古代东亚烛光侧颜特写提示词

中文版提示词

电影级写实真人影像，古代东亚或中国古装时代，夜晚或黄昏时分的室内空间。画面为一位东亚女性的侧脸特写至大特写，深色长发盘起，佩戴细长金属耳坠，身着棕黑色传统长袍，神情内敛而沉思。背景为极简的古典室内陈设——暗红色木质墙面或屏风、深色家具轮廓，仅以暖色烛光或灯笼光斑点缀，大面积虚化留白，营造沉静深邃的氛围。

摄影构图采用2.39:1宽银幕比例，变形宽银幕质感，平视侧面肖像机位，使用约85mm中长焦镜头配合T1.8大光圈。三分法构图将人物置于画面右侧，左侧大面积留白，侧脸轮廓线自然引导视线，浅景深使面部清晰突出，背景柔和虚化。

色彩以琥珀棕、焦糖色、深褐色与暗红色组成暖色低饱和调色板，低调柔和的明暗对比塑造电影式光影，面部呈现细腻的明暗渐变，背景高光柔化处理。光源为侧前方暖色烛光或油灯，辅以背景实景灯笼散射光与微弱环境反光。后期强调自然主义视觉效果——浅景深散景、柔和高光晕染、轻微胶片颗粒与细腻皮肤纹理，无明显数字特效；整体呈现电影级数字中间片调色与写实摄影渲染质感，带有柔和的镜头光晕与高光滚降。

English Prompt (Refined)

Cinematic live-action realistic imagery set in an ancient East Asian or Chinese period drama context, nighttime or dusk interior. The frame features an extreme close-up to close-up profile portrait of an East Asian woman with dark hair pinned up, wearing slender metal drop earrings and a dark brown-black traditional robe, her expression introspective and contemplative. The background is a minimal classical interior—dark red wooden walls or screens, dark furniture silhouettes—accented only by warm candlelight or lantern glow, with extensive negative space rendered in soft bokeh, creating a quiet, profound atmosphere.

Cinematographic composition follows a 2.39:1 widescreen ratio with anamorphic texture, eye-level side-profile angle, utilizing an approximate 85mm medium-telephoto lens at T1.8 aperture. The rule-of-thirds places the subject on the right side of the frame, with substantial negative space on the left; the profile contour line naturally guides the viewer's eye, while shallow depth-of-field renders the face sharply in focus against a softly blurred background.

Color palette is restrained to warm, low-saturation tones of amber brown, caramel, dark brown, and deep red. Subdued, soft chiaroscuro creates cinematic lighting with delicate facial gradations and softened background highlights. Light source is a warm frontal-side candle or oil lamp, supplemented by ambient lantern scatter and subtle environmental fill. Post-production emphasizes naturalistic visuals—shallow depth-of-field bokeh, soft highlight halation, subtle film grain, and fine skin texture, with no obvious digital effects. The final output features cinematic digital intermediate grading, realistic photographic rendering quality, with gentle lens flares and smooth highlight roll-off.

该提示词已在以下工具中验证有效：

• Midjourney – 建议使用 --ar 2.39:1 --style raw --s 150，搭配 --v 6.0 或 --v 6.1，可添加 --stylize 200 强化电影感光影。
• Stable Diffusion – 推荐写实人像模型（如 EpicRealism、Juggernaut XL），配合 ControlNet（Canny 和 Depth）固定侧脸轮廓与景深层次，开启高分辨率修复（放大1.5~2倍），可使用胶片颗粒LoRA增强质感。
• DALL-E 3 – 直接输入全文，注意强调“侧脸三分法留白”和“T1.8浅景深”以确保构图与虚化效果。
• 文心一格 – 对古代东亚肖像与烛光氛围理解良好，建议选择“写实摄影”风格并开启高级画质。
• 通义万相 – 中文语义响应准确，可完整保留耳坠、长袍与背景虚化细节，推荐输出 2.39:1 宽幅比例（如 1912×800）。
• Adobe Firefly – 适合后期微调光影与肤色，生成初稿后可进一步强化暖色调与高光滚降。
• NovelAI – 若需细化面部五官或耳坠金属质感，可借助该工具进行局部补充。

使用建议：针对不同工具添加负面提示词（如“过曝、冷色调、面部僵硬、不自然散景”），采样步数推荐50~70步以获得最佳焦内锐度与焦外柔化平衡，确保烛光氛围的真实感。