Stable Diffusion Prompts for Realistic Portraits

Include camera model, lens details, lighting setup, etc. to guide the AI towards a more photographic style. Some examples:

  • “Photograph taken with a Canon 5D Mark IV, 85mm lens, soft studio lighting”
  • “Nikon D850, 50mm lens, natural light portrait”

Reference Photographers or Styles

Mention a specific photographer’s style or aesthetic to help define the look you want, like:

  • “Portrait by Annie Leibovitz”
  • “Steve McCurry style portrait photo”

Describe Facial Features

Provide descriptive details on facial features, skin tones, expressions etc. This helps generate more realistic and defined portraits. For example:

  • “Woman with long wavy brown hair, bright green eyes, radiant skin”
  • “Middle aged man with weathered skin, greying beard, kind eyes”

Use Real World Context

Add background details and scene setting to ground the portrait, like:

  • “Woman sitting on a park bench in evening light”
  • “Toddler playing on the beach at sunset”

Add Rendering Quality Keywords

Use terms like “intricate details”, “lifelike”, “photorealistic” to push quality. But avoid overusing them.

Try Multiple Variations

Generating realistic portraits takes experimentation. Tweak prompts and try different combinations of details and styles. Review what works and build on it.

The key is to be very specific in describing what you want to achieve, while giving the AI enough room for interpretation. With practice, you’ll get better at crafting prompts that generate stunning, realistic portraits.