Grok Imagine Video Prompt - Halloween Parade
A video generation prompt for Grok Imagine featuring characters singing and dancing in a Halloween parade scene.
本页汇集归入“Crowd Group”分类的视频提示词示例。调整示例用于自己的项目之前,可先比较生成结果、构图或镜头选择以及原始写法。
1,130 prompts
主体
A video generation prompt for Grok Imagine featuring characters singing and dancing in a Halloween parade scene.
Design them dancing and singing the welcome to Halloween song in the Halloween parade
A detailed prompt for creating heartwarming, cinematic Pixar-style 3D animated family stories using Seedance 2.5.
Create a heartwarming, cinematic 3D animated family story in a high-quality Pixar-style aesthetic. A muscular, bearded father relaxes on a sofa using his phone while his adorable baby plays nearby. Show funny everyday family moments, the father interacting lovingly with his baby, and the mother joining them at home. Suddenly, an emotional moment unfolds as the parents realize how precious their child is. Use expressive big eyes, detailed facial animation, realistic lighting, warm sunlight, smooth camera movements, close-up emotional reactions, and beautiful storytelling. End with the father, mother, and baby embracing each other lovingly. Ultra-detailed 3D animation, cinematic depth of field, warm color grading, natural character movements, emotional music, consistent character designs, 4K quality, vertical 9:16 format.
A prompt for Seedance 2.0 to generate a 10-second hyper-realistic time-lapse video of miniature workers assembling Adidas football boots in a futuristic workshop.
Create an exactly 10-second vertical 9:16 ultra-realistic Adidas football boot assembly timelapse in a dark futuristic workshop. Exactly 8 miniature workers in blue uniforms operate robotic arms, conveyors, and precision tools. Assemble a black and electric-blue football boot with white Adidas stripes, black laces, textured panels, padded collar, and metallic studs. Show the complete process chronologically: workers arrange components, install studs, stitch fabric panels, attach the upper to the sole, thread laces, and inspect the finished boot. Machinery retracts as workers gather near the base. End with a premium stationary product hero shot. Fixed camera, consistent geometry, realistic physics, cinematic blue lighting, detailed textures, mechanical ASMR, and energetic 130 BPM music. No cuts, teleportation, extra workers, camera movement, text, or watermarks.
A high-detail cinematic combat prompt for Seedance 2.5 featuring a duel between an orc and a human rogue in a medieval tavern with realistic physics.
Style: Hyper-realistic dark fantasy tavern cinematic, grounded physical combat, realistic body momentum, medieval atmosphere, gritty lighting, practical effects, handheld cinematic camera, physically accurate movement, ultra-detailed textures, realistic collisions and environmental interaction. Duration: 15 seconds Aspect Ratio: 16:9 IMPORTANT: Keep the SAME characters fully consistent with the reference: — Nyssa: athletic woman, dark curly hair tied back, bronze leather armor, cloth wraps, sword on hip — Grok: massive muscular green orc, scarred skin, fur armor, heavy axes, large physical weight Combat MUST obey realistic physics: — no floating — no anime flips — realistic inertia and recovery — believable weight transfer — realistic impacts and exhaustion — grounded footwork — environment reacts physically to impacts [00:00-00:02] Wide cinematic shot inside a crowded medieval tavern filled with drunk mercenaries, candles, smoke, spilled ale, and loud cheering. Nyssa walks through the tavern cautiously while patrons stare. Grok sits at a massive wooden table drinking heavily. Deep bass-heavy tavern music and crowd ambience. [00:02-00:04] Close-up tension sequence: — Nyssa locks eyes with Grok — Grok slowly stands, towering over everyone — benches scrape across the floor — mugs shake from his weight — crowd forms a fighting circle cheering loudly Handheld camera movement feels realistic and grounded. [00:04-00:07] The fight erupts violently. Grok swings a heavy punch with believable momentum. Nyssa narrowly dodges while stumbling realistically into a wooden table. Wood cracks and splinters physically on impact. She counters with fast grounded strikes to Grok’s ribs and legs. Crowd roars and throws coins and mugs into the air. [00:07-00:10] Combat intensifies through the tavern: — Grok grabs Nyssa and throws her across a table — table explodes realistically beneath her weight — Nyssa rolls across the floor recovering naturally — Grok charges heavily, smashing furniture — patrons jump away realistically No exaggerated movement, only physically believable combat. [00:10-00:12] Nyssa uses speed and positioning intelligently. She dodges another heavy attack causing Grok to crash into a support beam. Dust and debris fall from the ceiling. She climbs briefly onto the bar counter and leaps down with controlled realistic momentum, locking Grok’s arm and using leverage to throw him off balance. [00:12-00:15] Final cinematic climax. Nyssa tackles Grok to the ground hard onto broken wooden debris. The floor shakes from the impact. She pins him down on top of him while holding a blade near his throat. Grok struggles realistically beneath her massive weight difference and exhaustion. Tavern crowd erupts cheering wildly, slamming mugs against tables. Camera slowly pushes inward on Nyssa breathing heavily while sweat, d
A detailed prompt for Seedance 2.0 to generate a realistic 15-second video of a person jumping from a helicopter onto a trampoline on the sea.
