É possível criar um ensaio fotográfico coeso com inteligência artificial sem deixar todas as imagens com a mesma aparência. O segredo está em dirigir cada cena com precisão: identidade, pose, enquadramento, luz, lente, cenário, figurino e tratamento precisam trabalhar juntos.
Neste projeto, as mesmas identidades foram levadas a cinco universos visuais: um estúdio monocromático, um campo dourado, um interior com luz de janela, um retrato masculino escultural e uma estação de metrô em movimento. Abaixo estão os prompts completos usados nas imagens finais.
Antes de começar
Os prompts abaixo usam imagens em funções diferentes. A primeira imagem indicada em cada prompt serve apenas como referência de composição, pose, iluminação e atmosfera. As demais são referências de identidade facial. Mantenha essa ordem ao anexar os arquivos e deixe explícito que o rosto da fotografia de composição não deve ser copiado.
Prompt 1: Estúdio monocromático: casal em alfaiataria preta
Prompt completo
Use case: identity-preserve
Asset type: premium editorial portrait, vertical 4:5 photograph
Input images: Image 1 is COMPOSITION AND STYLE REFERENCE ONLY; do not copy either original face. Image 2 is the sole facial identity reference for the man (REFERENCE IDENTITY A). Image 3 is the sole facial identity reference for the woman (REFERENCE IDENTITY B).
Primary request: Reconstruct the minimalist black-and-white studio fashion photograph from Image 1 with the exact same visual concept and the two identities from Images 2 and 3.
Scene/backdrop: seamless light-gray studio cyclorama, clean and empty.
Subjects and identity: one adult man in front and one adult woman close behind him. Preserve the man's recognizable facial geometry from Image 2: oval face, dark styled hair, dark brows, brown eyes, straight nose, defined mouth, trim mustache and goatee, natural skin, and his black rectangular eyeglasses. Preserve the woman's recognizable facial geometry from Image 3: oval face, dark almond eyes, defined brows, straight nose, full lips, warm medium-brown skin, and long dark hair. Do not blend their features, do not approximate them, do not copy the faces from Image 1.
Pose and interaction: man stands three-quarter body angled left, hands relaxed in trouser pockets, chin slightly raised, calm distant gaze off-camera left. Woman stands immediately behind his right shoulder, torso subtly angled, both hands behind her back, head resting gently on his shoulder, face turned toward camera with poised neutral expression. Maintain anatomically correct hands and shoulders.
Wardrobe: coordinated oversized matte-black editorial tailoring; man in loose black open-collar shirt and wide-leg trousers; woman in structured black blouse with subtle puff sleeves and high-waisted long black skirt or trousers. Elegant, modern, no logos.
Hair/presentation: retain each person's natural recognizable hair character; neat editorial grooming; understated makeup; retain the man's real mustache, goatee, and glasses.
Lighting: large soft key from camera-left and slightly above, restrained fill from right, smooth background gradient, sculpted cheekbones, gentle contact shadows, medium contrast.
Camera: vertical full-to-three-quarter portrait, camera at chest height, straight-on with slight left perspective, 70-85mm full-frame lens look, around f/5.6, crisp faces and clothing, background smoothly clean.
Composition: subjects occupy central-left vertical column, full heads and bodies not cropped awkwardly, balanced negative space above and to the right, elegant overlapping silhouettes.
Color/grade: true monochrome, deep blacks, luminous skin, fine tonal separation, subtle film grain, premium fashion magazine finish.
Realism and quality: photorealistic professional photography, accurate anatomy, real skin texture and pores, natural hair strands, realistic fabric folds, physically plausible light, high resolution.
Constraints: exactly two people; no extra limbs, fingers, people, jewelry clutter, text, watermark, signature, logo, plastic skin, uncanny eyes, distorted glasses, hybrid face, or artificial AI look. Identity preservation is the highest priority; composition and pose second.Como usar
Anexe primeiro a referência de composição, depois a foto masculina e a foto feminina. O prompt prioriza a separação das duas identidades, a pose sobreposta e a luz suave de estúdio.
