footage starts outside a photo booth. One woman announces, “We need a perfect picture!” She dramatically fixes her hair while another friend secretly messes it up from behind. She turns around in confusion. 00:03–00:06 All three squeeze into the tiny photo booth. The curtain gets stuck halfway open, leaving them awkwardly visible to the mall walkway. They try pulling it closed while laughing. 00:06–00:09 The countdown begins. They pose seriously for the first shot—then one woman sneezes exactly as the flash fires. Everyone looks at her in shock before bursting out laughing. 00:09–00:12 They attempt another photo. One friend changes poses too early, another closes her eyes, and the third accidentally bumps the booth wall. The camera outside captures their muffled screaming and laughter. 00:12–00:15 They emerge holding the photo strip. All three inspect it silently. Their faces slowly turn disappointed. One finally says, “We paid for THAT?” Everyone collapses laughing as the camcorder zooms aggressively into the terrible photo strip, loses focus, and cuts to black. Audio: Mall ambience, footsteps, distant arcade sounds, photo-booth countdown beeps, camera flashes, muffled voices behind the curtain, laughter, shopping bags, and natural conversation. No added music, no narration. Comedic Direction: Build the humor through increasingly bad luck: hair prank → stuck curtain → perfectly timed sneeze → ruined poses → terrible final photos. Keep reactions spontaneous and physically realistic. Goal: Make it feel like a genuinely forgotten 2005 MiniDV mall recording – the kind of embarrassing footage friends would rediscover years later and immediately start laughing at.

【最重要：参照画像の読み方】 参照画像は3×3、全9カットで構成された「時系列ストーリーボード」である。 グリッドそのものを映像として表示しない。 9画面を同時に動かさない。 各パネルを独立したフルスクリーン映像として再構築し、必ず以下の順番で連続した30秒の映像にする。 ①左上 → ②上中央 → ③右上 → ④中段左 → ⑤中央 → ⑥中段右 → ⑦左下 → ⑧下中央 → ⑨右下 登場人物の顔、髪型、髪色、メイド服、体型、屋敷の内装を参照画像から維持する。 主人公はピンク髪のメイド。 彼女が持つモップは掃除道具であり、それをマイクのように握って歌う。 物語が進むにつれて他のメイドたちが自然に歌へ加わり、最後は力強いゴスペルの大合唱になる。 【0〜3秒／CUT 1：左上】 静かな屋敷。 主人公が一人で床をモップ掛けしている。 掃除の手がふっと止まる。 モップの柄を見て何か思いつき、いたずらっぽく微笑む。 カメラは廊下の奥からゆっくりドリーイン。 【3〜6秒／CUT 2：上中央】 主人公がモップをマイクスタンドのように立て、歌い始める。 最初は一人だけ。 身体を左右に軽く揺らしながらリズムを取り、楽しそうに歌う。 斜め前からのミディアムショット。 歌声が勢いを増すにつれて緩やかにプッシュイン。 【6〜9秒／CUT 3：右上】 主人公の表情のアップ。 目を閉じて気持ちよく歌い、サビへ向かって声を強くする。 片手を胸から前へ大きく広げる。 カメラは顔の横を小さくアーク移動。 背景のメイドたちが「何をしているの？」という表情で振り返る。 【9〜12秒／CUT 4：中段左】 一人のメイドが主人公の歌に合わせて手拍子を始める。 もう一人も笑顔になってリズムを取る。 主人公はモップマイクを握ったまま歌い続ける。 カメラは主人公を画面端に置き、奥のメイドへラックフォーカス。 歌が周囲へ伝染していく瞬間。 【12〜15秒／CUT 5：中央】 3〜4人のメイドが主人公の後ろへ集まり、コーラスに参加。 主人公が歌う。 後ろのメイドたちが掛け声で応える。 call and response。 全員が同じ動きをするのではなく、それぞれ手拍子、身体を揺らす、腕を上げるなど自然に異なる動き。 カメラは正面から滑らかに後退し、人数が増えたことを見せる。 【15〜18秒／CUT 6：中段右】 音楽がさらに盛り上がる。 別の部屋や廊下からメイドたちが次々と駆け込んでコーラスへ加わる。 主人公はモップをマイクのように傾け、力強く歌う。 ローアングルから主人公を中心に半円状のアークショット。 スカートや髪が身体の動きに自然について動く。 【18〜22秒／CUT 7：左下】 屋敷の広いホール。 大勢のメイドによる本格的なゴスペル合唱。 手拍子、ステップ、身体を大きく揺らす。 主人公が中央でリードボーカル。 カメラは群衆の間を抜けながら主人公へ接近するダイナミックなドリーショット。 全員が楽しそうに、力強く歌っている。 【22〜26秒／CUT 8：下中央】 クライマックス。 主人公がモップを片手で高く掲げる。 周囲のメイドたちも一斉に腕を上げ、力強くコーラス。 カメラは主人公の周囲を大きくオービットしながら徐々に上昇。 豪華な屋敷のホール、シャンデリア、大勢のメイドたちまで一望できる。 歌声のエネルギーが最大になる。 【26〜30秒／CUT 9：右下】 最後の大合唱。 主人公がモップマイクを両手で握り、最後のロングトーンを全力で歌う。 メイドたちが背後で豊かなハーモニーを重ねる。 最後のフレーズで全員が笑顔になり、主人公がモップを高く掲げてフィニッシュ。 ローアングルから高速ではない力強いズームアウト。 最後は屋敷の大ホール全体と大合唱するメイドたちを見せて締める。 【映像全体】 明るく高揚感のあるミュージカル映画的な演出。 主人公一人の小さな歌から始まり、30秒かけて屋敷全体を巻き込む巨大なゴスペルへ発展する。 前半は比較的静かなカメラ、後半になるほどアーク、ドリー、オービット、ローアングル、ワイドショットを増やし、映像そのものも音楽と一緒に盛り上がる。 各CUTは参照画像の対応するパネルの構図・人物配置を出発点として使用する。 前のCUTの動きを次のCUTへ持ち込み、一本の連続した出来事として自然につなげる。 参照画像内のグリッド線、タイムコード、文字、ラベルは動画内に出さない。 画面分割、漫画コマ表示、ストーリーボード表示にしない。 完成映像は常に16:9のフルスクリーン。

