四模型游牧民族肖像对比

Goal: Create a wide horizontal model-comparison banner showing four side-by-side photorealistic portrait results of the same prompt, each labeled underneath. Canvas: Extra-wide landscape aspect ratio, approximately 4:1, with a thin white label strip along the bottom. The image area is divided into exactly 4 equal vertical portrait panels separated by clean straight boundaries. Layout: Four discrete panels from left to right, each occupying one quarter of the width. Beneath the panels, place one black text label centered or left-aligned within each panel’s bottom white strip. Use exactly these 4 labels: “GPT Image 2”, “Reve 2.1”, “Seedream 5.0 Pro”, and “Gemini 3.1 Flash 2K”. Subject details: Each panel depicts a realistic cinematic portrait of {argument name="subject" default="a Mongolian or Central Asian nomadic woman in traditional winter clothing"}. She wears a large fur hat with ornate embroidered fabric, heavy layered fur or wool garments, braids, weathered skin, and a serious calm expression. The portraits should look like different generations or variations of the same cultural theme: panel 1 is a close-up of a middle-aged woman facing slightly left with warm side lighting; panel 2 is a stern woman with two braids and a burgundy embroidered fur hat, looking off to the right, with a hint of an eagle at the lower left edge; panel 3 is an elderly woman with rosy weathered cheeks and gray braids, wearing a richly decorated fur hat against a sunset mountain valley; panel 4 is an older woman in a dark coat and fur hat holding a small object or glove, with a rider or eagle hunter and a white yurt/ger in the distant background. Environment: Use {argument name="setting" default="a rugged steppe and mountain landscape at golden hour"}, with shallow depth of field, muted earth tones, distant mountains, open grassland, and warm dramatic light. Visual style: Hyperrealistic documentary photography, cinematic color grading, natural skin texture, high detail in fur and embroidered textiles, medium-to-close portrait framing, soft bokeh backgrounds, subtle film grain, no cartoon style. Constraints: Include exactly 4 portrait panels and exactly 4 bottom labels. Keep the bottom label strip plain white with simple black sans-serif text. Do not add logos, watermarks, extra captions, borders, UI chrome, or additional panels.

目标：创建一个宽幅横向模型对比横幅，展示同一提示词生成的四个并排照片级真实肖像结果，并在下方标注对应模型。 画布：超宽横向比例，约为 4:1，底部带有细长的白色标签条。图像区域被精确划分为 4 个相等的垂直肖像面板，中间由清晰的直线分隔。 布局：从左到右四个独立的面板，每个面板占据宽度的四分之一。在面板下方，于每个面板底部的白色条带内居中或左对齐放置一个黑色文本标签。请务必使用以下 4 个标签：“GPT Image 2”、“Reve 2.1”、“Seedream 5.0 Pro”以及“Gemini 3.1 Flash 2K”。 主体细节：每个面板描绘一张写实的电影质感肖像，主体为 {argument name="subject" default="一位身着传统冬装的蒙古族或中亚游牧女性"}。她戴着饰有华丽刺绣的大皮帽，身穿厚重的多层皮草或羊毛服装，梳着辫子，皮肤饱经风霜，表情严肃而平静。肖像应呈现出同一文化主题下的不同世代或变体：面板 1 为一位中年女性的特写，面部略微朝左，带有温暖的侧光；面板 2 为一位表情严厉的女性，梳着两条辫子，戴着酒红色刺绣皮帽，看向右侧，左下角隐约可见一只鹰；面板 3 为一位老年女性，脸颊红润且饱经风霜，梳着灰白辫子，戴着装饰华丽的皮帽，背景为日落时的山谷；面板 4 为一位年长女性，身穿深色外套和皮帽，手持小物件或手套，远景背景中有骑手或猎鹰人以及白色的蒙古包 (ger)。 环境：使用 {argument name="setting" default="黄金时刻崎岖的草原和山地景观"}，具有浅景深、柔和的大地色调、远山、开阔的草地以及温暖且富有戏剧性的光影。 视觉风格：超写实纪录片摄影，电影级调色，自然的皮肤纹理，皮草和刺绣织物的高细节表现，中近景肖像构图，柔和的背景虚化，细腻的胶片颗粒感，拒绝卡通风格。 约束条件：必须包含 4 个肖像面板和 4 个底部标签。底部标签条保持纯白色，使用简洁的黑色无衬线字体。请勿添加任何徽标、水印、额外标题、边框、UI 元素或其他面板。