【優先参照画像&名称固定】 @ image1 =記述名称:主人公 主人公の人物同一性、顔、髪型、体型、衣装、装飾、装備、靴の外観基準。 以降の本文では、上記で名称固定した名称のみを使用する。 15秒、3:4、実写フォトリアル、明るい昼の青い海上を舞台にした、音声付きの一続きの映像。全編を通して画面の切替なし。人物参照は主人公の外見だけに用い、三面図の並列表現や文字・見本レイアウトは映像へ取り込まない。人物の同一性、衣装、装飾、体格は全編で維持する。BGM・字幕・テロップ・ロゴなし。 登場人物は主人公、クルーA、クルーB、クルーCの計4名。白い機体に青いラインが入った民間の中型撮影用ヘリコプター。黒いスキッド式着陸脚、開いた幅広の側面スライドドア、乗降用の外側ステップ、ダークグレーの機内。ヘリコプターは最初から最後まで海面から一定の高度を維持する。遠景には海岸線が見え、はるか下の海面上には円形トランポリンが設置されている。主人公は開いたドアの外側ステップに立ち、遠く下のトランポリンを見ている。クルーAはドア寄りに立ち、青いウインドブレーカーと黒い救命ベスト、航空用通信ヘッドセットを着用。クルーBは機内のドア近くに立ち、オレンジ色のウインドブレーカーと黒い救命ベスト、航空用通信ヘッドセットを着用。クルーCは機内の奥にいて、緑色のウインドブレーカーと黒い救命ベスト、航空用通信ヘッドセットを着用する。クルー3名の顔が見える。カメラはヘリコプターのドア付近の同じ撮影位置から、やや広角で主人公とクルー3名、ドアの向こうの海を捉えて開始する。カメラの設置位置は変えず、落下と上昇に応じて上下に向きを変える。風切り音とヘリコプターのローター音が全編に自然に続く。 0.0〜3.0秒:クルーAは主人公の方を見て笑顔で「じゅんびいい?」と問いかける。主人公はクルーAへ顔を向けてうなずき、それからカメラのレンズを直接見つめ、笑顔に緊張と高揚をにじませて「よし、いってきます!」と自然な日本語で勢いよく言う。両者の見えている口元はそれぞれのセリフ全文と一致する。クルーBは笑顔で拳を軽く上げ、クルーCは嬉しそうに笑い、身を少し乗り出してうなずく。主人公は発話後に大きく息を吸い、眼下の円形トランポリンへ目を向け、外側ステップを蹴って飛び出す。 3.0〜6.2秒:主人公がヘリコプターから飛び出した瞬間、クルーBが主人公に向けて「いけー!」と声を上げる。カメラはドア付近の撮影位置を保ったまま素早く下へ向きを変え、海へ落下する主人公を画面の中央付近で追う。主人公は身体を伸ばして落下し、距離が開くにつれて画面内で小さくなる。下方の円形トランポリンが見え、主人公はその中央を目指して近づいていく。カメラを下に向けた間、クルー3名は画面外となるが、機内側から楽しげな歓声が聞こえる。ローター音と風の音が続く。 6.2〜7.7秒:主人公が海面上の円形トランポリンの中央付近へ着地する。張った跳躍面が主人公の重みで大きく沈み、下降が止まった直後に強く押し返す。カメラは下を向いたまま、主人公の着地、跳躍面の沈み込み、反発による上昇への反転を一続きで見せる。着地の衝撃音と反発音が聞こえる。 7.7〜11.6秒:主人公は反発の勢いでヘリコプターへ向かって高く跳ね上がり、膝を胸へ引き寄せて小さく身体を折りたたみ、空中で勢いよく回転する。カメラは同じ撮影位置で上昇する主人公に合わせて上へ向き直る。主人公は近づくほど画面内で大きくなり、開いたドアと外側ステップが再び見えてくる。主人公はヘリコプターの外側ステップの直前まで勢いよく一気に上昇し、ステップに届く瞬間に素早く回転をほどいて脚を伸ばし、飛び出したのと同じ外側ステップへ間髪入れず両足を接地する。着地の瞬間に膝で衝撃を吸収して直ちに上体を起こし、冒頭と同じ位置・身体の向き・立ち姿に戻る。ヘリへの接近、回転解除、両足の接地、姿勢の復帰を通常速度で途切れなく続く、鋭く素早い一連の動きとして撮影する。ヘリコプターの海面からの高度は変わらない。 11.6〜15.0秒:カメラは冒頭と同じドア周辺の向きへ戻る。クルーAは主人公に向かって笑顔で拍手する。クルーBは喜んで拳を突き上げ、「やったーー!」と短く声を上げる。クルーCは主人公を見て笑顔で手をたたく。主人公は弾む呼吸のままカメラのレンズへ顔を向け、目を輝かせ、無事に戻れた興奮と喜びを込めて「やばい!やばい、さいこうだったよ!」と明るく自然な声で話す。クルーBと主人公の発話時には口元の動きをそれぞれのセリフ全文と一致させる。最後のフレームは、主人公が最初と同じ外側ステップに立ち、カメラへ話し終えて興奮した表情を見せ、背後でクルーAとクルーCが拍手し、クルーBが喜んでいる状態。クルーの拍手とローター音が続いたまま15秒で終わる。
A comprehensive video prompt for generating a comedy short film about marshmallow launcher pranks in a snowy Colorado landscape.
Create an ultra-realistic 16 9 cinematic comedy short film featuring a young American woman in her 20s attempting an absurd long-distance snack trick from a snowy Colorado mountain summit to a cozy ski resort lodge below. Use the uploaded reference image for her exact facial identity and appearance. She wears a fitted black winter outfit, gloves, and snow boots on the mountain, then a cream oversized sweater, dark jeans, and winter boots inside the lodge. Her friends wear colorful ski jackets: a red-haired woman in pink, a man in an orange jacket, a bearded man in blue, and a friend wearing a silver puffer jacket and ski goggles. On the mountain summit, she lies dramatically on a snow-covered viewing platform, carefully lining up a harmless oversized novelty marshmallow launcher toward a giant open snack bucket displayed on the lodge patio below. She bites her lip nervously, squints through a ridiculous toy-style scope, then gives the camera a confident grin. Extreme close-up of her eye, followed by a fast whip-pan and a bright white flash. Cut to a wide view of a beautiful Colorado ski resort, with pine forests, snowy slopes, chairlifts, and a rustic wooden lodge. A giant marshmallow sails through the crisp mountain air in exaggerated slow motion, wobbling comically toward the lodge. Inside the lodge café, her friends sit at a wooden counter with steaming mugs of hot chocolate. The marshmallow bounces harmlessly off the exterior window and lands directly in an open bowl of mini marshmallows on the patio table. Everyone looks outside in confusion. Cut to the woman inside the café, now wearing her cream sweater, sitting innocently with her mouth open as if she expected the snack to arrive. Her friends exchange suspicious looks. Suddenly, the lodge's outdoor snack bucket tips slightly in the wind, and a single marshmallow rolls through the open patio doorway and lands perfectly in her mouth. She closes her eyes and chews triumphantly while everyone stares at her in disbelief. She opens one eye, smiles, and says, “Delivery confirmed.” Her friends burst into laughter, and the final shot holds on their stunned expressions. CAMERA & STYLE: Photorealistic live action, crisp 4K detail, polished American outdoor-advert aesthetic, dramatic snowy mountain wides, extreme close-ups, smooth push-ins, whip-pan transitions, slow-motion snack shots, realistic snow and fabric textures, warm lodge lighting contrasting with the icy blue outdoors, and consistent character identities. AUDIO: Mountain wind, distant chairlift sounds, lodge chatter, a playful launcher pop, marshmallow bounce, hot chocolate cups clinking, surprised gasps, and spontaneous laughter. Light playful background music, with natural dialogue and accurate lip synchronization. NEGATIVE: Subtitles, on-screen text, watermarks, blood, real weapons, injury, dangerous stunts, beauty filters, CGI appearance, animation
A prompt for creating high-quality Pixar-style 3D animated family comedies, featuring a toddler playing with toy cars and interacting with parents in a cozy bedroom.
Create a high-quality 3D animated family comedy in a warm Pixar-style aesthetic. A cute toddler with big expressive eyes plays with toy cars on a soft bedroom rug while his loving parents watch him affectionately. The baby makes funny facial expressions, becomes stubborn, laughs joyfully, and interacts playfully with his mother and father. Transition to a cozy bedroom scene where the mother rests with the baby, followed by sweet, humorous moments between them. Show the toddler learning to walk, giggling, making adorable expressions, and sharing heartwarming moments with his family. Use cinematic lighting, soft golden sunlight, realistic skin and fabric textures, detailed character animation, expressive eyes, natural body movements, smooth camera transitions, close-up facial shots, medium shots, and wide family shots. Maintain consistent character designs throughout the video. The atmosphere should be wholesome, emotional, funny, and family-friendly, with professional 3D animation, vibrant yet soft colors, and a polished animated-movie look. Vertical 9:16 format, high resolution, smooth motion, and no visual glitches.
A detailed prompt for generating a realistic video of a combine harvester in an autumn rice field, focusing on camera and character movement.
秋の田んぼ。稲をかき分けて迫るコンバインの刈り取り部。運転席で目を丸くするさくら。畦のおじさんが腕を前へまっすぐ伸ばす。レバーを倒すと機体が奥へよれ、今度は手前へよれすぎる。土手の上から見下ろすと、刈り跡はへびのようにくねっていた。おじさんがしゃがんで、稲の株の列を手のひらでなぞる。 あごひもを結び直して、レバーを握る。分草板の先を株の列に合わせると、機体はぶれずにまっすぐ進む。最後の稲が吸い込まれる。見下ろすと、くねった跡の隣に一本だけまっすぐな刈り跡。軽トラの荷台に、金色のもみが滝のように流れ落ちる。運転席で両手を上げる。 一緒に来た友達が手持ちで撮った記録。1カット2秒前後、セリフ・字幕・BGMなし。画面に出るのは二人だけ。コンバイン1台、軽トラ1台。機体・服・帽子は文字なしの無地。コンバインの形と田んぼの配置は、人のいない参照画像で固定。
A prompt for generating a nostalgic 2003-style video of a Korean couple on a shopping date, emphasizing authentic camcorder aesthetics and natural interaction.
