neroidea
由 ChatGPT 根据所示提示词生成,点击图片可全屏查看。

摄影实景转手账风格上下对比海报

ChatGPT海报比例 3:4
提示词来自 @Hamburgerai

Usage: Style Transfer Asset Type: 3:4 vertical editing comparison poster Input Image: A real photo to be processed, and it is the sole source of truth. Create a vertical poster with a strict top-bottom comparison. The upper half must preserve the input photo's clear, recognizable high-fidelity realistic photographic effect: subject identity, quantity, pose, shooting angle, architectural structure, landscape relationships, lighting, and main colors all remain consistent, with only slight print-grade color grading. The lower half redraws the same scene as a warm beige paper page of a travel field journal; it cannot be just a filter, nor a generic travel illustration. Use dark brown pencil and ink outlines with a slight jittery feel, uneven hand pressure, dry brush marks, a small amount of low-saturation watercolor and crayon coloring, paper fibers, erasure marks, and slight ink-bleeding effects. Compress the scene into one main subject or landmark, plus one to two environmental anchors from the original photo's facts; simultaneously change the scale and information density so it is no longer a copy of the same crop and composition. Allow the key subject or landmark to have a touch of restrained surreal cartoonishness—only when facial features or small hands and feet can naturally grow from their real shapes, add simple eyes, mouths, or small limbs. Add 2–4 small hand-drawn icons; these icons must correspond to details actually visible in the original photo. The lower half must maintain about 20%–35% continuous warm-colored paper white space around the edges. Add a small handwritten title of 1–3 English words, the fixed field "FIELD NOTE", and a short observation sentence of 5–12 English words with a slight sense of rhyme on the paper; this sentence can only be written based on facts visible in the image. These texts should look like they are written directly into the paper, not floating digital fonts. Colors are extracted from the input photo and compressed into 2–5 low-saturation colors, plus 1 accent color from the original image; do not apply fixed travel color schemes. The top-bottom boundary must be clean and horizontal. Forbidden: Left-right layout, grids, multi-photo collages, "original/AI" labels, watermarks, logos, QR codes, fake Chinese, fictional landmarks, irrelevant people, changing characters' faces or poses, missing the original photo in the top half, making the entire image cartoonized, and redrawing the original photo with the same crop.

提示词解读

该提示会生成一幅 3:4 垂直排版的海报,将真实照片分成上下两部分:上半部分保留原图的高保真写实效果;下半部分则以温暖米色的手绘旅行记事本页面呈现同一场景。整体氛围兼具现场纪录与童趣速写。 视觉对比主要通过上半部分的轻度打印级色彩分级,下半部分的低饱和水彩与蜡笔渲染实现。底部使用深棕色铅笔与墨线勾勒,杂质感与风化纸纹、擦除痕迹、少量墨水渗染共同营造手绘质感。场景被压缩成主景点与少量环境锚点,允许在自然可放大的面部或手足上添加微小的卡通眼、嘴或小肢体。 想变更该风格,可保持严格的上下分割和上半部分轻微刷色;将旅行本颜色改为自定义低饱和调色板。若想更写实,可去除或限制卡通化点,只保留主体有限的造型。

#poster#opennana

觉得自己能猜中这条 prompt?来玩 Promptle