女人、怪物、双手、脸部近景与房间全部采用同一套明显电脑生成的AAA奇幻游戏CG美学；女人也必须具有精修游戏角色的雕刻体块、成束头发和统一材质着色，不让她变成真人演员而只有怪物是CG。不是真人实拍、cosplay、普通gameplay、2D动画或赛璐璐。画面原生填满整个横屏，重新组织主体构图，不保留竖屏黑底、嵌套视频框或圆角边框；全程没有任何文字、字幕、标题、标识或水印。环境：深夜一间破旧封闭的小屋，低矮天花板、暗褐色木地板、污旧土褐墙面，后墙两侧垂挂米灰色旧窗帘，中间是昏暗墙面，右后方有被旧布覆盖的家具。唯一明显实景光源是一盏放在地板上的旧式玻璃油灯，灯体有深色金属提架，位置在画面右前方、怪物腿脚旁。灯火持续轻颤，暖琥珀光从低处照亮人物，投下厚重但可读的长阴影；保持低饱和褐色、暗金与怪物红褐色，不加入现代实验室、冷蓝霓虹或室外背景。正好两个独立主体：A为约三十岁的成年女性，浅肤色，中等身材，棕色头发松散向后束起，脸侧有几缕垂发，穿朴素灰褐色连帽上衣、深色长裤，没有盔甲或武器。她始终跪在木地板上，身体在画面左侧，面对右侧的怪物，双膝着地，肩背因恐惧僵直，手停在身体两侧靠近地面；眼睛睁大，呼吸急促。B为同一个巨型瘦长人形怪物：开场脸部仍覆盖苍白的仿人外层，额头异常高且前额几乎无发，两侧垂着稀疏黑长发，双眼瞪大，嘴角拉出过宽的露齿笑容。揭开外层后，内部是一张暗红与肉粉色肌束交错的畸形长脸，深陷鼻腔、圆睁眼球、象牙色细齿与异常拉长的口腔，数层齿列沿纵向口腔和下颌深处排列；头、颈和胸部结构连续，不是几个独立头颅。全身呈红褐与肉粉色的雕刻生物组织，肩背脊线、肋部、长臂和大腿的起伏清晰，表面有克制的湿润高光，不是涂成红色的人类。四肢极长，双手粗大而手指细长，正常为两臂两腿；没有盔甲、武器、角、翼或可见生殖器。它蹲伏时仍明显高于跪着的A，躯干占据房间右半部，必须弯曲膝髋、压低背部才能在低屋顶下逼近；大小差通过女人、油灯、家具与窗帘共同体现。开场仿人脸和后面的红色巨怪是B揭开伪装后的连续状态，不增加第三个人。0.00–0.60秒：直接从B的正面脸部紧近景开始，苍白高额头、两侧黑发、瞪大的双眼与僵硬露齿笑容充满画面。暖油灯光从下方照上脸颊，背景小屋虚焦。它盯着镜头前方的女人，笑容持续轻微拉宽；摄影机接近固定，只做极缓推进，不先展示空房间。0.60–3.10秒：保持同一正面近景，一只手从上方扣住B自己额头与面部外层，另一只手从下方抓住自己的下颌外层；上手向上提，下手向下拉，把仿人面部从中间纵向撑开。先看见手指抓牢和外层绷紧，再看见苍白表面分离，露出下面连续的红色怪物面部；外层在两手之间拉出细长连接纤维，黑发随外层向侧面垂下。手指始终抓着同一块外层，不凭空出现第二张脸，不直接用闪光跳变成怪物。动作缓慢、连续、具有牵拉阻力，摄影机不切走遮住关键揭露过程。3.10–4.90秒：延续同一近景，B把松开的仿人外层继续向两侧带开，真正的红色畸形长脸完全露出；它张开异常纵长的嘴，内部几层细齿和拉长的下颌结构依次可读，眼球仍在同一个头部上。颈部与肩膀随动作轻微下沉，黏连纤维逐渐松开，嘴部发出低哑、无语言的生物声。保留形体揭露的恐怖感，不加入火焰、闪电或变身光柱，不复制头颅。4.90–7.20秒：硬切A正面略低角度的面部近景，位置与刚才B的视线相对。她跪着仰望，眼睛睁大，眉心绷紧，嘴因惊恐微张，下颌轻颤，头部只做很小的后缩；灰褐色连帽上衣领口和棕色发束清晰，脸侧由低处油灯照亮，另一侧陷入可读阴影。她有持续呼吸和细小反应，不是静止照片，不说台词，不把成年脸变成年幼角色。7.20–10.90秒：保持A面部构图，两只巨大的红褐色怪物手从前方进入画面。一只手从上方按住她的额头和发际，另一只手托住下颌，粗长手指压住下唇内侧。先明确建立手掌、额头与下颌的接触，再缓慢向相反方向施力，使她的头向后仰、嘴被迫越来越大地张开。她的眼睛仍盯着上方怪物，眼眶紧张，喉部起伏，肩膀因压力向下缩；手与头部关系始终清楚，怪物不捂住她的鼻眼，不把手穿过脸颊。她的脸仍属于同一个A，只在超自然力量下逐步扩大口部开合，不突然变成怪物脸。10.90–14.00秒：硬切房间侧面的低位宽景，完整建立A在screen-left跪地、B在screen-right深蹲的空间关系。怪物两只长臂向左伸出，一手压额、一手托颌；A的头被迫后仰，嘴朝向怪物。油灯固定在右前方，怪物双脚在灯后与灯侧踩住地板，长膝弯曲、背部拱起、肋部随呼吸扩张。B开始把自己的头和颈部沿左前方向压低，靠近A已经异常张开的嘴；摄影机保持几乎固定的侧面构图，让身高差、接触位置和头部接近过程都能看见，不能切成亲吻式脸部特写。14.00–17.40秒：持续同一个侧面宽景，B进一步折叠膝髋和长颈，额头及畸长面部逐渐对准A的口部开口，压住她额头的手稳住头，托颌的手持续维持开合；怪物的头沿连续路径开始探入这个超自然扩大的口部。A仍双膝着地，上身被迫后倾但没有滑走或换位，肩颈在压力下绷紧。B的头颈先靠近、接触、再逐步进入，庞大肩背和后半身仍清楚留在外面；不把A吞进怪物嘴里，不交换进入方向，不让两个身体直接透明穿模。17.40–20.00秒：继续侧面宽景，B的肩部向左前方压进，颈部与头脸在A口部位置逐步被遮挡，巨大背部拱起，大腿和小腿继续屈曲，脚掌与脚趾承重，身体重心随前移而调整。它的双手始终固定住A的额头与下颌，没有额外的手从身体长出；A颤抖、仰头、肩部下压，衣料出现受力褶皱，仍保持跪姿。油灯没有被碰倒，灯焰随附近动作轻颤，怪物与女人的阴影在后墙缓慢变化。20秒结束在怪物已经开始侵入、巨大躯干和腿仍留在外面的持续动作中，不补拍完全吞入、女人站起、变身成功、逃跑、死亡倒地或其他结局。节奏与摄影：本片是缓慢升级的超自然恐怖揭露，不改编成追逐或格斗。保留“怪物正面近景揭开外层—女人恐惧近景—巨手控制额头与下颌—侧面宽景展现巨大体型并持续逼近”的镜头顺序。前半段靠连续形变和人物反应推进，后半段以一个持续侧面宽景展现完整空间动作；不插入绕圈镜头、突然变焦、无关反应切镜或跳吓黑帧。所有慢动作感来自怪物缓慢施力与逼近，不额外加入ultra slow motion、speed ramp或冻结画面。保持同一暖琥珀调色、同一油灯方向和同一房间地理。连续性：正好A与B两个主体；A始终是成年女性、灰褐连帽上衣与深色裤子、棕色头发、跪地位置不变；B始终两臂两腿，外层剥开后的面部与后续巨型身体属于同一只怪物，体型和材质一致。人的头颈保持相连，怪物手掌接触部位清楚，进入方向始终是怪物头部朝女人口部，不反转关系。重点呈现生物形态、恐惧反应和逐步发生的超自然动作，不添加喷血、飞溅内脏或肢解镜头。严格禁止背景音乐：全片任何时段均不得生成背景音乐、电影配乐、音乐音轨、旋律、节奏性打击乐、合唱、歌曲、音乐性持续音或音调性氛围铺底。仅允许与画面同步且物理上属于场景的画内声（diegetic audio）：屋内轻微空气流动、旧木地板承重吱响、油灯微弱燃烧声、衣料摩擦、女人短促无语言喘息与受压气息、怪物低哑呼吸、外层组织牵拉和分离的湿涩声音、手掌压住额头与下颌的接触声、巨怪缓慢屈身时的低沉关节摩擦。没有对白、旁白、可懂喊词或字幕；不得用恐怖配乐、低频音乐持续音、心跳鼓点、弦乐尖鸣、管弦乐重击、旋律化音效或节奏性脉冲填补没有音乐的空缺。出现任何背景音乐或配乐即判定为不合格。music=0; no_music=1; strict_no_music=1; audio=diegetic_only; 全片原生满幅，无黑边、圆角框、任何文字、字幕、标题、logo、interface、播放器控件或watermark。

15s cinematic fantasy action. REFERENCE IMAGE 1 is the EXACT STORYBOARD. Follow its 20 frames, timing, shot order, camera angles, poses, actions, composition and transitions exactly. REFERENCE IMAGE 2 is ONLY the male character identity. Replace the female warrior in Image 1 with the exact male from Image 2. Preserve his face, wavy dark hair, full beard, skin tone, body proportions and black outfit. Keep the flaming spear, ruined fortress, cracked stone floor, fog and grey-blue cinematic lighting. 00:00–00:00.6 F1: Hero jumps from above with flaming spear. 00:00.6–00:01.2 F2: Hero lands in deep crouch, dust bursts. 00:01.2–00:01.8 F3: Hero stands as stone golems appear. 00:01.8–00:02.4 F4: Golems surround him; spear ready. 00:02.4–00:03.0 F5: Hero makes first powerful spear swing. 00:03.0–00:03.6 F6: Spear hits golem; sparks and rocks explode. 00:03.6–00:04.2 F7: Hero spins, flaming trail, hits multiple golems. 00:04.2–00:05.0 F8: Another golem falls among debris. 00:05.0–00:05.8 F9: Hero pauses and sees larger threat. 00:05.8–00:06.6 F10: Massive stone golem rises through fog. 00:06.6–00:07.4 F11: Giant close-up, blue-white eyes glowing. 00:07.4–00:08.2 F12: Hero faces giant, spear diagonal. 00:08.2–00:09.0 F13: Giant steps toward hero. 00:09.0–00:09.8 F14: Hero charges with spear raised. 00:09.8–00:10.6 F15: Spear hits giant; huge orange explosion. 00:10.6–00:11.4 F16: Giant staggers; rocks fly. 00:11.4–00:12.2 F17: Giant collapses; dust spreads. 00:12.2–00:13.0 F18: Hero stands in dust with burning spear. 00:13.0–00:14.0 F19: ENORMOUS stone face emerges behind ruins; blue eyes ignite. 00:14.0–00:15.0 F20: Final wide shot; hero faces colossal face. PRIORITY: Image 1 controls the complete video sequence and choreography. Image 2 controls ONLY male identity. Maintain exact identity, costume, weapon, environment, scale, lighting and geography. No female, extra heroes, modern objects, cartoon style, text, subtitles or watermark.