Create a 30-second ultra-realistic Korean home video from around 2003, casually recorded by a friend using an early-2000s consumer MiniDV camcorder. Capture a sweet, spontaneous shopping date between a young Korean couple at a busy Seoul shopping mall, with imperfect framing, natural interactions, and authentic everyday behavior. CHARACTERS: A young adult Korean woman in her early 20s with long black hair, natural makeup, a fitted pastel-pink top, low-rise blue jeans, sneakers, and a small shoulder bag. Her boyfriend is a young Korean man in his early 20s with slightly messy black hair, a casual striped T-shirt, loose jeans, and sneakers. Maintain consistent faces, outfits, and body proportions throughout. STORY: The couple walks through a crowded Seoul shopping mall, playfully teasing each other while the camera follows them from behind. They enter a toy store filled with colorful plushies. The woman excitedly picks up an oversized teddy bear, hugs it tightly, and makes her boyfriend take a picture of her with it. He jokingly holds a tiny stuffed rabbit beside his face, making her burst into laughter. They choose two soft toys, carry them around the mall, and stop in front of a large store mirror. She pulls him close and takes a mirror selfie with a small early-2000s digital camera while he makes a funny expression behind her. Later, they buy one large ice cream and share it while walking through the mall. She takes a bite, offers it to him, and laughs when he accidentally gets a little ice cream on his lip. She gently wipes it away with her finger as he smiles shyly. They enter an old-fashioned mall elevator, still holding their plush toys. Just before the doors close, he leans toward her and gives her a brief, sweet kiss. She smiles bashfully at the camera as the elevator doors slide shut, ending the recording abruptly. CAMERA & FOOTAGE: Authentic 4-3 early-2000s MiniDV footage, handheld shake, imperfect framing, accidental zooms, autofocus hunting, exposure fluctuations, faded colors, soft digital detail, mild compression noise, and natural motion blur. The camera behaves like a friend casually filming them, sometimes reacting late or briefly losing the action. No polished cinematography or smooth tracking shots. SETTING: A busy Seoul shopping mall circa 2003, crowded walkways, toy stores, colorful plushies, clothing shops, large mirrors, old-fashioned elevators, fluorescent lighting, and shoppers passing naturally in the background. AUDIO: Natural mall chatter, footsteps, shop announcements, toy-store ambience, camera shutter clicks, plastic shopping bags, laughter, ice cream wrapper sounds, elevator chimes, and quiet spontaneous conversation. Casual natural English dialogue. No background music or narration. REALISM & CONTINUITY: Natural couple chemistry, shy smiles, playful teasing, believable hand movements, realistic ice cream interaction, consistent character id
A detailed prompt for a 30-second cinematic military action video featuring a special ops team in a dark comic aesthetic with neon green and magenta tones.
一、基础视频规格 视频时长:30秒,一次性连续生成。 画面比例:16:9横屏,1920×1080,60fps。 视频类型:原创军事题材风格化战斗动画,电影级高强度动作短片。 视觉风格:暗黑军事漫画、表现主义战争动画、粗粝丝网印刷、动态泼墨、半色调颗粒、2.5D分层运动、真实军事动作与抽象视觉特效融合。 画面质量:高细节、高反差、强烈轮廓、电影级光影、清晰可辨的战斗动作。 全片无字幕、无解说、无水印、无品牌LOGO。 二、参考图像与视觉一致性 严格以两张上传图片为人物与美术风格参考。 人物设计: 一支由5名成年海军陆战特种作战人员组成的战斗小队,主角采用第一张参考图中的士兵形象。 全员穿着暗橄榄绿色战斗服、厚重战术防弹背心、模块化弹匣包、作战手套、战术护膝、重型作战靴,佩戴带有夜视仪安装座的战术头盔、护目镜及战术通信设备。 携带与参考图外形一致的现代模块化卡宾枪,保留枪身导轨、瞄准装置、弹匣、枪托和附件轮廓。 所有装备统一为深黑色与暗橄榄绿色,细节采用荧光绿色边缘反射及洋红色轮廓高光。 士兵脸部大部分隐藏于头盔阴影之中,面部表情冷峻,充满压迫感。 人物整体比例真实,人体结构准确,装备具有厚重的材质感和重量感。 绝对统一的色彩体系: 主背景:高饱和薄荷绿、灰绿色、荧光青绿。 人物主体:极深墨绿色、橄榄黑、纯黑色。 视觉强调:高饱和荧光洋红、玫红、少量亮粉色。 亮部:低饱和青绿色反射。 暗部:接近纯黑,但保留装备的结构层次。 禁止出现大面积橙黄色爆炸火光、蓝紫色霓虹灯、暖色夕阳或传统灰棕色战场调色。 动画材质: 人物具有厚重的手绘军事概念插画质感。 战斗服、头盔、武器及防弹背心使用粗糙笔触、黑色墨块、错位丝网印刷纹理和高反差阴影表现。 人物轮廓具有荧光洋红色油墨溢出效果。 每一次快速移动都伴随细小洋红色粒子、粉尘、墨迹拉丝和破碎的涂料颗粒。 背景由大面积绿色色块、黑色负空间、抽象烟尘和废墟剪影构成。 整体视觉保持参考图那种强烈、粗粝、危险、充满实验性的平面艺术特征。 三、30秒电影级逐秒分镜 镜头01|00:00至00:04|幽灵现身 纯黑画面中出现少量荧光绿色颗粒。 一条猛烈的洋红色墨迹从屏幕右侧划过,瞬间揭示一名全副武装的海军陆战特种士兵。 士兵正面面对镜头,缓慢向前迈步,低位持枪。 摄影机使用28mm广角镜头,从腰部高度轻微仰拍,快速推进至中近景。 头盔覆盖眉眼,面部隐藏在黑色阴影中。 人物右侧出现极其强烈的洋红色轮廓光。 背景为参考图同款纯净薄荷绿色,散布少量复古印刷颗粒。 随着脚步落地,黑色油墨从脚下向四周迅速扩散。 开场必须拥有强烈的视觉冲击力。 镜头02|00:04至00:08|黑色小队 快速切换至极低机位侧面跟拍。 五名海军陆战特种士兵沿着荒废工业区的残破通道快速推进。 主角位于队伍前方,其余队员以前后错落的队形穿过画面。 动作参考第二张图片中的士兵奔跑姿势。 士兵背包、通信天线、战术背心和武器随跑动自然震动。 摄影机与士兵保持平行高速跟踪,制造强烈的运动视差。 地面为粗糙破碎的黑色混凝土。 人物每一次脚步落地,地面都溅起洋红色粉尘及黑色液态墨迹。 背景从明亮薄荷绿逐渐过渡到大面积纯黑色负空间。 画面具有明显的漫画式速度感。 镜头03|00:08至00:13|火力遭遇 小队突然遭遇来自前方废墟的攻击。 黑色建筑剪影边缘出现急促的洋红色枪口闪光。 大量抽象粉色弹道光迹迅速划过画面。 主角快速侧身躲避,队员分散至废弃混凝土掩体后方。 一连串打击落在墙壁和地面,激起黑色碎片、绿色粉尘与荧光洋红色油墨飞溅。 摄影机快速摇移、甩镜、跟随主角下蹲。 插入一个0.5秒的极近特写,展示主角头盔边缘、护目镜、战术背心纹理和枪身细节。 枪口闪光以洋红色和白绿色抽象色块表现。 冲击瞬间采用短暂的漫画式抽帧与局部画面定格。 不出现血腥伤口。 镜头04|00:13至00:19|交叉掩护 进入全片最激烈的动作高潮。 主角从掩体后快速跃出,其他四名队员在不同纵深位置进行交叉掩护。 侧面广角镜头展现五名士兵在废弃工业走廊内高速穿插。 摄影机进行一次连续的低位横向移动,随后快速绕至主角肩后。 主角抬枪指向前方黑色敌方剪影,短促的洋红色枪口闪光照亮头盔和手套。 地面不断出现巨大的黑色泼墨爆裂。 大量洋红色粒子从背景向前景冲出。 废弃钢架和残破墙面以黑色几何图块的形式快速掠过。 所有士兵保持真实的人体重心和装备惯性。 运动节奏紧张、干净、猛烈。 使用少量0.2秒到0.4秒的速度渐变,强调冲刺、急停、转身和掩护动作。 镜头05|00:19至00:25|极限突破 摄影机紧贴主角右后方,以35mm镜头进行高速肩后追踪。 主角带领小队冲出狭窄通道,进入开阔的废弃集装箱区域。 背景突然转变为明亮的薄荷绿色大面积色块。 前景出现倾斜的黑色钢架剪影和巨大的残破集装箱。 士兵们快速穿过黑暗区域。 一道巨大的洋红色冲击波以抽象泼墨形态从画面深处猛烈扩散。 黑色碎片与粉红色油墨颗粒在慢动作中短暂悬浮。 主角跃过低矮障碍物,落地后继续向前推进。 镜头以高速推轨追逐主角,周围景物被拉伸成黑绿交错的抽象速度线条。 最后通过一次猛烈的黑色墨迹擦镜转场进入结尾。 镜头06|00:25至00:30|幽灵定格 运动突然减速。 主角从黑色烟尘中缓缓走出,正面面对摄影机。 其他四名海军陆战特种士兵逐渐显现在身后不同深度的空间中。 主角保持低位持枪姿态,缓慢迈出最后一步。 摄影机从膝盖高度缓缓仰拍,使用35mm电影镜头。 画面恢复第一张参考图的经典视觉构图。 中央人物为深黑与暗橄榄绿色。 背景为纯净、高饱和的薄荷绿。 人物肩膀、头盔及枪身右侧出现鲜明的洋红色轮廓光。 衣物和装备上呈现粗糙的油墨纹理。 少量黑色墨迹和荧光粉色碎片在人物周围缓慢漂浮。 最后1秒所有人物进入具有力量感的静态英雄构图。 画面突然切至纯黑,留下极短的低频声音尾响。 四、电影摄影与动画运动 全片以28mm和35mm电影广角镜头为主。 大量低机位仰拍、侧面高速跟拍、肩后追踪、极近距离装备特写、快速甩镜与猛烈推进。 镜头运动必须连贯、有重量感,保持主体明确可辨。 使用2.5D分层动画构建人物与背景纵深。 人物采用真实动作捕捉感的运动规律,同时呈现24fps手绘动画式动作节奏。 输出60fps,允许在关键冲击画面使用风格化抽帧和瞬间定格。 墨迹飞溅、粉尘扩散和粒子拖尾具有独立的动画层。 视觉特效与士兵身体动作精准同步。 保留强烈的印刷纹理、半色调网点、细小划痕、油墨偏移和动态胶片颗粒。 五、声音设计 音乐风格:黑暗工业电子、军事节奏、电影级低频冲击。 节奏:140 BPM左右。 00:00至00:04:沉重低频脉冲、军靴脚步、金属装备碰撞。 00:04至00:08:工业鼓点逐渐加速,加入
