Cyberpunk, Tokyo, 3 a.m.

A single 9:16 vertical photograph, a portrait shot in Tokyo's Kabukicho district at 3 a.m. on a Tuesday in early November. Light rain has been falling for hours; the street is wet but no longer puddled. The subject is the person in the uploaded photograph, photographed from across a small alleyway, the camera positioned about three metres away. Visual register: late-night cyberpunk realism in the lineage of Roger Deakins' Blade Runner 2049 cinematography — saturated colour driven by environmental light sources, not by colour grading.

The face — the hardest thing to preserve under this much colour, so be exact: carry over from the uploaded photograph the eye shape and spacing, the actual eye colour (which will catch and reflect the surrounding neon), the nose bridge profile and tip, the lip shape and natural fullness, the eyebrow shape and density, the jawline, the hairline, the skin tone (which will be tinted by the lighting but should remain identifiably theirs), and every visible mole, freckle, scar, or beauty mark. Avoid the model's tendency to render any cyberpunk subject with sharpened high cheekbones, a narrowed jaw, hooded smoky eyes, or "anime"-influenced features. Resist that gravitational pull completely. The subject is the actual person in the upload. The genre's lighting is on top of them; it does not replace them.

Setting — specific and immersive: the subject is standing in a narrow Kabukicho side-alley between two midcentury concrete buildings, both faces of the alley plastered with backlit signage in Japanese kanji and katakana. The most visible signage is for a small bar, a 24-hour ramen shop, and a karaoke parlor — names in saturated red, cyan, and acid green. Above the second-floor level, larger holographic advertisements project outward from the building face into the open air of the alley — half-translucent, slightly glitched at the edges, advertising things that don't quite scan (a soft drink with an unfamiliar logo, an idol whose face is partly resolved, a phrase in mixed Japanese and unfamiliar Latin script).

The pavement of the alley is wet asphalt with embedded grid markings from old roadwork. Standing puddles in the lower frame reflect the signage above — distorted, colour-saturated mirror images. Steam rises from a metal grate near the subject's feet. A row of beverage vending machines glows softly along one wall of the alley, their pale fluorescent insides adding a third colour temperature to the scene. Behind the subject and to one side, the alley opens onto a slightly busier street where blurred silhouettes of figures walk past — none of them clearly visible, all moving, the kind of background population a long exposure would render as motion-streaks.

Wardrobe — modern with a slight near-future edge:
Option A (universal): a long black technical-fabric raincoat with a high collar, the hood down, slight water-beading visible on the shoulders. Beneath the coat, a plain dark t-shirt visible at the throat. The coat is not a costume — it is the kind of jacket a real person would wear in this weather. Bottoms dark, technical-fabric trousers, no visible branding. One small detail: a small luminous patch or piece of reflective tape on the sleeve that catches the neon light differently than the surrounding fabric.
Option B (more fashion-leaning): an oversized black bomber jacket with the collar up, a thin silver chain at the throat, dark cargo trousers, sneakers visible at the lower edge of the frame. A pair of small wired earbuds with the cable visible against the jacket collar — analog technology, deliberately.
Option C (south-Asian-coded cyberpunk): a long structured asymmetric coat in deep black silk with subtle embroidery at the cuffs — Sabyasachi-meets-Kabukicho. A small jewelled piercing at the nose or ear catching the neon. Modern hair, no traditional styling. This option places an Indian subject naturally inside the cyberpunk visual language.

Pose: the subject is standing still, weight on the back foot, one shoulder slightly turned toward the alley opening behind them, head turned back toward the camera. One hand might be visible at the chest holding a small handheld device (the model can render as a sleek smartphone or a small futuristic-looking object — at its discretion), partially obscured by the coat. The other hand at the side, relaxed. The body language is *paused* — they were walking, they have stopped to look at something, they have not yet decided whether to continue.

Expression: read the upload. The genre defaults to "tired and unimpressed," but this is too easy a read. Better: the expression of someone who has been awake for a long time and is past tired into a kind of quiet alertness — calm, present, slightly distant. Eyes meeting the camera directly. Mouth softly closed. If the upload's resting face carries warmth, preserve some of it underneath the alertness — even cyberpunk subjects can be warm humans.

Light — the colour story is told through three competing sources:
— Source one (the dominant key): a holographic advertisement from above-camera-right, projecting magenta and warm pink across the right side of the face. The colour is not a flat wash — it has spatial structure, slightly varying intensity, the way a real holographic projection would.
— Source two (the rim): a cyan-and-green-coloured neon sign from camera-back-left, catching the left edge of the face, the left side of the jaw, the curve of the back of the head, and the wet collar of the coat. Pure neon glow, the colour of an old bar sign.
— Source three (the underlight): the wet asphalt of the alley acts as a giant uplit reflector, throwing soft pink-and-red bounced light upward into the underside of the jaw, the bottom of the cheekbones, and the underside of the coat collar. This is the secret of the genre — the wet ground is what stops the face from being side-lit-flat.
— A small fourth contribution: the cool fluorescent spill from the vending machines to one side, adding a touch of pale cyan to the shadow side of the face.
— The light is real and complex — no single direction, no clean key/fill geometry. The face sits inside a colour environment. This is what cyberpunk lighting is.

Atmospheric elements: light rain falling at a moderate angle, individual droplets visible in front of the holographic signage where they catch and momentarily reflect colour. Slight mist from steam, breath visible in the cold air at the mouth (it's November in Tokyo, it's that cold). The breath catches the magenta key and reads pink.

Camera language: 50mm prime lens at f/1.4 wide open, on a full-frame digital body (Sony A7S III for low-light) at ISO 3200. Some visible noise in the deepest shadows, but controlled and grain-like rather than digital-blocky. Subject distance roughly three metres. Three-quarter framing — from above the waist to a finger's width above the head. Slight motion blur on the background figures and on a few falling raindrops; the face is sharp. The depth of field is shallow enough that the holographic signage in the background reads as soft saturated colour blobs of light, not as readable signage.

Colour treatment: native digital sensor colour at the source — the colours of Tokyo at night, captured but not graded. The image has not been Photoshopped into cyberpunk; it has been *photographed* in cyberpunk. Some highlights from the brightest signage are slightly clipped to white (no information held), the way real signage at night clips. Skin tones are tinted by the lighting but preserve recognizable identity.

Skin texture preservation: pores visible across the face even under the coloured lighting. The wet sheen at the temple where rain has caught is rendered as actual skin moisture, not a digital highlight. Eyelashes individually rendered. Visible breath cloud at the mouth carries small details — slight imperfection in its edges, the way real breath dissipates. The lighting is on the skin, not replacing the skin.

Aspect ratio 9:16 vertical. No text in English, no readable subtitles, no logos other than the unfamiliar signage already described, no border, no caption, no watermark.