abf6edca-5ebc-4e43-90bd-bfdb9286f334

1.22 @Wren — identity locked from the approved 28-image reference set. Do not re-describe, restyle, beautify, correct asymmetry or alter facial structure. Reference images govern the face absolutely. STATE: fatigue 30% · hair wash-cycle day 3 · story clock 02:00, T-zone sheen on forehead and nose · hair tucked behind the right ear, loose on the left, parting 15mm right of centre · headphone dent present above both ears · @Headphones worn · @Chain at the suprasternal hollow · @Wren EP1 wardrobe look as locked on the S1.03–S1.21 contact sheet, wrinkles matching. EXPRESSION STATE: "Arrested." Ceiling 3. Nothing on the face comments. She has not noticed her own stillness. FRAME: vertical 9:16. Close-up, chest up. Her face fills the upper two thirds, eyes on the upper third line. She sits left of centre, facing screen-left. A narrow unreadable dark band occupies roughly 18 percent of frame width at frame-right — no legible room, no readable sodium bar. Both hands sit just inside the very bottom frame line, fully within frame, not cropped, frozen mid-typing position over @Keyboard, fingers still close to the keys, not withdrawn. Left index knuckle scar visible. CAMERA: 85mm, subject distance 1.3m, height 118cm, frontal. Steeper compression than the preceding coverage — the face reads more sculptural. Single-plane focus on her near eye (camera-left). Hands fall marginally soft at the bottom edge. POSTURE + GAZE: forward-head working posture. HEAD TILTED 5 TO 7 DEGREES TO THE RIGHT, CHIN DOWN 2 DEGREES — mandatory. Eyes have stopped tracking; gaze parked and unfocused roughly 30cm past the screen, not on the screen. Both feet flat on the floor. She is not looking at the camera. LIGHT: monitor key only — soft, close, front and slightly below, green (#3FD07A cast) at its most saturated point, extremely fast falloff. It lights the underside of the brow, the nose tip, the upper lip, the front of the throat, and abandons everything above the brow. Sodium rim barely present, a trace on the loose hair at frame-right. No fill. No second key. Blacks lifted and noisy. BACKGROUND: two planes only. Her, and unreadable falloff. No desk edge, no monitor, no geography, no legible object. RENDER: clean plate. No grain, no halation, no grade, no stylisation. Skin texture true: visible pores across nose and inner cheeks, fine lines at the outer eye, redness at nose and inner eye corners, under-eye trough present. Negative prompt beautiful, pretty, gorgeous, stunning, model, flawless, perfect skin, smooth skin, airbrushed, glamorous, youthful, cute, makeup, symmetrical face, corrected asymmetry, head straight, upright head, head tilted left, hands withdrawn, hands lifted off keyboard, hands cropped by frame edge, hands out of frame, missing fingers, fused fingers, extra fingers, mutated hands, eyes tracking, reading eyes, eye contact with camera, direct gaze, looking up, smile, tears, wet glassy eyes, blue-white key, neutral key, ring light, catchlight ring, three-point lighting, fill light, flattering rim light, haze, volumetrics, god rays, bokeh, lens flare, cinematic grade, film grain, halation, vignette, camera push, motion blur on camera, legible screen text, visible monitor, visible desk, readable room, wide shot, earrings, jewellery other than a thin steel chain, volumised hair, hair as a sheet, wind

Generate a single, hyper-realistic character reference image (strictly no collages, text, watermarks, or borders). The image must feature a cinematic, three-quarter portrait of the described character, maintaining a sharp, tack-focused emphasis on the face with natural skin textures, pore details, and realistic micro-expressions. The lighting should be cinematic and directional (e.g., subtle split or Rembrandt lighting) to give depth, blending seamlessly with the environment specified in the character description. The character must have a natural, living presence, looking slightly off-camera or directly at the lens with a subtle, context-aware expression. Ensure the full head, neck, and upper shoulders are completely unobstructed. This image serves as a high-fidelity identity blueprint for temporal-consistent video generation.