Prompt Seedance 2.0: 8,113+ contoh dan panduan

Cinematic horror-thriller short film set at a mountain ski resort lodge during a snowstorm, opens with an immediate hook, grounded realistic style, maximum dynamic camera work. Setting: a cozy wooden lodge interior with a fireplace, large windows looking out to falling snow, guests relaxing with hot drinks. CHARACTER LOCK: the main character is a lodge staff member, a woman in her late 20s, red ski jacket, consistent in every shot. Muted cold-toned color grading, natural diegetic sound throughout, tension score entering immediately. > > **[0-1s]** Hard cut: the large lodge window frosts over instantly from outside, ice crystals spreading unnaturally fast across the glass. > > **[1-2s]** It shatters inward, a humanoid creature of jagged ice and frost crystal stepping through, breath visible as steam in the sudden cold. > > **[2-3s]** Nearby guests gasp, backing away from their tables near the fireplace. > > **[3-4s]** It moves toward the nearest guest, a thin layer of frost spreading across anything it touches. > > **[4-5s]** He stumbles back, his sleeve frosting over where it brushed him. > > **[5-6s]** Panic spreads through the lodge, guests scrambling toward the far end of the room. > > **[6-7s]** The staff member grabs a fire poker from beside the fireplace, stepping between the creature and the fleeing guests. > > **[7-8s]** It turns toward her, crystalline features catching the firelight. > > **[8-9s]** She swings the poker; it connects, a crack splintering across its icy shoulder. > > **[9-10s]** It recoils, frost receding slightly from the point of impact. > > **[10-11s]** A second staff member grabs a thermos of hot water, flinging it directly at the creature. > > **[11-12s]** Steam erupts where the water hits, the ice visibly weakening at the contact point. > > **[12-13s]** It staggers, momentarily destabilized, frost patterns flickering across its form. > > **[13-14s]** The staff member presses the advantage, striking again with the poker. > > **[14-15s]** Wide shot: guests reach the far hallway, a few glancing back in fear. > > **[15-16s]** The creature lunges toward her, icy claws reaching out. > > **[16-17s]** She ducks low, narrowly avoiding the strike, rolling toward the fireplace. > > **[17-18s]** She grabs a burning log directly from the fire with gloved hands. > > **[18-19s]** SLOW MOTION insert: she swings it in an arc toward the creature's chest. > > **[19-20s]** Contact sends a shockwave of cracking ice through its entire body. > > **[20-21s]** It freezes mid-motion, fractures spreading rapidly across its crystalline form. > > **[21-22s]** SLOW MOTION: it shatters apart into scattering ice fragments and swirling cold mist. > > **[22-24s]** Wide shot: the fragments melt quickly near the fireplace's warmth, leaving only a damp patch on the floor. > > **[24-26s]** She stands breathing hard, log still smoking slightly in her hand, snow drifting in through the broken wi

15-second cinematic photorealistic science-fiction action short, 16:9. A massive human exploration fleet arrives above a completely ocean-covered alien planet. Suddenly the entire ocean begins rising from the planet. Not a wave. The WHOLE OCEAN lifts into space as an enormous spherical mass of water. The fleet scatters violently. Fighter spacecraft race between gigantic columns of water stretching from the planet into orbit. An enormous alien structure begins emerging from beneath the exposed ocean floor. Enemy ships suddenly appear and attack. The human fighters dive through waterfalls hundreds of kilometers long while missiles streak past them. One exploration ship races toward the emerging structure. The ocean continues rising around it. The ship enters the structure just before the entire planet becomes surrounded by a colossal floating sphere of water. Inside the structure, a single enormous artificial eye opens. CUT TO BLACK. Photorealistic, breathtaking planetary scale, extreme motion, spectacular space combat, enormous water physics, dynamic camera angles, cinematic lighting, intense escalation, no text, no logos, no anime, no cartoon.

