Extreme close-up portrait of a woman with voluminous wavy teal-turquoise hair, wearing oversized rectangular neon yellow-lime glasses with thin frames, silver septum nose ring piercing, glossy dark berry-stained lips slightly parted showing teeth, natural freckled skin with visible pores and sebum texture, dramatic split-colored lighting with cyan light 6000K from left and warm amber light 2700K from right creating 3300K color contrast on face, shallow depth of field, hyper-realistic skin texture with subsurface scattering, fashion editorial photography, 85mm lens, f/1.4, high detail, studio lighting setup --ar 2:3 --style raw
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 and customize to make it your own!
The core principle of neon cyberpunk portraiture is that the lighting must be defined by color temperature in Kelvin, not by color names. When an AI model reads "cyan and amber," it averages the two into a muddy neutral. When it reads "6000K and 2700K," it computes a specific chromatic gap and preserves the split across the entire face.
This distinction is not cosmetic. It changes how the model resolves skin, how it renders the specular highlights on the septum ring, and whether the freckles survive the render at all. The prompt in question works because every parameter is a measurable instruction rather than an aesthetic adjective.
The Kelvin Gap as the Primary Mechanism
Lighting in AI generation is essentially a white balance problem. The model receives a target color for the scene and distributes it based on the geometry of the light sources described. A 6000K source is slightly cooler than daylight; a 2700K source is a warm incandescent. Placed on opposite sides of a face, they create a 3300K differential that the model reads as directional contrast, not as a global tint.
This is why the original prompt specifies "creating 3300K color contrast on face" rather than simply saying "dramatic lighting." The phrase instructs the model to treat the gap as a measurable property of the scene. In practice, the cyan side of the face will render at roughly 6000K, the amber side at roughly 2700K, and the transition zone across the nose bridge will pass through a neutral point near 4300K.
That transition zone is where most prompts fail. Without the explicit Kelvin values, the model tends to apply a single global white balance and the split collapses into a uniform magenta or green cast. The freckles disappear first, then the pores, then the sebum texture. The face becomes plastic before the glasses even render correctly.
The lesson generalizes: any time a prompt asks for two colored light sources, it should also specify their color temperatures. The feathered portrait lighting approach uses the same principle with a narrower gap, usually 1500K to 2000K, to achieve a soft wrap rather than a hard split.
Skin Texture as a Specular Problem
"Hyper-realistic skin texture with subsurface scattering" is not a single instruction. It is three separate requests bundled into one phrase. Subsurface scattering controls how light penetrates the skin and bounces back; visible pores control the micro-geometry of the surface; sebum texture controls the specular sheen on top of that geometry. The model needs all three to produce skin that reads as physical rather than rendered.
In this image, the sebum highlight on the forehead and cheekbones is what sells the realism. It is a wet specular reflection that sits on top of the freckles and pores, not behind them. The prompt achieves this by placing "glossy dark berry-stained lips" and "sebum texture" in the same semantic cluster, which pushes the model toward a high-specular, high-detail surface treatment across the whole face.
The 85mm lens and f/1.4 aperture are not decoration. They determine the depth of field geometry. At f/1.4 on an 85mm equivalent, the plane of focus is roughly 4 to 6 centimeters deep at portrait distance. That is why the eyes and the septum ring are sharp while the hair behind the ears and the back of the glasses temples fall into soft blur.
If the aperture were f/4, the entire head would be in focus and the image would lose its editorial compression. If it were f/1.0, the eyes would blur and the glasses frames would smear. The 85mm f/1.4 combination is the specific window where a close-up portrait retains sharp facial features while still separating the subject from the background.
This same lens logic appears in the street portrait workflow, where a longer focal length is used to compress the background rather than blur it. The difference is intent: in a neon cyberpunk close-up, the background is irrelevant and the blur serves to isolate the face.
Why the Accessories Must Be Specified in Order
The prompt lists the glasses before the piercing, and the piercing before the lips. This ordering is not accidental. The model builds the scene hierarchically, with earlier tokens receiving higher attention weight in the composition. The oversized rectangular neon yellow-lime glasses are the largest visual element in the frame, so they are placed first.
The silver septum nose ring is second because it sits at the center of the face, directly on the vertical axis between the eyes. It is the smallest object in the frame but it carries the highest specular contrast, which is why the prompt specifies "silver" rather than leaving the metal unspecified. Silver produces a hard, cool reflection that survives the 6000K cyan light without turning green.
The glossy dark berry-stained lips are third because they occupy the lower third of the composition and provide the warmest color in the frame. The berry tone sits at roughly 340 degrees on the hue wheel, which is nearly opposite the cyan light source. This opposition is what makes the lips pop against the cool side of the face without the model needing to be told to increase saturation.
When accessories are listed in arbitrary order, the model tends to merge them into a single "accessories" concept and render all of them at medium detail. The ordering instruction forces separate attention allocations, which is why the glasses frames have crisp edges while the piercing retains a distinct metallic highlight.
The same hierarchical logic is visible in the cyberpunk robot streetwear composition, where the largest garment is specified first and the smallest hardware details last.
What the --style raw Flag Actually Changes
The --style raw parameter in Midjourney reduces the model's default aesthetic interpretation. Without it, the model applies a mild beautification pass that smooths skin, boosts saturation, and slightly softens contrast. With --style raw, those adjustments are suppressed and the prompt's literal instructions are followed more closely.
For this image, --style raw is essential because the prompt depends on visible pores and freckles. A standard style pass would remove the freckles entirely and flatten the sebum highlights into a uniform sheen. The raw style preserves the micro-texture because it does not apply the default skin-smoothing filter.
The --ar 2:3 aspect ratio is also load-bearing. A square or landscape ratio would force the model to include more of the hair and shoulders, which would reduce the relative size of the face. The 2:3 vertical ratio matches the natural crop of an editorial close-up and keeps the eyes on the upper third line.
External tools handle this differently. Midjourney uses the --style raw flag, while Adobe Firefly exposes a similar control through its "photo" versus "art" content type selector. The mechanism is the same: suppress the default aesthetic pass so the prompt's physical parameters dominate.
Where This Approach Stops Working
The boundary of this technique is the point where the Kelvin gap exceeds the model's ability to maintain skin tone continuity. Above roughly 4500K of differential, the transition zone across the nose bridge stops being a gradient and becomes a hard seam. The freckles on one side of the face no longer match the freckles on the other, and the skin reads as two separate materials joined at the midline.
The signal that you have crossed this boundary is the appearance of a visible edge rather than a gradient. If the cyan-to-amber transition can be traced as a line, the differential is too large. The fix is not to reduce the gap but to introduce a third fill source at 4300K to bridge the two sides and restore the gradient across the face.
Below roughly 800K of differential, the opposite failure occurs: the split collapses into a single global tint and the image loses its directional contrast entirely. The usable window is therefore approximately 800K to 4500K, with 3300K sitting near the center of that range as the most stable operating point for close-up cyberpunk portraits.
Label: Fashion
Key Principle: Split lighting works when you define color temperature in Kelvin, not color names. A 3300K differential between a 6000K cyan and 2700K amber source creates the split without muddying skin tones.
Comments
Post a Comment