使用分镜 @[storyboard ref] 作为编舞、节奏、镜头和运动规划的参考。在内部遵循动作流向、标注、箭头和运动指引，以服务于…

Use the storyboard @[storyboard ref] as choreography, timing, camera and motion-planning reference. Follow action flow, annotations, arrows and movement guides internally for staging and animation logic. Follow the storyboard shot by shot. Do not render storyboard elements in the final video. No arrows, colored lines, movement guides, notes, panel borders, numbers, text overlays, production marks, subtitles, logos or visible annotations. Use the provided character sheet @[character ref] as the final identity reference for Lian Zhi. Preserve face, hair, outfit layers, silhouette, war fan and color design exactly. No redesign. Emotion model: Valence: calm confidence → focused intensity → controlled ferocity Arousal: low → sharp burst peaks → sustained high → composed finish Laban movement: Glide → Float → Slash → Spin → Dab accents → Flow with moments of Bound control Performance language: graceful precision interrupted by explosive acceleration. Strong fan arcs, cloth trails, rotational body mechanics, readable sakuga motion. Camera: handheld cinematic energy, dynamic tracking, orbit moves, low hero angles, momentum-follow camera. Single character only. No dialogue. No subtitles. No visible storyboard elements. No background music or score.

使用分镜 @[storyboard ref] 作为编舞、节奏、镜头和运动规划的参考。在内部遵循动作流向、标注、箭头和运动指引，以服务于舞台调度和动画逻辑。逐镜头按照分镜执行。不要在最终视频中渲染任何分镜元素。不得出现箭头、彩色线条、运动指引、注释、分镜框、数字、文字叠加、制作标记、字幕、标志或可见标注。 使用所提供的角色设定图 @[character ref] 作为莲织（Lian Zhi）的最终身份参考。精确保留面部、发型、服装层次、轮廓剪影、战扇和色彩设计。禁止重新设计。 情绪模型： 效价（Valence）：平静自信 → 专注强烈 → 受控凶猛 唤醒度（Arousal）：低 → 尖锐爆发峰值 → 持续高位 → 沉稳收束 拉班动作（Laban movement）： 滑步（Glide）→ 漂浮（Float）→ 砍切（Slash）→ 旋转（Spin）→ 点触（Dab）重音 → 流畅（Flow）并穿插束缚（Bound）控制瞬间 表演语言：优雅精准中穿插爆发式加速。强烈的扇形弧线、衣料拖尾、旋转的身体力学、可读的作画张数动作（sakuga motion）。 镜头：手持电影感能量、动态跟随、环绕运镜、英雄低角度、跟随动量镜头。 仅单一角色。无对白。无字幕。无可见的分镜元素。无背景音乐或配乐。