生成一支【30秒、16:9横版、4K、真人实拍、高预算电影摄影质感】的“中国文明时间旅行”短片，深度参考我上传的参考视频真正的镜头结构和时间推进逻辑：这不是多个历史场景的剪辑，而是一台摄影机从第0秒到第30秒始终跟随【同一个中国男性】持续向前运动，从远古一路穿越中国几千年文明，并且像参考视频一样最终继续越过当代进入未来；使用我上传的男性照片作为全片【唯一且绝对的人脸参考】，照片只负责锁定人物本人，必须严格保持他的脸型、五官比例、眉眼、眼距、鼻梁、鼻尖、嘴唇、下颌骨、面部骨相、肤色基础、年轻年龄感和人物辨识度，绝对不要参考照片里的蓝色西装、领带、餐厅、吊灯、玫瑰、首饰、坐姿和背景，全片0—30秒必须始终让观众明确认出就是参考图里的这个男人，禁止每个时代换演员、禁止五官漂移、禁止变成不同的古装帅哥或AI模板脸，时代变化只能改变他的发型、头发长度、束发方式、胡须、衣服材质、鞋履、帽饰、腰带、工具、配饰、身份状态和动作习惯；整支影片视觉上必须是真正的“一镜到底”，禁止硬切、黑场、闪白、淡入淡出、普通溶解、突然换场、人物瞬移、背景刷新，人物也绝不能站在原地让世界自己变化，他必须始终主动向前，奔跑、快步前进、跨越、跳下台阶、侧身避让、抓住车辆、穿过人群、推门、登桥、穿建筑、经过交通工具，摄影机持续与他一起运动，在侧前方跟拍、正面倒退跟拍、侧面平移、背后追拍、低机位脚步、大远景和人物近景之间像参考视频那样自然改变景别；所有时代变化优先利用真实物理运动完成隐藏转场，例如岩石、陶轮、车轮、战车、柱子、旗帜、布料、卷轴、门板、桥柱、马车、蒸汽、列车、公交车、人群、玻璃幕墙、透明屏障等从摄影机极近处擦过并短暂遮挡0.2—0.8秒，在遮挡过程中人物造型、手中物品、脚下地面和周围世界必须同时演化，遮挡结束后人物的步幅、身体重心、运动方向、速度和空间位置完全延续上一瞬间，绝不能出现“背景已经换时代但人物没换”的情况；0—3秒必须极具抓力，清晨日出，中国远古山崖与巨大河谷，低角度晨光穿过空气薄雾，远处是辽阔山川、河流、森林、野生动物和极小的原始聚落，山洞附近残留微弱篝火，风吹荒草与人物粗糙长发，参考图中的男人此时是一名远古中国先民，脸依然必须保持本人，只是披着粗糙兽皮、植物纤维和原始编织物，皮肤有真实毛孔、汗液、尘土与轻微晒痕，手持石制矛头木矛，他突然听见远方声音，快速转身并沿山坡向前奔跑，摄影机迅速由大全景下降并绕到侧前方与他一起高速运动，他跨过一块巨大岩石，镜头压低跟住脚步，落脚瞬间不要切镜，干燥岩石纹理从脚底向前扩散并连续变成湿润泥土、田埂和早期农田，兽皮随着跑动和身体自遮挡逐渐变成麻布、葛布等早期织物，木矛在换手过程中缩短并变成木耒、石镰等农具，散乱长发逐渐简单束起，周围开始自然长出半地穴式或木土房屋、陶罐、谷物、编织篮、磨盘、牲畜、晒粮架以及真正正在生活和劳动的人群，约3—5秒进入新石器农耕文明，人物继续高速穿过村落，侧身绕开抱陶罐的居民和动物，一只巨大的旋转陶轮从摄影机极近处横向经过并占满大半画面，陶轮持续旋转时泥土纹理逐渐变得坚硬、深沉并产生金属光泽，圆形结构自然演化成青铜礼器纹饰和战车车轮，约5—7.5秒遮挡离开后空间已经连续进入夏商周青铜文明，人物跑动动作完全没有中断，但身上的粗麻衣物已经变成真实、朴素、结构更明确的早期华夏服饰，绝不是仙侠、古偶或影楼汉服，周围出现夯土城墙、木构建筑、青铜铸造、炉火、车马、旗帜、礼器和繁忙居民，人物跑上略高道路，一辆战车快速横穿，他侧身避让并顺手抓住战车旁被风吹起的长旗借力继续向前，巨大的旗布猛烈扫过镜头形成完整遮挡，布料覆盖摄影机期间人物发式、衣领、革带、鞋履和道路结构同步向秦汉演化，约7.5—10秒布料离开后人物已经奔跑在真实秦汉城门与驿道之中，同一张参考脸绝不能改变，人物为普通秦汉旅行者或城市男性，真实束发、交领服饰、革带和布履，不是皇帝或将军，摄影机从人物侧面追随他穿过城门、驿站、人群与车马，他短暂抓住一辆向前行驶的马车边缘，被惯性带着向前两步后松手继续跑，巨大木车轮从镜头近处飞速滚过，轮子离开时原本较空的道路沿人物前进方向自然扩张、变宽并迅速增加建筑、商铺、商品、人群和色彩，约10—12.5秒连续进入盛唐大型城市，人物同步变成真实唐代普通城市男性装束，脸完全保持参考男性本人，街道不要影视城空景，而是充满真实木构楼阁、商贩、马匹、驼队、不同地区商旅、货物、食物、布匹、器皿、旗幡和拥挤生活流，人物在人群间高速穿行，低头躲过抬过来的木梁，再侧身从马匹与行人之间穿出，摄影机由侧拍自然运动到正面倒退拍摄，一大片市场彩色布幡从头顶被风吹落并扑向摄影机，布幡靠近镜头后纤维越来越薄并自然转化为展开的大幅宣纸书画卷轴，摄影机贴着卷轴表面向前滑行，纸面墨迹、河道、桥梁和屋宇线条随着透视连续获得真实体积，不要魔法发光，而是让画中的桥梁木线逐渐成为真实桥栏、山水线条成为河面与城市结构，约12.5—15秒摄影机越过纸张边缘后已经进入极其繁华的宋代城市，同一个男人动作不中断，服装与发式同步宋代化，真实河道、拱桥、商铺、茶肆、船只、摊贩、瓷器、书画、印刷、货物和密集市民生活从他身体周围向前延展，他跑上桥面，摄影机降低靠近桥板倒退跟拍，人物一手扶栏侧身避过迎面的大型货箱，货箱贴近镜头制造短暂遮挡，摄影机重新找到他时他已经冲下数级台阶并推开前方巨大木门，厚重门板贴着镜头划过，在门板完全遮挡期间街巷、建筑、衣料和人物头饰同步继续演化，约15—17秒穿门后自然进入明清城市，仍是参考图中的同一个男人，服装符合真实普通城市居民身份，不要清宫戏，不要满街官服和华服，周围有成熟街巷、牌坊、商铺、院落、仓栈、手工业、小贩、居民、货物和车马，人物从其中持续快步向前，一辆木制车辆擦过他身边，摄影机贴近车轮，木质轮缘随着高速旋转逐渐出现铁箍、钢铁、铆钉与机械结构，约17—20秒木车轮自然变成蒸汽机车的巨大铁轮，木构城市也沿运动方向连续变成中国近代的铁路、钢梁、砖石建筑、电线杆、电灯、报摊、自行车、近代火车站、码头和工厂，必须明确仍然是中国近代而不是英国工业革命，人物传统服饰同步过渡为中国近代长衫、短褂和早期西式剪裁混合的城市男性造型，他沿运行中的蒸汽列车快速向前，大量真实白色蒸汽突然喷向摄影机并完全覆盖画面约半秒，人物在蒸汽里仍保持同样步频，约20—22秒蒸汽逐渐散去时火车、道路和城市已经连续演化成二十世纪中国，自行车流、公交、电车、普通居民楼、商店、电线和城市人群不断出现，人物发型变成短发，衣着从近代长衫自然变为中山装、衬衫、夹克和长裤，一辆公交车从前景横向掠过并遮挡人物，公交离开后街道明显更宽，自行车逐渐变成汽车，低层建筑不断向上发展，玻璃、钢筋混凝土和现代公共交通的比例越来越高，人物仍然持续向前，约22—24秒进入真正的当代中国城市，摄影机贴着一整面巨大玻璃幕墙与人物平行前进，玻璃倒影中的旧城市首先开始加速演化，自行车、公交、低层建筑逐渐变成现代道路、地铁、高铁、电动汽车、现代商业街、绿化、公园和玻璃建筑，摄影机绕过玻璃边缘后真实当代中国城市完全展开，视觉气质可以融合上海、深圳、杭州等中国现代都市但不要直接复制著名地标，人物也已经成为现代年轻男性，保持参考图本人脸

