CGI / VFX
To reconstruct Rome in 350AD, our team combined traditional CGI tools with historical research. Leveraging Paul Bigot’s “Model of Rome,”, Blender, Unreal Engine 5, and topographical data from OpenStreetMaps, we recreated the city’s essence during Constantinian era. The Colosseum stood out as a timeless icon - reimagined through both historical insight and artistic expression.
generative AI
The scale of recreating living ancient cities demanded innovative solutions beyond traditional VFX capabilities. By incorporating generative AI into our established CGI pipeline, we faced both opportunities and unique challenges. While AI offered efficient solutions for environmental details and crowd dynamics, it required extensive human oversight to maintain accuracy. Each generated element needed careful curation, which oftentimes required further refinement with traditional VFX tools. This hybrid approach ultimately allowed us to enhance our core CGI work while maintaining the integrity of our reconstructions.
PRODUCTION
Managing the complexity of historical reconstructions required more than just technical precision—it demanded an adaptive iteration process. By integrating our custom visualization framework, we ensured seamless collaboration between CGI artists, AI specialists, and researcher team. This iterative approach allowed for continuous feedback loops, enabling us to refine visual details, maintain historical integrity, and deliver high-quality outputs more efficiently.