A prompt for creating cinematic K-drama visuals focusing on a tense silent interaction between two women in an apartment hallway.
Create a 28-second ultra-realistic cinematic Korean drama featuring two beautiful young Korean women in a quiet, dimly lit apartment hallway. One woman has long, straight black hair and wears an elegant black outfit, while the other wears a sophisticated white pantsuit with a simple handbag. Begin with a mysterious rear shot of the woman in black walking slowly through a dark corridor illuminated by soft blue light. Transition into a beautifully detailed apartment hallway where both women walk past each other, creating a subtle atmosphere of tension and curiosity. Use cinematic close-ups, natural walking movements, realistic facial expressions, and smooth camera tracking to build suspense. Show a close-up of a wooden doorway and a subtle movement near the floor, followed by an empty hallway and a metallic elevator interior. Capture the woman in black standing quietly inside the elevator with a serious expression. End with the woman in white walking alone through the hallway, followed by a breathtaking close-up of her face outdoors, with her black hair moving gently in the breeze. Use soft natural lighting, muted colors, shallow depth of field, realistic skin texture, cinematic composition, smooth transitions, and professional Korean drama cinematography. Maintain consistent facial identities, hairstyles, outfits, and body proportions throughout the video.
A detailed prompt for generating a photorealistic cinematic martial arts battle scene by Seedance 2.5, set in a white stone quarry featuring a female warrior fighting multiple male warriors.
Create an ultra-realistic, cinematic martial arts battle scene set inside a massive abandoned white stone quarry. A fierce young Asian woman with long black hair tied in a high ponytail, wearing a fitted black martial arts outfit, fights against a large group of muscular male warriors dressed in dark red traditional combat uniforms. Scene 1 (0–3 seconds): Wide-angle establishing shot of multiple warriors surrounding the female fighter across a vast white stone arena. The fighters rush toward her as dust rises from the cracked ground. The camera moves low and dynamically through the battlefield. Scene 2 (3–6 seconds): Medium close-up of the female warrior blocking a powerful punch, twisting her body, and delivering a fast counterattack. Her long black ponytail swings naturally through the air. Use realistic martial arts choreography, precise body movements, and intense facial expressions. Scene 3 (6–9 seconds): The female fighter performs a spectacular spinning kick, sending several opponents flying backward. One warrior leaps into the air while another crashes onto the dusty stone floor. Add dramatic slow motion, realistic physics, flying dust, and subtle debris. Scene 4 (9–12 seconds): Fast tracking shots follow the woman as she dodges punches, sweeps an opponent's legs, and executes a powerful spinning attack. Show dynamic camera movement, low-angle shots, and brief close-ups of her determined expression. Scene 5 (12–15 seconds): A high aerial crane shot reveals the entire white quarry battlefield, with numerous defeated warriors scattered across the cracked ground and the female fighter standing confidently in the center. Finish with a dramatic cinematic pullback. Visual Style: Photorealistic live-action martial arts movie, Hollywood action cinematography, epic wuxia-inspired combat, natural overcast daylight, desaturated colors, realistic skin textures, detailed black and dark-red costumes, volumetric dust, subtle motion blur, high dynamic range, 4K detail, professional anamorphic lens look. Camera: Dynamic tracking shots, low-angle perspectives, wide establishing shots, fast pans, smooth crane movements, controlled slow motion during major impacts. Sound Design: Heavy footsteps, realistic punch and kick impacts, rushing wind, stone debris, clothing movement, deep cinematic impacts, subtle epic orchestral background music. Consistency Requirements: Maintain the same female character, facial features, hairstyle, black outfit, male opponents, red uniforms, and white quarry environment throughout the entire video. Natural anatomy, realistic balance, continuous motion, coherent fight choreography, no duplicated limbs, no distorted faces, no sudden scene changes, no text, no subtitles, no logos. Duration: 15 seconds. Aspect ratio: 16:9. Quality: Ultra-realistic cinematic 4K.
A prompt for testing Seedance 2.5 and Kling 4.0 using simple gray-box previz animations, covering office interactions, cycling, and cavalry battles.