创建一段30秒、4:3横向画幅、高度写实的个人健身Vlog，呈现朋友用2000年代初家用DV随手记录一次日常锻炼的感觉。不是健身广告，也不是精致的电影短片。30秒是一次健身过程的剪辑摘要，通过自然跳剪串联出门、准备、训练和结束，不要让人物在30秒内匆忙完成所有事情，不要加速播放。 【人物与服装】 使用上传的这张灰色运动装全身人物图【@图片1】作为唯一角色参考。主角为参考图中的成年女性，严格保持她的五官长相、脸型、眼睛、鼻子、嘴唇、细框眼镜、发际线和整体年龄感一致。不同角度、运动状态、说话和笑起来时，依然必须是同一个人，不能变成另一张相似的脸。参考图只用于确定人物身份、服装和身体外观，不沿用白色摄影棚背景或原图的固定站姿。 人物具有自然、健康的健身力量感，肩臂、腰腹和腿部有适度肌肉线条，运动时能够看到真实的肌肉收缩和稳定的身体控制。保持参考图的整体身高和身材比例，不要夸张增肌、过度收腰或拉长双腿。皮肤保持白皙细腻，但保留自然皮肤质感，运动后可以出现轻微泛红，不能过度磨皮或呈现塑料感。 全程穿着参考图中的同一套运动服装：麻灰色宽领短袖T恤，领口自然倾斜，露出她的右肩，也就是正面面对镜头时画面左侧的肩膀；同侧腰部衣摆打结，形成露出腰腹的短款造型，保留胸前原有粉紫色印花；里面是粉紫色细肩带运动内衣，肩带和少量边缘自然露出；下身为麻灰色高腰贴身运动短裤，搭配白色罗纹中筒运动袜和参考图同款白色系带运动鞋。衣服随运动产生真实褶皱和轻微位移，但不能改变款式、颜色、材质、打结位置或露肩方向。不要换成黑色上衣、长款紧身裤、蓝色袜子或长靴。 出门时保留参考图中的披散黑色长发，进入健身房前亲手扎成马尾，此后始终保持同一条马尾，保留自然额前碎发和鬓发，不要突然换刘海或改变发量。眼镜全程佩戴，运动中保持合理位置。道具固定为一个小号黑色健身包、一只白色瓶盖的透明水瓶和一条小白毛巾，不要在不同镜头中改变外观。 【镜头流程】 0—3秒｜去附近健身房。 安静的社区街道，自然日光。朋友在她侧前方边走边拍，她穿着上述灰色运动套装，单肩背着小号黑色健身包，向附近健身房走去。黑色长发随着脚步轻轻晃动。她短暂看向镜头，露出一个自然的小笑容，随后继续看路，不刻意摆拍。镜头有走路带来的轻微起伏，构图略偏，像随手记录。 3—6秒｜扎好头发。 在健身房入口旁停下，侧面中近景。她抬起双手，将长发收拢，用黑色发圈扎成利落的马尾，随后扶一下眼镜、调整肩上的包带。可以用自然跳剪略过部分扎发过程，但必须让观众看到从披发到马尾的实际动作，不能让发型凭空变化。动作放松熟练，手指和头发之间的接触真实。 6—9秒｜放下东西，简单热身。 切入普通社区健身房，黑色橡胶地面、哑铃架、训练凳和墙面镜子，环境真实、有人使用的生活痕迹，不要豪华空旷的商业摄影棚。她把黑色健身包放在靠墙训练凳旁，将水瓶和白毛巾放在凳子一端。自然跳剪到她轻轻活动肩膀、舒展手臂，调整呼吸。镜头像朋友站在旁边随手拍摄，不围绕身体刻意展示。 9—15秒｜简单力量训练。 两个正常速度的训练片段：先拍她进行轻哑铃弯举，完整呈现一次平稳抬起、短暂停顿、控制放下的过程，肩膀不乱晃，手掌真实握住哑铃，手臂随着发力出现自然肌肉线条；随后切到侧前方中全景，拍她完成一次自重深蹲，从站立到下蹲再稳定站起，双脚接触地面，膝盖和髋部运动合理，躯干稳定。她专注呼吸，发力时轻轻呼气，不对镜头摆健身造型。运动服出现自然拉伸和褶皱，马尾随着动作轻微摆动。 15—20秒｜跑步机慢跑。 侧前方手持中景，她在跑步机上轻松慢跑，步幅适中，左右脚交替接触跑带，手臂自然摆动，马尾随步伐有节奏地晃动。呼吸稍微变重，额头、鼻翼和颈侧逐渐出现少量汗水，不是突然全身湿透。她看了镜头一眼，带着一点累又觉得好笑的表情轻轻笑一下。随后自然剪切到跑步机已经减速、停止，她确认站稳后再走下来，不跳下运行中的跑带。 20—25秒｜喝水、擦汗。 回到同一张训练凳旁，她拧开同一只水瓶，喝两小口水，再用小白毛巾轻按额头和颈侧。毛巾与皮肤有真实接触和遮挡，不穿过眼镜，不直接擦掉全部汗水。她偶然看到镜子里有些出汗的自己，轻轻笑起来，是运动后放松的真实反应，不是夸张表演。镜中倒影的动作、服装和位置必须与本人准确对应。 25—30秒｜练完收工。 她坐在训练凳上，呼吸仍稍微急促，肩膀随着呼吸轻轻起伏，手里拿着毛巾。朋友把镜头稍微凑近，她抬眼看向镜头，用自然的中文口吻笑着说：“今天练得还挺爽！”声音略带运动后的气息，嘴型准确，说完轻笑一声。最后自然切到她背上同一个黑色健身包走向出口，马尾轻晃，镜头稍慢半拍跟上，在生活化的动作中结束，不做广告式定格。 【摄影与画面质感】 整段呈现2000年代初消费级MiniDV录像质感：轻微手持抖动、不完美构图、偶尔跟拍不及时、轻微自动对焦搜索、室外进入室内时短暂曝光和白平衡变化、柔和的数字细节、轻微暗部噪点、正常运动模糊，以及一次不太熟练的小幅变焦。这些瑕疵应克制、随机，不要每个镜头都故意失焦或剧烈晃动。人物脸部始终足够清楚，可以辨认出参考图中的同一个人。以平视中景、中全景和自然面部近景为主，训练镜头保留必要的身体和器械信息，不刻意对胸部、臀部或其他局部进行特写。不要稳定器式丝滑运镜、大片调色、慢动作、炫光、胶片烧灼、夸张颗粒或伪造的磁带故障。画面中不出现DV机本体。 【音频】 仅保留同期环境音：社区脚步声、轻微街道声、推门声、包带与衣料摩擦、健身房空调嗡鸣、远处模糊交谈、哑铃轻放声、鞋底接触地面的声音、跑步机马达声、逐渐变重的呼吸、水瓶开盖和喝水声、毛巾摩擦声、自然笑声，以及结尾的一句人物同期对白。音质带有家用DV内置麦克风的轻微底噪和空间感，不要录音棚式人声。没有背景音乐，没有解说旁白，没有字幕。 【负面约束】 禁止身份漂移、变脸、眼镜消失或变形、发型无故切换、衣服换色换款、露肩方向左右翻转、身体比例变化、夸张肌肉、塑料皮肤、美颜滤镜、异常皮肤纹理、不合理汗水、扭曲手掌、多余手指、器械变形、哑铃穿手、脚底滑动、跑步动作失真、物品漂浮或瞬移、错误镜面倒影、重复主角、突然出现的道具。除参考服装已有印花和标记外，不新增品牌标志、水印、日期时间戳、界面文字或其他画面文字。

Create a 15-second vertical 9:16 photorealistic cinematic video of a cute orange tabby cat inside a muddy World War-style trench, wearing a realistic olive-green military helmet and holding a small vintage rifle. Start with the camera focused on the dark trench entrance, with stacked sandbags and rough wooden planks clearly visible above. At first, only the top of the helmet slowly appears from the trench. The cat gradually rises into view while holding the rifle naturally with both paws. Keep the cat’s face, fur, helmet, rifle, trench, and background consistent throughout the entire video. Around the middle, the cat looks directly toward the camera with a serious but adorable expression and slowly raises one paw in a small salute. Maintain realistic fur movement, natural paw motion, subtle head movement, and cinematic depth of field. The camera should remain mostly stable with a very gentle handheld cinematic movement. Near the end, the cat lowers its paw and slowly ducks back down into the trench, leaving only the helmet visible. Use realistic outdoor lighting, soft overcast sky, detailed muddy soil, natural textures, shallow depth of field, smooth motion, and no text, watermark, distortion, or extra characters.