Observação técnica: Referência estimada como estúdio com luz suave lateral, lente equivalente de 70-85 mm e profundidade de campo moderada. A imagem final manteve a sobreposição dos corpos, a pose de apoio no ombro e o acabamento monocromático. Identidade, anatomia e óculos foram considerados adequados na primeira geração.
Prompt 2: Campo dourado: contraste entre marfim e preto
Prompt completo
Use case: identity-preserve
Asset type: cinematic outdoor fashion editorial, vertical 4:5 photograph
Input images: Image 1 is COMPOSITION, POSE, LIGHTING AND WARDROBE REFERENCE ONLY; ignore its original faces and any watermark. Image 2 is the sole identity reference for the man (REFERENCE IDENTITY A). Image 3 is the sole identity reference for the woman (REFERENCE IDENTITY B).
Primary request: Recreate the quiet luxury couple portrait in a vast golden wheat field, replacing only the people with the exact identities from Images 2 and 3.
Scene/backdrop: rolling field of tall dry golden grass reaching mid-thigh, distant soft hills and pale clean sky, late-afternoon rural landscape, no buildings.
Subjects and identity: exactly one adult woman at camera-left and one adult man at camera-right, back-to-back with a small gap and opposite orientations. Preserve the man's facial geometry from Image 2: oval face, dark styled hair, strong dark brows, brown eyes, straight nose, defined mouth, trim mustache and goatee, natural skin, black rectangular eyeglasses. Preserve the woman's facial geometry from Image 3: oval face, dark almond eyes, sculpted brows, straight nose, full lips, warm medium-brown skin, and long dark hair. Never blend identities and never use faces from Image 1.
Pose/expression: woman closer to camera, body three-quarter turned left with right hand hidden in a side pocket, shoulders relaxed, face turned toward lens, serene confident expression. Man slightly farther behind, body facing right, arms relaxed at sides, head gently lowered and gaze off-camera right, introspective expression. Wind subtly moves hair and fabric.
Wardrobe: woman in an ankle-length ivory architectural dress with high neckline, softly belted waist and voluminous bishop sleeves; man in a black tailored long coat or suit over a black turtleneck. Keep the man's recognizable eyeglasses. No logos or text.
Lighting: golden-hour sun low from rear-left, warm rim light through grass and hair, soft frontal sky fill, realistic long shadows, gentle highlight rolloff, luminous but not overexposed.
Camera: vertical medium-full portrait from upper-thigh to head with some field context, camera around waist height, slight telephoto compression, 85mm full-frame lens look, approximately f/2.8, subjects sharp and foreground/background grasses softly blurred.
Composition: woman on left third and man on right third, horizon high-mid frame, back-to-back diagonal tension, golden grass framing foreground, generous pale sky negative space.
Color/grade: warm wheat gold, ivory, matte black and pale blue; restrained saturation; cinematic soft contrast; subtle cream highlights and honey shadows; fine natural film grain.
Realism and quality: premium fashion campaign photography, authentic skin pores and texture, physically plausible sunlight, detailed fabric and individual grass stalks, coherent anatomy, high resolution.
Constraints: exactly two people; no extra limbs, deformed hands, duplicate bodies, additional figures, hats, text, signature, watermark, logos, plastic skin, artificial eyes, distorted glasses, face fusion, or AI artifacts. Identity preservation first, anatomy second, composition third.Como usar
Use uma referência de pose em campo aberto e mantenha as fotos de identidade na mesma ordem indicada. O contraluz de golden hour e a compressão de teleobjetiva são decisivos para a atmosfera.
Observação técnica: Referência interpretada como golden hour com contraluz, compressão de teleobjetiva e abertura visual próxima de f/2.8. A imagem final preservou o contraste preto/marfim, a relação costas com costas e a separação limpa das identidades. Sem artefatos anatômicos relevantes.
