Weird camera mappings?

Hi all,
I’m looking to learn more about rendering using weird shapes of camera maps. Toroids, regular polygons, rings, etc. I feel like I’ve seen people render using weird maps like this to render custom outputs but unfortunately I possibly am missing the language to search for what I need. In other words, I’m searching for a workflow to build custom “lenses” in weird shapes. for rendering out content, and what the limitations are (I’m aware this is probably a deep rabbit hole). Any advice? This is not supposed to be in lieu of reprojection for a room or anything this is for experimenting with weird renders.

Hello,
Your intentions are not very clear but, if I understand a little, would be to have the impression the image is mapped to an arbitrary volume ?
For that, I use a 3xD software as Blender to build a wrapping map (how the image is wrapped on the volume, some tutorial on the web), export that map and the fbx to TD.
Building a viewing system in TD (fbx, constant mat, camera, render) you can have the image of your file wrapped to the volume.
Be more explicit for your ask with some examples.
Jacques

Yeah maybe I shouldnt have asked, since this is such a vague idea i’m working with, but i appreciate your input for sure. What I’m looking for is different from the standard reprojection workflow for xr and also different from uv mapping etc. its more like: imagine you have a 3d pretzel shape uv mapped to a super high res 2d texture output. For each point on that 2d map, find the corresponding 3d point and get your pixel value by marching outward along the normal of that point on the shape. Its unfortunately difficult for me to visualize this. But I’m looking to be able to express this as a camera matrix. Am I way off in crazy land?