共通「動画1は灰色の人形と箱の仮組み。取るのは人の立ち位置・道筋・タイミング・カメラの動きだけ。人形は画像の人物に、箱は本物の場所に置き換える」 01「扉から入った人が、座る人の後ろで止まる→座っていた人が椅子から横にずれて立ち、向き合う。正面の引き→左へ回り込み→2人へ寄る」 02「後ろから自転車が近づき5秒で並び6秒で追い越す→減速して止まり、上半身ごと振り返る。道の同じ側から並走」 03「キャラは最後まで2Dのセル画のまま実写の交差点を渡る。歩行者はちょうど6人。真上の俯瞰→目の高さ」 04「両軍が向かい合って駆け寄り4.6秒ですれ違う→大将が馬を返して刀を掲げる。刀と槍は体に触れず空を切る」
A video generation prompt focusing on limited animation, handheld camera effects, and an EDM party atmosphere.
unchanging style, perfect visual consistency, edm, limited animation, handheld camera, feeling the music, techhouse, party, bouncing, shaky camera, dynamic, windy, foreshortening, hand dancing, non verbal, camera rocks side to side, unstable camera,...
A prompt for creating a heartwarming 3D animated short depicting a story of a grandmother, baby, and parents in a suburban setting, mimicking Pixar's animation style.
Create a high-quality, cinematic 3D animated short film in a beautiful suburban neighborhood on a sunny day. An elderly grandmother wearing a brown floral hijab, a modest brown dress, and a cozy beige cardigan lovingly holds her adorable baby grandchild, who wears a light green dinosaur-print onesie. A young couple, the baby's parents, stand beside a silver SUV, smiling and waving goodbye before leaving. Show the parents getting into their vehicle while the grandmother gently holds the baby and watches them leave. Use warm sunlight, a peaceful residential street, beautiful houses, green trees, and soft golden-hour lighting to create an emotional, heartwarming atmosphere. Next, show the grandmother standing outside her front door with the baby in her arms. The baby looks curious and cheerful. Transition to a funny scene where the baby, now standing independently in the doorway, smiles brightly while holding a cookie. The grandmother stands behind the baby, looking surprised and amused. Finally, cut to a close-up of the young parents reacting with exaggerated shock and wide eyes, creating a humorous ending. Visual Style: High-quality 3D animation, Pixar-inspired family-friendly character design, expressive faces, large sparkling eyes, realistic fabric textures, detailed suburban environments, cinematic camera movements, smooth transitions, warm natural lighting, vibrant colors, and professional animation. Character Consistency: Keep the same facial features, clothing, hairstyles, and body proportions for every character throughout the entire video. Format: Vertical 9:16, 4K quality, smooth animation, cinematic storytelling, emotional expressions, family-friendly comedy, and no unwanted text or watermarks.
A prompt for generating a 1990s-style talk show scene where the host examines a guest's Elon Musk-shaped pendant.
Create a 17-second photorealistic late-night talk-show scene in horizontal 16:9 format using the four attached references. All spoken dialogue must be in English. REFERENCE PRIORITY Image 1: opening composition, studio, furniture placement and character positions. Adapt the framing to 16:9 by extending the studio naturally on both sides, without stretching the image or cropping either character. Image 2: male host’s face, hairstyle and clothing. Image 3: adult female guest’s face, copper-red hairstyle, glasses, red dress and necklace. Image 4: exact design of the small gold pendant shaped like Elon Musk’s head. Preserve both characters’ identities throughout. Elon Musk appears ONLY as the miniature pendant, never as a person in the studio. FORMAT AND VISUAL STYLE Horizontal 16:9 throughout. No vertical framing, black bars or aspect-ratio changes. An old television broadcast from the late 1990s: slightly soft standard-definition detail, subtle analog video noise, restrained colors, warm studio lighting and a blue nighttime city backdrop. Preserve this archival television texture within the 16:9 frame. Avoid modern cinematic grading, glossy commercial lighting and excessive VHS distortion. Faces and pendant must remain readable. TIMELINE — EXACTLY 17 SECONDS 0.0–3.0 seconds: Begin with the scene from Image 1, reframed naturally into 16:9. A steady two-shot shows the host and guest seated, looking at each other. The host notices her necklace, shifts his eyes toward the pendant and asks with natural curiosity: HOST: “What’s that you’re wearing?” The guest follows his gaze and gives a small welcoming smile. 3.0–6.0 seconds: Keep the action visible in the two-shot. The host rises from his chair and moves toward the guest around the nearest open end of the desk. He bends at the waist to inspect the pendant closely. The guest remains seated and slightly raises her chin. He carefully slides his left hand, palm facing upward, beneath the pendant and lifts it only a few centimeters. The pendant settles onto his open palm, still attached to the necklace around her neck. The chain remains loose. His attention stays on the jewelry. 6.0–8.5 seconds: Cut to a tight jewelry detail shot, maintaining 16:9. The tiny sculpted gold head rests face-up on the host’s open left palm. Its shape and facial details match Image 4 exactly. The attached chain remains visible, leading back toward the guest’s neck. The host’s curious face is visible close behind and to the side of his palm. GUEST: “It’s a pendant shaped like Elon Musk’s head.” 8.5–10.5 seconds: Cut to a medium two-shot from the same side of the scene. The host remains standing and leaning beside the seated guest, still supporting the pendant in the SAME left palm. He looks up from the pendant toward her face. HOST: “Why do you wear it?” 10.5–15.0 seconds: Hold the medium two-shot. The guest looks at him with a sincere, dreamy smile and says: GUEST: “Because he’s my idol, and I want to fly in a Tesla!” She means “fly” literally; do not replace it with “drive” or “ride.” The host listens without interrupting. He keeps his hand still, supporting the pendant without pulling the chain. 15.0–17.0 seconds: A brief comic pause. The host slowly turns his head toward the camera with raised eyebrows and a puzzled, deadpan expression. The guest smiles proudly. A short burst of studio audience laughter begins after her final word. End on their reactions, with the pendant still resting on his palm. AUDIO English dialogue only. Speak exactly the four specified lines, in the exact order, with natural conversational delivery and accurate English lip synchronization. One speaker at a time. Conversational male host voice; warm, sincere female guest voice. Subtle studio room tone, quiet chair and clothing movement. Audience laughter only during the final two seconds. No music, narration, translations or additional dialogue. CONTINUITY AND PHYSICS Exactly two people. Keep faces, hairstyles, glasses, clothes and studio layout unchanged. The host begins seated, stands once, and remains standing afterward. The guest remains seated throughout. Show the host’s movement before cutting to the pendant; no teleportation or movement through the desk. Keep both heads within the frame during the standing and leaning movement. The pendant stays attached to the same chain around her neck throughout. One pendant only, constant miniature size, rigid gold material, no morphing or animation. The host supports it with his LEFT palm in every shot. Five natural fingers; no fused hands or fingers. Keep the chain slack and the pendant close to its original hanging position. Preserve screen direction and matching hand positions across cuts. The focus is the jewelry, dialogue and facial reactions. No subtitles, captions, logos or watermarks.
A prompt for a 30-second video showing a Korean girl enjoying cherry blossoms and coffee in a dramatic setting.
