AI 5 min read

Marcos shares a GPT Image 2 prompt for fashion portrait in black and white with double exposure

@arsalannazir07 publishes a JSON prompt for ChatGPT in X: 4:5 monochrome editorial portrait with wool coat, ghost silhouettes and analog aesthetics.

Verified Editorial standards
GPT Image 2 portrait in black and white with double exposure: dark wool coat, translucent ghost silhouettes and analog editorial aesthetic
Monochrome portrait with double exposure, structured coat and white studio background. Source: @arsalannazir07

On June 19, Marcos (@arsalannazir07) posted on X a JSON-structured prompt block — marked “created on @ChatGPTapp” — for GPT Image 2. The template generates a hyperrealistic fashion portrait in black and white with double exposure: it preserves the facial identity of the uploaded photo, fixes a dark wool coat on a ribbed sweater, adds two translucent ghost silhouettes behind the face and emulates analog editorial photography with grain, vintage border and 85mm lens.

AI Prompt · How to do it

Prompt result: B/W portrait with double exposure, sharp centered face and soft ghost silhouettes

Example posted by Marcos on June 19, 2026. @arsalannazir07

Steps

  1. Open chatgpt.com with GPT Image 2 active.
  2. Upload your reference photo — the prompt replaces only the original subject.
  3. Copy the main prompt below (in English).
  4. If the result comes out with color or artifacts, add the negative prompt to the end of the message.
  5. Ask for 4:5 vertical format if the proportion is not respected.

Prompt to copy

Create a hyper-realistic black and white double-exposure fashion portrait based on the uploaded reference image. Keep the exact same vertical framing, centered composition, camera angle, medium portrait crop, subject distance, body posture, coat shape, background, lighting, ghosting effect, double exposure placement, contrast, grain, and vintage film atmosphere. Replace only the original person with the target character from the reference image. Preserve the target character's facial identity, face shape, hairstyle, facial proportions, eyebrows, eyes, nose, lips, jawline, facial hair, skin-tone values ​​translated into monochrome, and serious natural expression with maximum accuracy. The target character stands upright facing forward with relaxed shoulders, wearing a premium dark wool coat with a structured high collar over a light ribbed knit sweater. The main face remains centered, sharp, and dominant. Behind the subject, include two elegant double-exposure ghost silhouettes: a semi-transparent side-profile silhouette slightly offset behind the left side of the face and a larger soft translucent silhouette behind the right side of the head. Blend both naturally into the pale background with medium-low intensity, realistic photographic layering, and no digital-looking artifacts. Maintain soft bright white studio background, gentle frontal lighting, deep shadows on one side of the coat, smooth grayscale transitions, subtle analog film grain, soft edge falloff, slight vintage border texture, and premium editorial fashion aesthetics. Shot with an 85mm realistic portrait lens, shallow depth of field, monochrome skin texture, cinematic black-and-white analog film look, soft focus edges, high-end magazine photography quality. No text, no logos, no watermark, no color, no surreal distortions, no extra objects.

Negative prompt (optional)

color, oversaturated tones, cartoon style, anime rendering, low resolution, blurry face, distorted facial features, extra eyes, extra limbs, duplicated body parts, unrealistic ghosting, strong digital effects, harsh HDR, heavy vignette, readable text, logo, watermark, poster layout, frame artifacts, exaggerated double exposure, background clutter, low detail skin, plastic skin, over-sharpening

Text in English according to the original publication in JSON format. Only use your own photos or with explicit permission.

What makes this prompt special

It doesn't ask for a simple "black and white portrait": it sets a complete photographic template - centered vertical frame, medium distance, upright posture, coat shape, soft front lighting and exact placement of double exposure layers - and only allows you to change the subject to the reference one. This reduces variability and brings the result closer to a reproducible study session.

The JSON block from the original post also specifies aspect_ratio 4:5, 85mm lens, shallow depth of field, analog grain, vintage edge, and soft blur edges. The negative prompt blocks color, anime style, exaggerated ghosting, and plastic skin—signs that Marcos fine-tuned the result by iterating on the ChatGPT app.

Original JSON parameters

  • aspect_ratio: 4:5
  • style: hyper-realistic editorial fashion photography
  • lighting: soft frontal studio light with subtle rim light
  • camera: 85mm · shallow DOF · medium portrait · centered
  • color_mode: black_and_white
  • film_effect: analog grain, vintage border, soft focus edges

Variations

  • No double exposure: eliminates ghost silhouette lines for a clean B/W portrait.
  • Other clothing: replace the wool coat with a trench coat or leather blazer.
  • More contrast: add “high contrast noir lighting” at the end.
  • Different expression: change “serious natural expression” for a subtle smile or side look.

Before publishing

Only use your own photos or those with consent. Portraits generated from third parties may violate OpenAI policies and image rights. The post does not include branded text; avoid results with legible logos in commercial use.

Most read

What readers are engaging with right now.