style: id: film-sketch name: 电影素描 name_en: Film Sketch category: [编辑视觉, 场景, 建筑] visual_dna: line: {organic: 78, precision: 22} color: {saturation: 32, warmth: 68} lighting: {softness: 34, contrast: 66} texture: {material: 72, clean: 28} detail: {density: 69} depth: {cinematic: 88} mood: {energy: 45} composition: depth: cinematic framing: 宽幅电影式取景，水平线条主导 subjects: [场景, 建筑] models: [GPT Image, Midjourney, Flux] recipe: identity: 电影分镜转译为铅笔与色粉 line: 手势驱动的铅笔排线 color: 去饱和的胶片色域 lighting: 方向性主光，强明暗对比 material: 铅笔、色粉、粗纹纸 texture: 模拟颗粒与纸纹叠加 character: 写实但被手绘柔化 atmosphere: 沉静、戏剧性、克制 negative: [cartoon, flat vector, oversaturated] reference_notes: 电影分镜稿 / 黑泽明构图分析 constraints: anchor: 主视觉即唯一锚点：明暗关系与水平线取景以它为准；换主体不换光影结构。 bans: # 按优先级，前置到提示词 - no hard ink outlines - no evenly spaced parallel hatching covering the frame - no flat frontal lighting fail_if: - no clear terminator between the brightest and darkest areas - horizontal lines do not dominate the framing - pure color blocks beyond the desaturated film range pass: 应像分镜稿放大到银幕尺寸仍然成立。