GPT Image 2.5 的街头摄影提示词。我已经比最近分享的其它一些提示词更喜欢这个了。, 图片 1
1/1

PROMPT RECORD · 公开作品详情

GPT Image 2.5 的街头摄影提示词。我已经比最近分享的其它一些提示词更喜欢这个了。

公开 AI 图像创作记录,包含原图、提示词、使用模型、作者、标签与来源。

使用模型
GPT-Image-2.5
创作者
@Gdgtify
发布时间
Wed Sep 23 15:26:00 +0000 2026
图片数量
1 张

完整提示词

源提示词按原始换行与标点完整保留
Street photography prompt for GPT Image 2.5. I already like it better than some of the other ones I have shared lately. 

2x2 grid, 16:9, AI picks 4 stunning subjects: // ═══ LEXER: Scene Elements ═══ tokens = [   ENV({{CITY}}, {{TIME_WEATHER}}),   SUBJ({{SUBJECT_ACTION}}, velocity=1.2m/s, awareness=FALSE),   BG({{BACKGROUND_IRONY}}, distance=8m),   FG({{CLIPPING_ELEMENT}}, distance=0.5m),   GEAR(Leica_M6, Summicron_35mm_f2, {{FILM_STOCK}}) ]  // ═══ PARSER: Exposure & Optics Calculator ═══ AST = ExposureTriangle(   // To freeze 1.2m/s at 35mm, we need minimum 1/250s   shutter = "1/250s",       // To keep BG (8m) and Subj (2.5m) in acceptable focus:   aperture = "f/5.6",       // To compensate for 1/250 @ f/5.6 in {{TIME_WEATHER}}:   iso = "800 (pushed_1_stop_from_400)",      // Hyperfocal calculation:   dof_near = 1.8m,   dof_far = 12m,      composition = Composition(     tilt = -3deg,     parallax_shift = Y:-4mm, // Rangefinder viewfinder vs lens axis     fg_clip = TRUE   ) )  // ═══ CODEGEN: Render Pipeline ═══ function render_street_photo(AST):   layer_1 = trace_rays(AST.env, AST .bg)   layer_1.apply_dof_blur(focus_plane=2.5m, aperture=5.6)      layer_2 = trace_rays(AST.subj)   layer_2.apply_motion_blur(vector=AST.subj.velocity * 1/250s)      layer_3 = trace_rays(AST.fg)   layer_3.apply_dof_blur(focus_plane=2.5m, aperture=5.6, clip=TRUE)      composite = z_composite(layer_3, layer_2, layer_1)   composite.apply_parallax_shift(AST.composition.parallax_shift)   composite.apply_horizon_tilt(AST.composition.tilt)      // Film Emulation Pass   composite.apply_silver_halide_grain(iso=800, distribution=Poisson)   composite.apply_push_processing(shadows=crushed, contrast=high)      return composite  execute(render_street_photo(AST))

提示词译文

译文与源提示词分开显示,不替代原文
GPT Image 2.5 的街头摄影提示词。我已经比最近分享的其它一些提示词更喜欢这个了。

2x2 网格,16:9,AI 挑选 4 个惊艳的主体:// ═══ LEXER:场景元素 ═══ tokens = [   ENV({{CITY}}, {{TIME_WEATHER}}),   SUBJ({{SUBJECT_ACTION}}, velocity=1.2m/s, awareness=FALSE),   BG({{BACKGROUND_IRONY}}, distance=8m),   FG({{CLIPPING_ELEMENT}}, distance=0.5m),   GEAR(Leica_M6, Summicron_35mm_f2, {{FILM_STOCK}}) ]  // ═══ PARSER:曝光与光学计算器 ═══ AST = ExposureTriangle(   // 为了在 35mm 定格 1.2m/s 的运动,我们需要至少 1/250s   shutter = "1/250s",       // 为了让 BG(8m)和 Subj(2.5m)都处于可接受的清晰范围内:   aperture = "f/5.6",       // 为了在 {{TIME_WEATHER}} 下补偿 1/250 @ f/5.6:   iso = "800 (pushed_1_stop_from_400)",      // 超焦距计算:   dof_near = 1.8m,   dof_far = 12m,      composition = Composition(     tilt = -3deg,     parallax_shift = Y:-4mm, // 旁轴取景器与镜头光轴的视差     fg_clip = TRUE   ) )  // ═══ CODEGEN:渲染管线 ═══ function render_street_photo(AST):   layer_1 = trace_rays(AST.env, AST .bg)   layer_1.apply_dof_blur(focus_plane=2.5m, aperture=5.6)      layer_2 = trace_rays(AST.subj)   layer_2.apply_motion_blur(vector=AST.subj.velocity * 1/250s)      layer_3 = trace_rays(AST.fg)   layer_3.apply_dof_blur(focus_plane=2.5m, aperture=5.6, clip=TRUE)      composite = z_composite(layer_3, layer_2, layer_1)   composite.apply_parallax_shift(AST.composition.parallax_shift)   composite.apply_horizon_tilt(AST.composition.tilt)      // 胶片模拟 Pass   composite.apply_silver_halide_grain(iso=800, distribution=Poisson)   composite.apply_push_processing(shadows=crushed, contrast=high)      return composite  execute(render_street_photo(AST))
35mm 胶片动态模糊浅景深暖色调多分屏多面板布局文字主导群像电影感摄影风格迁移文生图