电影级写实质感，纯古风中国仙侠美学。 以成熟的身份冲突、克制的武侠人物表演、依靠潜台词和人物站位完成反转的资深编剧式对白为核心。 采用： Arri Alexa 电影机质感 自然体积光 细腻胶片颗粒 核心反转 这次跳出以下这些已经使用过的结构： 抗命 替罪 认输 营救 忠诚测试 把冲突直接放在 “继承” 本身。 所有人都认为： 对小师妹最大的认可，就是让她接过剑仙师姐的剑、名号和位置。 真正的反转却是—— 师姐恰恰因为真正认可她，才不允许她的人生最终只变成： “第二个自己。” 参考与环境 将 @图片 1 和 @图片 2 严格作为人物身份锚点。 本轮上传的全部背景与地点参考图共同决定同一套环境 DNA。 正式构图前，无声重组彼此兼容的： 真实地形 建筑语言 材质年代 植被 水体 天气 山雾 倒影 主光方向 空气纵深 形成一个全新的统一地点。 背景原则 背景始终持续鲜活，但在叙事上绝对中立。 角色设定 角色 ID A｜剑仙师姐 始终是同一个 @图片 1 剑仙师姐： 25–30 岁东亚女性 椭圆白皙面庞 深色杏眼 黑色长发半挽，以白玉簪固定 身材高挑纤细 穿同一套白色刺绣真丝汉服 半透明宽袖 银色腰封 玉佩 白色布靴 持唯一一柄银色长剑 角色 ID B｜小师妹 始终是同一个 @图片 2 小师妹： 20–25 岁东亚女性 圆润灵动的脸 黑发编辫 身形娇小 穿同一套青绿色亚麻汉服 深色腰带 木簪 黑色布鞋 持唯一一柄已经有使用磨损的深色钢剑 其他次要角色 一名年长师父 两名宗门弟子 一名外宗使者 这些角色只作为次要见证者存在。 分段结构 0–5s｜全景或远景 一名年长师父主持正式传位仪式。 他双手递出一柄象征剑首身份的礼仪银剑，沉声说道： “接剑。从今日起，你继她的位。” 同一个小师妹下意识伸手。 手指还没有碰到剑， 同一个白衣剑仙突然出剑，只用一次极干净、克制的剑身碰撞把礼仪银剑拨向侧面，绝不能攻击任何人。 她只说： “她不接。” 小师妹立即转头看她： “为什么？” 5–10s｜中景或牛仔景 保持： 同样两名女子 同样服装 同样长剑 同一柄礼仪剑 完全一致的地理空间 师父明显不悦： “你不认她？” 师姐真正看向小师妹，回答： “正因为认。” 完整停顿半拍，再落下一句： “我的位置，只够站一个我。” 浅景深后方，同一个外宗使者忍不住轻笑： “那她算什么？” 师姐甚至没有转头回应这个问题，只继续看着小师妹： “问你自己。” 这一刻不要煽情音乐强推情绪，只保留自然空间声与低而克制的配乐底色。 背景中的以下元素继续自然运动，却完全不介入剧情： 风 水体 山雾 衣袖 发丝 倒影 远处人物 10–15s｜特写或极近特写 同一个小师妹先低头看一眼被拨开的礼仪银剑，又看向自己手里那柄陪伴她多年的旧深色钢剑。 她不再问自己能不能继位，而是换了一个问题，低声问： “那我站哪儿？” 同一个师姐没有替她决定身份，只回答： “别站我后面。” 完整停顿半拍。 小师妹真正听懂之后，从原先落后师姐半步的位置走出来，第一次主动与师姐并肩站立，侧头问： “这儿？” 师姐始终看着前方，没有赞美，也没有安慰，只再说三个字： “再往前。” 极近特写跟随小师妹的眼神变化： 她短暂愣住，随后真正向前走出完整一步，第一次站到师姐前方。 师姐没有伸手拉她，也没有叫她回来。 浅景深后方： 同一个年长师父彻底沉默 那个刚才还在轻笑的外宗使者也慢慢收起笑容 最终画面明确停在： “小师妹拿着自己的旧剑站在最前方” 被拨开的传承银剑仍留在后方，没有任何人再去捡它。 这意味着： 不是小师妹继承了师姐的位置， 而是师姐第一次把前面的位置，真正让给她自己去走。 戏剧核心 整段真正完成的不是“拒绝传位”，而是： 拒绝把小师妹定义成“下一个师姐” 拒绝把认可等同于复制 让她不再作为谁的延续存在 让她第一次站到属于自己的位置上 表演原则 剑仙师姐 不能煽情 不能说教 不能像在做牺牲式成全 她的态度必须稳定、克制、坚定 她不是否认小师妹，而是拒绝小师妹被自己的名号吞没 小师妹 前半段必须处于“顺着大家期待去接位”的自然状态 中段听不懂，但开始动摇 后段真正听懂“别站我后面”的含义 情绪不能爆发式外放 必须主要靠眼神变化和站位变化完成成长 年长师父 起初代表制度与传统秩序 最后沉默本身就是态度变化 不需要认错，也不需要解释 外宗使者 起初带一点旁观者的轻慢 最后收起笑容，说明他也意识到这不是简单的拒绝，而是更高层级的认可 硬性要求 严格总时长 15 秒 16:9 横屏 严格 三个连续清晰镜头 原生同步普通话对白 精准口型 精准视线关系 人物身份全程稳定 发型全程稳定 服装全程稳定 两人原有长剑全程稳定 礼仪银剑全程稳定 配角位置全程稳定 地理空间全程稳定 人物动作数量严格克制 以下三个动作必须清楚可读： 剑身拨开 回头 迈步 微表情必须真实克制 真丝衣料、金属长剑和头发物理效果自然 前景、中景、远景保持真实视差和同步空间环境声 不生成字幕 Seedance 2.0 Mini 执行重点 针对 Seedance 2.0 Mini，重点强化： 人物参考一致性 多人物视觉主次 少而精准的动作节点 稳定摄影机连续性 光影连续性 核心道具连续性 对白、金属轻碰声、脚步声、衣料声和环境声的精准音画同步 Negative blurry, bad quality, low quality, low resolution, noisy, jpeg artifacts, watermark, text, error; deformed, mutated, bad anatomy, poorly drawn hands, bad composition, out of frame, disfigured; inconsistent character, changing clothes, face morphing, background shift, glitching cuts, disappearing props

Create a 60-second ultra-realistic cinematic lifestyle vlog featuring a beautiful young Korean woman with long dark brown hair, soft natural makeup, and a gentle smile. She wears a delicate white floral dress with a light cream cardigan and carries a small white handbag. **Scene 1:** She stands outside a modern aesthetic café surrounded by lush green plants and glass windows. She holds an iced coffee, takes a sip, gently adjusts her hair, and smiles naturally at the camera. **Scene 2:** Close-up cinematic shots of her face as she enjoys her coffee. Soft sunlight illuminates her face. She looks relaxed and happy, with subtle natural movements. **Scene 3:** She walks along a beautiful city street during golden hour, wearing the same outfit. The camera follows her from behind and captures elegant side-profile shots. **Scene 4:** She arrives at a rooftop viewpoint overlooking a beautiful city skyline during sunset. The golden sun slowly sets behind the buildings. She sits on a bench, gazes at the sunset, and enjoys the peaceful atmosphere. **Scene 5:** At blue hour, she walks along a modern urban street with glowing streetlights and passing cars. The camera captures her beautiful side profile and flowing hair in a cinematic tracking shot. **Scene 6:** Nighttime outdoor café with warm fairy lights hanging overhead. She sits at a cozy wooden table with her iced coffee and handbag. She smiles at the camera, raises her drink, takes a sip, and waves goodbye with a cheerful expression. **Visual style:** Ultra-realistic 4K, photorealistic skin texture, natural facial expressions, cinematic depth of field, soft lighting, warm golden-hour tones, beautiful bokeh, smooth camera movements, professional color grading, realistic hair physics, natural body movements, aesthetic Korean lifestyle vlog, dreamy peaceful atmosphere. **Camera:** Mix of wide shots, medium shots, close-ups, slow-motion details, smooth tracking shots, and gentle cinematic transitions. **Mood:** Peaceful, romantic, cozy, elegant, dreamy everyday life. Maintain perfect character consistency, identical facial features, hairstyle, clothing, and accessories across all scenes. No distorted hands, no flickering, no face changes, no unnatural movements, no text, no watermark.

10秒，9:16，真人写实搞笑短视频。 中国古建筑景区，一位年轻亚洲女性穿古装，站在一朵白色云上，离地约2米，踩着云缓慢向前飞行。 前4秒：女生和白云稳定同步向前移动，女生神态轻松自然。 第4秒开始突然吹来一阵横向大风，女生头发和衣袖被吹动。脚下的白云被风单独向右侧慢慢吹走，女生不要跟着云移动。 白云完全离开女生脚下后，女生仍在原位置悬空停顿半秒。她先低头看自己的脚下，再抬头看向飘远的云，露出懵住的表情。 随后女生失去支撑，垂直向下掉出画面。白云继续悠闲地向远处飘走。 重点：云是被风横向吹走，人留在原地；云离开后人物先悬空停顿半秒，再掉下去。不要出现复杂法术特效。

