mozgi
Seedance 2.0
Two swordswomen in hanfu, one in white and one in green, prepare a leap across a supposed abyss in mountain mist, but the fog clears to reveal a narrow canal
Prompt
1. SHOT LOCATION Cinematic photorealism. Use a “reference-image-driven living world system,” combining poetic xianxia tension, spatial-misdirection comedy, restrained performances, a cool silver-blue dawn, warm sunrise reflections, natural volumetric mountain mist, and realistic wet stone and water physics. The final twist must be achieved entirely through a change in visibility. The geography must remain continuous at all times; sudden location changes and terrain deformations are strictly prohibited. 2. REFERENCE ROLES Seedance 2.0 supports image, audio, and video references, so all reference images in this prompt must be treated as a single world-building library, not as isolated background images. The reference images collectively define: continuous terrain architectural logic spatial scale material age ecology and vegetation water level and waterways weather conditions cloud movement mist density key light direction reflection relationships perceived distance of sounds realistic paths the characters can travel Internal environment simulation before generation First, run one silent environment simulation. Combine all reference images and rearrange them into a new space unique to this iteration. This space must be: physically coherent structurally continuous realistically navigable by the camera anchored by stable foreground, midground, and background elements consistent with the environmental DNA of the same world 3. CHARACTER LOCK Character ID A｜White-clad sword immortal The same East Asian sword-wielding woman, aged 25–30. Strictly preserve: the same face as @图片1 long black hair slender body proportions white embroidered silk hanfu jade ornaments white cloth boots one and only one silver longsword Character ID B｜Green-clad rival The same East Asian rival woman, aged 20–25. Strictly preserve: the same face as @图片2 black braided hair petite body proportions blue-green hanfu dark cloth shoes one and only one dark steel sword 4. STORYBOARD Shot 1｜0–5s｜wide or long shot Shot objective Establish a misleading spatial relationship, making the audience believe along with the two characters that an enormous, bottomless ravine lies between them. Scene content Both women are shown full-length, from shoes to head, standing on two ancient stone platforms covered in wet moss. Dense clouds and mist obscure the space between them, making it look as though they are separated by a ten-thousand-zhang-deep ravine. Environmental anchors that must be preserved A waterfall disappears into the layer of mist Pine branches at different depths sway in the wind by different amounts Water droplets slowly trickle across the stone surface Roofed buildings in the distance appear and disappear in the mountain mist The rising sun continues to push the mist sideways Performance Both women remain cautious and restrained, as if facing a formidable enemy. They do not rush into action yet; first, let the spatial misdirection take hold. Shot 2｜5–10s｜medium or cowboy shot Shot objective Both women misjudge the scale of the space, believing a real ravine lies between them, so they simultaneously prepare to use a formal, energy-intensive sword-flying move to cross it. Continuity requirements Preserve: the same two women the same outfits the same longswords exactly the same geographical space the same mist and stone-platform layout the same environmental anchors, including the pine forest, waterfall, and roofed buildings Action Both women simultaneously enter a preparatory stance: wide sleeves are gently lifted by the airflow the blades emit a low resonance mountain mist swirls around their ankles postures are solemn, formal, and restrained Dialogue The white-clad sword immortal says in a low voice: “This strike will carry me across the ravine.” The woman in green replies: “I’ll be waiting.” Camera The camera slowly orbits the two women from the side. Throughout the orbit, environmental anchors such as the stone platforms, pine forest, waterfall, and roofed buildings must remain visible so the spatial layout can be verified. Shot 3｜10–15s｜close-up or extreme close-up Shot objective Deliver the twist through a change in visibility. The truth must come from previously established environmental features becoming visible, not from generating a new