Sep 27 13:56 · AI content
Промпт №7
Промпт для создания селфи
Можно настроить:
- персонажа: прикрепить референсы с внешностью
- колор
ПРОМПТ для Sora GPT image 2, Sora GPT image 2.5
UNIFIED COLOR PALETTE — set the background and rim light color scheme with this single word: [ Powder Blue ]. Only the background environment and the dramatic rim lighting use this color — clothing, skin, and foreground elements remain naturally colored and distinct from the background.
CHARACTER REFERENCE TRANSFER — ABSOLUTE PRIORITY: The attached reference image defines the ONLY face and body that may appear in this generation. This is not a style reference or a mood board — it is a precise biometric template. Reproduce the reference subject with forensic accuracy: identical bone structure, facial geometry, eye shape and exact color, nose bridge and tip shape, lip shape and fullness, jawline angle, cheekbone height and prominence, brow shape, skin tone including undertones and any visible texture or marks, exact hair color and curl pattern, facial hair if present, body type and build, and any visible identifying features. If the reference shows a woman, generate that exact woman; if it shows a man, generate that exact man. The AI must analyze and lock every measurable facial landmark from the reference and replicate them without deviation — do not blend, average, idealize, or substitute features. The reference supplies identity only: facial structure, proportions, skin tone, and unique features. Do not copy the lighting, flatness, shadows, or skin rendering from the reference — re-light the same face entirely for this scene. Think of this as generating a photograph of the same person in a new scene with a completely different lighting setup, not copying how the face looked in the reference photo. Do not generate any other person.
CLOTHING AND STYLING — MANDATORY FASHION STANDARD: The outfit must be naturally colored and contrast clearly with the background — never tint or color-match the clothing to the background. The clothing must be genuinely fashion-forward and current for 2026 — the kind of look a top-tier social media influencer, fashion week attendee, or creative director would wear. Think: architectural outerwear with unexpected proportions, premium layering with texture contrast, avant-garde streetwear silhouettes, high-fashion separates with editorial edge, bold asymmetric cuts, or luxury sport-luxe combinations. The look must feel intentionally styled by a professional — not assembled from basics. Explicitly forbidden: plain white t-shirts worn alone, simple jeans-and-top combinations, generic hoodies, office basics, normcore, anything that could have been worn in 2015–2020 without looking dated. The styling must be something you would stop scrolling for on Instagram in 2026. Invent the styling freely each time — never repeat the same pieces across generations. Match gender and age from the reference. Keep accessories editorial but not overpowering.
POSE AND EXPRESSION: The subject is alone in frame, playing with the camera in a candid flirtatious way — looking directly into the lens with a mischievous, self-aware gaze as if taking a selfie or posing for a friend standing right in front of them. The pose is non-trivial and spontaneous: a hand gesture near the face, a wink, a tilted head, a tongue-out tease, a playful grimace, a finger-gun, or a confident lean toward the camera, one arm extended as if holding the phone for a selfie. Feeling: caught mid-play, clowning and fully aware of the lens. Vary pose creatively between generations.
SKIN, FLASH LIGHTING, AND COLOR GRADING — CRITICAL: Do not copy the lighting, flatness, shadows, or skin rendering from the face reference. The reference supplies identity only. Re-light the same face from scratch for this scene using a soft frontal on-camera flash placed close to the lens axis and slightly above eye level, combined with a subtle directional key light from one side. The flash must create clean, controlled, luminous highlights on every raised facial plane: the center of the forehead, brow bone, upper cheekbones, bridge and tip of the nose, inner corners of the eyes, cupid's bow, center of the lower lip, chin, collarbones, and shoulders. These highlighted planes must be visibly and clearly lighter than the surrounding skin — not subtle, not implied, but actually brighter, as if a soft flash has caught those surfaces directly. The transition from highlight to shadow is smooth but definite: soft falloff into clearly darker areas beneath the cheekbones, along the sides of the nose, under the jawline, in the eye sockets, and under the chin. The face must never look flat, evenly lit, or tonally uniform — every viewer should immediately see the three-dimensional volume of the face through the contrast between illuminated raised planes and shadowed recessed areas. The skin finish is polished editorial beauty: subtle diffusion, a soft inner glow, delicate specular shine on the highest highlights, and smooth tonal gradients. Natural skin texture, pores, freckles, and distinctive marks remain faintly visible — do not over-smooth into plastic. Apply cohesive cinematic color grading to the entire image: rich contrast, lifted luminous highlights, deep soft shadows, gentle highlight roll-off, and a unified color treatment that makes the face, clothing, rim light, and background belong in the same image. The result sits between a high-end retouched fashion photograph and a digital painting — polished, dimensional, and atmospheric, not a raw snapshot.