Create a video of beautiful Korean girls in 16:9 landscape format, exactly 30 seconds long, featuring a beautiful young Korean woman with black hair tied in a ponytail with a burgundy ribbon, wearing an olive-green top, a camouflage jacket, matching trousers, and carrying a black shoulder bag. She walks along a modern Korean roadside on a beautiful spring day surrounded by blooming cherry blossom trees. Three stylish Korean female friends wait beside a pale-yellow SUV, smiling and waving at her. She approaches them, and they greet each other warmly while holding takeaway coffee cups. One friend opens the SUV's rear hatch as the girls gather around, laughing and chatting naturally. Another friend takes a selfie with the group using her smartphone. The main woman playfully poses while holding her coffee, surrounded by beautiful pink cherry blossoms. Finally, show a cinematic close-up of her holding the coffee cup, wearing a small white cherry blossom hair clip, gently sipping her drink and smiling peacefully in the sunlight. Use photorealistic Korean drama cinematography, natural skin texture, realistic facial expressions, soft daylight, smooth camera movements, detailed clothing, and consistent faces and outfits throughout the video. Add gentle birdsong, a spring breeze, friendly laughter, and soft background music. Maintain smooth transitions, natural body movements, and high-quality cinematic visuals throughout the entire 30-second video.
A detailed multi-shot video prompt simulating 1990s camcorder footage of a lake monster attacking a family, featuring VHS artifacts and specific camera movements.
1990s camcorder found footage, VHS tape artifacts, tracking lines, chromatic aberration, auto-focus hunting, overexposed highlights, shaky amateur handheld camera, a date timestamp burned into the corner. Color grade: washed out gray-green, faded skin tones, blown out pale sky. A colossal dark lake creature with a ridged back and a long neck, mostly seen in silhouette through fog. A family on a wooden dock with a canoe at a northern lake. SHOT 1 (0 to 3s): Wide handheld shot matching the reference image, the camera zooms in clumsily as the creature's back breaks the surface far out on the lake, water rolling off in sheets. SHOT 2 (3 to 6s): Handheld medium shot, the camera whips toward the family on the dock, a child pointing, a father pulling the canoe out of the water, the image glitching with tracking lines. SHOT 3 (6 to 9s): Wide shot, auto-focus hunting, the long neck rises out of the fog, impossibly tall, a wave rolling toward the dock. SHOT 4 (9 to 12s): Running handheld shot, the camera bouncing as the operator flees up the wooden stairs, the lake visible in flashes behind, the wave slamming into the dock. SHOT 5 (12 to 15s): Static shot, the camera dropped on the grass and tilted sideways, the creature's silhouette slowly sinking back into the fog, then the tape glitches to blue screen. Loons calling, tape hiss, wind on the microphone, a deep foghorn-like call from the lake, a child shouting, panicked breathing, wood cracking, a final VHS tracking screech into silence.
High-detail script and visual prompts tailored for Seedance 2.5 to generate photorealistic cinematic videos of family bedtime interactions.
Create a photorealistic, cinematic 30-second vertical featuring a loving family. Daddy is a 30-year-old man wearing a soft navy-blue T-shirt. Mommy is a 28-year-old woman wearing a cream-colored nightdress. Their adorable 4-year-old daughter wears pink pajamas with tiny stars. Keep faces, hairstyles, clothing, voices, and proportions consistent throughout. Setting: a cozy child’s bedroom at night with a warm bedside lamp and soft blankets. SCENE 1 — ONE LAST THING (0–7 seconds) Camera: Medium shot of daddy tucking his daughter into bed and pulling up her blanket. Daddy: “Goodnight, princess. Time to sleep!” The daughter gently grabs his hand. Daughter: “Wait, daddy! I have one more wish!” Daddy smiles and sits beside her. SCENE 2 — THE UNEXPECTED WISH (7–14 seconds) Camera: Close-up of the daughter looking into daddy’s eyes. Daddy: “What do you wish for, sweetheart? A new toy?” She shakes her head. Daughter: “No. I wish you could stay little with me forever!” Daddy’s smile softens as he listens. SCENE 3 — MOMMY HEARS HER (14–22 seconds) Mommy appears in the doorway and hears the conversation. Mommy: “Why do you want daddy to stay little?” The daughter looks at mommy. Daughter: “Because then nobody would tell me to go to bed!” Mommy laughs. Daddy raises his eyebrows in playful surprise. SCENE 4 — THE SWEET ENDING (22–30 seconds) Daddy gently kisses his daughter’s forehead. Daddy: “I may grow older, but I’ll always be your daddy.” The daughter hugs him around the neck. Daughter: “Good. Then you can carry me even when I’m big!” Mommy sits beside them and joins the hug. End with the daughter smiling sleepily between her parents. STYLE: Photorealistic faces and natural skin textures, realistic child expressions, accurate lip-sync, soft warm lighting, gentle bedtime piano music, close-ups during emotional dialogue, and smooth cinematic camera movements. Natural conversational pacing. Keep the bedroom and characters consistent.
A prompt for a 15-second video of a cosplayer at Comiket changing poses every 3 seconds with character consistency.
🎬 Gemini監督への指示 アップロードした3枚の画像(全身、アップ、ミディアム)を上手く使って5シーンの切り替えて「コミケ会場でコスプレイヤーが撮影会に決めポーズや映えるポージングを3秒毎(1シーン内で2ポージング)に変える」、15秒の動画の日本語プロンプトを作ってください。最後はクローズアップして決めポーズで片目でウインクさせてください。 BGM「(画像の雰囲気に合ったBGMを考えて)」 人物の顔や髪型は一貫性を維持させてください。画像生成や動画生成は不要。 プロンプト中の画像1(全身)は@Image1、画像2(アップ)は@Image2、画像3(ミディアム)は@Image3 に置き換えてください。 🎥 動画生成プロンプト(Grok Imagine) 【基本設定・一貫性】 * 人物・衣装の参照: @Image1(全身)、@Image2(アップ)、@Image3(ミディアム)の女性コスプレイヤー。銀髪の三つ編みヘアー、エルフ耳、青と白のファンタジー衣装、顔立ちおよび髪型の一貫性を維持。 * ロケーション: 賑わうコミケ会場(背景にカメラマンや展示ブースのボケ感)。 * BGM提案: 爽やかで明るいアップテンポなEDM/アニソン風ポップス(ファンタジー感のあるキラキラしたシンセ音入り)。 【タイムライン・構成(全15秒)】 * シーン1(0:00 - 0:03)/ 参照: @Image1 * 画角: フルショット(全身) * ポーズ1(0〜1.5秒): 衣装のスカートを軽く持ち上げて優雅におじぎをするようなポーズ。 * ポーズ2(1.5〜3秒): 両手を広げて華やかに微笑み、全身の衣装を見せるポーズ。 * シーン2(0:03 - 0:06)/ 参照: @Image3 * 画角: ミディアムショット(腰上) * ポーズ1(3〜4.5秒): 胸の前で両手でハートマークを作り、首を少し傾けて微笑む。 * ポーズ2(4.5〜6秒): 片手を顎に当て、カメラを目線で追いかけながらクールな表情を作る。 * シーン3(0:06 - 0:09)/ 参照: @Image1 * 画角: アングルを変えたフルショット(斜め全身) * ポーズ1(6〜7.5秒): 身体を斜めにひねり、髪に軽く手を添えて振り返るポーズ。 * ポーズ2(7.5〜9秒): 片足を一歩前に出し、ファンタジー作品のヒロイン風に手をカメラへ差し出すポーズ。 * シーン4(0:09 - 0:12)/ 参照: @Image3 * 画角: ミディアムショット * ポーズ1(9〜10.5秒): 両手を頬に添えて、驚いたような可愛らしいポーズ。 * ポーズ2(10.5〜12秒): カメラに向かって爽やかに手を振るポーズ。 * シーン5(0:12 - 0:15)/ 参照: @Image2 * 画角: クローズアップ(顔〜肩上) * ポーズ1(12〜13.5秒): カメラをじっと見つめ、優しく微笑む。 * ポーズ2(13.5〜15秒): カメラに向かって決めポーズを取り、片目で可愛らしくウインクして笑顔で静止。
A prompt for generating realistic Korean school fight scenes featuring a confident female student confronting bullies with cinematic lighting and choreography.