— REFERENCE DEFINITIONS — @[girl]: Young East Asian woman in her early twenties. Layered short black hair (wolf cut, wispy brow-length bangs, ends grazing the shoulders with a slight outward flip), cool and aloof features, natural no-makeup skin. Club look: black choker necklace, dark semi-sheer sequined mesh long-sleeve top. Apartment look: champagne pure-silk slip nightdress, a thin gold chain at the collarbone. Slim build — Character appearance only. Reference. @[club]: Underground nightclub dance floor, low ceiling, mixed magenta/purple/deep-blue light sources, a strong red key light raking down from high to one side, old wooden floorboards, dense crowd. Dancers wear black cross-print jackets, studded leather belts and frayed fringe mini skirts. Background full of out-of-focus bokeh and haze — Location and mood reference only. Reference. @[apart]: Retro American apartment living room, warm daytime sun. Grey-green fabric sofa with a pink velvet cushion in the foreground. In the background, white grid-pane windows with semi-sheer teal curtains, and a rounded vintage TV on a mid-century wooden sideboard beneath the window. A rattan-shade floor lamp on the left, a metal-arm floor lamp and houseplants on the right, a low white coffee table. Overall warm amber-yellow tone — Location and mood reference only. Reference. — TECHNICAL BLOCK — Ultra filmic, photoreal. 16:9. 30s. Music + SFX. Original instrumental score, no vocals: dark driving industrial club track at 136 BPM, minor key, half-time heavy distorted kick, tight fast 8th-note hi-hats, gritty saturated bass synth, abrasive bright mid-range synth stabs, stuttering 4-hit accent bursts every few bars, played loud as if heard from inside the dance floor. Music hard-stops dead at 22.5s with no fade, followed by silence and room tone only. Every cut lands on the kick, accent bursts get flash frames. 35mm film look, heavy Kodak 500T grain, strong halation, crushed blacks, red/magenta/blue neon palette in club switching to warm amber daylight in apartment. Handheld, shallow depth of field, rapid-cut stream-of-consciousness editing, flash frames, light leaks, motion blur smears. NON-IP. No text, no logo. — PROMPT — Out of black, a distorted kick drum slams down and the picture bursts open. From here on, one cut on every beat: an extreme close-up of @[girl]'s eyes and cheeks soaked in blood-red light, cold blue leaking in at the edges; a fisheye view through the bottom of a whisky glass, blurred figures swaying in a pale cyan glow; a pure black flash frame; cold blue bokeh sweeping across the lens; two magenta neon light trails streaking sideways through the dark; gold-orange light dust drifting slowly down like embers. The instant the gritty bass synth surges in, hard cut into the @[club] dance floor: low-angle handheld weaving through the crowd at waist height, a dancer in a black cross-print jacket flinging their arms, magenta and purple lights flickering with the hi-hats, bodies in the foreground constantly blocking the frame to black. Bar close-up: a whisky on the rocks is set down on the wooden counter under red light, the ice rocking. A 4-hit burst in the music and the picture cuts 4 times in sync: @[girl] swaying out of focus in the crowd; her eyes lowered as she raises the glass and sips; her sequined cuff glinting in the red light; another sip. Pair after pair of shoes stamping on the old wooden floor. Mouth close-up: @[girl] breaks into a toothy grin in the red light, the smile slightly out of control. Second 4-hit burst: the glass is lifted right up to the lens, amber liquor and ice sloshing hard, blue-violet light shining through the glass; she tips her head back and drinks it down. Cut into a surreal flashback: a silhouetted figure in a suit falls backward in slow motion, limbs spread, against churning, bursting blood-red fire with a marbled texture, debris hanging in the air, a 500fps feel, the music roaring on and driving the fall. Back on the dance floor: a flash close-up of her eyes in red light; on the third burst, rapid cuts of deep-red and purple whip-pan smears, focus completely lost. Close-up of the glass on the bar, the ice catching a glint of firelight. On a heavy downbeat, an overexposed milky-white flash, strobe-cut against the black silhouette of a shoulder in red light, the synth getting harsher and harsher. @[girl] suddenly stops dead in the middle of the crowd and stares straight into the lens with no expression, her choker outlined in red light, the crowd behind her still swaying. Cut to the studded leather belt and white frayed fringe at a dancer's waist, then a single blurred teal flash frame that feels underwater: the music stops dead on that frame, no fade at all, the sound sucked out in an instant. In absolute silence, hard cut to @[apart], daytime, warm amber natural light. Medium close-up from the side: @[girl] in the champagne silk nightdress, slumped against the sofa arm fast asleep, lips slightly parted, loose strands of hair stuck to her face, the camera pushing in extremely slowly, only quiet room tone. Cut to a locked-off wide shot: she jolts awake and sits up, turns to look behind her in a daze, catches her breath, eyes empty; one hand presses her temple and her fingers rake back through her hair, which falls loose and messy; her whole body sinks back into the sofa, head tipped back, eyes closed, mouth open in a long exhale, the hangover weighing on every part of her. The light from the window sits perfectly still, and the vintage TV screen is a blank grey-white. SFX (under the score): crowd murmur, glass set down on wooden bar, ice cubes clinking, liquid sloshing, whoosh on flash frames, low rumbling roar under the falling silhouette, underwater muffle at the music cut-off, quiet daytime room tone, soft sleeping breath, sharp inhale on waking, silk rustling against sofa, long heavy exhale.

Create a short, ultra-realistic Japanese high-school horror-action sequence featuring the female teacher from the reference image. She is teaching normally when a single drop of blood falls onto a student’s notebook. Everyone looks up as the ceiling suddenly cracks and a zombie crashes into the classroom, sending students into panic. As they rush for the exit, another student suddenly reveals he is already infected, turning the classroom into chaos. The teacher grabs a heavy wooden bench and fights back while several students use chairs and desks to defend themselves. They eventually knock down the attackers and barricade the door. Just as the room becomes silent, a deep zombie groan echoes from the hallway. The teacher slowly turns toward the blocked door as the handle begins to move. Photorealistic live-action, practical-effects horror, realistic Japanese classroom, natural handheld camera, authentic reactions, realistic physics, muted cinematic grading, diegetic sound only. No music, narration, excessive gore, graphic injuries, animation, anime, cartoon, CGI look, distorted anatomy, subtitles, text, or watermark.

Reference the original video, keeping its choreography, camera movements, rhythm and music completely unchanged. Replace only the people and scene with those from Image 1. Preserve the identities, faces, hairstyles, body proportions, outfits and overall appearance of the people in Image 1, along with its environment and setting. All character movements, choreography, accents, camera work and soundtrack must strictly follow the original video without changes. For close-up shots, make the red-haired girl from Image 1 the subject. Ensure the replacement blends naturally with consistent lighting, realistic details and seamless visual continuity.

@ Image1 の女性が、@ Image2 の場所で、[オーディオ1]の曲でドープなヒップホップダンスを踊っている。あなたの創造性を爆発させて、この世で誰も見たことがないMusic Videoを30秒に濃縮して表現して。マルチカットなのか、長回し1カットなのかは神のみぞ知る。見る人を興奮させる、野獣派のアーティストやシュルレアリズムのアーティストの影響を受けたアーティスティックな30秒のMV.

image = Main Subject. Preserve her exact identity, face, hairstyle, skin tone, body proportions, and outfit from the reference throughout. Create a raw early-2000s Southeast Asian/Korean home-video aesthetic in a quiet residential neighborhood: narrow concrete alleys, simple houses, terraces, clotheslines, potted plants, parked motorcycles, large trees, and tangled overhead wires. No shops or commercial activity. Strong authentic DV-camcorder realism: heavy handheld shake, imperfect framing, frequent reframing, subject often near frame edges, occasional face cropping, autofocus hunting, focus breathing, exposure fluctuations, motion blur, rolling shutter, faded/desaturated colors, washed highlights, digital noise and compression artifacts. No stabilization, cinematic lighting, glamour, posing, or modern color grading. 00:00–00:03 — Sitting beside the street, she fixes her messy ponytail with both hands and smiles naturally. 00:03–00:06 — Loose close-up; she looks away and smiles as sunlight falls across her face. 00:06–00:09 — Walks into a narrow alley, crouches and feeds a neighborhood cat beside a house terrace. 00:09–00:12 — Three-quarter side shot; she waves and softly says, “Hai.” 00:12–00:15 — Walks through the neighborhood while lighting a cigarette, glancing at camera with a playful smile. 00:15–00:18 — Low handheld shot of her walking as a cat follows behind. 00:18–00:22 — At her house, she hangs freshly washed clothes on a clothesline. 00:22–00:26 — Sits on the terrace drinking coffee, looking toward the quiet street. 00:26–00:30 — Walks with coffee, notices camera, turns and smiles naturally; abrupt cut to black. Natural ambience only: birds, breeze, distant motorcycles, neighborhood sounds, cat, footsteps, leaves, and coffee-cup sounds. No music. 16:9.

30-second cinematic rooftop rescue: An overweight man hangs from a rooftop ledge, desperately shouting, “HELP ME!” A Siberian Husky grabs a fabric strap on his jacket and tries to help pull him up. Suddenly, a cardboard food box opens nearby. The Husky instantly hears it, releases the strap, and sprints toward the food, leaving the shocked man shouting, “YOU’RE LEAVING ME FOR FOOD?!” The Husky grabs the meat and unexpectedly returns. It drops the food, grabs the rescue strap again, and helps the man climb safely onto the rooftop. The man laughs, hugs the Husky, and says, “YOU LEFT ME FOR FOOD... AND CAME BACK A HERO.” The dog licks his face as they sit together overlooking the city at sunset. Style: ultra-realistic cinematic 3D, natural physics, consistent Husky, realistic sounds, dramatic-to-heartwarming mood, no cartoon effects.

[Image 1](image_1) is the face and identity reference. A young Korean woman in her early twenties with EXACTLY the face of @[Image 1](image_1) — same facial structure, same features, natural Korean skin, no glasses. Long pastel pink hair, wearing a 1940s teal-orange wool coat-dress with white collar, a brown leather satchel worn cross-body, grey knee socks, black leather shoes, holding a folded yellow envelope. Every shot of her matches @[Image 1](image_1). 1940s Venice at golden hour. Amber and teal cinematic grade, anamorphic 35mm film look, shallow depth of field, volumetric god rays, fine film grain. Inception-style folding architecture — the Venetian city curls upward and hangs inverted overhead like a mirrored ceiling. The woman always stays under normal gravity while the world folds around her. Continuous camera motion, no hard cuts. [TIMELINE PROMPT] 0–3s: Lateral tracking shot along a stone quay by the Grand Canal at sunset. The young woman @[Image 1](image_1) in a teal coat sprints past camera, clutching a yellow envelope. Above her the entire city hangs upside down, mirrored — a man in a bowler hat walks a dachshund across the inverted street. A red umbrella drifts weightlessly between the two worlds. Pigeons scatter through the amber light. 3–6s: Camera follows behind her into a narrow Venetian alley, dry leaves swirling in her wake. At the far end the street peels upward into the sky, buildings stacking vertically like a wall of windows. Her footsteps echo between the stone walls. 6–9s: The camera slowly rolls 90 degrees. The alley wall becomes the ground beneath her feet; she keeps running, unfazed. The roll continues into a wide orbit as she races up a floating spiral stone staircase suspended in mid-air, fragments of terracotta Venetian rooftops rotating around her like a kaleidoscope. 9–12s: Extreme wide silhouette. She runs across the top of an arched stone bridge over a wide canal against the blazing setting sun. The mirrored city hangs both above and below the arch. On the inverted plane, a lone pedestrian walks the opposite direction. Birds burst across the frame in slow motion. 12–15s: Low angle looking straight up between two towering building walls, the sky a thin bright strip. She leaps across the vertical chasm, arms spread wide, coat and hair flying, satchel swinging. Camera holds on her against the sky. Wind roars. 15–18s: Handheld push-in through a crowded 1940s Rialto market street — vendors stacking crates of apples, laundry strung overhead between the buildings. She weaves between blurred foreground shoppers, running away from camera. An apple rolls loose across the cobblestones. 18–21s: Camera tilts up a grand brick bell tower — a Venetian campanile — as the surrounding city folds and curls around it. The woman appears tiny on the tower ledge, pauses against the sun, then steps off toward the rooftops. Bells begin to ring. 21–24s: Rooftop terrace garden framed by a rose-covered arbor. An elderly woman in a grey cardigan waters a bed of vivid flowers with a tin can. Behind her, Venice and the domes of St Mark's Basilica glow at sunset over the lagoon while the inverted city hangs overhead. The young woman balances along the stone balustrade, arms out, then jumps down onto the terrace. 24–27s: She holds out the yellow envelope. The old woman turns, sets down the watering can, and takes it. Both smile warmly at each other. Rose petals drift upward past them, falling toward the inverted sky. 27–30s: Camera pulls back and rises fast. The whole city folds and rotates until it becomes a top-down aerial of Venice's canals and streets curling into a sphere. The two tiny figures remain on the terrace. Slow fade to warm light.

