monkeycloud opened this issue on May 06, 2012 ยท 90 posts
bagginsbill posted Mon, 07 May 2012 at 4:00 PM
Speaking of alternative mappings - there's no reason we need super high resolution for the whole sphere.
There are at least two approaches I can think of to better make use of space and time.
Both involve having higher resolution for the part the camera can see directly, and lower resolution for the rest.
Use two maps - one that is the full 360 by 180 (lighting and reflections), the other that is perhaps only 120 by 60 (background). Both can still be equirectangular format. I would set up a double image shader that overlays one on the other. I see from screen shots that in View you can render any subsection at will. So you could do the full image at 4K x 2K, and the concentrated background also at 4K by 2K, but the background would look like 12K x 6K. It would have the same appearance but 1/9th the space (and time to render it).
Use a single map that has a new mapping that isn't homogeneous. There are many possibilities for this mapping, but the construction of it might require a new tool. I can build a decoder for it in Poser nodes. The resulting image could probably be smaller than option 1 but still have the same information as option 1. Option 1 has redundant info - the full sphere is reproducing the partial image.
Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)