environmental explanation. Key event A natural mountain breeze, consistent with the weather established earlier, blows away the last layer of dense mist between the women. Truth revealed Clearly reveal that the so-called “ten-thousand-zhang-deep ravine” is actually just an old mountain irrigation channel about three steps wide. The dark water, reflections of the vertical stone walls, and dense mist together created the false impression of depth. Hard constraints When the truth is revealed, all of the following must remain completely continuous: the same edges of the stone platforms the same waterfall flowing into the channel the same moss the same pine trees the same roofed buildings the same direction of the rising sun the same movement logic of the mountain mist Nothing may suddenly deform. Performance and action Both women freeze with their raised longswords, ready to perform their ultimate moves. The woman in green slowly looks down at the channel and says: “How about… we just walk across?” The white-clad sword immortal silently lowers her sword. She calmly walks to the opposite bank by stepping on three stones in the channel that were there all along. As she passes the woman in green, she says without changing expression: “Save our energy.” Ending An extreme close-up captures the woman in green’s mouth as she struggles to suppress a smile. Meanwhile, once the spatial relationship is revealed, the previously grand “ravine roar” of water suddenly sounds like nothing more than a small nearby channel gently trickling. 5. AUDIOVISUAL AND PHYSICS REQUIREMENTS Native, synchronized dialogue in Standard Mandarin Spatial environmental audio must remain consistent throughout Mountain wind, mist, flowing water, waterfall, channel, reflections, and the sense of wet stone must all follow a unified physical logic Foreground, midground, and background must remain in continuous, natural motion Realistic parallax is required Natural light changes continuously, but the changes must be restrained and smooth Silk fabric, hair, water droplets, vegetation, and mist movement must look believable 6. TECHNICAL SPECIFICATIONS Strict total duration: 15 seconds Aspect ratio: 16:9 landscape Number of shots: 3 clear, continuous shots Only two visible characters at all times Do not generate subtitles 7. 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
1. 片段定位 电影级写实质感。 采用“参考图驱动的活世界系统”，融合诗意仙侠紧张感、空间误导式喜剧、克制表演、清冷银蓝色清晨、温暖日出反光、自然体积山雾，以及真实湿润石材和水体物理效果。 最终反转必须完全通过 可见度变化 完成。 地理结构始终连续，绝不允许突然换景，绝不允许地形变形。 2. 参考职责 Seedance 2.0 支持图像、音频、视频参考，因此本次所有参考图统一视为 同一个世界设定库，不得把它们当作彼此孤立的背景图。 参考图共同负责 连续地形 建筑逻辑 空间尺度 材质年代 植被生态 水位与水系 天气状态 云层运动 雾气密度 主光方向 反射关系 声音距离感 人物真实可行的行进路线 生成前的内部环境推演 先执行一轮无声环境推演。 综合全部参考图，重新排列出一个本轮独有的新空间。 这个空间必须： 物理统一 结构连续 可被摄影机真实调度 拥有稳定前景、中景、远景锚点 保持同一世界的环境 DNA 3. 角色锁定 角色 ID A｜白衣剑仙 同一个 25–30 岁东亚剑仙女子。 严格保持： @图片1 同一张脸 黑色长发 纤细身材比例 白色刺绣真丝汉服 玉石饰品 白色布靴 唯一一柄银色长剑 角色 ID B｜青衣宿敌 同一个 20–25 岁东亚宿敌女子。 严格保持： @图片2 同一张脸 黑色编发 娇小身材比例 青绿色汉服 深色布鞋 唯一一柄深色钢剑 4. 分镜设计 Shot 1｜0–5s｜全景或远景 画面目标 建立误导性的空间关系，让观众与两位角色一起相信前方隔着一座深不见底的巨大峡谷。 场景内容 两人从鞋到头完整站在两块布满湿润青苔的古老石台上。 她们之间被浓重云雾遮挡，看起来仿佛隔着一座万丈深谷。 必须保留的环境锚点 瀑布落入雾层后消失 不同纵深的松枝以不同幅度随风摇动 石面水珠缓慢流动 远处瓦顶建筑在山雾中时隐时现 正在升起的朝阳持续从侧面推动雾层横向移动 表演 两人都保持谨慎、克制、如临大敌的状态。 暂不急着动作，先让空间误导成立。 Shot 2｜5–10s｜中景或牛仔景 画面目标 两人都误判空间尺度，以为彼此之间隔着真正深谷，于是同时准备使用高消耗的正式御剑跨谷招式。 连续性要求 必须保持： 同样两名女子 同样服装 同样长剑 完全相同的地理空间 同一雾层与石台逻辑 同样的松林、瀑布、瓦顶建筑等环境锚点 动作 两个人同时进入蓄势状态： 宽袖被气流轻轻托起 剑身发出低沉共鸣 山雾沿脚踝旋转 身体姿态庄严、正式、克制 对白 白衣剑仙沉声说道： “这一剑，跨谷而来。” 青衣女子回答： “我等着。” 摄影机 摄影机沿两人侧面缓慢环绕。 环绕过程中必须始终保留石台、松林、瀑布和瓦顶建筑等环境锚点，保证空间可验证。 Shot 3｜10–15s｜特写或极近特写 画面目标 通过可见度变化完成反转。 真相必须来自前面已经存在的环境结构被看清，而不是新生成一个环境答案。 核心事件 一阵符合前面天气逻辑的自然山风，吹散两人之间最后一层浓雾。 真相揭示 清楚揭示所谓“万丈深谷”，其实只是一条 宽约三步的古老山间引水渠。 深色水面、垂直石壁倒影和浓雾共同造成了错误深度判断。 硬约束 揭示时必须保持以下内容完全连续： 同样的石台边缘 同样的瀑布汇入水渠 同样的青苔 同样的松树 同样的瓦顶建筑 同样的朝阳方向 同样的山雾运动逻辑 任何东西都不能突然变形。 表演与动作 两个人举着准备施展绝技的长剑，同时僵住。 青衣女子缓慢低头看一眼水渠，说： “要不…… 直接过去？” 白衣剑仙沉默收势。 她踩着水渠中三块原本就存在的石头，平静走到对岸。 经过青衣女子身边时，面不改色地说道： “省灵力。” 收尾 极近特写捕捉青衣女子努力憋住笑意的嘴角。 与此同时，先前恢宏的“峡谷水声”在空间关系揭晓后，突然显得只是近处一条小水渠的潺潺声。 5. 声画与物理要求 原生同步普通话对白 空间环境声音必须前后一致 山风、雾、水流、瀑布、水渠、倒影、石材湿润感都要符合统一物理逻辑 前景、中景、远景始终持续自然运动 必须有真实视差 自然光持续变化，但变化必须克制、连续 真丝衣料、发丝、水珠、植被、雾气运动真实可信 6. 技术规格 严格总时长：15 秒 画幅：16:9 横屏 镜头数：3 个连续清晰镜头 始终只出现两名可见角色 不生成字幕 7. 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
- Needs your photo: attach an image to the request — the prompt works with it.
The rights to this prompt belong to its author @Soranlan. You are responsible for any commercial use — promoting AI tools, aggregators or Telegram bots: mozgi.io accepts no liability. Details →
