Avoid the "AI plastic look" on generated human portraits and architectural scenes by structuring your prompts with specific optical and lighting cues:
markdown
[Subject] candid street portrait of an elderly watchmaker in a sunlit Prague workshop, intricate brass clock gears, fine wrinkles, natural skin texture with subtle imperfections [Optics] 35mm anamorphic lens, f/1.8 shallow depth of field, natural morning volumetric dust rays [Film Stock] shot on Kodak Portra 400, fine grain [Parameters] --ar 16:9 --v 6.1 --style raw --stylize 120