Prompt 3: Intimidade à janela: retrato cinematográfico de casal
Prompt completo
Use case: identity-preserve
Asset type: intimate cinematic luxury couple portrait, vertical 4:5 photograph
Input images: Image 1 is POSE, COMPOSITION, LIGHTING AND MOOD REFERENCE ONLY; do not retain its original faces. Image 2 is the sole identity reference for the man (REFERENCE IDENTITY A). Image 3 is the sole identity reference for the woman (REFERENCE IDENTITY B).
Primary request: Reconstruct the intimate low-key window portrait from Image 1 with the exact male and female identities from Images 2 and 3.
Scene/backdrop: refined dark interior beside a tall window with a sheer ivory curtain at camera-right; soft charcoal-brown wall fading into shadow; quiet timeless atmosphere; no distracting objects.
Subjects and identity: exactly one adult man in front at camera-right and one adult woman immediately behind/left. Preserve the man's recognizable facial structure from Image 2: oval face, dark styled hair, dark brows, brown eyes, straight nose, defined lips, trim mustache and goatee, natural skin, black rectangular eyeglasses. Preserve the woman's recognizable facial structure from Image 3: oval face, dark almond eyes, defined brows, straight nose, full lips, warm medium-brown skin, and long dark hair. Do not merge identities or copy faces from Image 1.
Pose/interaction: man in right-facing profile, torso upright toward window, chin level, calm serious gaze into the light. Woman stands close behind his left shoulder, eyes softly closed, cheek near his shoulder blade, both hands resting gently across his upper shoulder and chest; fingers elegant, clearly separated and anatomically correct. Their posture communicates quiet trust and intimacy, not smiling.
Wardrobe: man in impeccably tailored black tuxedo jacket with crisp white shirt and narrow black tie, eyeglasses retained. Woman in a fluid pearl-ivory satin blouse or dress with gathered cuff and understated drop earring; modest, elegant, no logos.
Hair/presentation: man retains recognizable dark swept hair, mustache and goatee; woman retains long dark hair pulled into a low polished style while preserving her face; luminous editorial makeup, natural skin.
Lighting: soft directional daylight from the tall window on camera-right, grazing both profiles; deep low-key shadow on camera-left; gentle reflected warm fill; dramatic chiaroscuro with protected highlight detail; subtle rim on hair and hands.
Camera: vertical tight medium portrait from upper torso to head, camera at shoulder height, slightly left of subjects, 85mm full-frame lens look, approximately f/2.0, precise focus on the woman's near eye area/face and man's profile, shallow background falloff.
Composition: two layered profiles form a diagonal from woman's face and hands to man's illuminated profile; subjects occupy right two-thirds; large dark negative space on left; curtain is a vertical leading line at right.
Color/grade: muted espresso brown, black, ivory and warm skin; fine-art cinematic grade, slightly desaturated, amber highlights and neutral shadows, subtle 35mm grain.
Realism and quality: premium editorial photography, authentic pores, natural fine facial hair, realistic hands, fabric weave, soft curtain texture, physically plausible light, high resolution.
Constraints: exactly two people, no extra fingers or arms, no duplicate hands, no additional figures, no text, watermark, signature or logo, no distorted eyeglasses, no plastic skin, no hybrid faces, no oversmoothing, no artificial eyes. Identity fidelity and hand anatomy are highest priority.Como usar
A cena depende da posição das mãos, dos perfis sobrepostos e da luz lateral. Se o gerador aceitar imagens de referência, use a fotografia de composição apenas para pose e iluminação, nunca para os rostos.
Observação técnica: Referência tratada como retrato low-key com luz lateral de janela, lente equivalente de 85 mm e profundidade de campo rasa. A imagem final conservou a leitura de perfis sobrepostos, o gesto da mão e o chiaroscuro. As mãos, os óculos e os perfis foram aprovados sem regeneração.