TITLE: THE KOREAN SCHOOL BADDIE — ONE GIRL AGAINST THE ENTIRE CLASS DURATION: 30 seconds STYLE: Ultra-realistic Korean action drama, cinematic school fight scene, photorealistic characters, realistic physics, dramatic lighting, high-quality choreography, 4K film quality. SETTING AND CHARACTERS: A realistic South Korean high school classroom during the daytime. Large windows cover one side of the classroom, allowing soft natural sunlight to enter through cream-colored curtains. Wooden student desks and chairs are arranged in rows. A green chalkboard, classroom bulletin boards, school supplies, scattered notebooks, and backpacks create an authentic Korean school environment. The main character is a beautiful Korean high school girl with a sleek, straight black bob haircut, pale natural skin, sharp expressive eyes, and a cold, fearless expression. She wears a perfectly realistic Korean school uniform consisting of a dark navy blazer with white piping, a white collared shirt, a dark tie, a dark pleated skirt, black socks, and black school shoes. Several Korean male students wear matching navy school uniforms. They behave like arrogant school bullies, lounging in their chairs and laughing among themselves. IMPORTANT: Maintain the same face, hairstyle, uniform, body proportions, and character identity throughout the entire video. Every movement must be physically believable. No cartoon effects, no supernatural powers, and no exaggerated CGI. SCENE 1 — ESTABLISHING THE CLASSROOM (0–3 SECONDS) Begin with a cinematic wide shot of a sunlit Korean high school classroom. The Korean girl sits quietly at her desk near the large windows. Her notebook lies open, and a pink pen rests on the wooden desk. She looks forward with a calm, serious expression. Behind her, several male students are laughing and behaving disrespectfully. One boy leans back in his chair with an arrogant smile while his friends exchange amused looks. The camera slowly moves toward the girl, establishing the contrast between her quiet composure and the disruptive students behind her. Natural daylight, realistic classroom ambience, subtle curtain movement, authentic Korean school atmosphere. SCENE 2 — THE COLD STARE (3–6 SECONDS) Cut to a close-up of the girl's face. She slowly raises her eyes and looks directly toward the troublesome students. Her expression changes subtly: her eyes become sharper, her lips tighten, and her face remains completely composed. She does not shout or display exaggerated anger. Her silence communicates confidence and intimidation. Cut briefly to a male student smirking at her. He leans forward, believing she will not react. Use shallow depth of field, realistic skin texture, natural eye movements, and cinematic close-up framing. SCENE 3 — THE FIRST EXPLOSIVE MOVE (6–10 SECONDS) Without warning, one of the male students provokes her and moves aggressively into her personal space. The gir
A Seedance 2.5 multi-scene video prompt depicting a teenage girl releasing supernatural powers after being mocked, leading to a car accident.
SCENE 1 (0:00 - 0:02): A cinematic medium shot of a teenage blonde girl with a messy updo, wearing an off-shoulder light grey long-sleeve top with a pink paint stain on the front, walking past a beige vintage Volkswagen Beetle in a high school parking lot during autumn. Three teenagers sit on top of the car, chatting and drinking pink smoothies. Golden hour lighting, fallen autumn leaves on the ground. SCENE 2 (0:02 - 0:04): Close-up shot of the three teenagers laughing and enjoying their drinks. Cut to a wide shot from behind the teenagers showing the blonde girl walking away across the asphalt parking lot towards the brick school building steps, surrounded by autumn trees. SCENE 3 (0:04 - 0:06): A low-angle close-up of a large white semi-truck's front wheel rolling over autumn leaves. Cut to the blonde girl standing on the concrete steps of the school entrance, turning around to look back over her shoulder with a serious expression. SCENE 4 (0:06 - 0:08): Over-the-shoulder shot of the blonde girl looking back at the teenagers on the car in the distance. Cut to a dramatic close-up of the blonde girl's face with intense glowing pink eyes, radiating a supernatural power. SCENE 5 (0:08 - 0:11): A fast-moving red light trail/laser beam streaks across the parking lot past the car. Cut to other students sitting on the school steps as a red light flash sweeps by them, creating a sudden gust of wind. SCENE 6 (0:11 - 0:15): Close-up of a ginger-haired teenage girl screaming with shock and wide eyes. Cut to a pink smoothie cup on the roof of the beige car as a large white semi-truck violently crashes into the side of the car, sending debris flying. SCENE 7 (0:15 - 0:20): Wide shot of the teenagers sitting on the concrete curb in front of the school in absolute shock and panic as smoke rises from the wrecked car in the background. Close-up of the ginger girl looking bewildered and speaking in disbelief. SCENE 8 (0:20 - 0:27): The blonde girl stands calmly, looking back at the smoky car wreck with a subtle expression, then turns around and walks up the school steps. Final close-up shot of her black Converse sneakers walking up the concrete steps.
