The quality of that env map could easily generate about 10-15 amazing splat reference images. I imagine that would be more than enough. (?) I'm no expert. I just find all of this fascinating tbh. Also, I heard that an image2depth process of some kind would help out a lot with the final splat. I forget if I actually read about that - or if just wildly speculated about it one bar night. lol. The env map itself is GREAT! It does a REALLY good job at the polar parts of the map. It's very impressive.

Have you heard that we will be launching Text2Image on Monday. This model has it's own flavor and we would love for you to try it out. We want to see your wild imaginations. #GenAI #text2image #image2depth #creator #creative #workflowoftheweek #workflow #imaginethepossibilities

I'm comparing image-to-depth models. huggingface.co/spaces/shariq… vs convert.leiapix.com Do you know any other image2depth models I can try?

I don’t think we’re far out. It will be crude at first, but soon it will be very capable. There are already image2depth AI so full geometry is soon here

Depth2Image in StableDiffusion 2... how about Image2Depth?