PROMPT FOR IMAGE GENERATOR (use image conditioning) Inputs: - Con 1 = base scene (two-person studio portrait). - Ref 1 = reference person A (outdoor close-up in black ribbed mock‑neck sweater). - Ref 2 = reference person B (mirror selfie, ash‑blond hair, glasses, shirtless with gold necklaces). Task: - Start from Con 1 and keep the background, studio environment, camera angle, framing, crop, and composition exactly as in Con 1. Do not change any objects or structures from Con 1. - Replace the two original people with the people from the references, keeping the same positions, scale, and poses as the original figures in Con 1. Placement and likeness: - Left/back figure (eyes closed, head resting on partner’s shoulder/back): replace with the person from Ref 1. Exact likeness: same facial structure, brow shape, eyelids, nose, lips, jawline, age, and skin tone. Same hairstyle as Ref 1 — short platinum‑blond pixie with soft wispy ends and darker brows. Clothing: the same black ribbed knit mock‑neck sweater from Ref 1. Preserve the exact pose from Con 1 (eyes closed, gentle smile, cheek pressed to partner’s back). Add natural contact shadow on the partner’s back where the face rests. - Right/front figure (facing camera in profile/three‑quarter): replace with the person from Ref 2. Exact likeness: same facial structure, eyes, nose, mouth, jawline, light mustache, and skin tone. Same hairstyle — short ash‑blond side part with slight volume. Keep the round wire‑frame glasses and the two thin gold necklaces from Ref 2; shirtless athletic torso like Ref 2. Maintain the identical head angle, body turn, and arm position as in Con 1. Lighting and integration: - Match Con 1 lighting perfectly: soft studio key from camera-left with gentle roll‑off to the right, soft fill, subtle catchlights, smooth shadow transitions on faces and bodies, and accurate contact shadows where bodies touch and at the neck/shoulders. - Preserve Con 1’s tonal style: if Con 1 is black‑and‑white, keep a clean monochrome conversion with identical contrast curve and dynamic range; if color, keep Con 1’s color palette. No HDR glow or oversaturation. - Blend edges seamlessly: hair wisps, sweater collar, necks, glasses, and hands should integrate without halos, ghosting, double features, or mismatched grain. Maintain consistent film‑like grain across subjects and background. Photographic spec: - Ultra‑realistic high‑resolution full‑frame look, 35mm or 50mm prime, natural depth of field with soft background bokeh. Clean detail in shadows and highlights, crisp but natural skin texture (visible pores and individual hair strands), subtle micro‑contrast, no plastic skin, no painterly or cartoon look, no visible AI artifacts. Negative/guardrails: - Do not alter background, framing, or camera perspective from Con 1. - No extra people, no added props (no phones/earbuds), no duplicated or fused features, no warped anatomy, no banding, no oversharpening. Output: - A single ultra‑realistic image where the two original figures in Con 1 are perfectly replaced by the individuals from Ref 1 and Ref 2 in the exact poses and positions, fully matched to Con 1’s lighting and style.