LIGHTING SETUP: Three distinct lighting layers must be visible and separate. Layer one: the soft frontal flash near the camera axis creates the bright sculpted highlights on all raised planes of the face and body — this is the dominant facial light. Layer two: a subtle directional warm key light from one side adds depth, shadow, and volume without overpowering the flash. Layer three: the colored rim light from the background wraps only around the outer contour of the hair, ears, shoulders, arms, and clothing edges in the color defined at the top of this prompt — it does not recolor the face or change the skin tone. These three layers must feel distinct: the face is brightly and sculpturally lit by the flash, the volume comes from the side key, and the colored edge light connects the figure to the environment.
BACKGROUND — CRITICAL BLUR REQUIREMENT: The background must contain real architectural structure — walls, windows with visible frames, doorways, columns, brick or concrete surfaces, corridors — rendered entirely in deep dark saturated shades of the color defined at the top. These architectural elements must be blurred to the point where they are felt rather than read: heavy cinematic bokeh at the equivalent of f/1.2 fully open, so that straight lines dissolve into soft glowing streaks and color gradients, window frames become luminous rectangular halos of light, and surfaces melt into atmospheric color fields. A viewer should sense real interior space but should not be able to read any specific detail, texture, or sharp edge in the background. The depth-of-field separation must be extreme — the subject is razor-sharp, the background is a soft glowing atmospheric wash of color with only the ghost of structure visible beneath the blur. The background emits ambient colored glow that bleeds into the rim light on the figure, making the subject feel embedded in the environment, not pasted onto it.
TECHNICAL SPECS: f/1.2–1.4, 85mm portrait lens equivalent, eye level or slightly below, high-end fashion editorial post-production.
MOOD: Confident, playful, effortlessly cool, charismatic — a perfect selfie-magic moment with editorial polish.
FINAL REMINDER: Face, body, gender, age, ethnicity, and every identifying feature must be an exact 1:1 match to the reference. Do not generate a different person.
Настройка
прикладываете изображения внешности (в промпте указывать порядковые номера изображения не нужно. Они все считаются как референсы внешности)
колор задается сверху промпта в блоке “UUNIFIED COLOR PALETTE — set the background and rim light color scheme with this single word: [ Powder Blue ]. Only the background environment and the dramatic rim lighting use this color — clothing, skin, and foreground elements remain naturally colored and distinct from the background.”
Powder Blue - меняем на свой
Генерирую с помощью https://syntx.ai/welcome/JPmpIYam
Можно настроить:
- персонажа: прикрепить референсы с внешностью
- колор
ПРОМПТ для Sora GPT image 2, Sora GPT image 2.5
UNIFIED COLOR PALETTE — set the background and rim light color scheme with this single word: [ Powder Blue ]. Only the background environment and the dramatic rim lighting use this color — clothing, skin, and foreground elements remain naturally colored and distinct from the background.
CHARACTER REFERENCE TRANSFER — ABSOLUTE PRIORITY: The attached reference image defines the ONLY face and body that may appear in this generation. This is not a style reference or a mood board — it is a precise biometric template. Reproduce the reference subject with forensic accuracy: identical bone structure, facial geometry, eye shape and exact color, nose bridge and tip shape, lip shape and fullness, jawline angle, cheekbone height and prominence, brow shape, skin tone including undertones and any visible texture or marks, exact hair color and curl pattern, facial hair if present, body type and build, and any visible identifying features. If the reference shows a woman, generate that exact woman; if it shows a man, generate that exact man. The AI must analyze and lock every measurable facial landmark from the reference and replicate them without deviation — do not blend, average, idealize, or substitute features. The reference supplies identity only: facial structure, proportions, skin tone, and unique features. Do not copy the lighting, flatness, shadows, or skin rendering from the reference — re-light the same face entirely for this scene. Think of this as generating a photograph of the same person in a new scene with a completely different lighting setup, not copying how the face looked in the reference photo. Do not generate any other person.
CLOTHING AND STYLING — MANDATORY FASHION STANDARD: The outfit must be naturally colored and contrast clearly with the background — never tint or color-match the clothing to the background. The clothing must be genuinely fashion-forward and current for 2026 — the kind of look a top-tier social media influencer, fashion week attendee, or creative director would wear. Think: architectural outerwear with unexpected proportions, premium layering with texture contrast, avant-garde streetwear silhouettes, high-fashion separates with editorial edge, bold asymmetric cuts, or luxury sport-luxe combinations. The look must feel intentionally styled by a professional — not assembled from basics. Explicitly forbidden: plain white t-shirts worn alone, simple jeans-and-top combinations, generic hoodies, office basics, normcore, anything that could have been worn in 2015–2020 without looking dated. The styling must be something you would stop scrolling for on Instagram in 2026. Invent the styling freely each time — never repeat the same pieces across generations. Match gender and age from the reference. Keep accessories editorial but not overpowering.