Prompt 4: Retrato masculino escultural com cadeira de palhinha
Prompt completo
Use case: identity-preserve
Asset type: moody male fashion editorial portrait, vertical 4:5 photograph
Input images: Image 1 is COMPOSITION, POSE, CHAIR, LIGHTING AND MOOD REFERENCE ONLY; do not use its original face or ethnicity. Image 2 is the sole facial identity reference for the subject (REFERENCE IDENTITY A).
Primary request: Reconstruct the sculptural seated chair portrait from Image 1 using the exact recognizable male identity from Image 2.
Scene/backdrop: minimal dark olive-brown studio room, seamless textured wall, subtle atmospheric haze, vintage cane-back wooden chair facing backward.
Subject and identity: exactly one adult man. Preserve the precise facial geometry from Image 2: oval face, dark thick swept hair, strong dark eyebrows, brown eyes, straight nose, defined mouth, trim mustache and goatee, natural warm-medium skin tone, and black rectangular eyeglasses. The result must unmistakably be the same person, not a lookalike. Do not copy the face from Image 1.
Pose/expression: seated backward astride the cane chair, torso leaning forward; left forearm folded across the top rail and raised partly near the lower face without obscuring the eyes; right arm drapes diagonally down the chair back, wrist loose, fingers naturally separated and anatomically correct. Head tilted slightly downward but eyes looking directly into camera with quiet intensity. Shoulders relaxed, body compact and sculptural.
Wardrobe: soft off-white premium cotton T-shirt with slightly rolled sleeves, dark trousers mostly outside frame, one understated ring permitted, black rectangular eyeglasses retained. No logos.
Hair/presentation: retain the subject's recognizable styled dark hair, mustache and goatee; natural grooming and skin texture.
Lighting: single warm soft key high from camera-left, feathered across face and forearms; deep but detailed shadows; faint neutral fill from camera-right; subtle halo of haze; chiaroscuro, low-key and cinematic.
Camera: vertical seated medium portrait from hips to head, camera at chest height and nearly straight-on, slight downward perspective, 70-85mm full-frame lens look, approximately f/2.8, sharp eyes and face, gently softer hands and background.
Composition: chair cane panel fills lower center; face positioned near upper third behind layered arms; diagonals of arms create a graphic frame; generous dark negative space; no awkward crops at joints.
Color/grade: muted tobacco, olive, cream and warm skin; low saturation; lifted film blacks; subtle amber highlights; refined 35mm grain.
Realism and quality: premium editorial photography, real pores, fine facial hair, accurate eyeglass reflections without glare over eyes, realistic cane weave and cotton fabric, coherent anatomy, physically plausible light, high resolution.
Constraints: exactly one person; do not change identity; no extra limbs or fingers, no fused hands, no duplicated chair parts, no text, signature, watermark or logo, no plastic skin, no distorted glasses, no face occlusion, no artificial eyes, no AI artifacts. Identity fidelity first, anatomy second, pose and lighting third.Como usar
Envie a referência de pose antes da referência facial masculina. Verifique principalmente os dedos, os reflexos dos óculos e a leitura dos dois braços sobre a cadeira.
Observação técnica: Referência estimada como retrato de estúdio low-key, câmera à altura do tórax, lente de 70-85 mm e luz principal quente alta pela esquerda. A versão final manteve a moldura criada pelos braços, a cadeira de palhinha e o olhar direto. Dedos, estrutura facial e reflexos dos óculos foram aprovados.
Prompt 5: Movimento urbano: editorial feminino no metrô
Prompt completo
Use case: identity-preserve
Asset type: cinematic urban female fashion editorial, vertical 4:5 photograph
Input images: Image 1 is COMPOSITION, ENVIRONMENT, MOTION AND LIGHTING REFERENCE ONLY; do not use its original face or hairstyle. Image 2 is the sole facial identity reference for the woman (REFERENCE IDENTITY B).
