Verified against Nano Banana / Gemini 3.1 Flash Image · 2026-07-24
Combine two separate photos into one believable scene
Blends two independently-uploaded reference photos into a single coherent image with unified lighting and perspective — Nano Banana's signature multi-image compositing feature, briefed so the merge doesn't read as two photos stitched together.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are compositing two separately uploaded photos into one single, coherent scene, as if both subjects had actually been photographed together in the same place at the same time. FIRST IMAGE a headshot of a woman in a navy blazer, photographed indoors under soft, even light. Preserve this subject's identity, proportions, and any defining details exactly as shown. SECOND IMAGE a headshot of a man in a grey sweater, photographed outdoors under bright overcast daylight. Preserve this subject's identity, proportions, and any defining details exactly as shown. COMBINED SCENE both of them standing together in a bright, modern office lobby, as if photographed for a company leadership page. Decide which of the two source images' settings (if either) the combined scene should be based on, or describe a new setting entirely if neither original background fits — but be explicit about that choice rather than defaulting to whichever image happened to be uploaded first. INTERACTION BETWEEN SUBJECTS standing side by side, slightly angled toward each other, both facing the camera with a relaxed, professional posture. Render this interaction physically — correct relative scale between the two subjects, believable eye-line and body orientation if they're meant to be looking at or acknowledging each other, and any physical contact (a handshake, an arm around a shoulder) rendered with anatomically plausible hand and arm placement rather than overlapping shapes that don't actually connect correctly. LIGHTING UNIFICATION use the indoor soft even lighting from the first photo as the base — it suits an office lobby setting better than the outdoor daylight from the second. This is the detail that most often gives away a composite: pick one lighting scenario for the combined scene and render both subjects under it consistently — same light direction, same color temperature, same shadow hardness — even if their original source photos were lit completely differently. Do not let one subject visibly carry over their original photo's lighting into a scene where it no longer matches. PERSPECTIVE AND SCALE both should be framed at the same medium-shot distance, roughly waist-up, standing at a natural conversational distance apart. Both subjects need to sit at a camera distance and angle that makes physical sense relative to each other — a subject who was originally photographed close-up next to one photographed from far away needs to be re-rendered at a consistent apparent distance for the combined shot, not simply resized to match height. OUTPUT One composited image where, if a viewer didn't know the two subjects came from separate source photos, nothing about the lighting, scale, or interaction would tip them off. If the two source images' lighting or perspective are too different to unify convincingly without a major style choice, name that trade-off explicitly rather than delivering a mismatched composite silently.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Merging two independent reference images into one coherent output is Nano Banana's most distinctive and heavily marketed capability relative to single-image generators — it can genuinely condition on more than one uploaded photo at once rather than only accepting one reference plus a text prompt — but that capability doesn't automatically resolve conflicts between the two sources, which is why the lighting-unification instruction is the single highest-leverage part of this brief: two photos shot under different light sources, color temperatures, and shadow hardness are the most common and most visually obvious tell of a composite, more so than any scale or interaction mismatch, because human perception is unusually sensitive to inconsistent light direction even when it can't immediately articulate why an image looks fake. Second, explicitly requiring a decision about relative camera distance rather than simply resizing subjects to match height addresses a documented artifact of naive photo compositing: two people who were originally photographed at different distances from the camera have different amounts of lens perspective compression on their features, so scaling one to match the other's height without correcting for that underlying distance difference produces a subtle but real proportion mismatch that reads as "something's off" about the combined image even when the two subjects are technically the same height. Third, requiring anatomically plausible physical contact — a handshake or an arm around a shoulder rendered with hands and arms that actually connect at the right point, rather than overlapping silhouettes — targets a specific and common compositing failure where two independently-generated figures are placed near each other convincingly but any point of actual physical contact between them looks approximate rather than physically grounded, which is exactly the detail a viewer's eye goes to first in a photo depicting two people interacting, making it the place a merge is most likely to visibly fail if it isn't named as its own explicit requirement.
Verified against
Nano Banana / Gemini 3.1 Flash Image Gemini 3.1 Flash Image · 2026-07-24
Changelog
- 2026-07-24 — Initial publish, verified against Nano Banana (Gemini 3.1 Flash Image) on merging two differently-lit headshots into one office-lobby scene.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
