← AI prompts · Portrait

Red cinematic studio portrait — dramatic AI prompt

A studio portrait where the red is in the light itself, not on the wall behind you — which makes this prompt a fight between that colour and your own skin tone. Free prompt below.

By Naresh Basude · Founder, Dressing School

Photorealistic AI portrait of a man in thin metal glasses and a dark jacket, lit by dramatic red and orange studio light on a scarlet gradient — Made with the prompt below — 3:4 vertical.

The one line that decides this: “The red lighting should influence the skin naturally without turning the entire face unnaturally red.” Repeat it rather than weakening it.

Match the pose →

Create a highly detailed, photorealistic editorial portrait using the uploaded face-reference photographs as the only source of facial identity. Replace the person in the composition with the person from the identity references while preserving their exact age, facial structure, face shape, natural skin tone, forehead, hairline, hairstyle or baldness pattern, eyebrows, eyes, nose, lips, ears, jawline, facial-hair pattern, and all permanent visible features. The final subject must unmistakably look like the person in the uploaded identity references, not the individual in the style reference. Recreate the reference composition as a vertical 3:4 portrait with no text, captions, logos, or graphic elements. Position the subject slightly right of centre, filling most of the frame from the upper chest and shoulders to just above the head. The face should be the dominant visual element, occupying approximately the middle and upper-right portion of the image. Leave a controlled amount of negative space around the upper-left side and above the head. Photograph the subject from a dramatic low-angle perspective, with the camera positioned below chin level and tilted gently upward. Use a medium close-up crop showing the complete head, neck, upper chest, and the tops of both shoulders. Place the camera approximately 80–110 centimetres from the subject, using a realistic 50–65mm full-frame lens perspective. Preserve the subtle perspective emphasis on the chin, nose, glasses, and forehead without producing wide-angle distortion. Turn the subject's upper body slightly toward frame-right while keeping the shoulders relaxed and naturally uneven because of the low camera position. The near shoulder should enter prominently from the lower-left corner, while the opposite shoulder remains partially visible along the lower-right edge. Arms and hands remain completely outside the frame. Keep the back upright, with the neck extended naturally and the head tilted slightly backward. Turn the head approximately 20–30 degrees toward frame-right. Raise the chin modestly to create a confident, authoritative posture. The subject's eyes should look slightly upward and beyond the camera toward the upper-right area of the scene. Use a calm, serious, contemplative facial expression: relaxed forehead, naturally closed lips, no smile, no visible teeth, and a focused, self-assured gaze. Avoid aggressive tension or exaggerated emotion. Preserve the subject's real hairstyle and facial grooming from the identity references. Do not copy the reference person's bald head, grey beard, wrinkles, age, or facial proportions unless those characteristics are genuinely present in the identity photographs. Render every individual hair, eyebrow strand, beard or moustache hair, skin pore, subtle crease, and natural tonal variation with realistic detail. Keep the skin authentic and textured rather than airbrushed. Add thin, circular metal-frame eyeglasses similar to those visible in the reference: understated dark gunmetal rims, narrow temples, a delicate bridge, and clear lenses. Position them naturally according to the subject's facial anatomy. Include controlled reflections on the lenses from the orange-red studio lights while keeping both eyes visible and sharply defined. Do not use sunglasses or heavily tinted lenses. Dress the subject in a minimal, dark navy-to-black structured jacket with a matte premium fabric finish. The garment should have a clean, understated silhouette with no visible logos, patterns, buttons, jewellery, or decorative details. Show realistic fabric density, subtle folds around the shoulders and collar area, and very limited reflected light. Most of the jacket should fall into deep blue-black shadow, creating strong contrast with the illuminated face. Use a seamless studio background consisting only of a smooth, saturated colour gradient. The lower-left and middle-left areas should glow with intense burnt orange and fiery red-orange, gradually transitioning into rich scarlet, crimson, and deep burgundy toward the top and right side. The gradient must be clean and atmospheric, with no visible horizon, furniture, architecture, texture, objects, or environmental details. Create dramatic cinematic studio lighting dominated by intense red and orange tones. Place the main warm key light slightly in front of and to frame-right of the subject, positioned near eye level or marginally below, producing a strong orange-red highlight across the forehead, cheekbone, nose bridge, upper lip, and front edge of the face. Allow the highlight to reveal authentic skin texture and facial contours without clipping. Keep the frame-left side of the face in deep, sculpted shadow with cool navy-black undertones. Add a narrow saturated red rim light from behind and frame-left, tracing the outer edge of the head, ear, cheek, jaw, neck, and shoulder. Include a subtle warm reflection along the lower beard or jaw area and a restrained specular highlight on the glasses. Preserve a very high lighting ratio with bold shadows, luminous highlights, and a strong separation between the subject and background. Use a colour palette of fiery orange, saturated red, crimson, burgundy, deep navy, charcoal black, and natural warm skin tones. Apply premium cinematic colour grading with deep blacks, rich contrast, controlled saturation, and smooth highlight roll-off. The red lighting should influence the skin naturally without turning the entire face unnaturally red. Keep the subject's face, eyes, glasses, facial hair, and front facial planes extremely sharp. Use a shallow-to-moderate depth of field around f/2.8–f/4 so the ears and rear shoulder soften slightly while the essential facial features remain clearly resolved. Render the background as a perfectly smooth, softly defocused colour field with no visible texture or noise. The final image should feel like a powerful contemporary magazine portrait or premium cinematic character study: bold, intelligent, serious, mysterious, sophisticated, and professionally art-directed. Use realistic anatomy, natural proportions, physically accurate lighting, high dynamic range, detailed skin texture, subtle film grain, clean tonal transitions, and true high-resolution photographic quality. NEGATIVE PROMPT Different person, copied identity from the reference subject, loss of identity, altered face shape, altered skin tone, incorrect age, copied baldness, copied grey beard, artificial ageing, changed hairstyle, changed hairline, changed facial-hair pattern, generic face, face morphing, beauty-filtered face, plastic skin, waxy skin, excessive skin smoothing, oversharpening, exaggerated wrinkles, distorted forehead, asymmetrical eyes, crossed eyes, incorrect gaze, closed eyes, missing eyes behind glasses, opaque lenses, sunglasses, oversized glasses, crooked glasses, duplicated glasses, warped metal frames, excessive lens glare, smiling, visible teeth, open mouth, exaggerated expression, aggressive expression, lowered chin, front-facing passport pose, high camera angle, eye-level camera, wide-angle facial distortion, fisheye effect, oversized nose, tiny chin, distorted ears, stretched neck, incorrect anatomy, extra shoulders, visible hands, extra fingers, cropped head, cut-off chin, extreme head tilt, incorrect body posture, bright clothing, patterned clothing, logos, jewellery, necklaces, hats, additional accessories, furniture, architecture, scenery, objects, textured background, visible studio equipment, flat lighting, soft beauty lighting, evenly lit face, weak shadows, missing rim light, blue background, pastel colours, washed-out colours, blown highlights, crushed facial detail, colour banding, excessive bloom, artificial glow, heavy fog, motion blur, low resolution, pixelation, compression artifacts, painterly rendering, illustration, cartoon, anime, 3D-rendered appearance, CGI skin, text, captions, typography, watermark, signature, border, poster layout, or logo.