FORMAT 10 seconds, 2 shots. Shot 1: Indoor bedroom, 0.0–5.0s. Shot 2: Outdoor night street, 5.0–10.0s. Transition: seamless hard cut exactly on shoe impact. Style: photorealistic live-action, premium cinematic commercial, realistic movement/physics, authentic handheld cinematography. SHOT 1 — INDOOR BEDROOM | 0.0–5.0s ENVIRONMENT Minimalist young woman's private bedroom in daylight; clean, modern, intimate, naturally lived-in. Large curtain several meters behind Sinta, simple desk/console farther behind, minimal personal objects and subtle decoration. Enough depth for subject separation. Background softly out of focus; Sinta remains sharp and readable. CAMERA / LIGHTING Eye-level handheld, medium-wide/full-body framing with room for lateral movement and shoe actions, moderate shallow depth of field. Soft natural daylight, natural skin tones, soft highlights, gentle realistic shadows, warm clean atmosphere. Camera is operated physically with Sinta: lateral tracking, short push-in, reactive pull-back, quick pans, tilts and reframing. Never perfectly smooth or automated; movement is quick, light, imperfect and physically motivated by her actions. ACTION + CAMERA 0.0–1.1s — Sinta enters from RIGHT holding TWO SHOES, one in each hand, and walks toward center. Camera: handheld lateral tracking LEFT, physically stepping sideways while keeping medium-wide framing. 1.1–1.8s — Sinta stops near center, raises one shoe toward camera as if showing it, then brings it toward her nose. Camera: short handheld push-in; keep face and shoe sharp. 1.8–2.4s — Sinta smells the shoe, immediately recoils with surprised disgust, pulls her face away and gives a short head shake. Camera: small reactive pull-back, then immediately settle toward her face. 2.4–3.0s — Sinta casually throws the first shoe toward the side of the room. Camera: quick reactive pan following the shoe briefly, then fast pan back to Sinta; believable handheld imperfection. 3.0–3.8s — Sinta shifts sideways, focuses on the second shoe, and throws it vertically upward. Camera: tilt up following the shoe; as it reaches the top of its trajectory, quickly tilt back down toward Sinta while keeping awareness of her and the descending shoe. 3.8–4.4s — Sinta tracks the descending shoe with her eyes, adjusts footing, rotates slightly sideways and prepares a high side kick. Camera: short lateral handheld adjustment keeping body and falling shoe in frame; subtle natural Dutch tilt may develop. 4.4–5.0s — Shoe descends into kicking range. Sinta performs a fast, precise HIGH SIDE KICK: leg extends sharply sideways, torso rotates slightly, supporting leg grounded, arms naturally counterbalance, eyes locked on the shoe. Full real-time. Camera: quick handheld lateral reposition and reactive pan following the kick; frame momentarily shifts with the force. EXACT MATCH CUT — 5.0s Hard cut at the exact moment Sinta's extended foot contacts the airborne shoe. The kick is still at full speed and the cut happens DURING the action. Outdoor shot continues the same body direction, leg extension, torso rotation and momentum. SHOT 2 — NIGHT / QUIET CITY STREET | 5.0–10.0s ENVIRONMENT Relatively quiet modern urban street at night with broad paved roadway, streetlights, scattered storefronts, illuminated windows, trees, signage and distant buildings creating layered practical lighting. Sparse distant traffic. Spacious, cinematic, quiet. A WHITE SPORT MOTORCYCLE is behind Sinta and slightly toward one SIDE of frame, not centered directly behind her. Open street creates a clear diagonal path from Sinta to the motorcycle. LIGHTING / OUTFIT Naturalistic night lighting: warm streetlights mixed with cooler ambient city illumination, realistic highlights on black riding outfit, natural reflections on white motorcycle and asphalt, cinematic contrast with realistic exposure. Sinta wears a coordinated black-and-white motorcycle riding set: fitted black racing jacket with clean white shoulder/sleeve panels; fitted black short riding pants with subtle white detailing; compact black knee/shin protection; black motorcycle gloves with small white accents; clean white low-top sport sneakers; white crew socks to mid-calf; loose natural hair; no helmet. Motorcycle: white sport motorcycle. CAMERA Handheld kinetic cinematography. Operator physically follows Sinta using diagonal tracking, short arc movement, reactive panning, slight push-in and quick reframing. Never a perfectly programmed path; camera reacts to her changing body direction. ACTION + CAMERA 5.0–5.5s — HARD CUT. Sinta appears outdoors continuing the indoor side kick: kicking leg extended, torso rotated, momentum continuing. White sport motorcycle partially visible behind her toward one side. Camera immediately reacts. Camera: quick handheld pan/reframe catching the continuing kick trajectory. 5.5–6.3s — Sinta lowers her kicking leg naturally and does not stop. She moves DIAGONALLY across frame toward the motorcycle, forward-diagonal rather than straight backward/sideways, gradually turning torso and hips toward it while moving. Camera: handheld diagonal tracking with slight forward tracking plus lateral shift; keep Sinta dominant while motorcycle remains visible near the side. 6.3–7.2s — Sinta continues diagonally; body progressively rotates toward motorcycle. This is a traveling diagonal turn, NOT a stationary turn-in-place. Feet keep carrying her forward while torso rotates. Camera: short handheld arc around Sinta while continuing diagonal tracking; slight angle change and subtle pan to keep her centered; motorcycle remains partially visible at the side. 7.2–8.0s — SLOW MOTION BEGINS. Sinta is already mid-diagonal movement toward motorcycle. Body continues diagonal rotation; loose hair swings naturally; jacket moves with rotation; movement becomes elegant and cinematic. Camera: slightly smoother handheld character, subtle handheld push-in combined with short arc; stay physically close; motorcycle remains visible toward the side. 8.0–8.4s — Gradually return to normal speed. Sinta completes diagonal rotation and now faces motorcycle more directly. Camera: normal handheld movement; quick reframing keeps Sinta and motorcycle visually connected. 8.4–10.0s — Sinta moves confidently toward the white sport motorcycle, purposeful and energetic. Motorcycle stays slightly off-center. Camera: handheld forward tracking, maintaining medium-wide framing, subtle push-in, then small reactive pan toward motorcycle as Sinta reaches it. End with Sinta immediately beside the motorcycle. SPEED PRIORITY 0.0–5.0s: full real-time; shoe throw and high side kick are fast; match cut at full-speed impact. 5.0–7.2s: full real-time immediately after cut. 7.2–8.0s: subtle slow motion only during diagonal body rotation toward motorcycle. 8.0–10.0s: normal real-time. MATCH CUT PRIORITY Cut exactly on impact. Preserve across the cut: leg extension, torso rotation, body orientation, movement direction, screen position and continuous physical momentum. Location, time of day, lighting, wardrobe and motorcycle may change entirely through the hard cut. The transition must feel instantaneous, physical, stylish and seamless.