Primary request: Reconstruct the dramatic subway-platform portrait from Image 1 using the exact recognizable female identity from Image 2.
Scene/backdrop: clean arched European-style underground station with cream ribbed tiled ceiling, fluorescent strip lights and a vivid yellow train moving behind the subject; platform edge lines visible; no readable signs or ads.
Subject and identity: exactly one adult woman. Preserve the precise facial geometry from Image 2: oval face, warm medium-brown skin, dark almond-shaped eyes, defined dark brows, straight nose, full lips, softly contoured jaw, long dark hair. She must unmistakably be the same person, not a lookalike. Do not copy the face from Image 1.
Pose/expression: woman stands calmly in the foreground, body angled three-quarters right while shoulders remain open to camera; hands naturally inside coat pockets; head turned toward lens; confident neutral expression with relaxed lips and direct gaze. Stable body contrasts with the rushing train.
Wardrobe: long minimalist black wool coat with clean lapels over a black top, understated small gold earrings, no logos. Retain her long dark hair but style it in loose shoulder-length-looking layers gathered behind the coat so identity remains faithful; a few strands lifted by the train breeze. Natural refined makeup.
Lighting: mixed practical station light and soft warm reflection from the yellow train; key from upper front-left, subtle fill from platform; realistic edge light along hair and coat; face properly exposed while background retains motion streaks.
Camera: vertical three-quarter portrait, camera near waist-to-chest height, slight low angle, 35mm full-frame lens look, around f/2.0, shutter approximately 1/20-1/30 second aesthetic so subject stays crisp while train has horizontal motion blur, shallow-to-medium depth of field.
Composition: subject centered slightly right of middle; yellow train creates strong horizontal leading lines; repeating ceiling arches frame her head; upper lights form diagonals; enough negative space around face; no awkward crop at hands or head.
Color/grade: bold yellow, deep black and warm cream; cinematic contrast; gently muted skin highlights; cool-neutral shadows; subtle analog grain and slight halation around lights.
Realism and quality: premium location fashion photography, authentic skin pores and fine hair, realistic wool texture and train motion blur, physically plausible mixed lighting, coherent body proportions, high resolution.
Constraints: exactly one person; no commuters, extra people, limbs or fingers, no distorted hands, no text, logos, watermark or signature, no plastic skin, no artificial eyes, no facial asymmetry artifacts, no hybrid or substituted identity, no frozen train. Identity fidelity first, face sharpness second, anatomy third, environmental motion fourth.Como usar
A força do resultado vem do contraste entre a modelo nítida e o trem em movimento. Preserve o rosto como prioridade e use a descrição de obturador lento para criar o rastro horizontal no fundo.
Observação técnica: Referência interpretada com lente equivalente de 35 mm, ângulo discretamente baixo e obturador visualmente lento para registrar o trem em movimento com a modelo nítida. A imagem final preservou o contraste preto/amarelo, as linhas da estação e o rosto da referência feminina. Sem pessoas extras ou problemas anatômicos aparentes.
O que mantém o ensaio consistente
Mesmo com cenários e figurinos diferentes, a série permanece coesa porque os prompts repetem três prioridades: fidelidade facial, anatomia correta e iluminação fisicamente plausível. A direção de câmera e a paleta mudam em cada fotografia, mas a identidade não é reinventada.
Ao adaptar os prompts, altere primeiro o cenário ou o figurino. Evite mudar muitas variáveis ao mesmo tempo e sempre revise olhos, mãos, óculos, proporções corporais e quantidade de pessoas antes de aprovar o resultado.
Quer criar imagens profissionais com IA?
Se você quer aprender a estruturar prompts profissionais, preservar identidades e desenvolver seus próprios ensaios com inteligência artificial, conheça o treinamento completo.
