OpenZhenZhen AI 提示词画廊 · MiniMax H3 · video
首尔夜街音乐定格与追逐
Seoul Night Street Music Freeze and Chase
首尔夜街音乐定格与追逐，使用同名原始提示词配对并提供中英文版本
提示词 / Prompt
中文提示词
subject_definitions: <Subject 1> 是 <Picture 1> 中的年轻女生，作为目标视频唯一女生的五官与身份基准：她的脸型、眉形、眼型、鼻唇轮廓、肤色、深棕色长直发、小巧耳钉与黑色上衣均需严格一致；目标视频中她穿带裙摆的黑色连衣裙，表情从参考图的微笑改为气怒，但绝不换脸。 summary: [reference generation] 目标视频以 <Picture 1> 作为 <Subject 1> 的五官和身份参考，生成一段20秒第一人称浪漫喜剧片段：女生在首尔韩文招牌街道赌气离开，手机按下播放键后世界定格，镜头环绕并捏脸，音乐停止时她恢复动作追向镜头；全程保持同一女生、不换脸、不出现字幕。 retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 4], [Shot 5], [Shot 6], [Shot 7], [Shot 8]): fully_preserved - 参考图的五官、脸型、肤色、深棕长发、耳钉和黑色上衣特征全程保留，表情、姿势与拍摄角度按剧情变化，身份始终唯一。 detailed_description: 影片采用第一人称写实浪漫喜剧风格，手持视角贴近人眼，首尔夜街的灯光在密集韩文招牌间闪烁。全片无字幕、无标题，<Subject 1> 的五官严格锁定 <Picture 1>，全程不换脸。 [Shot 1] 正面平视构图中，<Subject 1> (S1) 双手叉腰站在热闹街道中央，深棕长发垂落，黑色连衣裙裙摆轻垂；她眉头紧皱、腮帮鼓起，直视镜头。周围行人正常穿行，韩文招牌层层后退。<Subject 1> (S1) 用气鼓鼓的清晰声音说，<d>[Chinese] 再也不理你了。</d> [Shot 2] At 00:02.000, 第一人称跟拍视角中，<Subject 1> 猛地转身，长发随转身甩动，随即气冲冲迈开大步向前走。镜头贴在她身后跟随，脚步急促，背影占据画面中心，行人依旧在两侧正常移动。 [Shot 3] At 00:04.000, 视角快速下移并切到自己的手部特写：一只手掏出黑色智能手机，屏幕显示音乐播放界面，食指稳稳落下，精准按下白色播放键，伴随清脆的点击声。 [Shot 4] At 00:06.000, 轻快的音乐突然响起，全场时间瞬间暂停。所有路人完全定格，<Subject 1> 迈步前冲的姿势僵在半空，发丝和裙摆都悬停不动，周围车辆与行人也全部静止。 [Shot 5] At 00:08.000, 镜头沿360度环绕轨迹平滑移动半周，从 <Subject 1> 的侧面缓慢绕到她的正前方。全程所有人物和物体保持绝对静止，发丝、裙摆与背景行人没有任何晃动，运镜稳定无抖动。 [Shot 6] At 00:10.000, 第一人称的手缓缓伸入画面，轻轻捏了捏 <Subject 1> 气鼓鼓的脸颊。她的脸颊被捏得微微变形，眉头和眼神依旧保持生气表情，身体与周围人群仍处于定格状态。 [Shot 7] At 00:12.000, 镜头向 <Subject 1> 的脸部平滑推近，形成脸部特写。她鼓着腮帮子，眼神凶狠地瞪着镜头，表情凝固在又气又可爱的状态。 [Shot 8] At 00:13.000, 镜头开始快速向后退，手机里的音乐戛然而止，时间瞬间恢复。<Subject 1> 眼神一怔，随即眉头压得更紧，挥起拳头朝镜头冲来，背景行人也恢复走动。At 00:16.000，她一边逼近一边喊，<d>[Chinese] T8，等你做完教程，看我怎么收拾你！</d> 镜头持续快速后退，她鼓着腮帮追至画面近处，视频在追赶状态中结束。 overall_soundscape: 前两秒充满首尔街道的人声、车流声、脚步声和店铺环境音。手机按键发出清脆点击声；定格瞬间环境声骤然抽空，只剩手机传出的轻快音乐，恢复时人声、脚步、衣料摩擦和急促呼吸一起回来。后退拍摄期间，女生追赶的脚步持续逼近。 non_diegetic_music: N/A
English prompt
subject_definitions: <Subject 1> is the young woman in <Picture 1>, the sole woman's facial and identity reference for the target video. Keep her face shape, eyebrows, eyes, nose and lips, skin tone, long straight dark-brown hair, small stud earrings and black top strictly consistent. In the video she wears a black dress with a flowing skirt. Her smile changes into an angry, pouting expression, but her face and identity never change. summary: [reference generation] Use <Picture 1> to anchor <Subject 1>'s identity in a 20-second first-person romantic comedy. On a Seoul street with Korean signs, she walks away in a huff. Pressing play on a phone freezes the world; the camera moves around her and playfully pinches her cheek. When the music stops, time resumes and she chases the camera. Exactly one woman, no face changes and no subtitles. retention_analysis: <Subject 1> appears in Shots 1, 2 and 4–8. Preserve her reference facial structure, complexion, long dark-brown hair, studs and black top details while expression, posture and angle follow the story; keep one consistent