日本のフルカラーアニメ映画、シネマティック高品質映像。 手描きセルアニメ、精密でシャープな線、暖かい砂金色のハイライト、落ち着いたティールの影、細かいフィルムグレイン。 BGMなし、セリフなし、字幕なし、文字なし、環境音あり。 フルサイズ換算35mm F1.8のシネマレンズで撮ったような自然な画角と浅い被写界深度。 手前の前ボケは大きく柔らかく、背景は自然に溶ける。 料理の美味しさと、成人女性の静かな官能性を、湯気・汗・呼吸・素足・首筋・鎖骨・衣装の揺れで表現する。 # 制約 すべてのカットで画角・アングル・焦点距離を変えて単調にしない。 各カットの角括弧内の数値は、そのカットの開始時刻と終了時刻を秒で示す。 この時刻配分に従って全カットを描き、最後のカットまで到達すること。 人物は全カットを通して同じ成人女性にする。 人物の顔、髪型、髪色、肌色、体型、衣装、装飾は @Image1 を最優先で再現する。 @Image1 が複数ポーズの設定資料の場合は、人物の外見把握に使い、動画は単一の連続した空間として構成する。 衣装は @Image1 の状態を全カットで維持する。 足元は全カットを通して素足で統一する。 肌の表現は @Image1 の衣装の範囲で行い、上品で静かな大人の色気を保つ。 配置の定義：木の作業台は画面左、コンロとパエリア鍋は画面右。作業は画面左から右へ移っていく。 光の定義：右奥の大きな窓から差し込む夕日と、コンロ上の暖色ランプを全カットで維持する。 奥行きの定義：すべてのカットで前景・中景・遠景を明確に分離する。 カメラが移動するカットでは、前景を最も速く、中景を中程度、遠景を最も遅く動かす。 汗は微細な粒と薄い艶の膜として表現する。 パエリアの具材は、米、海老、イカ、ムール貝、アサリ、赤パプリカ、レモン、パセリに限定する。 熱い魚介出汁を注いだ後は、米を均した状態を保ち、静かに煮る。 画質が粗い場合はリファレンスとして画像を一から作り直すこと。 # 登場人物 ## 女性料理人 @Image1 20代後半の成人女性。 顔、髪型、髪色、肌色、体型、衣装、装飾は @Image1 を厳密に参照する。 表情は静かで集中しており、余裕のある大人のまなざし。 動作はゆっくりで手慣れている。 腰に簡素な濃色の作業用エプロンを着けてよい。 足元は素足。 微細な汗が首筋や鎖骨に薄い艶を作る。 視線は常に食材、鍋、手元へ向けられている。 # 情景 スペインの港町にある古い食堂のオープンキッチン。夕暮れ。 窓の外には漁船、石積みの防波堤、オレンジ色に染まる海が見える。 木の作業台に、洗っていないボンバ米、サフラン、赤パプリカ、にんにく、海老、イカ、ムール貝、アサリ、レモン、パセリが並ぶ。 パエリア鍋は浅い鋼鉄製で、左右に小さな金属の取っ手がある。 オリーブオイルは細い注ぎ口のある透明なガラス瓶。 熱いサフラン入り魚介出汁は素焼きのピッチャーに入っている。 料理の定義：完成するのは海鮮パエリアひと鍋。黄色い米、海老、イカ、口を開いたムール貝とアサリ、赤パプリカ、レモン、パセリが載っている。 照明：窓からの暖かい夕日を主光源にし、コンロ上のランプが鍋と人物を照らす。 奥行き構成：前景に木のカウンターの縁・湯気・ハーブ、中景に彼女とパエリア鍋、遠景に窓外の港と夕日。 # シーン 夕暮れの港町の食堂で、女性料理人が一人で海鮮パエリアを作る。 食材の準備、米の加熱、出汁の吸収、貝が開く瞬間を追いながら、 素足、ローアングルの身体のライン、首筋、鎖骨、胸元、呼吸、汗の艶を静かに交える。 人物を見せるカットと料理の変化を交互に配置し、品のある大人の色気を保つ。 cut1: [0.0-1.7] ワイドショット、波打ち際の低い位置から。 夕暮れの港町。手前の古いロープが大きくボケ、中景に食堂のテラス、遠景に漁船と防波堤が見える。 カメラがゆっくり右へパンし、暖かい窓の光が海風に揺れる。 環境音：小さな波、遠いカモメ、木の桟橋がきしむ音。 cut2: [1.7-3.1] ミディアムショット、正面斜め左から。 彼女がオープンキッチンの作業台の前に立つ。 @Image1 の髪、顔、衣装が夕日に照らされ、腰のエプロンを静かに結ぶ。 素足の指先がテラコッタの床に触れている。カメラ固定。 環境音：布が擦れる音、遠い波、食器が触れ合う微かな音。 cut3: [3.1-4.4] エクストリームクローズアップ、床スレスレのローアングル。 素足の甲、足首、かかとが暖かい夕日に照らされている。 彼女が作業台からコンロへ一歩進み、体重の移動で足指と土踏まずが静かに沈む。 床の淡い反射が足の輪郭をなぞる。手前に床の木目が大きくボケる。カメラ固定。 環境音：素足が床に触れる柔らかな音、遠い波。 cut4: [4.4-5.7] クローズアップ、真上からの俯瞰。 木の作業台に米、サフラン、海老、イカ、ムール貝、アサリ、赤パプリカが並ぶ。 彼女の手がサフランの小鉢をコンロへ近づける。 貝殻と金属ボウルに夕日が点状に反射する。カメラがごくゆっくり前進する。 環境音：貝殻同士が触れる硬い音、米粒が鉢の中でこすれる音。 cut5: [5.7-7.0] マクロ、斜め上から。 まな板の上で赤パプリカが等間隔の細切りになっていく。 包丁が一定のリズムで落ち、濡れた赤い断面が夕日に照らされる。 手前のまな板の木目が大きくボケる。カメラ固定。 環境音：包丁がまな板を叩くリズミカルな音。 cut6: [7.0-8.2] クローズアップ、斜め横から。 レモンが半分に切られ、さらに小さな楔へ分けられる。 果汁の細かい霧が夕日に光り、まな板の表面が薄く濡れて輝く。 カメラ固定。 環境音：ナイフがレモンへ入る湿った音、果皮がまな板に触れる音。 cut7: [8.2-9.5] マクロ、パエリア鍋の縁スレスレの真横から。 ガラス瓶の注ぎ口から黄金色のオリーブオイルが細く落ち、熱した鍋底にゆっくり広がる。 油膜が揺れ、窓の夕日を映して波打つ。カメラ固定。 環境音：オイルが熱面に広がる微かな音、遠い波。 cut8: [9.5-10.8] クローズアップ、斜め上から。 鍋の中でにんにくと赤パプリカが木べらで静かに炒められる。 にんにくの縁が薄い金色になり、赤パプリカの皮がつややかに縮む。 手前を薄い油煙が柔らかく横切る。カメラ固定。 環境音：具材が鍋底をこする音、油の小さな泡立ち。 cut9: [10.8-12.1] クローズアップ、鍋の内側から見上げるようなローアングル。 海老とイカの輪切りが油へ落ち、表面が透明な灰色から白と淡い朱色へ変わる。 細かい油粒が跳ね、海老の殻が鍋に触れて鳴る。カメラ固定。 環境音：魚介が熱面に触れる音、油が跳ねる音。 cut10: [12.1-13.4] クローズアップ、真上からの俯瞰。 両手に持った鉢からボンバ米が細い流れになって鍋へ降る。 白い米粒が油と具材に絡み、一粒ずつ透明感を帯びていく。 手前に赤パプリカの端が大きくボケる。カメラ固定。 環境音：米が鍋へ落ちる乾いた音、木べらが米を均す音。 cut11: [13.4-14.9] クローズアップ、背後の高角度から肩越し。 彼女の肩と髪が手前にボケ、その先の鍋へ素焼きのピッチャーから琥珀色のサフラン出汁が注がれる。 出汁が米の隙間へ沈み、白い湯気が勢いよく立ち上がる。 鍋の表面が黄金色に変わり、米は均一に沈んだ状態を保つ。カメラ固定。 環境音：液体が熱い鍋へ注がれる音、湯気が噴き上がる音。 cut12: [14.9-16.4] ローアングルの全身ショット。カメラが素足から顔へゆっくりティルトアップする煽り上げ。 手前に湯気が大きくボケて流れ、素足、足首、すね、腰、胸元、鎖骨、首筋、顔へ順に夕日の光が移っていく。 衣装の薄い布が熱気でわずかに揺れ、彼女は鍋を静かに見下ろしている。 表情は集中し、口元にかすかな余裕がある。 環境音：鍋の煮立つ音、静かな呼吸、遠い波。 cut13: [16.4-17.7] クローズアップ、真上からの俯瞰。 ムール貝とアサリが円を描くように米の上へ置かれ、海老とイカが間に整えられる。 黒い貝殻、朱色の海老、黄金色の米が幾何学的に並ぶ。 手前に鍋の金属縁が浅くボケる。カメラ固定。 環境音：貝殻が米に沈む硬い音、汁が煮立つ音。 cut14: [17.7-19.0] マクロ、鍋の縁からほぼ水平に。 煮立つ出汁の中でムール貝とアサリの殻がゆっくり開き、内側の乳白色の身が現れる。 貝の間で金色の出汁が細かく泡立ち、海老の表面が濡れたように輝く。 カメラ固定。 環境音：殻が開く硬い音、出汁が泡立つ音。 cut15: [19.0-20.4] クローズアップ、胸元から鎖骨へかけて、やや斜め前から。 湯気が胸元をゆっくり通り過ぎ、衣装の薄い布が呼吸に合わせて静かに上下する。 鎖骨のくぼみと首筋に微細な汗の粒が光り、夕日が肌の輪郭を細く縁取る。 布の上端が呼吸でわずかに揺れ、彼女が細い吐息をこぼす。顔の一部だけがフレーム上部に入る。カメラ固定。 環境音：静かな吐息、鍋の煮立ち、湯気の音。 cut16: [20.4-21.7] エクストリームクローズアップ、真横から。 小さな木のスプーンが金色の出汁をすくい、彼女の唇へ運ばれる。 唇が閉じ、喉がわずかに動く。目が静かに細くなり、味を確かめる。 スプーンが離れると、下唇に薄い出汁の艶だけが残る。カメラ固定。 環境音：遠い煮立ち、衣擦れの微かな音。 cut17: [21.7-23.0] マクロ、鍋の縁の斜め上から。 出汁がほぼ吸われ、鍋底に近い米が乾き始める。 鍋の縁の米粒が薄い琥珀色に変わり、小さな泡が弾けては消える。 表面にレモン色の油膜が細かく光る。カメラ固定。 環境音：鍋底で米がパチパチと焦げる微かな音。 cut18: [23.0-24.5] ミディアムクローズアップ、斜め後方の低い位置から。 布で包んだ両手がパエリア鍋の取っ手を掴み、火から外して木の鍋敷きへ静かに置く。 力が入った腕、肩、背中のラインが動きに合わせて滑らかに連動する。 湯気に湿った髪や衣装の裾がわずかに揺れる。カメラがゆっくり下へティルトする。 環境音：金属の取っ手を握る音、鍋が木へ置かれる鈍い音。 cut19: [24.5-26.7] クローズアップ、真上からの俯瞰。 完成間近のパエリアの上にレモンの楔が放射状に置かれる。 続いて刻んだパセリが指先からふわりと落ち、黄色い米、朱色の海老、黒い貝殻の上に緑が散る。 夕日がレモンの果汁と貝殻の濡れた光沢を強調する。カメラ固定。 環境音：パセリが落ちる微かな音、鍋底の小さなパチパチ音。 cut20: [26.7-30.0] クローズアップ、浅い被写界深度、斜め前方からゆっくり回り込む。 完成した海鮮パエリアが木の鍋敷きの上で静かに湯気を立てている。 手前のレモンが大きくボケ、中景の米と貝にピントが合い、遠景に窓外の夕暮れの海が溶ける。 彼女の手が鍋の縁からゆっくり離れ、フレームから消える。 カメラが止まり、料理だけが暖かい光の中に残る。静かな余韻。 環境音：鍋底の焦げが止まっていく微かな音、遠い波、港の静寂。