POSE AND EXPRESSION: The subject is alone in frame, playing with the camera in a candid flirtatious way — looking directly into the lens with a mischievous, self-aware gaze as if taking a selfie or posing for a friend standing right in front of them. The pose is non-trivial and spontaneous: a hand gesture near the face, a wink, a tilted head, a tongue-out tease, a playful grimace, a finger-gun, or a confident lean toward the camera, one arm extended as if holding the phone for a selfie. Feeling: caught mid-play, clowning and fully aware of the lens. Vary pose creatively between generations.
SKIN, FLASH LIGHTING, AND COLOR GRADING — CRITICAL: Do not copy the lighting, flatness, shadows, or skin rendering from the face reference. The reference supplies identity only. Re-light the same face from scratch for this scene using a soft frontal on-camera flash placed close to the lens axis and slightly above eye level, combined with a subtle directional key light from one side. The flash must create clean, controlled, luminous highlights on every raised facial plane: the center of the forehead, brow bone, upper cheekbones, bridge and tip of the nose, inner corners of the eyes, cupid's bow, center of the lower lip, chin, collarbones, and shoulders. These highlighted planes must be visibly and clearly lighter than the surrounding skin — not subtle, not implied, but actually brighter, as if a soft flash has caught those surfaces directly. The transition from highlight to shadow is smooth but definite: soft falloff into clearly darker areas beneath the cheekbones, along the sides of the nose, under the jawline, in the eye sockets, and under the chin. The face must never look flat, evenly lit, or tonally uniform — every viewer should immediately see the three-dimensional volume of the face through the contrast between illuminated raised planes and shadowed recessed areas. The skin finish is polished editorial beauty: subtle diffusion, a soft inner glow, delicate specular shine on the highest highlights, and smooth tonal gradients. Natural skin texture, pores, freckles, and distinctive marks remain faintly visible — do not over-smooth into plastic. Apply cohesive cinematic color grading to the entire image: rich contrast, lifted luminous highlights, deep soft shadows, gentle highlight roll-off, and a unified color treatment that makes the face, clothing, rim light, and background belong in the same image. The result sits between a high-end retouched fashion photograph and a digital painting — polished, dimensional, and atmospheric, not a raw snapshot.
LIGHTING SETUP: Three distinct lighting layers must be visible and separate. Layer one: the soft frontal flash near the camera axis creates the bright sculpted highlights on all raised planes of the face and body — this is the dominant facial light. Layer two: a subtle directional warm key light from one side adds depth, shadow, and volume without overpowering the flash. Layer three: the colored rim light from the background wraps only around the outer contour of the hair, ears, shoulders, arms, and clothing edges in the color defined at the top of this prompt — it does not recolor the face or change the skin tone. These three layers must feel distinct: the face is brightly and sculpturally lit by the flash, the volume comes from the side key, and the colored edge light connects the figure to the environment.
BACKGROUND — CRITICAL BLUR REQUIREMENT: The background must contain real architectural structure — walls, windows with visible frames, doorways, columns, brick or concrete surfaces, corridors — rendered entirely in deep dark saturated shades of the color defined at the top. These architectural elements must be blurred to the point where they are felt rather than read: heavy cinematic bokeh at the equivalent of f/1.2 fully open, so that straight lines dissolve into soft glowing streaks and color gradients, window frames become luminous rectangular halos of light, and surfaces melt into atmospheric color fields. A viewer should sense real interior space but should not be able to read any specific detail, texture, or sharp edge in the background. The depth-of-field separation must be extreme — the subject is razor-sharp, the background is a soft glowing atmospheric wash of color with only the ghost of structure visible beneath the blur. The background emits ambient colored glow that bleeds into the rim light on the figure, making the subject feel embedded in the environment, not pasted onto it.
TECHNICAL SPECS: f/1.2–1.4, 85mm portrait lens equivalent, eye level or slightly below, high-end fashion editorial post-production.
MOOD: Confident, playful, effortlessly cool, charismatic — a perfect selfie-magic moment with editorial polish.
FINAL REMINDER: Face, body, gender, age, ethnicity, and every identifying feature must be an exact 1:1 match to the reference. Do not generate a different person.
Настройка
прикладываете изображения внешности (в промпте указывать порядковые номера изображения не нужно. Они все считаются как референсы внешности)
колор задается сверху промпта в блоке “UUNIFIED COLOR PALETTE — set the background and rim light color scheme with this single word: [ Powder Blue ]. Only the background environment and the dramatic rim lighting use this color — clothing, skin, and foreground elements remain naturally colored and distinct from the background.”
Powder Blue - меняем на свой
Генерирую с помощью https://syntx.ai/welcome/JPmpIYam
ии
генерации
промпт