identity. detailed_description: Realistic first-person handheld romantic comedy, close to human eye level. Seoul night-street lights shimmer among dense Korean shop signs. No subtitles or titles; lock <Subject 1>'s face to <Picture 1> throughout. [Shot 1] Frontal eye-level frame. <Subject 1> stands in the middle of the lively street with both hands on hips, dark-brown hair hanging loose and the black dress skirt falling naturally. She frowns, puffs her cheeks and looks directly at the camera. Pedestrians walk normally around her; layers of Korean signs recede into the distance. In a clear annoyed voice she says exactly, <d>[Chinese]再也不理你了。</d> [Shot 2] At 00:02.000, she turns sharply away, long hair swinging, and strides forward angrily. The first-person camera follows close behind. Her back stays central while pedestrians continue to move normally on both sides. [Shot 3] At 00:04.000, the point of view tilts quickly down and cuts to a close-up of the viewer's own hand pulling out a black smartphone. Its screen shows a music player. An index finger lands precisely on the white play button with a crisp click. [Shot 4] At 00:06.000, bright playful music starts and time freezes immediately. All pedestrians stop perfectly. <Subject 1> is suspended mid-stride, with hair and dress hem motionless; nearby vehicles and people are completely still. [Shot 5] At 00:08.000, the camera makes a smooth half-orbit along a 360-degree path, travelling from her side to her front. Every person and object remains absolutely frozen, including hair, skirt and background pedestrians. Stable motion without camera shake. [Shot 6] At 00:10.000, the viewer's first-person hand slowly enters and gently pinches one of <Subject 1>'s puffed cheeks. The cheek deforms slightly under the fingers. Her brows and gaze retain the angry expression, while her body and crowd remain frozen. [Shot 7] At 00:12.000, push smoothly into a facial close-up. Her cheeks stay puffed and her eyes glare fiercely at the camera, her expression frozen between anger and cuteness. [Shot 8] At 00:13.000, the camera rapidly backs away. The music cuts off at once and time resumes. She is startled for an instant, furrows her brows harder, raises a fist and runs toward the lens; background pedestrians begin walking again. At 00:16.000, while approaching, she shouts exactly, <d>[Chinese]T8，等你做完教程，看我怎么收拾你！</d> Keep retreating quickly as she chases close to frame, still puffing her cheeks. End while the chase is in progress. overall_soundscape: For the first two seconds use Seoul street voices, traffic, footsteps and shops. The phone button clicks sharply. When time freezes, environmental sound vanishes and only the upbeat music from the phone remains. On resumption, voices, footsteps, clothing movement and quick breathing return, with her chasing steps drawing closer as the camera retreats. non_diegetic_music: N/A. The music in the frozen interval comes from the phone within the scene.
