
咖啡桌邊的法鬥,街景全在糊
人和狗都清楚,背景的路人與車流被拖成線,黑白顆粒很重。
✍️ ‘誰清楚、誰糊’分得很明白:人和狗鎖死對焦,其餘一律拖成方向性的線,這是整張圖的關鍵變數。想複製這種效果,就得在提示詞裡把清晰區和模糊區列成兩張名單。
📋 提示詞 自然語言856 字元
Candid, kinetic, overcast: a square documentary street photograph in black and white, with clearly visible fine film grain, deep charcoal shadows and bright luminous highlights.
Sharp subjects: a blonde woman with a short bob sits at the left in a cream cardigan and jeans, chin resting on her hand. Opposite her, on the right, a tousled-haired man in an overshirt over a white undershirt and tan trousers holds a coffee cup. Dead centre in front, on wet stone paving, sits a merle French bulldog with upright ears. All three faces are crisply focused.
Setting: between two checkered rattan chairs stands a round marble bistro table with two cups, a croissant on a plate and a small flower vase.
Everything else moves: hair tips, leaves, pavement, the street architecture around them, passers-by and traffic are all dragged into directional motion blur.
📋 中文譯寫版 自然語言
抓拍感、動勢、陰天:一張方形黑白紀實街拍,清晰可見的細膩膠片顆粒、深炭色陰影和明亮通透的高光。
清晰的主體:左邊坐著一位金色短波波頭的女士,穿奶油色開衫和牛仔褲,下巴託在手上。她對面右側,一位頭髮蓬鬆的男士,襯衫外套裡是白色內搭,配淺棕色長褲,手裡端著咖啡杯。正前方正中、溼潤的石板路上,坐著一隻立耳的梅爾勒色法國鬥牛犬。三張臉都對焦清晰。
場景:兩把格紋藤椅之間擺著一張圓形大理石小酒館桌,桌上有兩隻杯子、一盤可頌和一小瓶插花。
其餘一切都在動:髮梢、樹葉、地面、周圍的街道建築、路人和車流,全都被拖成有方向的動態模糊。