30 SECONDS | 16:9 | CINEMATIC TRAVEL DOCUMENTARY | SEEDANCE 2.5 | MULTISHOT CORE CONCEPT A young woman travels through Turkey without a fixed itinerary. Instead of chasing famous landmarks, she lets small unexpected moments guide her: a bakery she notices while walking, a wrong turn through an old neighborhood, a ferry she almost misses, a spontaneous conversation with locals, and finally an unplanned journey toward the coast. The film must feel like REAL TRAVEL FOOTAGE captured during an actual trip — not a music video, not a tourism commercial, not a polished montage. Every shot must naturally lead into the next. CHARACTER Young adult international female traveler, mid-20s, long dark-brown hair, natural facial features, minimal makeup, contemporary modest travel clothing, comfortable sneakers, small crossbody travel bag. She must remain visually consistent across every shot. Natural behavior only. She is curious, observant, slightly spontaneous and never performs for the camera. SETTING Realistic contemporary Turkey across Istanbul, a quieter residential neighborhood, local bakery, Bosphorus ferry, Cappadocia countryside, a small roadside tea stall and a Mediterranean coastal town. Real Turkish streets, local businesses, pedestrians, traffic, shop signs, imperfect sidewalks, parked cars, ordinary architecture and authentic environmental details. Avoid postcard-perfect environments. STORY — 12 MULTISHOT SEQUENCES 00:00–00:02.5 — HOTEL EXIT She steps out of a modest Istanbul hotel carrying her small crossbody bag. She checks her phone, looks at a saved list of places, hesitates, then locks the phone and smiles slightly. She starts walking with no clear destination. CAMERA: handheld 35mm documentary framing, following slightly behind her. AUDIO: distant traffic, footsteps, city ambience. 00:02.5–00:05 — THE BAKERY While walking, she notices a small neighborhood bakery through an open doorway. She stops. The camera moves closer as she looks inside at fresh bread being pulled from the oven. A baker notices her looking and gestures for her to come in. She smiles and enters. CAMERA: imperfect handheld push-in, natural autofocus adjustment. AUDIO: street noise transitioning into bakery ambience, oven door, quiet Turkish conversation. 00:05–00:07.5 — UNEXPECTED MOMENT She receives a small piece of freshly baked bread from the baker. She takes a bite and reacts naturally — pleasantly surprised. The baker laughs. She gives a small thumbs-up. No staged posing. CAMERA: close handheld 50mm, slight focus breathing. 00:07.5–00:10 — WRONG TURN Back outside, she turns down a narrow street instead of following the main road. The camera follows. She realizes she has taken the wrong direction, checks her phone, then decides to keep walking anyway. The street opens into a lively local neighborhood. CAMERA: handheld 24mm, walking movement, occasional framing imperfections. 00:10–00:12.5 — THE FERRY Cut to her arriving at a Bosphorus ferry terminal. She sees the ferry beginning to board and quickly walks toward it. She reaches the entrance just before departure. CAMERA: handheld tracking from behind, natural movement, slight motion blur. AUDIO: ferry engine, footsteps, people talking, distant announcements. 00:12.5–00:15 — BOSPHORUS She stands near the ferry railing. Wind moves her hair and jacket naturally. Istanbul passes behind her as she watches the shoreline. She looks toward the camera and quietly says: “Okay… this wasn’t planned.” CAMERA: 50mm handheld close-up, natural background compression. 00:15–00:17.5 — A NEW DIRECTION Later, she arrives in Cappadocia. She steps out onto a quiet rural road surrounded by dry valleys and distant rock formations. Instead of immediately heading toward a tourist attraction, she notices a small tea stand beside the road. She walks toward it. CAMERA: 24mm wide handheld establishing shot transitioning into a 35mm follow. 00:17.5–00:20 — TEA WITH LOCALS An older local man pours Turkish tea into a small glass. She sits nearby. They exchange a few words despite the language barrier. He points toward the distant landscape. She looks where he points. CAMERA: observational 50mm handheld, shallow depth of field, natural focus transitions. AUDIO: wind, tea glasses, distant birds, quiet conversation. 00:20–00:22.5 — FOLLOWING THE SUGGESTION She walks toward the direction the man indicated. The landscape gradually opens. She reaches a viewpoint overlooking a vast Cappadocian valley. She stops completely. No dramatic pose. Just genuine observation. CAMERA: slow handheld reveal from behind her shoulder into a natural wide shot. 00:22.5–00:25 — THE COAST Hard geographic transition to a small Mediterranean coastal town in Turkey. She walks beside the water carrying the same bag. A fisherman is pulling a small boat toward shore. She briefly stops to watch. The fisherman looks over and casually nods. She smiles back. CAMERA: 35mm handheld street-documentary perspective. 00:25–00:27.5 — KEEP WALKING She continues along the waterfront as evening approaches. No destination visible. She passes cafés, locals sitting outside, children playing nearby and small fishing boats. The camera follows naturally from several meters behind. A slight autofocus hunt occurs as a person crosses the foreground. 00:27.5–00:30 — FINAL MOMENT She reaches the end of the waterfront and looks toward the open sea. She takes out her phone. Instead of checking a map, she puts it back into her pocket. She smiles and keeps walking toward an unfamiliar street. CUT TO BLACK. No text overlay. CAMERA Photorealistic travel-documentary cinematography. Primarily handheld 24mm, 35mm and 50mm lenses. Natural walking shake. Small framing imperfections. Occasional autofocus hunting. Natural exposure adaptation when moving between interiors and exteriors. Realistic motion blur. No floating camera. No impossible tracking shots. No perfect gimbal movement. No excessive shallow depth of field. The camera should feel physically operated by a real person documenting an actual trip. LIGHTING Natural available light. Soft Istanbul daylight. Warm bakery interior lighting. Cooler ferry daylight with natural atmospheric haze. Strong but realistic Cappadocian sunlight. Soft late-afternoon coastal light transitioning toward evening. No artificial cinematic lighting. No exaggerated golden-hour glow. PHYSICS Accurate human movement. Natural hair and clothing movement from wind. Realistic walking speed. Correct ferry movement and water behavior. Natural crowd movement. Authentic Turkish tea handling. No synchronized background pedestrians. No floating objects. No warped architecture. No impossible reflections. HUMAN PERFORMANCE All reactions must feel spontaneous and unscripted. The traveler does not constantly smile at the camera. She looks around, gets distracted, hesitates, makes small mistakes and reacts naturally to unfamiliar situations. Locals behave independently and do not perform for the camera. No influencer posing. No commercial-model behavior. No exaggerated facial expressions. AUDIO Diegetic location audio throughout. Footsteps. Street traffic. Turkish conversations. Bakery sounds. Oven door. Ferry engines. Water movement. Wind. Tea glasses. Distant birds. Coastal ambience. Very subtle natural room tone between locations. NO MUSIC-VIDEO FEEL. NO CONSTANT MUSICAL MONTAGE. Dialogue should sound naturally recorded in the environment. VISUAL CHARACTER Photoreal cinema with authentic documentary texture. Slight filmic grain. Mild halation on highlights. Natural skin tones. Teal-cyan skies. Punchy crimson reds. Deep forest greens. Warm skin. Crushed-but-not-crushed blacks. Cinematic teal-and-warm grade. Greens slightly cyan. Natural atmospheric haze. 24fps. 180-degree cinematic shutter. Natural motion blur on hair and fabric. NO CGI LOOK. NO PLASTIC SKIN. NO BEAUTY-FILTER SMOOTHING. NO AI-SMOOTHED FACES. NO EXTRA FINGERS. NO WARPED ARCHITECTURE. NO PERFECTLY SYMMETRICAL COMPOSITIONS. REALISM RULE The result must look like footage captured during a real woman’s spontaneous trip through Turkey and later assembled into a coherent documentary. Prioritize believable human behavior and environmental realism over visual spectacle. Each shot should feel observational rather than designed. The locations must feel connected through the traveler’s decisions. This is a STORY, not a music video. CONTINUITY Same traveler throughout. Same face, hair, clothing, crossbody bag and physical proportions. Natural time progression from daytime Istanbul → ferry → Cappadocia daylight → coastal late afternoon/evening. Geographic transitions should feel motivated by the story. No random location jumps. No wardrobe changes unless naturally justified. No duplicate characters. No invented logos or watermarks. No random UI.

