
Documento originale
GPT Image 2 · en
Scene:
A square wide-angle scene inside a fictional after-midnight arcade that has been transformed into a vaporwave dream room. The camera faces a receding perspective grid floor, with glossy tiles reflecting purple, coral, and cyan lights. Brightness is high at the horizon and softer in the foreground, like an analog video still paused on a CRT. The image should feel cinematic but organized, with strong depth lines leading inward.
Subject:
Rows of empty arcade cabinets, chrome palm trees planted in square tile cutouts, pastel plastic stools, and a central glowing sun-disc sculpture at the far end of the room. The cabinet screens show only abstract moving color fields, waveforms, checkerboards, and unreadable symbol blocks. Overhead, mirrored ceiling panels fracture the neon palms and grid floor into repeated reflections. The air contains mild video static, but the architecture remains crisp.
Important details:
The visual language must balance vaporwave nostalgia with plausible physical space. Keep perspective lines coherent, reflections aligned with the glossy floor, and CRT screen curvature visible on each cabinet. Scanlines and RGB edge fringing should be stronger on screens and bright reflections, weaker on matte objects. Avoid any real game branding; all machines are fictional forms with blank controls and abstract displays.
Use case:
Y2K Vaporwave CRT single study 4/8, testing deep vaporwave composition, reflections, and CRT screen styling.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no real arcade titles, no illegible fake lettering, no people, no random cables cluttering the perspective grid.Come usare
- Copia tutto il prompt originale, inclusi i parametri.
- Adatta soggetto, testo e composizione alle tue esigenze.
- Usa uno strumento esterno compatibile; verifica accesso e risultato.
Prima di iniziare
- Controlla se serve un’immagine di riferimento. La compatibilità non prova il modello usato per l’esempio.
- Verifica nomi, numeri e testi; i risultati cambiano con le versioni del modello.





