原则：达芬奇式的主体描述必须包含比例和结构的暗示
✅ 好：
"A seated woman in three-quarter view, hands gently folded,
following classical eight-head proportion, shoulders relaxed
in contrapposto stance"
（一位坐着的四分之三侧面女性，双手轻柔交叠，
遵循古典八头身比例，肩膀以对立平衡姿势放松）
— 包含了姿态、比例标准、身体语言
❌ 差：
"a beautiful Renaissance woman portrait"
（一位美丽的文艺复兴女性肖像）
— 空洞、AI会填入默认的平庸模板
✅ 好（复杂场景）：
"The Last Supper moment: Christ at center of long table,
twelve apostles arranged in four groups of three,
each showing distinct emotional reaction through gesture,
one-point linear perspective with vanishing point
behind Christ's head"
（《最后的晚餐》时刻：基督在长桌中央，
十二门徒分为四组三人排列，每人通过手势表现独特的情绪反应，
单点透视消失点位于基督头部后方）
— 包含了构图逻辑、人物安排、透视方案
Component 2: 光线与空间（Light & Space）— sfumato的灵魂
达芬奇的光线不是"照明"——它是"揭示"和"消融"的工具。
常用光线/空间模板：
【柔和漫射光（最适合sfumato）】
"soft diffused light from an unseen window on the left,
gentle modeling of forms without harsh shadows,
light gradually diminishing into deep shadow areas,
no sharp shadow edges anywhere,
luminous half-tones bridging light and dark seamlessly"
（来自左侧不可见窗户的柔和漫射光，
形体上的温柔塑造没有生硬阴影，
光逐渐减弱进入深邃的阴影区域，
任何地方都没有锐利的阴影边缘，
明亮的中间调无缝连接着光明与黑暗）
【明暗对照法（chiaroscuro）— 戏剧性更强】
"strong chiaroscuro lighting with single light source,
dramatic contrast between illuminated form and deep shadow,
figures emerging from darkness like sculptural volumes,
Rembrandt-influenced but with Leonardo's softer transitions"
（强烈的明暗对照法单光源照明，
被照亮的形体与深邃阴影之间的戏剧性对比，
人物如雕塑般体积从黑暗中浮现，
受伦勃朗影响但带有达芬奇更柔和的过渡）
【大气透视（atmospheric perspective）— 空间的深度】
"deep atmospheric perspective in background,
distant mountains dissolving into blue-gray haze,
color temperature cooling as distance increases,
detail progressively lost in aerial perspective layers,
sense of vast receding space behind the subject"
（背景中的深层大气透视，
远山消融在蓝灰色薄雾中，
颜色温度随距离增加而冷却，
细节在层层大气透视中逐渐丢失，
主体后方巨大的后退空间感）
【莱昂纳多式特殊光效】
"leonardesque lighting with characteristic
subtle smile of illumination on faces,
light that seems to emanate from within the subject,
shadows that are not black but composed of
layered transparent colors (ultramarine, ochre, vermilion)"
（莱昂纳多式光照，特征是脸上
微妙的照明之笑，
光似乎从主体内部发出，
阴影不是黑色的而是由
分层的透明颜色组成（群青、赭石、朱红））
Midjourney v6 推荐配置：
--ar 4:5 或 3:4 （竖构图更适合肖像和宗教题材）
--stylize 200-300 （中等偏高艺术强度，保持古典优雅）
--chaos 10 （低变化度，达芬奇追求精确可控）
--v 6.0 （使用最新版本）
DALL-E 3 建议：
无特殊参数，但在描述末尾加：
"in the style of Leonardo da Vinci,
High Renaissance oil painting,
sfumato technique, golden ratio composition,
anatomical precision combined with ethereal atmosphere"
Stable Diffusion 建议：
CFG Scale: 7-8
Steps: 35-50 （达芬奇风格需要更多步数来渲染细腻渐变）
Sampler: DPM++ 2M Karras 或 Euler a
推荐模型：DreamShaper / RealisticVision / MajicMix
LoRA 推荐：Da Vinci Style LoRA / Renaissance Painter LoRA
Negative prompt: "cartoon, anime, outline, thick lines,
harsh shadows, modern clothing, bright colors,
oversaturated, photorealistic, digital smooth,
3d render, plastic texture"
完整示例
示例1：文艺复兴肖像（标准难度）
Prompt:
A young woman in three-quarter view, seated against a dark
verdant landscape with winding imaginary river,
soft sfumato lighting from upper left, gentle modeling across
facial features without any perceptible outline or contour line,
the famous "ambiguous smile" created by subtle shadows
at the corners of the mouth,
hands delicately folded in foreground, fingers rendered with
anatomical precision yet softened by atmospheric haze,
background featuring atmospheric perspective with distant
blue mountains dissolving into golden horizon mist,
painted in High Renaissance oil painting style,
Leonardo da Vinci inspired, golden ratio facial proportions,
contrapposto posture suggesting quiet dignity,
extremely smooth tonal transitions like smoke dissolving,
inner luminosity of skin, aged oil paint texture with fine craquelure,
--ar 4:5 --stylize 250 --chaos 10 --v 6.0
光线：sfumato from upper left, no outlines, ambiguous smile
达芬奇关键词：High Renaissance / da Vinci / golden ratio / contrapposto / sfumato / atmospheric / inner luminosity / craquelure
参数：--ar 4:5 --stylize 250
示例2：维特鲁威式人体研究（学术向）
Prompt:
An anatomical study of a male figure in the Vitruvian pose,
simultaneously inscribed within a circle and square,
figure with idealized eight-head proportion, arms extended
to touch both circular and rectangular boundaries,
nude male form showing classical muscular definition derived
from careful anatomical observation, yet rendered with artistic
beauty rather than clinical coldness,
the circle representing cosmic harmony, the square representing
earthly order, their intersection at the navel as divine center,
drawn in red chalk style reminiscent of Leonardo's Codex pages,
with proportional annotations and construction lines faintly visible,
sepia-toned aged paper background, Renaissance scientific illustration
aesthetic meets fine art sensibility,
precise yet poetic, mathematical beauty expressed through human form,
--ar 1:1 --stylize 200 --chaos 8 --v 6.0
示例3：《最后的晚餐》式群像（高难度）
Prompt:
A dramatic religious scene with thirteen figures arranged
along one side of a long table,
central figure radiating calm authority while surrounding
twelve figures react with varied emotions — shock, anger,
confusion, grief — each captured in distinct gestural expression,
one-point linear perspective with all orthogonal lines
converging precisely at the central figure's head creating
powerful visual focus,
architectural setting with coffered ceiling receding into depth,
each window revealing different quality of light,
chiaroscuro lighting from the right side of the composition,
figures emerging from shadow into selective illumination,
painted in Leonardo da Vinci's late High Renaissance manner,
pyramidal subgroupings within the overall horizontal arrangement,
psychological narrative conveyed through body language alone,
deteriorated fresco surface texture with visible crack patterns,
sense of historical weight and spiritual gravity,
--ar 16:9 --stylize 280 --chaos 12 --v 6.0