KAZEMIRU: THE JADE-WIND BLADE: ONE-TAKE CASTLE ASSAULT Duration: 30 seconds | Aspect Ratio: 16:9 | No music ONE-TAKE LOCK: This is ONE SINGLE UNBROKEN CONTINUOUS SHOT lasting the full 30 seconds. ZERO cuts. ZERO dissolves. ZERO fades. ZERO match-cuts. ZERO time jumps. ZERO scene resets. The camera never leaves the action and every location change happens physically inside the shot: the camera passes through torn doors, over railings and up the roof with Kazemiru. Time runs in real time. STYLE: Ultra-photorealistic live-action Japanese dark-action cinema with restrained subtle VFX. Impossible anime-inspired dual-sword choreography performed by real human bodies. Large-format cinema look, natural daylight, muted earthy color palette, soft realistic shadows, no oversaturated colors. @Image1 is KAZEMIRU. Preserve her exact face, silver-white ponytail, torn ivory kimono, charcoal split hakama, wind-swirl forearm tattoo and gold hoop earrings in every frame. @Image2 shows the five elite guards. @Image3 shows Captain Tetsugan. POSITIVE LOCKS (true for all 30 seconds): - Exactly ONE Kazemiru with exactly TWO real steel katanas, one in each hand, both always visible and independently controlled. - Her face always matches @Image1. - Every enemy comes from @Image2 or @Image3 and stays visually consistent. - Her body physically crosses every distance, with no teleportation. - It is always daytime with bright natural light. PERSONALITY: Cold, calm, battle-hungry. Minimal expression, steady breathing, focused eyes. Her hunger shows through relentless speed and escalating ferocity, never through grinning or shouting. JADE-WIND EFFECT: Every fast movement leaves a thin, soft jade-green wind trace that follows her exact physical path and fades almost instantly: sword arcs leave fine blade trails, spins bend the air, footsteps kick tiny wakes, supersonic bursts add a brief compressed-air ripple. Always thin, subtle and realistic. No aura, no smoke, no glowing body, no effects moving on their own. CAMERA (one continuous handheld-style pursuit): Constant dutch angle of 25–35°. Wide field of view around 75° in the interior and courtyard, tightening to about 50° for the rooftop duel. Kazemiru moves first and the camera follows 0.1–0.3 seconds late: whip-pan, overshoot, correct, reacquire at the moment of contact. The camera flows through rear over-the-shoulder, low-angle pursuit, side tracking, ground-level chase and front three-quarter retreat, always in one connected movement. Never static, never perfectly smooth. ══ 0:00–0:10: THE GRAND AUDIENCE HALL ══ Setting: bright daylight through large shoji panels, gold folding screens, tall dark timber pillars, tatami sections, raised platform, armor displays, low lacquer tables, silk curtains. 0:00–0:03: Rear over-the-shoulder, five guards closing in. Her fingers tighten on both hilts. BURST: a thin jade streak snaps across the hall and the silk curtains whip. The camera loses her, whip-pans, and finds her inside Guard 1's reach. Left blade parries, right blade slashes diagonally, a pivot, a reverse cut. He drops. 0:03–0:06: Guards 2 and 3 converge. She ducks under a swing, plants a foot on a pillar and pushes off, dust bursting from the wood. Mid-air her left sword deflects Guard 2 and her right sword cuts down. She lands on the edge of a lacquer table, which slides back, and rebounds straight toward Guard 3. 0:06–0:10: Aerial right slash, left reverse slash, dual cross-slash with two separate jade trails. She lands low and bursts across the floor. Guard 4's spear thrusts: she runs two steps up a wall, flips off, and cuts him down through a folding screen. Guard 5 charges: she catches his blade with one sword, vaults the railing above him, and finishes with a cross-slash as she lands behind him. ══ 0:10–0:20: THE PALACE COURTYARD ══ Transition inside the same shot: she does not pose. Both swords still in hand, she drives straight through the paper doors in one burst. The torn paper flutters past the lens as the camera follows her through the doorway onto the wooden veranda and out into a stone courtyard with lanterns, a koi pond, a red lacquer arched bridge and pine trees. Eight new guards: swordsmen below and archers on the roofs. 0:10–0:13: A volley of arrows. She cuts two out of the air with a spinning double slash and a third grazes her sleeve. She sprints along the veranda, pivoting off each post, a jade ripple following her feet. 0:13–0:16: She kicks off a stone lantern, which tips over, and leaps onto the bridge railing. Three swordsmen charge. She runs along the rail, slashes the first, flips over the second, and lands in the koi pond with a huge splash. In the water she parries two blades and cuts both guards. Droplets scatter across the lens. 0:16–0:20: She bursts out of the pond and sprints up a pine trunk in three steps. The camera tilts up with her. She launches onto the roof among the archers: left sword disarms, right sword strikes, roof tiles scatter. Two archers fall. ══ 0:20–0:30: THE ROOFTOP DUEL ══ Transition inside the same shot: she keeps climbing over the roof ridge and the camera rises with her, tightening to a 50° field of view. The tiled castle rooftop opens up with wind in her ponytail and the town far below. Captain Tetsugan (@Image3) waits with his naginata, calm and unflinching. 0:20–0:23: Tetsugan sweeps the naginata in a wide arc. Kazemiru ducks, slides down the sloping tiles, and crosses both swords to stop the blade. Sharp metal sound as his strength pushes her back a step. Her expression stays cold. 0:23–0:27: Fast exchange: his thrust, her left parry, her right slash that cuts his armor cord. He spins the pole and sweeps for her feet: she jumps, kicks off the shaft, and flips over him. Mid-air she cuts both swords across his back armor. Tiles crack under their feet and a faint jade wake curves around her path. He staggers, strikes back, and she narrowly dodges as the blade cuts a roof beam. 0:27–0:30: Supersonic burst along the ridge line, the jade streak slightly brighter but still thin. She slides under the naginata, and as she rises both swords slice through the shaft in a clean cross-cut. Tetsugan drops to one knee, defeated but alive. Kazemiru stands, both swords still in hand, and the last jade filaments disappear. She turns her cold eyes toward the courtyard far below, where dozens more guards pour in. The camera slowly pulls back and up to a wide shot of the whole castle. She lowers her body for the next burst. THE FIGHT CONTINUES. AUDIO: Sword resonance, compressed-air whooshes, metal clashes, footsteps on wood and tiles, arrows, water splashes, fluttering silk and paper, enemy shouts, wind over the rooftops. No dialogue, no music. HUMAN-FIRST LOCK: Kazemiru is always a real human: natural skin pores, real eyes, individual hair strands, believable anatomy and cloth physics. Only her choreography is anime-level impossible. NEGATIVE: cuts, edits, dissolves, fades, scene transitions, time jumps, multiple shots, single-sword fighting, missing or duplicate sword, duplicate Kazemiru, clones, teleportation, thick aura, smoke, glowing body, neon or oversaturated colors, detached VFX, random flips, decorative acrobatics, idle pose, static camera, perfectly smooth tracking, slow combat, floating, broken anatomy, CGI or plastic skin, anime rendering, blood, gore, subtitles, logos, watermark.

CHARACTER LOCK: the main character is a man in his 30s, dark casual jacket, , short dark hair — consistent into Part 2. Muted color grading, natural sound throughout, tension score entering early. [0-1s] A woman leans close to a mannequin to inspect an outfit — its head turns toward her unexpectedly. [1-2s] It steps off its stand and reaches for her, pulling her out of frame. [2-3s] Wide shot: the four other mannequins nearby begin moving too, stepping down from their platforms. [3-4s] They move toward nearby shoppers with quick, unnatural motion. [4-5s] People scatter and call out in alarm through the section. [5-6s] One figure reaches a shopper near a clothing rack, pulling him off balance. [6-7s] Wide shot: concern spreads through the main corridor, people moving quickly away. [7-8s] The main character ducks behind a support pillar, assessing the situation. [8-9s] He spots a storage room nearby and waves urgently for others to follow. [9-10s] A group of shoppers hurries toward him and into the room. [10-11s] He pulls the last few inside and shuts the door firmly. [11-12s] Interior: the room full of worried people, some holding each other close. [12-13s] He leans against the door, then notices a wall mirror reflecting the corridor. [13-14s] Through the mirror he sees one figure standing still over a man on the floor. [14-15s] Close-up: his expression tightens with concern. [15-16s] He spots three more people hiding behind a distant pillar through the mirror. [16-17s] He glances at the room, then carefully opens the door a crack. [17-18s] The sound draws a nearby figure's attention toward the doorway. [18-19s] He grabs a life-size cardboard standee and sets it out in the corridor. [19-20s] The figure moves toward the standee instead, momentarily distracted. [20-21s] He slips out and moves low along the wall, away from it. [21-22s] Tracking shot: he hurries toward the three people behind the pillar. [22-23s] Close-up: their worried faces as he reaches them, signaling quiet. [23-24s] He motions for them to follow, checking the corridor both ways. [24-25s] Wide shot: the four of them move quickly back toward the storage room. [25-26s] Behind them, another figure's attention turns toward the movement. [26-27s] It begins following, picking up speed. [27-28s] He urges the group to move faster, glancing back. [28-29s] They reach the storage room door, hands on the handle. [29-30s]: the door swings open, light spilling out, the figure closing the distance behind them. Cut to black. Natural sound throughout — fabric movement, calls of alarm, footsteps, door sounds, breathing — with a tense instrumental score building steadily. No text overlay, no watermark. Varied camera work — close-ups, wide shots, mirror reflections, tracking shots. Natural retail lighting.