Create a 30-second ultra-realistic AAA open-world gameplay video set inside a massive unfinished skyscraper at a modern Indian construction site. It must look like authentic playable next-generation game footage—not a movie, trailer, cutscene, or pre-rendered cinematic. PLAYER & CAMERA A young Indian man wears a dark work jacket, dusty cargo pants, safety boots, gloves, and a yellow safety helmet. Keep the same face, outfit, proportions, and appearance throughout. Use a consistent third-person camera slightly behind and above him, with responsive player-controlled movement. No first-person view, cinematic cuts, slow motion, or scripted hero shots. HUD Original realistic AAA HUD throughout: - Minimap bottom-left - Health/stamina top-right - Objective top-left - Context-sensitive prompts bottom-center - Interaction indicators when needed Controls: L3 MOVE | R3 CAMERA | X JUMP/CLIMB | ○ DODGE | □ INTERACT | △ ENTER/EXIT CONTINUOUS 30-SECOND GAMEPLAY 0–2s: Inside a temporary construction elevator. Player presses the control panel. □ INTERACT. Elevator rises rapidly through the unfinished 50-floor skyscraper. Objective: REACH FLOOR 50. 2–4s: Cables snap. Elevator drops several floors and shakes violently. Player grabs the emergency railing. ○ GRAB / STABILIZE. 4–6s: Emergency brakes activate. Elevator slams to a stop between floors. Dust falls. Warning: ELEVATOR FAILURE. □ OPEN EMERGENCY DOOR. 6–8s: Player forces open the doors, climbs onto the elevator roof through the dangerous gap. X CLIMB. Exposed shaft, concrete floors, steel beams, cables, and construction lights visible. 8–10s: Player climbs a steel maintenance ladder upward. L3 — CLIMB. Huge vertical shaft visible above and below. 10–12s: A metal pipe falls from above. Player dodges sideways. ○ DODGE LEFT. Pipe crashes into the structure and falls into darkness. 12–14s: Player reaches an unfinished floor and jumps onto narrow scaffolding. X JUMP. Scaffolding shakes; distant workers evacuate. 14–16s: A crane swings a suspended steel beam across the path. ○ DUCK. Player crouches as it passes inches above his helmet. 16–18s: Floor cracks and begins collapsing. Warning: FLOOR COLLAPSING. Player sprints forward. L3 + R2 — SPRINT. Concrete falls behind him. 18–20s: Player jumps across a missing floor section and barely catches the opposite steel beam. X — JUMP. He hangs above the shaft. 20–22s: Player pulls himself up, stands, and runs toward a temporary emergency staircase. X — CLIMB. Red construction alarms flash. 22–24s: Player sprints up the unfinished staircase. L3 + R2 — SPRINT. Workers move downward, with dust, sparks, cables, and emergency lights. 24–26s: Player reaches the rooftop door and opens it. □ OPEN. Camera remains gameplay-controlled as the huge Indian city skyline, cranes, traffic, and buildings appear. Objective: GET TO THE ROOFTOP EXTRACTION POINT. 26–28s: Player runs across the unfinished rooftop. Scaffolding collapses behind him. ○ DODGE. He narrowly avoids falling debris. 28–30s: A realistic rescue helicopter approaches. Rotor wash throws dust and plastic sheets into the air. Player reaches the marked extraction zone. △ ENTER HELICOPTER. Mission notification: MISSION COMPLETE — ELEVATOR FAILURE SURVIVED. End immediately like a completed gameplay mission. VISUAL STYLE Photorealistic AAA game-engine graphics, realistic Indian metropolitan construction environment, unfinished concrete, steel scaffolding, cranes, cables, machinery, workers, dusty surfaces, warning lights, distant traffic, and believable destruction. Realistic climbing, jumping, falling, collisions, and character animation. STRICT CONTINUITY One character, one elevator, one continuous building, one uninterrupted mission. No duplication, outfit changes, teleportation, disappearing or floating objects, impossible physics, random spawning, fantasy technology,

Ultra-photorealistic cinematic animated adventure video following a young boy in an orange explorer jacket and backpack through a magical fantasy world → he discovers and gently interacts with a cute blue bird surrounded by glowing particles → explores a beautiful enchanted forest filled with lush trees, flowers, fruit trees, rocks and friendly gorillas → discovers a mysterious glowing blue portal in a forest clearing while colorful fantasy creatures watch nearby → seamless transition into a futuristic greenhouse filled with plants, robotic arms and adorable green plant creatures → the boy reunites with his blue bird and companions → final warm cinematic shot of the group together in the magical futuristic garden. Detailed expressive characters, realistic fabric and fur textures, natural character movement, warm sunlight, volumetric lighting, glowing magical particles, cinematic depth of field, smooth camera movement, dynamic framing, believable environmental interaction, consistent character appearance, polished high-quality 3D animation, seamless transitions, no text, no watermark, no distorted faces, no flickering.

CONCEPT Cinematic live-action rooftop superhero fight featuring the EXACT male hero from Reference Image 1. Faithfully recreate the reference video's action, pacing, camera language, environment, and movement. SETTING Bright daytime urban rooftop with concrete walls, stairs, wires, palm trees, buildings, dust, and hard sunlight. Preserve realistic physics. CAMERA Dynamic cinematic action: wide angles, rapid reframing, low angles, close tracking, handheld movement. CAMERAMAN MOTION Start with fast lateral tracking and purple energy. Track and rotate during wall-running. End wider for the aerial impact. SUBJECT — REFERENCE IMAGE 1 IDENTITY LOCK The hero MUST be the EXACT male character from Reference Image 1. Preserve face, eyes, nose, lips, jawline, skin, hairstyle, hairline, body proportions, physique, age appearance. Do not replace or morph him. He wears a fitted blue-and-white martial-arts outfit. Opponents are adult males in dark tactical clothing. Purple energy surrounds attacks. ACTION TIMELINE 0–3s HOOK: Fast rooftop tracking with purple energy. Reveal the exact Reference Image 1 hero rushing toward attackers, then wall-running. 3–7s BUILD UP: Hero runs horizontally along the wall, transitions into combat, and strikes attackers with purple energy. 7–10s PAYOFF: Hero performs a powerful aerial move and descends with force. The rooftop cracks on impact, sending dust and debris outward. End as dust settles. AUDIO Action: wind, whooshes, impacts, energy, footsteps, concrete breaks, debris crashes. Build intensity to the final impact. Maintain consistent character scale, anatomy lighting, impact timing, and camera direction throughout the sequence. NEGATIVE PROMPT No identity drift, different face, female hero, morphing, face swap, hairstyle/body changes, costume redesign, duplicates, extra limbs, teleportation, impossible physics, random cuts, new locations/events, cartoon/anime style, plastic skin, uncontrolled purple glow. Preserve live-action realism and Reference Image 1 fidelity.

- fabric friction - strings stretching and contorting - boots clumpy and hard against concrete - reverberation of all diegetic audio

