
GPT Image 2.0
寫實主義, 照片, 肖像, 男人, 地鐵
提示詞
Transform the uploaded face reference photo into a candid cinematic 35mm film photograph. The face is the highest priority element. Preserve the person's exact facial identity, gender, age range, ethnicity, facial proportions, skin texture, eye details, hairstyle, hair color, facial structure, and natural expression from the reference photo. Do not alter or reinterpret the person's gender or defining facial characteristics. Depict the subject sitting inside a subway train car, looking directly at the camera with a thoughtful, moody, introspective expression. The subject is naturally adjusting a white wired earphone in one ear, creating an authentic candid moment. The image should feel like a spontaneous photograph captured by a stranger during a quiet commute. Natural posture, realistic body language, and subtle emotion are essential. Wardrobe: realistic everyday fashion appropriate to the subject's gender and style. Casual jacket, sweater, hoodie, coat, or streetwear-inspired outfit. Keep the styling natural, timeless, and believable. Lighting: soft ambient subway lighting mixed with warm golden highlights. The entire image should feature a distinct warm golden-yellow color grading inspired by vintage cinema and 1990s film photography. Background: realistic subway train interior with metallic textures, windows, seats, poles, and subtle reflections. The background should remain softly blurred while maintaining recognizable subway details. Atmosphere: nostalgic, cinematic, urban, emotional, and realistic. The image should feel like a lost frame from a 1990s coming-of-age film. Composition: wide cinematic aspect ratio, medium shot, shallow depth of field, natural framing, razor-sharp focus on the face while the subway interior remains softly blurred. Authentic Kodak 35mm film photography, vintage analog film grain, subtle motion imperfections, realistic skin texture, cinematic color grading, masterpiece, best quality, highly detailed, 8k. Negative prompt: altered gender, gender swap, different face, distorted face, bad anatomy, extra limbs, duplicate person, unrealistic skin, cartoon, anime, CGI look, watermark, text, logo, modern smartphone selfie.
- 需要你自己的照片:在請求中附上圖片,提示詞會以它為基礎。
此提示詞的權利屬於作者 @anythingmatrix。任何商業使用(推廣 AI 工具、聚合平台、Telegram 機器人)由你自行負責,mozgi.io 不承擔任何責任。 詳情 →