Create a smooth, realistic fashion walk video using the reference image as the exact character and outfit reference. Preserve her facial features, hairstyle, clothing, body proportions, colors, and overall identity throughout the entire video. Full-body framing, cinematic fashion photography, natural realistic movement, soft daylight. Scene 1 – Walk Forward: She begins standing confidently, then slowly walks toward the camera with natural elegant steps. Her hands remain casually inside the varsity jacket pockets, hair gently moving with each step. Scene 2 – Side Pose: While walking, she slightly turns her body toward the side, looks briefly toward the camera, and gives a soft confident smile. Her hair naturally follows the movement. Scene 3 – Jacket Pose: She stops for a moment, shifts her weight onto one leg, slightly pulls her shoulders back, keeps one hand in the jacket pocket, and casually adjusts the jacket with the other hand. Scene 4 – Final Walk & Turn: She resumes walking past the camera, then gently turns her head over her shoulder with a subtle smile before continuing forward. Her skirt and hair move naturally with the motion. Smooth camera tracking, subtle cinematic camera movement, realistic walking physics, natural facial expressions, realistic fabric and hair motion, no sudden movements, no body distortion, no face changes, no extra fingers or limbs, consistent outfit and appearance, premium fashion campaign aesthetic, vertical 2:3, 8–10 seconds.

生成一条10秒真实旅行Vlog风格短视频。主角始终是同一位年轻亚洲女性，人物五官、发型、服装、身材比例全程保持一致，不换脸、不换衣服。 整个视频分为3个独立场景，北京、上海、广州。城市切换使用“响指后直接硬切”的方式完成，不做背景融合，不做空间扭曲，不做建筑叠化。 0-3秒，北京。 女生站在北京具有代表性的城市地标附近，神态自然放松，双手空闲。她看向镜头，抬起右手打一个响指。响指完成后，画面立刻硬切到上海。 3-6秒，上海。 女生已经站在上海外滩或陆家嘴附近，右手拿着一杯咖啡，自然向前走两步，然后喝一口咖啡。 喝完后，她走到旁边可以放置物品的栏杆平台或小桌边，明确地把右手中的咖啡杯放下。咖啡杯必须真实接触平台，手指完全松开，右手彻底离开咖啡杯。停顿约半秒，确认右手已经空出来。 然后她再抬起右手打一个响指。响指完成后画面立即硬切到广州。咖啡杯留在上海原位置，不随人物瞬移，不允许咖啡杯凭空消失、悬空、变形或突然换手。 6-10秒，广州。 女生空手出现在广州塔附近。她自然向前走两步，转身看一眼广州塔，然后站到合适位置和广州塔合影，最后轻轻看向镜头微笑结束。 全程保持真实手机旅行Vlog质感，轻微手持感，动作自然简单，人物重心和肢体运动符合真实物理规律。 特别强调： 上海段右手拿咖啡 → 喝一口 → 把咖啡杯放到固定平台 → 完全松手 → 右手空出来 → 再打响指。 不允许跳过“放下咖啡”这个动作，不允许咖啡杯直接消失。 避免：人物换脸、服装变化、身体比例变化、人物复制、多余手臂、手指畸形、咖啡杯消失、咖啡杯变形、杯子悬空、杯子换手、人物拿着咖啡同时打响指、建筑融合、北京上海广州地标同时出现、城市背景残影、过度复杂运镜。

Create a cinematic, ultra-realistic live-action sequence set in a warm, dusty neighborhood at golden hour. Start with an aerial view of a lively residential area → close-up of a child placing a worn football on the dusty ground → kids begin playing an intense street football match → dynamic low-angle shots of quick dribbles, footwork and the ball flying through the air → end with a cinematic slow-motion shot of the kids celebrating after scoring. Warm golden sunlight, realistic dust particles, natural movement, handheld camera feel, shallow depth of field, detailed environments, authentic emotions, smooth cinematic transitions, photorealistic live-action, nostalgic coming-of-age atmosphere. No text, subtitles, logos or watermark.

KAZEMIRU — WILD-WIND NIGHT HUNT — 5 SAMURAI Duration: 15s | 16:9 | No Music STYLE: Ultra-photorealistic REAL LIVE-ACTION, AAA Japanese dark-action cinema, restrained semi-CGI VFX. ARRI ALEXA 65, IMAX, Panavision anamorphic. Nighttime Edo Japan, cold moonlight, warm lanterns, deep shadows, mist, realistic atmosphere. @Image1 — STRICT IDENTITY LOCK: Preserve exact male face, facial geometry, white hairstyle, body proportions, torn-white outfit, tattoo, accessories and identity. Exactly ONE male Kazemiru, TWO real metal swords, FIVE adult samurai. Kazemiru ALWAYS remains a photorealistic real human male. COMBAT: Wild, aggressive, acrobatic anime-inspired dual-sword choreography translated into live action. BURST → CROSS-SLASH → PIVOT → REVERSE CUT → WEAVE → DOUBLE SLASH → RE-BURST. Both swords remain physical and clearly visible. ZERO idle, posing, waiting or neutral resets. SUPERSONIC WIND: Kazemiru physically crosses every distance; NO teleportation. Each short-range burst creates restrained emerald-green wind trails, subtle particles, air distortion, dust, leaves, bending lantern flames and realistic clothing reaction. LOCATION: Edo castle district at night: broad dirt road, wooden houses, stone walls, gates, tiled roofs, lanterns and trees. Five samurai surround Kazemiru. CAMERA: Strong 25–35° Dutch angle. Camera is always moving and intentionally slower than Kazemiru. Kazemiru moves first → camera reacts 0.1–0.3s late → whip-pan → overshoots → corrects → reacquires at contact. Use rear OTS, 3/4 tracking, aggressive side tracking and low-angle pursuit. NEVER static, floating or robotic. 00:00–00:03 — #1: Five samurai rush from different directions. Kazemiru bursts forward, air blast throws dust and leaves. Camera loses him, whip-pans, reacquires. He reaches #1 and delivers a brutal dual-sword CROSS-SLASH. Clear contact. #1 collapses. No blood/gore. 00:03–00:06 — #2: #2 attacks. Kazemiru weaves beneath the blade, redirects it with one sword and reverse-slashes with the second. Clear contact. He plants one foot on a stone wall and explosively pushes off. Wind and dust erupt. Camera overshoots toward #3. 00:06–00:09 — #3: Kazemiru flies into frame. Camera catches him mid-air. Real human body twists through a descending dual-blade attack. Both swords remain readable. Clear contact. He lands low and instantly bursts again. 00:09–00:12 — #4: Low Dutch tracking above dirt. #4 charges. Kazemiru weaves past by centimeters, pivots and performs two distinct reverse-slashes. Clear contact. Pressure wave throws dirt outward. Camera loses him and violently reacquires. 00:12–00:15 — #5: #5 searches desperately. Emerald wind crosses background. Kazemiru bursts from the side. He ducks the sword, rotate

Ultra-realistic cinematic action scene inside a high-school basketball gym, a fierce young woman in a stylish school-uniform-inspired outfit unleashes powerful glowing red and purple supernatural energy from her hands, intense confrontation with another girl, dramatic energy trails and particles, dynamic movement, wooden basketball court, bleachers in the background, cinematic lighting, realistic facial details, dramatic atmosphere, motion blur, volumetric light, shallow depth of field, high contrast, photorealistic VFX, 8K detail, vertical 9:16 composition, epic supernatural action movie aesthetic.

