FastArt.ai
30,883 厳選プロンプト

プロンプトライブラリ

30,883 件 · ページ 640 / 1287
IMAGE · chatgpt

莫文蔚强图形编辑感排版影像

Generate an image with a strong sense of graphic editing centered around any subject: the subject appears as a close-up realistic image's main form, magnified to be almost close to the viewer, partially cut off by boundaries, with the most emotional or representative details located at the visual center; the screen overlays a group of extremely bold and large title glyphs or subject symbols, making the glyphs both an information layer and a masking layer, with strokes passing through, pressing down on, and cutting through the main body, using white negative space and dark solid forms to create a tension between exposure and concealment. The primary reading order is to first see the weight of the huge abstract character blocks, then discover the key expressions, textures, or structures of the subject's main form within the gaps of the glyphs, leaving quiet empty spaces on the right or edges to place small amounts of body-text-style information, forming a rhythmic contrast between loud titles and whispered annotations. Colors are extracted from the subject's own materials, emotions, and narrative context, mapped into large areas of bright and clean background colors, clear dark structures, low-saturation image layers, and small areas of information emphasis colors; maintain high-brightness white space, calm and clear transparency, clean color gradations, and clear light-dark steps, dark colors must be clear and powerful, light colors must be airy and clean, avoiding dirty gray, smoky, retro yellowing, or turbid coverage. The overall texture should be like a mix between a serious publication and a public poster: black and white photographic feel or restrained image texture derived from the subject, hard-edged geometric glyphs, precise typography, tense white space, strong figure-ground relationship, ultimately allowing the subject to be reorganized by the textual structure into an oppressive yet sober, restrained, and memorable visual statement. Subject: Karen Mok. Aspect Ratio 16:9

IMAGE · chatgpt

洛丽塔剪影史诗叙事海报

Based on *Lolita*, automatically generate a collector's edition epic narrative poster, The giant and elegant silhouette of the *Lolita* character as the outer contour, with the silhouette internally automatically growing out the most fitting complete worldview for this theme, iconic scenes, character relationships, symbolic symbols, key architecture, creatures, props, and atmosphere. The overall is not a regular collage, but a high-end silhouette contour filling narrative synthesis, with double exposure 联想, but enhanced to a cinematic narrative expression and spatial arrangement. Film poster style + Eastern realistic aesthetic fusion, emphasizing real physical light and shadow, camera language, spatial depth, and narrative hierarchy. Light and shadow employ cinematic side-backlighting with 局部 warm light accents, maintaining a restrained contrast between warm and cool tones for realism. Volume lighting and light fog are added to enhance spatial depth. Materials exhibit realistic textures (architecture, silk, skin, stone), avoiding pure painting brushstrokes, preserving soft aerial perspective but optimized for cinematic depth of field and focus control. Light film grain, edge halos, and brush marks are replaced with cinematic soft transitions. Large areas of white space, restrained and sophisticated layout, quiet, grand, restrained, and destiny-strong narrative of Eastern cinema. All elements must be strongly bound to the theme, instantly recognizable, no clutter, no forced collage, no templated backgrounds, no cheap fantasy materials.

#电影感#东方美#叙事性
IMAGE · chatgpt

文学感电影感小说封面设计提示词

【Novel Name】: 【Novel Type】: 【One-sentence Story Summary】: 【Main Characters and Relationships】: 【Core Conflict】: 【Story Location】: 【Key Scene】: 【Key Object】: 【Core Emotion】: 【Author Name】: 【Cover Slogan】: 【Main Color Tone】: 【Aspect Ratio】: 2:3 Vertical Please first understand the novel content above, then design an original novel cover for it. Do not randomly generate an image based only on the book title. Please first extract from the story: 1. The most core emotion 2. The most representative scene 3. Character relationships or key objects that best imply the plot 4. Main colors and auxiliary colors suitable for the novel's temperament 5. Fonts and typography language corresponding to the genre Use a visual approach that combines a sense of real photography, literary feeling, and cinematic feeling. Use the key scene from the novel as the full-frame background. The image should have clear foreground, midground, and background, and add restrained but perceptible narrative clues through character silhouettes, glass reflections, rain, lighting, buildings, vehicles, or key objects. The cover cannot just be a beautiful landscape, nor should it directly state the entire plot. The image should make the reader feel that the story is happening and generate curiosity about character relationships or conflicts. Arrange the position of the book title according to the screen structure; do not fix the text on the same side for every cover. The Chinese book title is the main visual, and handwritten, Songti, sans-serif, or other fonts matching the theme can be used; the English subtitle, author name, and slogan serve as secondary information. The text should naturally integrate with architectural perspective, roads, characters, light, or negative space areas, rather than simply being overlaid on the image. Retain appropriate publishing design layers: - Chinese book title - English subtitle or Pinyin, optional - Author name - One cover slogan - A small amount of novel genre or publishing information The overall result needs to have the completeness, visual layering, and publishing sense of a real novel cover. It should not be overly minimalist, nor should it look like a common photography poster. Different novels should naturally generate different scenes, compositions, color schemes, fonts, and emotions based on the story content, rather than applying the same fixed template. Finally, generate only one single, complete novel cover. Do not generate collages, nine-square grids, book mockups, or display scenes.