Haunting Dark Portrait — Black Smoke Rises Behind Him
A full-body cinematic dark portrait prompt engineered for face-consistent character reference generation — combining luxury all-black styling with a dramatic smoke entity rising from shadow.
Getting the smoke entity to stay behind the subject — not bleed into the face — took serious work. I ended up front-loading the face-match instruction and reducing the smoke description weight. The warm rim / cool fill split is what keeps the face reading sharp even against all that dark atmosphere
Tags
Prompt
Expected Output
Images from this prompt create a full-body portrait with strong visual impact. The subject stands centered against a dark industrial setting, with wet reflective ground catching warm amber light on one side and cool blue-grey fill on the other. Behind him, swirling black smoke forms a subtle humanoid silhouette with faint amber eyes—powerful yet not distracting from the subject. Cinematic fog adds depth. Technically, the subject remains sharp while the background falls into soft bokeh, similar to an 85mm f/1.8 look in Midjourney v6 with --style raw. Skin detail is highly realistic, especially when face-match instructions appear early. The --q 2 flag enhances fine details like beard texture, fabric grain, and smoke. In practice, this prompt works well for personal brand visuals needing a dark, luxury aesthetic without studio costs. Ideal for creators seeking high-end editorial imagery quickly.
- Atmospheric fog and deep shadow create natural background separation that enhances subject presence
- All-black luxury outfit renders with fabric texture differentiation — coat, turtleneck, and trousers read as distinct materials
- Wet reflective ground adds a cinematic layer without requiring post-production compositing
- Full-body composition at 2:3 ratio produces print-ready portrait dimensions without crop adjustments
- Smoke entity scales visually with the subject without distorting the face or foreground clarity
- Warm amber rim light and cool fill create genuine volumetric depth rather than flat studio contrast
- Face-matched character reference holds skin tone, beard, and bone structure across multiple scenes
Parameters & Variables
| Variable Token | Meaning | Examples | Effect |
|---|---|---|---|
| FACE REFERENCE | Controls character identity — facial structure, skin tone, beard, and expression anchored to a source image | character sheet referencepreviously upscaled portrait URLMidjourney-generated headshot URL | Determines how closely the generated face matches the provided reference; anchor image quality directly affects face retention consistency |
| OUTFIT DESCRIPTION | Specifies the subject's clothing — silhouette, fabric, formality, and colour system | navy double-breasted coat with white turtleneckcharcoal tactical overcoat with cargo trousersall-black slim suit with open collar | Shifting from monochromatic black to multi-tone outfits changes the visual weight balance and can draw attention away from the smoke entity dynamic |
| SMOKE ENTITY | Controls the supernatural background element — density, scale, eye presence, and humanoid definition | minimal smoke wisps with no eyesbright white light entity with angelic silhouettefaint grey mist with no humanoid shapedense black smoke with glowing red eyes | Increasing entity definition and eye brightness pushes the image toward supernatural concept art; reducing it keeps the output closer to editorial fashion photography |
| LIGHTING SETUP | Defines the key and fill light sources — colour temperature, direction, and intensity split | neon purple rim + green fillnatural daylight side-key + deep shadow no fillorange-red cinematic splitwarm golden rim + cold blue fill (current) | Changing the colour temperature pairing shifts the emotional register dramatically — warmer splits read as powerful/heroic, cooler splits read as sinister/isolated |
| --ar | Controls the aspect ratio and compositional format of the output | 16:9 (cinematic wide)1:1 (square)9:16 (Reels/Stories)4:5 (Instagram post)2:3 (portrait print) | Switching from 2:3 to 9:16 shifts the model toward a taller composition that emphasises vertical atmospheric depth; 16:9 introduces environmental context at the cost of subject scale |
Pro Tips / Best Practices
- 🔁 Iterate Fast: Generate your base image first without --cref using just the outfit and environment description to nail the lighting and composition. Once you have a generation you like, retrieve the seed, lock it, then add the character reference on the next run. This two-pass approach eliminates the variable contamination of trying to nail face, lighting, and smoke simultaneously.
- 📐 Aspect Ratio Guide: The current --ar 2:3 is optimised for print and Pinterest. Switch to --ar 9:16 for Instagram Reels covers and Stories. The model will extend the industrial environment upward, increasing smoke entity height and making the background atmospheric depth feel more immersive.
- 🎛️ Customize It: Swap the smoke entity description for a wings of black feathers unfurling or a constellation of light particles to shift the supernatural register entirely while keeping the portrait architecture and luxury outfit intact.
- 💬 My Personal Take:: The smoke entity behaves unpredictably at --cref weight 100 — the model sometimes lets the smoke bleed into the face area when the character reference pulls too strongly toward the lighting of the original image. Drop to --cw 65 and re-run if you see that happening. The face stays locked but the smoke renders as a cleaner, more separated element.
- 👤 Who Should Use This: Personal brand photographers, high-end Fiverr portrait sellers, and Instagram creators building a dark-luxury content identity. This prompt is specifically built for anyone delivering a series of character-consistent editorial portraits — the face-match architecture means you can place the same person across ten different scenes without visible identity drift.