Main Subject: Young Korean woman, 24, naturally attractive, realistic skin, minimal makeup, long dark hair loosely down. Wearing an oversized maroon hoodie and loose casual pants, carrying a simple canvas shoulder bag. Preserve her exact identity, facial features, hairstyle, body proportions and appearance throughout. Location: Quiet older Seoul neighborhood on a pleasant morning. Narrow residential street with low-rise apartment buildings, small local shops, parked bicycles, potted plants and modest storefronts. A little Korean girl, around 5–6 years old, is standing safely near the sidewalk with a small ice cream cone in one hand. Her shoelaces have come completely undone. Lighting & Atmosphere: Warm, cheerful and cozy morning atmosphere. Soft sunlight filters between the buildings, creating gentle warm highlights while some areas remain naturally shaded. Muted surroundings, soft shadows, slightly faded colors and a nostalgic early-2000s feeling. The scene should feel wholesome and ordinary rather than sentimental or cinematic. Style: Ultra-realistic early-2000s Sony MiniDV home video filmed by another person casually walking with her. Multi-shot but completely candid and unstaged. Natural handheld movement, subtle camera shake, imperfect framing, gentle reframing, occasional autofocus hunting, mild exposure shifts, faded colors, soft contrast, authentic DV compression, subtle digital noise and natural microphone noise. Shot 1 — 00:00–00:04: Medium-wide handheld shot as the young woman walks casually down the quiet street. She notices a little girl standing near the sidewalk, holding an ice cream cone while looking down at her untied shoelaces with a confused expression. Shot 2 — 00:04–00:07: The camera moves slightly closer as the woman slows down. The little girl looks at her shoelaces, then at her ice cream, clearly unsure how to deal with both at once. The woman understands the problem and gently gestures toward the laces, asking if she can help. Shot 3 — 00:07–00:11: The woman crouches down beside the child and carefully ties the loose shoelaces into a secure bow while the child stands still, carefully holding her ice cream away from them. The woman's movements are gentle and efficient. Shot 4 — 00:11–00:14: She gives the finished shoelaces a quick little check and stands back up. The child looks down at her shoes, then looks up at the woman with a small grateful smile. Shot 5 — 00:14–00:17: The woman gives the child a warm little smile and a tiny nod, then continues walking down the street. The child happily takes another bite of her ice cream. Shot 6 — 00:17–00:20: As the woman walks away, she briefly turns toward the camcorder and gives a quiet, contented smile before looking forward again and continuing down the street. Audio: Natural sound only—soft footsteps, distant traffic, subtle neighborhood ambience, children's distant voices, light breeze, fabric movement and the natural sounds of the street.

Create a 30-second ultra-realistic fictional home-video featuring an adult Korean man in his late 20s spending a cheerful afternoon in an older Seoul residential neighborhood. CHARACTER: A fictional adult Korean man in his late 20s with medium-to-long dark hair tied in a casual low ponytail, natural masculine facial features, realistic skin texture and a friendly expressive personality. Keep his appearance consistent throughout the entire video. OUTFIT: He wears a bold, fashionable summer outfit consisting of a vivid red fitted T-shirt, high-waisted relaxed-fit white denim shorts, clean casual sneakers, a simple bracelet, small understated earrings and a small shoulder bag. The outfit is youthful, colorful and stylish, appropriate for a warm summer day. SCENE 1 — LEAVING HOME He steps out of an older Seoul apartment building carrying a small reusable shopping bag. He closes the gate behind him, notices the camera and gives a spontaneous smile before walking down the quiet residential lane. SCENE 2 — LOCAL BAKERY He enters a tiny neighborhood bakery, looks through the pastry display and chooses a small baked treat. He pays the shopkeeper with coins, thanks them and walks back outside. SCENE 3 — FRIENDLY ENCOUNTER While walking through the narrow lane, he suddenly sees his adult male friend approaching from the opposite direction. They smile, greet each other warmly and share a quick friendly hug. They chat and laugh for a moment before waving goodbye and continuing in opposite directions. SCENE 4 — FRUIT STAND He stops at a small roadside fruit stall and chooses a peach. He pays the vendor, places it inside his shopping bag and continues walking. SCENE 5 — EVERYDAY NEIGHBORHOOD MOMENT He walks beneath trees while eating his pastry. A bicycle passes nearby, and he naturally steps aside. He looks around at the quiet houses and smiles as the camera follows him from a short distance. SCENE 6 — ENDING He reaches the end of the lane, turns toward the camera and gives a cheerful wave. He smiles naturally and says, “Bye bye!” He then turns around and walks away down the residential street while carrying his bag and peach. The camera stays behind him until he disappears around the corner. CAMERA STYLE: Authentic early-2000s consumer MiniDV home-video footage. Handheld camera shake, imperfect framing, autofocus hunting, automatic exposure fluctuations, soft digital detail, mild digital noise, slight compression artifacts, realistic motion blur and occasional awkward zooms. It should feel like a genuine recording made by a friend or family member, not a professional production. AUDIO: Only natural neighborhood sounds: footsteps, bicycles, distant traffic, bakery activity, casual conversations, birds, children playing and light wind. No music, no narration, no subtitles, no text overlays, no beauty filter, no sexualized

Create a cinematic, ultra-realistic 9:16 school classroom scene. A young male student with styled red hair suddenly reacts dramatically as a large burst of flames erupts from a backpack in front of him. Another student in a school uniform sits at a desk in the background, shocked and watching. Warm classroom lighting, realistic fire and smoke effects, dynamic movement, natural facial expressions, detailed environment, cinematic camera motion, shallow depth of field, realistic textures, dramatic atmosphere, high detail, 4K, photorealistic.

Create a high-quality cinematic 3D animated short film featuring adorable, stylized baby otter-like woodland creatures in a beautiful sunny meadow. Scene 1: A cute brown baby otter stands in a lush green meadow, happily holding a large shiny blue berry with both paws. It has soft brown fur, a round chubby body, huge glossy black eyes, rosy cheeks, and a tiny dark scarf around its neck. The character looks innocent and curious. Warm morning sunlight, soft grass, rolling green hills, tall trees, blue sky, cinematic depth of field. Scene 2: The blue berry slips from the otter's paws and rolls quickly down a gentle grassy slope. The little brown otter runs after it, trying desperately to catch it. The camera follows the rolling berry with a smooth dynamic tracking shot while the otter runs behind it. Scene 3: The berry rolls toward a large hollow fallen tree trunk lying in the meadow and stops near its entrance. The brown otter reaches the tree and looks around curiously, wondering where the berry went. Scene 4: The camera moves upward through the surrounding trees toward the bright sunlight shining through the leaves. A magical, peaceful atmosphere fills the scene with glowing sun rays, floating dust particles, and gentle wind moving the leaves. Scene 5: A second adorable baby otter appears — smaller and gray-colored, with soft gray fur, a round chubby body, huge expressive black eyes, rosy cheeks, and an innocent expression. It comes out near the fallen tree and notices the brown otter. Scene 6: The gray otter and brown otter interact playfully near the hollow tree. They look at each other with curiosity and surprise, then move around the tree together. Keep their character designs consistent throughout the entire video. Scene 7: The gray otter climbs onto a nearby tree branch and looks around. The camera follows from behind and slightly below, showing the bright green forest canopy and warm sunlight filtering through the leaves. Scene 8: The gray otter rests against a large moss-covered rock, looking sleepy and exhausted. Its eyes become heavy as it relaxes in the warm sunlight. Scene 9: The gray otter suddenly notices the shiny blue berry again. The camera cuts to a close-up as it happily picks up the berry with both paws. Its huge eyes sparkle with excitement. Scene 10: The gray otter proudly holds the blue berry close to its face and smiles innocently. The brown otter approaches from behind. Final scene: The brown otter gently jumps onto the gray otter in a playful, affectionate hug. They tumble softly onto the grass together beside the blue berry, laughing and cuddling. Tiny glowing sparkles appear around them as they lie peacefully in the meadow. Visual style: premium cinematic 3D animation, adorable family-friendly characters, soft detailed fur, expressive oversized eyes, rounded proportions, vibrant natural environment, warm golden sunlight, volumetric lighting, soft shadows, realistic grass an

