Luma - Video to Video with Ray 3.2 - Create endless worlds

One performance. Three completely different worlds. For this experiment, I started with a single 10 second performance: a woman sitting beside a window, noticing something outside, leaning forward and eventually standing to get a better look. Then I used Ray 3.2 Modify inside @lumalabsai to reinterpret that exact performance three different ways: - A moving night train - An underwater observation station with something enormous outside - A futuristic spacecraft where even the character herself becomes alien The interesting part for me wasn’t simply changing the background. I wanted to see how far I could change the environment, lighting, production design, wardrobe and character identity while still keeping the core performance and camera movement recognizable. In the video, I also show part of the actual workflow inside Luma creating reference keyframes, using Modify Video, adjusting prompts and working with adherence controls to get the movement right. And yes, there was testing involved. Getting the exterior environments to actually move while preserving the performance took a few iterations. That’s part of working with professional AI filmmaking tools: the direction still matters. Same performance. Different context. Completely different story. Try Luma today: https://lumalabs.ai/claudiaai This video was created in partnership with Luma. #LumaPartner #AIFilmmaking #Ray32 #AIVideo #Filmmaking #GenerativeAI