About this prompt

This site has a second red portrait prompt and the two are not interchangeable. On the crimson overcoat portrait the red is a painted backdrop under warm white light, so your skin tone survives. Here the red is in the lamps — an orange-red key, a narrow saturated red rim light, a gradient in the same family. Nothing is lit neutrally. What that costs is skin tone: a face is recognised through small tonal shifts, and a coloured key compresses them toward one hue.

Which is why the colour paragraph ends the way it does. The red lighting should influence the skin naturally without turning the entire face unnaturally red. That line does more work than anything else here: it is the only one arguing against the rest of the text — which escalates from intense red and orange tones to a strong orange-red highlight to a very high lighting ratio. Delete it and nothing tells the model where to stop, and AI image tools have a well-known habit of taking a colour instruction as far as it will go.

The framing then works off the gradient, which is not one flat colour: the lower-left and middle-left areas glow burnt orange, cooling to deep burgundy toward the top and right side. So the near shoulder enters from the lower-left corner, dark jacket against the hottest part of the frame, while the face sits slightly right of centre and high, against the darkest. Two separations, opposite directions — which is what re-centring the subject costs you.

How to use it

Copy the prompt. Copy the full red cinematic studio portrait prompt from this page.
Attach face references without heavy colour. Open Google Gemini, Nano Banana, ChatGPT in image mode, Midjourney with a character reference or Leonardo, and attach two to four sharp photos of your face taken under neutral light rather than coloured light, so the identity reference is not already tinted.
Generate at 3:4 vertical. Generate at a 3:4 vertical aspect ratio so the off-centre composition and the negative space above the head both survive the crop.
Check the skin, not the mood. Look at the lit half of the face first. If it has gone the same colour as the background, repeat the line saying the red lighting should influence the skin naturally without turning the entire face unnaturally red, and lower the key from intense to moderate.

None of ChatGPT, Gemini or Nano Banana exposes a transformation-strength, identity-strength or beautify control, so there is nothing to set in those three — you get the same effect by saying it in words: keep my exact face and skin tone, relight and re-dress me, do not restyle my features. If your tool does surface a denoise or strength slider, as some image-to-image apps do, a low starting point is the reasonable one.