主题与风格
探索更多图片和视频提示词 · 生成图片 · 生成视频 · 在 AI 画布中创作
相关作品
- 烈焰环绕的透明瓶
制作一段20秒、16:9的电影感产品艺术视频。画面中央是在偏远森林空地、焦黑地面上的半透明玻璃伏特加瓶，瓶身呈丙烷气瓶形状。受控的森林大火在与瓶子保持安全距离的位置环绕燃烧，呈现橙色火焰、发光余烬、飘散烟雾和热浪；瓶子保持完整清晰，真实捕捉火光与反射。开场从烟雾中以低机位缓慢推进，随后进行轻柔的180度环绕，展示独特的气瓶轮廓，最后定格在余烬掠过的稳定特写。采用深炭色阴影衬托高对比金橙色光线，表现逼真的玻璃折射、无标签表面、自然火焰运动、电影级景深、细微变形宽银幕高光和稳定流畅的镜头移动。不要人物、文字、标志或多余物体，不要爆炸，不提供不安全的操作性内容。时长必须正好20秒；画幅比例必须正好1
- 玻璃穹顶下的维多利亚哥特小屋
MiniMax H3 视频生成规格：生成一段20秒、16:9的电影感镜头：一座微型桌面哥特式房屋被安置在华丽的玻璃穹顶下，置于昏暗维多利亚风格房间中的厚重天鹅绒桌布上。苍白的光线从唯一的窗户照入，呈现褪色的色彩、尘埃、精细的陈旧纹理和戏剧性阴影。采用超写实、极度细致的微缩工艺，营造古董般的氛围、克制的低饱和色调与优雅的维多利亚风格。开场展示铺着天鹅绒桌布的桌面较宽视角，随后镜头缓慢稳定地向穹顶推进并轻微环绕，最后停在微型房屋及玻璃反光的三分之四近景。全程保持真实的玻璃折射、柔和的悬浮尘埃、浅景深和稳定的几何结构。不要文字、标志、人物、源视频或音频输入。
- 玻璃穹顶下的微型森林
生成一段20秒、16:9画幅的电影感视频。画面展示一座微型桌面森林，在华丽的玻璃穹顶下发光，穹顶置于铺着奢华天鹅绒桌布的昏暗、积满尘埃的维多利亚式客厅中。开场以缓慢而富有氛围感的推进镜头掠过桌面，逐渐展现穹顶内的微型树木、苔藓、蕨类植物和温暖的生物荧光。随后镜头轻柔地环绕穹顶，玻璃反光与尘埃微粒在戏剧性的光束中漂浮，最后以弧形玻璃后方发光的微型生态系统特写收尾。采用超写实、极致细节的纹理，褪色古董照片般的色彩分级，华丽材质，深邃阴影，体积光，浅景深，细腻玻璃反射和高端精致的3D渲染。全程保持场景连贯、优雅且氛围浓郁；不要文字、标志、人物或依赖音频的动作。
- 玻璃穹顶下的微型森林
生成一段20秒、16:9的电影感视频：一座微型桌面森林被安置在华丽的玻璃穹顶下。展现层次丰富的林地微缩景观，包括细小树木、苔藓、蕨类、微型石块和自然细节，置于奢华的天鹅绒桌布上。采用超写实材质、极其精细的细节、维多利亚式装饰设计和戏剧性棚拍灯光，让玻璃高光与微型植被呈现幽光。镜头先缓慢靠近穹顶，再环绕移动以展示森林的纵深，最后轻柔停在透过玻璃观察的主体画面。保持奢华优雅的氛围、符合物理规律的反射与折射、浅景深、稳定流畅的摄影机运动，以及高端Octane渲染般的真实感。不要出现文字、标志、人物、艺术家姓名，也不要使用外部源视频或音频。
- 苏格兰黑屋的中世纪风貌
生成一段20秒、16:9画幅的电影感建筑短片，展现一座坐落于崎岖高地景观中的中世纪苏格兰黑屋。房屋具有厚重深色石墙、低矮草皮屋顶、狭小内凹开口、粗粝风化的表面纹理和细微的岁月痕迹。开场从风吹荒原与戏剧性阴天构成的广角全景开始，随后镜头缓慢稳定地向建筑推进，最后停留在入口与石砌结构的三分之四细节视角。采用自然冷色日光、克制的土色调、柔和空气雾感、真实石材与草地纹理，并表现周围植被随风轻动。使用具有19–45毫米变焦镜头观感的电影化摄影，保持稳定构图、真实透视和纪录片级建筑细节。不要出现人物、文字、标志、水印、源视频或音频。
- 废弃实验室中的玻璃穹顶微型森林
生成一段20秒、16:9的电影感视频：一片桌面微型森林被封存在废弃维多利亚风格实验室里的华丽玻璃穹顶下。微型生态系统散发柔和的生物荧光绿与温暖琥珀色高光，展现苔藓、纤细蕨类、微型树木、潮湿土壤和漂浮的尘埃微粒。开场以缓慢的镜头展示荒废实验室，随后镜头轻柔向穹顶推进并进行细微环绕，最后透过玻璃定格在亲密特写上，呈现真实的凝结水汽、反射和精致自然纹理。采用超写实材质、极其丰富的细节、戏剧性体积光、浅景深、符合物理规律的阴影和精致的实时三维电影级渲染。全程保持几何结构与连续性稳定；不要文字、标识、人物、品牌或依赖音频的元素。