Cyberpunk Sunglasses Portrait: The Exact AI Prompt Revealed
Quick Tip: Click the prompt box above to select it, then press Ctrl+C (Cmd+C on Mac) to copy. Paste directly into Midjourney, DALL-E, or Stable Diffusion!
Why Most Cyberpunk Portraits Fail at the Lens
The reflective sunglasses shot is one of the most requested cyberpunk portrait formats—and one of the most consistently botched. The failure isn't aesthetic. It's structural. When you ask an image generator for "reflective sunglasses," you're describing a material property without giving the model any reason to resolve what exists behind that surface. The result is inevitable: opaque mirror lenses that reflect generic neon cityscapes, with no hint of the face beneath.
The breakthrough comes from understanding how diffusion models handle transparency. These systems don't "see" layers. They predict pixel values based on statistical patterns in their training data. "Reflective" correlates with mirror-like surfaces in most photographic training examples. To get partial transparency—an eye visible through a tinted lens—you must explicitly describe the relationship between foreground and background elements. The model needs anatomical anchors: eyelashes, the curve of an eyelid, the specific color of an iris. Without these, it defaults to the simpler solution: solid reflection.
This prompt uses "visible green eye with dramatic black eyelashes behind left lens" as a forcing function. The eyelashes are critical. They create depth cues that help the model resolve the spatial relationship: the lens is a surface, the eye is behind it, and the eyelashes exist in the intermediate space where both are partially visible. This construction mirrors how photographers actually capture this effect—using angle, lighting, and lens coating to create that precise balance of reflection and transmission.
The Physics of Color Temperature in AI Rendering
Color temperature is where most cyberpunk prompts collapse into visual noise. The genre's association with neon produces a common error: "neon lighting" treated as a color choice rather than a physical condition. The model interprets this as "make everything pink and cyan," resulting in flat, saturated images without dimensional form.
The technical solution is to specify light as an event with direction, quality, and consequence. This prompt uses "cool blue cinematic lighting from left creating soft shadows on right side of face." The mechanism matters. "Cool blue" establishes Kelvin temperature (roughly 8000K-10000K, typical of overcast sky or certain LED sources). "From left" fixes the vector. "Creating soft shadows on right side" links cause to effect, which helps the model maintain coherent geometry—the shadows must correspond to the light source.
The "soft" modifier is equally important. Hard light produces sharp, defined shadows with high contrast between lit and unlit areas. Soft light diffuses, creating gradual transitions. In portrait work, soft light is generally more forgiving for skin texture, but the choice should be intentional. Hard light from the same left position would emphasize pore structure and create more dramatic sculpting of cheekbones. The prompt's choice of soft light prioritizes skin texture visibility over dramatic contouring—a decision that aligns with the "fashion photography style" specification.
Alternative approaches often fail by specifying color without physics. "Blue and purple lighting" produces color patches without dimensional logic. "Cyberpunk colors" triggers the model's most generic associations—saturated magenta and cyan gradients applied as overlay effects rather than integrated illumination. The correct approach treats light as a physical actor with position, quality, and consequence for every surface it touches.
Skin Texture: The Specificity Threshold
Perhaps no element reveals the gap between generic and precise prompting more clearly than skin. The default tendency of image generators—especially at standard stylization levels—is toward idealized, smoothed skin that erases pores, fine lines, and surface variation. This isn't a bug. It's the statistical center of beauty photography in training data, where retouching conventions have systematically removed these details.
To override this default, you must name specific imperfections. "Hyper-realistic skin" is insufficient. The term "realistic" in model training correlates with lighting accuracy and proportion, not necessarily with unretouched texture. "Visible pores and subtle freckles" crosses the specificity threshold. Pores are named. Freckles are named. The model now has concrete texture elements to render rather than a quality to approximate.
The mechanism extends to other skin properties. "Pale skin" establishes base tone. The combination with "cool blue" lighting produces the characteristic cyan shift in shadows that defines this image's color grading. Warm light on pale skin would push toward peach and gold. The cool direction maintains the cyberpunk color palette while preserving natural skin behavior under that light condition.