Create an EXACTLY 30-SECOND, 16:9 LANDSCAPE (1920×1080) ULTRA-PHOTOREALISTIC LIVE-ACTION VIDEO. Create an original viral rooftop rescue comedy with genuine physical realism, natural human expressions, believable animal behavior, and cinematic but grounded camera work. The video must look like authentic footage filmed with a real overweight American actor and a real adult Siberian Husky on a physical urban rooftop. Never CGI, animation, cartoon, video-game footage, or visibly AI-generated imagery. The story: An overweight American man hangs from a tall rooftop ledge. A Siberian Husky initially tries to rescue him but abandons him instantly after hearing a food box open. The dog then unexpectedly returns carrying food and helps rescue the man, ending with a funny, heartwarming reunion. MAIN CHARACTER — DETAILED APPEARANCE Create an overweight white American man approximately 38 years old with a distinctive ordinary, slightly rugged face. His face has a broad, rounded shape, full cheeks, a slightly wide jaw, a broad forehead, and natural facial asymmetry. He has short, messy dark-brown hair with irregular strands falling loosely across his forehead. His hair is slightly untidy and flattened in places, as if affected by wind. He has light skin with realistic natural pores, mild redness around the cheeks and nose, subtle under-eye creases, faint forehead lines, and slightly uneven skin texture. He has natural short brown stubble across his cheeks, chin, and upper lip. His eyebrows are moderately thick and slightly uneven. His eyes are medium-sized, naturally expressive, and slightly tired-looking, with realistic moisture and subtle wrinkles around them. His nose is broad and naturally proportioned. His mouth is relatively wide, with expressive lips that stretch naturally when he shouts. His face should look like an ordinary real man rather than a conventionally perfect model. His expression can transition from intense panic to disbelief, frustration, surprise, relief, and genuine happiness. Preserve consistent facial structure throughout the video. No face morphing, plastic skin, excessive smoothing, beauty-filter effects, or artificial facial symmetry. He has a heavy, realistic body with natural weight distribution, a broad upper body, thick arms, a substantial torso, and believable physical strain while hanging from the rooftop. WARDROBE: Weathered brown utility jacket over a dark-gray hoodie, dark jeans, and worn sneakers. The jacket has natural creases, slightly faded fabric, and realistic wear. Maintain consistent clothing throughout every shot. SIBERIAN HUSKY Exactly ONE adult Siberian Husky. Realistic black-and-white thick double-layer fur, natural blue eyes, erect triangular ears, detailed muzzle, realistic teeth, whiskers, paws, claws, and anatomically correct canine proportions. The Husky is highly food-motivated and behaves like a real trained dog. It can bite and pull a thick fabric rescue strap with its mouth but never uses its paws like human hands or behaves like a human. The dog’s food reaction must be sudden and unmistakable, followed by an unexpected return and a physically believable second rescue. ENVIRONMENT & LIGHTING A large, weathered concrete apartment rooftop overlooking a dense American city. Dark-gray rooftop surfaces, rough concrete ledges, small gravel particles, distant streets, neighboring buildings, and a pale blue-gray evening sky. A small rooftop access room is located several meters behind the ledge. An open door leads to the room. A metal table beside the door holds a cardboard food box containing dog-safe cooked meat. Natural late-evening ambient light. No artificial cinematic lighting, orange-and-teal grading, fake fog, glossy skin, excessive sharpening, or unrealistic visual effects. Use natural exposure, realistic lens depth, subtle handheld camera movement, accurate architectural scale, and physically correct shadows. CAMERA & MOTION Original cinematic compositions created specifically for this story. Real 24fps live-action appearance, natural motion blur, realistic camera movement, clean direct cuts, and believable physical timing. Use tight frontal shots, low-angle views, side close-ups, wider rooftop shots, and interior access-room shots. The camera must remain physically grounded, with no floating movement, digital zoom, speed ramps, slow motion, fades, dissolves, or montage-style rapid cutting. --- 0–4 SECONDS — IMMEDIATE VIRAL HOOK TIGHT FRONTAL ROOFTOP SHOT. Open immediately with the overweight man hanging from the vertical concrete face of the rooftop. His face, shoulders, both complete forearms, and both complete hands are clearly visible. The city skyline and pale evening sky remain recognizable behind him. His left and right hands separately grip the same rough concrete ledge. All ten fingers are anatomically correct, naturally wrapped around the edge, and consistently visible. His arms shake under his body weight. His sneakers scrape the concrete wall. Wind moves his brown jacket and messy hair. His face shows genuine fear and physical strain. His mouth opens naturally as he shouts in American English: “HELP ME! SOMEBODY HELP ME!” Natural strained breathing, distant city ambience, and subtle handheld camera movement. 4–8 SECONDS — HUSKY STARTS THE RESCUE HARD CUT TO A LOW-ANGLE POINT OF VIEW FROM THE MAN. The Siberian Husky stands on the rooftop ledge and looks directly down at him. The man's two hands remain visible at the bottom of the frame, gripping the same concrete positions. The Husky carefully crawls forward, bites a thick fabric strap attached to the man's jacket near his left shoulder, and begins pulling backward. The dog bites ONLY the fabric strap, never the man's hands or skin. All four paws remain on the rooftop. The Husky braces against the concrete while pulling with visible effort. The man actively pulls upward with both arms and pushes against the wall with his feet. The dog provides only a small stabilizing pull. The man rises several centimeters through his own effort. The man's face shows desperate hope as he says: “COME ON, BOY! YOU GOT ME!” Show realistic fabric tension, natural fur movement, believable canine effort, and accurate body weight. 8–11 SECONDS — THE FOOD BOX OPENS SIDE CLOSE-UP OF THE RESCUE CONTACT. The Husky's teeth remain attached to the same fabric strap. The strap stretches naturally under tension. Its front legs brace against the concrete, and its rear legs push backward. The man's two hands remain fully visible and continuously attached to the ledge. A distinct cardboard box-opening sound comes from the nearby access room. HARD CUT INSIDE THE ROOFTOP ACCESS ROOM. A realistic cardboard food box rests on a metal table. No person's hand is visible. An automatic spring mechanism lifts the lid. The lid suddenly snaps open with a sharp cardboard movement. Fresh dog-safe cooked meat is visible inside. One piece shifts slightly inside the box. The sound is natural and unmistakable. The tense background music abruptly stops. No electronic beep, notification sound, cartoon boing, artificial whoosh, or exaggerated comedy effect. 11–15 SECONDS — THE HUSKY ABANDONS HIM HARD CUT BACK TO THE EXACT SAME ROOFTOP AND RESCUE POSITIONS. In the first frame, the Husky is still biting and pulling the fabric strap. Within LESS THAN HALF A SECOND of hearing the food box: - Both ears snap completely upright simultaneously. - Its eyes widen naturally. - Its head whips sharply toward the access room. - It immediately releases the fabric strap. - Its hindquarters drop for a fraction of a second. - Both rear paws compress against the concrete. - Its entire body launches backward away from the rooftop edge. The reaction must be sudden, large, and unmistakable. No hesitation, cautious pause, slow head turn, or delayed response. At the exact instant the dog releases the strap, the man tightens both hands around the concrete ledge. All ten fingers remain consistent and anatomically correct. His body drops only a few centimeters and swings naturally. He never loses his grip. He looks at the escaping Husky in complete disbelief and shouts: “WAIT! WHERE ARE YOU GOING?!” The Husky does not look back. 15–19 SECONDS — FULL-SPEED FOOD ESCAPE WIDER SIDE ROOFTOP SHOT. The Husky lands securely on all four paws approximately one body length behind its previous position. It pivots nearly 180 degrees during the landing and immediately accelerates into a full-speed sprint toward the open access-room door. Its claws scrape the concrete briefly. Its tail moves naturally for balance. Its thick fur responds to the wind. Its paws strike the rooftop with realistic impacts. The dog never walks, trots, pauses, hesitates, or checks the man again. The man remains hanging with both hands gripping the ledge. His expression changes from panic to stunned frustration. He shouts: “YOU'RE LEAVING ME FOR FOOD?!” The Husky physically runs through the open doorway into the access room. It does not teleport or vanish. Hold briefly on the man's incredulous expression while the dog's footsteps fade toward the room. The comedy must come from the dog's extreme food motivation, not cartoon effects. 19–22 SECONDS — UNEXPECTED RETURN WITH FOOD HARD CUT INSIDE THE ACCESS ROOM. The Husky reaches the open cardboard food box and picks up a large piece of dog-safe cooked meat with its mouth. The food is clearly visible and securely held between its teeth. The dog immediately turns toward the rooftop exit. HARD CUT TO A MEDIUM-WIDE ROOFTOP SHOT. The Husky bursts back through the open doorway carrying the food and runs toward the hanging man. The man looks up, startled. Both hands remain firmly attached to the concrete ledge. His expression transitions from frustration to confusion, surprise, and cautious hope. He says: “YOU CAME BACK?” The Husky continues toward him carrying the food. The dog must remain the same individual, with consistent fur patterns, body size, and appearance. No duplicated dog, teleportation, unexplained location change, or sudden appearance. 22–26 SECONDS — THE SECOND RESCUE LOW-ANGLE RESCUE SHOT. The Husky approaches the rooftop edge carrying the food. It drops the food safely on the concrete, then bites the same thick fabric rescue strap attached to the man's jacket near his left shoulder. It braces all four paws against the rooftop and pulls backward. The man actively pulls himself upward with both arms and pushes against the wall with his feet. The Husky's body strains naturally. Its paws grip the rough concrete. Its fur moves with the effort. The man climbs onto the rooftop in a believable, controlled motion. The dog assists with stability and traction but does not magically lift the man's entire body weight. Ensure the man's hands, arms, and fingers remain anatomically correct throughout the climb. Show a realistic transition from hanging vertically to getting his upper body over the ledge and safely rolling onto the rooftop. The man exhales sharply with relief. 26–30 SECONDS — HEARTWARMING VIRAL ENDING WIDE ROOFTOP SHOT TRANSITIONING INTO A NATURAL MEDIUM CLOSE-UP. The man safely rolls onto the rooftop and sits on the concrete, breathing heavily. The Husky stands beside him, tail wagging naturally. The man looks at the dog, then notices the piece of food on the concrete. He laughs in exhausted disbelief. He picks up the food and offers it back to the Husky. The Husky gently takes it from his hand. The man wraps one arm around the dog while smiling with genuine relief. He says warmly in American English: “YOU LEFT ME FOR FOOD... AND CAME BACK A HERO.” The Husky licks his face. The man laughs naturally. The final shot captures both of them sitting together on the rooftop, the city skyline and pale evening sky behind them. The man gently pets the Husky while the dog enjoys its food. A distant city siren passes far below. Wind moves the man's jacket and the dog's fur. End immediately at 30 seconds with a genuine, heartwarming moment. SOUND DESIGN Only realistic environmental and action sounds: - Rooftop wind. - Distant traffic and city ambience. - Distant siren. - Strained breathing and man's shouted dialogue. - Sneakers scraping concrete. - Gravel and dust shifting. - Fabric stretching. - Dog claws scratching concrete. - Cardboard food-box lid opening. - Food movement inside the box. - Rapid paw impacts as the dog sprints. - Natural dog breathing and subtle panting. - Gentle final interaction sounds. The cardboard food-box opening is the EXACT TRIGGER for the Husky's immediate reaction. Use restrained tense background music during the rescue. Cut the music abruptly when the box opens. Allow the dog's rapid footsteps and man's frustrated shout to create the comedic punchline. Introduce gentle warm background music only during the final reunion. No electronic beeps, notification chimes, cartoon boings, meme sounds, artificial whooshes, bubble sounds, gurgling, or exaggerated comedy effects. ABSOLUTE ACTION PRIORITY The Husky's food reaction must be sudden and unmistakable: 1. Instant ears upright. 2. Immediate wide-eyed reaction. 3. Sharp head turn toward the food. 4. Immediate release of the strap. 5. Powerful backward leap. 6. Rapid pivot. 7. Full-speed sprint to the access room. 8. No hesitation or final look back. 9. Quick return carrying food. 10. Second rescue attempt. 11. Happy reunion with the man. The comedic twist must be clear to the audience. The dog initially prioritizes food, then returns unexpectedly and helps the man. The dog's return should feel like believable trained-dog behavior rather than human-level reasoning. It should not talk or perform human gestures. STRICT CONTINUITY & QUALITY CONTROL Exactly ONE man and ONE Siberian Husky throughout the video. The man maintains two anatomically correct hands and ten consistent fingers gripping the ledge until he safely climbs up. No missing hands, disappearing arms, fused fingers, extra limbs, changing grip, body distortion, or unnatural hand deformation. The man's face remains consistent throughout every shot. No identity drift, facial morphing, artificial smoothing, or changing facial proportions. The man's wardrobe remains consistent. The Husky maintains the same fur pattern, size, and appearance. No teleportation, duplicated dog, changing clothes, plastic skin, flickering details, shifting rooftop geometry, fake city scenery, CGI wind, subtitles, captions, text, logos, or watermarks. The man never falls. The rescue concludes with him safely back on the rooftop. Prioritize authentic live-action realism, natural facial expressions, physically believable action, clear storytelling, strong comedic timing, and a satisfying heartwarming ending. FINAL OUTPUT: EXACTLY 30 SECONDS, 16:9 LANDSCAPE, 1920×1080, ULTRA-PHOTOREALISTIC LIVE-ACTION, NATURAL CINEMATOGRAPHY, BELIEVABLE PHYSICAL ACTION, AND A FUNNY, HEARTWARMING VIRAL ENDING.