FORMAT: 15s / 720P / 16:9. REFERENCES: Image1 is the storyboard and sequential visual guide. SCENE: One career woman races through the final hours of Friday in a modern city office, finishes her last task, and steps into an inviting golden evening. STYLE: contemporary live-action cinematic realism, brisk pacing, cool blue-gray office shifting into amber sunset and warm city lights. LENS / LIGHT / GRADE: 35mm full-frame with 85mm close-ups; cool overhead office light shifting to golden-hour backlight from screen-right; subtle film grain. STATIC LOCKS: same adult woman, face, tailored workwear, hair and accessories throughout; coherent office geography, screen direction and light progression. DYNAMIC MOTION: quick clock glance, purposeful typing, notifications, brisk lateral tracking, decisive laptop close, forward stride through glass doors, gentle city drift, settle into stillness. TIMELINE: 0-1.5s tight clock and anticipatory eyes; 1.5-3.2s wide office typing as colleagues wind down; 3.2-5s notifications then controlled exhale; 5-7s tracking as she finishes a call and gathers her coat; 7-8.8s final email sent, laptop shuts, relief; 8.8-10.8s she exits into golden evening; 10.8-13s city walk opens into warm weekend atmosphere; 13-15s she smiles, phone goes into her bag, holds a calm liberated final frame. AUDIO / VOICE: ticking clock and soft keyboard taps build over light rhythmic instrumental music; laptop click, office ambience recedes into evening street sounds; music opens into warm buoyant weekend energy. No dialogue or voiceover. RULES: one continuous cinematic video, not a storyboard slideshow; no panel borders, captions, masthead, arrows, overlays or readable screen text; no extra featured characters; preserve natural movement and identity. NEGATIVE: no identity drift, frantic unreadable cuts, garbled text, split-screen or frozen montage.

Create a 20-second ultra-realistic cinematic camping vlog featuring a young Japanese woman with straight black hair and bangs, wearing a dusty pink outdoor jacket. Scene 1 (0–4 sec): Selfie-style vlog in a beautiful pine forest during golden hour. She smiles at the camera while holding a beige tent bag and excitedly introduces her solo camping adventure. Natural handheld camera movement, realistic facial expressions. Scene 2 (4–7 sec): Wide cinematic shot of her setting up a beige camping tent among tall pine trees. She happily raises both hands to celebrate after pitching the tent. Warm sunlight filters through the forest. Scene 3 (7–12 sec): Nighttime campfire scene. She sits on a folding camping chair, roasting a marshmallow on a stick. Close-up of her gently blowing on the toasted marshmallow, with realistic firelight illuminating her face. Scene 4 (12–16 sec): She sits beside the campfire holding a glowing marshmallow, enjoying the peaceful night in the forest. Natural expressions, cozy atmosphere, realistic shadows and warm orange lighting. Scene 5 (16–20 sec): Inside her illuminated tent at night, she smiles at the camera while holding a camping lantern. Suddenly, she hears a mysterious noise outside, opens her eyes wide, and pulls her sleeping bag up to her nose with a cute, surprised expression. Visual style: Photorealistic live-action, 4K, cinematic depth of field, natural skin texture, realistic lighting, smooth camera transitions, authentic outdoor atmosphere, detailed facial expressions, cozy solo camping aesthetic. Audio: Natural forest ambience, birds chirping, gentle wind, crackling campfire, subtle fabric sounds, and soft nighttime forest ambience. No background music, no subtitles, no text, no watermark. Consistency: Maintain exactly the same woman, hairstyle, outfit, facial features, and tent throughout all scenes.

Generate a cinematic live-action superhuman fight scene. MAIN CHARACTER: Use image1 as the only main character. A superhuman fighter with extreme speed, strength, and reflexes. Calm expression. Precise martial artist. Every movement is controlled and powerful. OPPONENTS: Use image2 and image3 as main enemies. Add multiple opponents surrounding the main character. Aggressive street fighters. LOCATION: Modern subway station underground. Large platform. Concrete pillars. Subway signs. Metal structures. Train tracks in background. Bright artificial lighting. Passengers watching from a safe distance. FIGHT STYLE: Superhuman close combat. Fast boxing combinations mixed with Taekwondo. Continuous flow. No pause between attacks. Every move transitions naturally. 0–2 seconds: Opponents rush from different directions. Image1 instantly moves forward. Fast combo: left jab → right cross → body punch → spinning back fist. Before the opponent reacts: low kick → side kick → jumping knee strike. Camera follows every impact. 2–5 seconds: The fight becomes extremely fast. Image1 chains: double punch combination, elbow movement, hook punch, roundhouse kick, spinning heel kick, flying kick. Every hit creates powerful cinematic impact. Opponents are pushed backward by the force. Dust and small debris fall from damaged pillars. 5–8 seconds: Multiple enemies attack together. Image1 uses superhuman reflexes: dodges, counters, redirects attacks. Combination: block → counter punch → sweep kick → aerial kick → instant landing. The camera rotates around the action. Motion blur from extreme speed. 8–10 seconds: FINAL MOVE: Image1 charges energy through movement. Explosive combo: rapid punches followed by a powerful spinning kick. A shockwave pushes opponents away. They crash into the environment. Walls crack. Dust fills the subway platform. Final shot: Image1 stands alone in the center. Damaged station. Dust floating. Enemies scattered. Calm expression. CAMERA: Dynamic blockbuster camera. Handheld action style. Fast tracking. Low angles. 360° movement. Slow motion on strongest impacts. STYLE: Ultra realistic live action. High budget action movie. Superhero cinematic quality. Realistic physics. SOUND: Punch impacts. Kick impacts. Footsteps echoing in subway. Air whoosh. Metal vibrations. Crowd reactions. NEGATIVE: cartoon, anime, CGI look, plastic skin, fake physics, static camera, blood, gore, graphic injury, weapons, comedy, low quality,

Create a realistic cinematic superhero speed sequence in a busy downtown city street during autumn, with natural daylight, pedestrians, cars, crosswalks and detailed buildings. Scene 1 — Wide street-level shot: pedestrians cross a busy intersection while cars wait in the background. Camera tracks smoothly from left to right. Scene 2 — A female superhero in a sleek fitted red superhero suit suddenly appears and accelerates across the crosswalk at extreme speed, powerful running posture, flowing hair. Scene 3 — Side-tracking close shot: she races past outdoor café tables and pedestrians, surrounded by intense glowing red energy trails and realistic motion blur. Scene 4 — Low-angle tracking shot: she rapidly approaches a small white dog in the crosswalk, crouches down while still moving at super speed and safely grabs the dog. Scene 5 — Close dynamic shot: she carries the dog securely while sprinting forward, red energy streaks wrapping around her body, strong cinematic motion blur. Scene 6 — Wide establishing shot: she races across the entire intersection at incredible speed, leaving long red light trails behind her as pedestrians react naturally. Scene 7 — Fast lateral camera pan: she blasts past people sitting on a city bench reading newspapers, red energy trails sweeping through the frame. Scene 8 — Final wide-to-medium tracking shot: she continues running rapidly down the street and toward the camera, then passes close to the camera with powerful speed, realistic cinematic lighting and motion blur. Photorealistic, cinematic superhero VFX, realistic human movement, natural city environment, dynamic camera tracking, shallow depth of field, realistic red energy trails, high detail, smooth motion, dramatic speed-ramp feeling, no text, no logos, no watermark.

Created a 20 seconds cinematic Japanese-style scene of a young schoolgirl walking alone through a quiet Japanese residential street during warm golden-hour sunlight, shown from behind as she slowly approaches a railway crossing with a train passing in the background, then transitioning to a close-up of her face and flower-decorated hair as she watches the train, followed by a wider shot of her standing near a quiet canal with soft evening light reflecting on the water, then showing her gently touching small white flowers in a golden field, with close-up shots of her hand and the flowers moving naturally in the breeze, gradually transitioning to a wide sunset landscape with the girl standing peacefully in the field, and finally ending with a warm close-up of her face as she softly looks toward the camera, maintaining the same character, hairstyle, school uniform, realistic skin details, natural expressions, smooth cinematic camera movement, soft depth of field, warm sunset lighting, realistic Japanese environment, subtle film grain, peaceful emotional atmosphere, and a dark fade-out at the very end.