This approach connects to broader portrait techniques explored in dramatic feathered portraiture, where specific texture language similarly overrides default smoothing. The principle extends across subject matter: name the physical property you want visible, not the quality you want it to have.
Material Physics: From "Metallic" to Coated Glass
The sunglasses in this prompt demonstrate how material description determines render quality. "Round metallic-framed sunglasses with iridescent yellow-green gradient lenses" breaks the specification into components: frame material, lens optical property, and lens color behavior.
"Iridescent" is the critical term. It specifies a coating that shifts color with angle of view—different from "reflective" (mirror behavior) or "tinted" (uniform color reduction). The model's training includes iridescent coatings on sunglasses, oil slicks, butterfly wings. The term activates these associations, producing the color shift visible in the image where yellow-green transitions toward orange at lens edges.
The "digital grid reflections" add environmental context. Pure reflection without content produces mirror surfaces that feel empty. Specifying what reflects—digital grids, city lights, data streams—gives the model concrete patterns to incorporate. This connects to techniques in cyberpunk robot portraiture, where environmental reflection similarly grounds synthetic subjects in plausible spaces.
The original prompt used "reflective yellow-green gradient lenses." The revision to "iridescent" sharpens the material definition. "Reflective" is ambiguous between mirror and coated transmission. "Iridescent" commits to the interference-coating physics that produces this specific visual effect.
Composition and the Extreme Close-Up
The "extreme close-up" framing is doing substantial work. It justifies shallow depth of field—only the plane of the face is sharp, with ears and background falling off. It enables the tight crop that makes the sunglasses dominate the frame. And it triggers the model's associations with beauty photography and editorial portraiture, which helps align the "fashion photography style" specification.
The 9:16 aspect ratio reinforces this. Vertical orientation in close-up portraiture emphasizes facial structure, draws the eye upward through the composition, and creates the smartphone-screen-native format that has become standard for AI-generated portrait work. The Midjourney model processes aspect ratio as a strong compositional prior—9:16 activates different framing assumptions than 1:1 or 16:9.
Shallow depth of field here isn't merely aesthetic. It serves the technical goal of isolating the sunglasses and face from environmental context. In a genre associated with dense urban backgrounds, this selective focus keeps attention on material and lighting rather than narrative setting. The "moody atmosphere" specification reinforces this—emotional tone without environmental specificity.
The Role of Stylization and Raw Mode
The parameter combination—--style raw --s 250—represents a deliberate choice about how much the model should impose its default aesthetic. Raw mode removes Midjourney's training toward "beautiful" images: the automatic smoothing, the pleasing color harmonies, the idealized proportions. At stylization 250, some coherence preservation remains, but the plastic uniformity of higher values (750-1000) is avoided.
This matters specifically for skin texture and material detail. Standard mode at default stylization would likely soften pores and homogenize the iridescent lens effect into a simpler color gradient. The raw mode preserves the specific imperfections named in the prompt. The 250 value prevents the chaos of very low stylization (0-50) where image coherence can fragment.
Alternative approaches might use higher stylization with heavier texture prompting, but this produces diminishing returns. The model's default beauty bias becomes harder to override as stylization increases. Raw mode with moderate stylization offers more direct control over specific visual properties.
For related exploration of how stylization parameters affect technical subjects, street portrait techniques demonstrate similar parameter choices for documentary-style rendering.
The completed image succeeds because every element is specified as a physical property rather than an aesthetic quality. Light has direction and temperature. Skin has pores and freckles. Lenses have coatings and partial transparency. The model doesn't interpret "cyberpunk mood." It renders specific conditions that collectively produce that mood—conditions that could be verified by examining the actual physics of the scene if it existed. This is the difference between prompting for association and prompting for construction.
Label: Fashion
Key Principle: Always describe what light does, not what it is. "Blue light from left creating shadows on right" produces dimensional form; "blue lighting" produces flat color.