Why these lines are in the prompt

as the only source of facial identity … not the individual in the style reference. The prompt expects two attachments: your face, plus an image supplying pose and light. Which is also why it bars copying the reference person's bald head, grey beard, wrinkles, age — an odd line to write unless the identity kept drifting toward the man in that second image.

the camera positioned below chin level against the key light positioned near eye level or marginally below. Two heights, easily merged. The camera is low; the lamp is not. Collapse them and you get uplighting — nostrils lit from underneath, shadows climbing the forehead.

Add a narrow saturated red rim light from behind and frame-left. The key sits at frame-right, so the rim is deliberately opposite. On that side it is the only thing separating a blue-black jacket from a burgundy background — hence missing rim light in the negatives.

Three ways this goes wrong

Your face comes back the same colour as the background. That takes the likeness with it: the detail that makes a face yours has flattened into one hue. Asking for less red overall drains the gradient too. Add after the colour-grading sentence: “The orange-red light sits as a highlight on top of the subject's natural skin tone and must not replace it. The lit cheek and forehead should still read as skin.” If it persists, change intense red and orange tones to moderate — which lands on the key light, not the backdrop.

It turns into uplighting. Lit nostrils, shadow travelling up the forehead, an enlarged chin: the camera height and the lamp height have merged. Add: “The camera is below chin level but the key light is not — the key stays at the subject's eye level and above the lens axis, so shadows fall downward across the face.”

The glasses eat the eyes. One clause asks for controlled reflections on the lenses and for both eyes to stay visible and sharply defined, and those pull against each other under a hard coloured source. The negatives guard it four ways — opaque lenses, missing eyes behind glasses, excessive lens glare, sunglasses. Replace the reflections sentence with: “The lenses are optically clear and almost invisible. Allow only a thin crescent reflection at the outer edge of each lens. Both irises and pupils must be fully visible through the glass.”

Making it yours

Swap the colour. Teal, violet and cold cyan all suit it — but red here is not one word: it is in the background paragraph, both lighting paragraphs and the palette line.
Different wardrobe. Replace the jacket sentence with “a plain black high-neck merino knit, clean shoulder line, no visible seams or logos”. Do not go pale — a light garment catches the orange key and becomes the brightest thing in the frame, which is the face's job.
Drop the glasses. Delete the glasses paragraph rather than adding “no glasses”, and remove the glasses entries from the negatives. Then add “a small, distinct catchlight in each eye” — the lens highlight was carrying some of the sparkle.
3:4 to 9:16 for a Story. You lose composition, not just edges: the off-centre placement and the negative space around the upper-left side need frame width, so the subject re-centres and the lower-left shoulder crops first. 4:5 keeps most of it.
An eye-level version. Change below chin level and tilted gently upward to “at the subject's eye level” and drop the raised chin — then delete eye-level camera and lowered chin from the negatives, or it argues with your edit.

For the same colour handled the opposite way, the crimson overcoat portrait keeps the red on the backdrop and the key warm white. Away from the prompts, the face shapes guide covers which frames suit you, and the beard styling guide the grooming this one preserves.

Frequently asked questions.

Which tools does this work in, and are there settings to change?

Anywhere you can attach a face reference: Gemini, Nano Banana, ChatGPT in image mode, Midjourney with a character reference, Leonardo. None of ChatGPT, Gemini or Nano Banana exposes a transformation-strength, identity-strength or beautify slider, so say it in words instead: keep my facial structure and skin tone exactly as in the attached photos.

Do I need a second reference image, or just my face?

Just your face is enough. The prompt is written as though a style reference is attached too, mentioning the individual in the style reference and glasses similar to those visible in the reference. Attaching only your own photos, delete those two mentions and the sentence starting Do not copy the reference person's bald head.

I already wear glasses. Should my reference photos have them on?

Send at least a couple without them. The prompt supplies its own pair, thin circular gunmetal frames, positioned by your facial anatomy, so it needs a clear view of your eyes, brows and nose bridge.

Does the red lighting work on deeper skin tones?

It needs a gentler hand: the lit half of the face carries all the tonal detail here. Change intense red and orange tones to moderate and very high lighting ratio to strong but controlled, then add: keep the shadow side of the face readable rather than fully black.

The background comes out in bands, not a smooth gradient.

That is why colour banding sits in the negative prompt. Generate at the largest size available, then add to the background paragraph: render the gradient with smooth continuous tonal transitions and fine film grain, no stepping or contour lines. Grain is the standard cure for banding.

Can I use this as a work headshot?

Not as it stands, for most workplaces. A raised chin shot from below chin level under a red key is a magazine cover, not a directory photo. A work version needs an eye-level camera, no chin raise and a neutral key — most of this prompt rewritten.

The pose, on your own camera

Chin up, camera low, head turned right.

A modest chin raise and a 20–30° turn, held on screen.

Open Pose Mirror →