并排 3D 角色半身像预览
根据参考角色创建两个匹配的正面 3D 风格半身像预览，用于测试头部与身体的拼接工作流。
gpt-image-2 · 3D 渲染 · 人像 / 自拍 · 极简主义 · 游戏素材 · 角色
中文提示词
以 REFERENCE_0 为角色基础，将其转换为适合头部与身体拼接测试的简洁 3D 模型预览。将角色调整为基本正面的上半身半身像姿态，简化轮廓，并使头部与颈部的连接清晰可见，同时保留参考图中令人不安的玩偶般面孔、角状侧刺、包裹式织物服装、悬垂的绳索以及磨损的织物纹理。将生成的角色渲染图放置 2 个完全相同的副本，左右并排，比例和姿势保持一致以便对比。使用纯色 {argument name="background color" default="深炭灰色"} 背景，采用柔和的摄影棚灯光，呈现出略带渲染感或摄影测量风格的效果，而非精修插画。裁剪为宽屏横向画幅，仅展示头部、肩部和上躯干。避免添加文字、Logo、道具或额外角色。
原始提示词
Using REFERENCE_0 as the character base, convert the subject into a clean 3D-model preview suitable for head-and-body joining tests. Repose the character to a mostly front-facing upper-body bust, simplify the silhouette, and make the head/neck connection clearly visible while preserving the unsettling doll-like face, horn-like side spikes, wrapped textile outfit, dangling cords, and worn fabric textures from the reference. Place exactly 2 identical copies of the resulting character render side by side, one on the left and one on the right, with matching scale and pose for comparison. Use a plain {argument name="background color" default="dark charcoal gray"} background, soft studio lighting, and a slightly rendered/photogrammetry look rather than a polished illustration. Crop to a wide landscape frame showing the head, shoulders, and upper torso only. Avoid adding text, logos, props, or extra characters.