A detailed prompt for generating a realistic, handheld-style video of a sudden crocodile attack in a safari park, including specific character references and camera movements.
【视频规格与整体风格】 9:16竖屏,15秒,一镜到底。中国野生动物园内,游客用普通手机偶然拍到的惊险现场。起初平静,巨鳄突袭后摄影者晚半拍受惊,镜头抖动、拍偏,再慌忙找回人物。保留业余手机录像的真实感。 【角色与参考锁定】 女主使用预设成年中国女性角色#1@598c0e35-7c27-4b7c-8bac-19bb206ff1c3,穿着预设服装#2@d57ec59a-c0a5-4eae-84f4-4f02ef21a92d,严格保持参考中的面貌、长发、服装款式、颜色和材质。#1是人物,#2是服装,编号不出现在画面中。 管理员是一名四十岁左右的中国男性,短黑发,穿卡其色短袖工作服、深色长裤,胸前佩戴工作证,站在女主侧后方。 后方有少量成年中国游客,穿夏季便服,自然错落站在栈道上。 【场景与机位】 热带鳄鱼观赏区,旧木栈道沿深绿水池延伸,对岸是竹子、棕榈和灌木。粗木栏杆从画面右下方斜向背景延伸。 外部摄影者站在女主侧后方约一米,略高于她手中手机的位置,竖屏向下拍摄。画面右侧是女主、管理员和栈道,左侧及下方是水池。 潮湿白天,自然日光穿过植被,水面反射轻微绿色。普通手机1倍主摄,轻微手抖、歪斜、压缩涂抹、暗部噪点,急转时短暂模糊。无电影调色、无人机机位或人像虚化。 【女主手机朝向——必须准确】 女主是在使用手机后置摄像头拍摄水里的鳄鱼,不是自拍,也不是把手机展示给外部摄影者。 女主双手横握手机两端,手指避开后置摄像头。手机位于她胸口以下、栏杆外沿附近,机身明显向下俯倾,后置镜头的拍摄方向准确落在左下方的水下巨大暗影上。 手机屏幕朝上,并略微朝向女主的脸;手机背面及后置镜头朝下,面对水面。女主低头看屏幕,通过屏幕观察水中的目标。 从外部摄影者略高的侧后方角度,应主要看到手机屏幕和机身侧边,不应正面清楚看到整个手机背部及摄像头模组朝着外部摄影者。 屏幕取景内容与真实水面一致:能看见同一片水面及同一条模糊暗影,不出现女主自拍脸、无关风景或错误方向的画面。 女主轻微调整手腕时,手机后置镜头始终追随水下暗影。直到巨鳄突袭,她才因受惊而收回手机、打乱朝向。 【巨鳄与首帧状态】 只有一只异常庞大的深灰黑色鳄鱼。跃起时,头部接近成人上半身大小,张开的嘴占据画面左侧近半幅,尺寸参照原视频,不生成普通小鳄鱼。 首帧中巨鳄完全潜在水下,只呈现一条巨大、模糊的长形暗影,像沉木。水体反光和浑浊遮住细节,不能清楚辨认头部、眼睛、鼻孔、鳞片或牙齿,没有任何部位露出水面。 巨鳄从始至终保持相同体积,不在跃起瞬间突然膨胀。 【0—2秒:对准水下暗影拍摄】 女主站在栏杆内侧,身体微微前倾,双手横握手机,将手机后置镜头朝下对准左下方水中的巨大暗影。 她低头看朝向自己的屏幕,屏幕内的暗影与水池里的暗影位置对应。外部摄影者从她侧后方拍到手机屏幕、女主侧脸和下方水面。 管理员严肃提醒: “女士,退后一点,别靠栏杆!” 水面只有轻微波纹,鳄鱼仍未露头。 【2—4秒:目标靠近,手机跟随】 水下巨大轮廓缓慢朝女主所在位置接近。女主小幅转动手腕,让手机后置镜头继续对准正在移动的暗影,屏幕朝向自己,不把手机背面转向外部摄影者。 管理员靠近一步,伸手准备扶她,语气变急: “那不是木头,是活的!别探身!” 女主开始收回身体,但手机仍在拍水面。本段末尾,鳄鱼鼻端和眼睛才刚刚破水。 【4—6秒:巨鳄突袭】 巨鳄猛然顶开水面,巨大头部沿栏杆外侧升起,张开的嘴瞬间占据画面左侧近半幅,带起大片水花。 女主尖叫,急忙往回收手机和手臂。手机朝向因惊吓而失去稳定,此前必须一直正确对准鳄鱼。 巨鳄短暂咬住她伸在栏杆外的一侧前臂及附近衣料。管理员立即扶住她的上臂和肩背,稳住她在栏杆内侧的重心。 外部摄影者受惊后退半步,镜头短暂上抖,没有切镜或伤口特写。 【6—9秒:挣脱与落水】 巨鳄短暂松口、头部回落,女主趁机收回受伤前臂,管理员顺势将她拉到栈道内侧。动作必须有明确松口,不能让手臂穿过闭合牙齿。 手机从女主手中滑落,翻转后掉进水池,只掉落一次。 巨鳄砸回水面,激起白色水花。女主捂住前臂,伤处有少量红色液体渗出,形成短而细的流痕,沾到接触位置的衣料,并有一两滴落向木栈道。 不展示深层伤口,不喷射,不大量涌血。 管理员喊: “退后!别伸手!别捡手机!” 【9—11秒:游客慌乱】 女主退到栈道中间,护住受伤前臂,肩膀发紧,呼吸急促。 近处游客先缩肩后退,拉住同伴;后方游客听到尖叫才转身,有人收手机,有人侧身让路,先后惊呼: “快退!” “别靠栏杆!” 人群反应有时间差,不整齐同步,不穿模,不跌入水池。巨鳄留在水下,不再进行第二次攻击。 【11—13秒:管理员疏散】 管理员站在女主与栏杆之间,护着她继续退开,伸手示意游客让路: “往后退!大家别挤!” 一名附近游客递来纸巾,女主接过按住同一处伤处,纸巾接触位置逐渐染上少量红色。 摄影者小幅后退并修正构图,同时拍到女主、管理员和正在慌乱退让的游客。 【13—15秒:惊魂余波】 女主捂住前臂,喘着气、声音发颤地说: “吓死我了……它真咬啊!” 管理员仍护着她,警惕地看向水池。背景游客低声提醒: “离远点,别拍了!” 结束在女主惊魂未定、人群退开、水面余波尚未平息的现场。 【现场声音】 全部为手机现场收音,中文对白,无配乐、旁白和字幕。 开头有游客低声交谈、竹叶摩擦和轻微水声。管理员声音近而清楚,随危险发生从严肃提醒变为急促制止。 巨鳄出水时有沉重水声、女主尖叫、近处吸气声;随后是手机落水、水花砸落、急促喘气和凌乱脚步。最大水花与尖叫使麦克风短暂削波,随后自动增益恢复。 远处游客声音更散、更轻,不添加怪兽咆哮。 【连续性与重点约束】 开头女主手机必须正确拍摄水下目标:屏幕朝上并朝向女主,后置镜头朝下对准鳄鱼所在位置;外部摄影者主要看到屏幕与侧边。禁止手机背部镜头朝外部摄影者,却表现成正在拍摄水下目标。 首帧只有水下模糊巨影,不出现清楚鳄鱼头、眼睛、鳞片或牙齿。红色液体只在咬伤之后出现。 保持人物#1、服装#2、管理员、巨鳄和栈道连续。受伤始终是同一侧前臂,红色流痕、衣料损伤与纸巾按压位置对应。 只有一部手机,落水后不回到手中。女主与管理员始终在栏杆内,鳄鱼@d57ec59a-c0a5-4eae-84f4-4f02ef21a92d始终在池内。全过程保持接触、重量、咬合、松口和后退动作的因果关系。 无伤口特写、无断肢、无深层组织展示。全片一镜到底,无慢动作、回放、摄影棚揭底、界面覆盖或水印。
Detailed prompts for generating realistic cinematic scenes of calligraphy practice, including specific character actions and camera instructions.
板張りの大広間。先生が巨大な筆を全身で振り下ろし、墨が扇形に飛ぶ。目を見開くさくら。受け取った筆の重さで、腕ががくんと下がる。墨を吸った穂先から、しずくがぼたぼた落ちる。筆を頭上に掲げて、紙の上を一歩ずつ。上の真ん中に下ろそうとした瞬間、重さに体を持っていかれて、筆がどすんと落ちる。大きな墨だまり。頬に、墨の点がひとつ。先生はしみをじっと見て、うなずき、「続けて」と手を流す。 頬に墨をつけたまま、口を結んでうなずく。墨だまりのすぐ下に穂先を下ろし、後ろへ下がりながら線を引く。全身で振り下ろして、まっすぐ下へ。手首を返して、はねる。腕を組んだ先生が、小さくうなずく。体をひねって左へ払い、右へ抜く。肩で息をする。二人で紙の端を持って、立ち上げる。あの墨だまりは、『永』の一画目の点になっていた。見上げて、笑う。 友達が手持ちで撮った記録。1カット2秒前後、セリフ・字幕・BGMなし。画面に出るのは二人だけ。筆は1本、紙は1枚、字は『永』一文字だけ。書いている途中は字の全体を映さず、全体が見えるのは紙を立てたときだけ。完成した字は、前半の墨だまりのコマから作った参照画像どおり。
先选择与目标相符的示例,清楚描述主体和预期结果,再补充视频生成所需的构图、镜头、光线、音频、运动或版式细节。生成前请替换产品、品牌、参考素材和对白变量。
打开卡片,对照预览与原始写法,再将原始 Prompt 复制到 Topview AI 或兼容的生成器中。请根据项目调整参考素材、限制条件和风格说明。
请将每条提示词视为经过展示的示例或起点,而不是结果保证。应根据所选模型检查品牌准确性、版权、安全性和生成能力。
可以。媒体主页会链接到独立分类页;每个分类页只保留一种分类条件,使示例范围和搜索意图更加清晰。
当前提示词库包含 Seedance 2.5, Seedance 2.0, Wan 3.0, GPT Image 2, Nano Banana Pro, Gemini Omni, Grok Imagine 的示例,具体可用内容会随媒体类型和筛选条件变化。
当前浏览“Crowd Group”主体分类下的视频提示词。