Looking for advice/feedback.Have you looked at how that's usually implemented? Is there some limitation with those that you are trying to avoid?
For my RPG, the final look of the game is a 2D fantasy map. I'm not sure how I can render this.
Essentially it's like Google maps where the camera can show the whole world or zoom in close to the player.
65,536×65,536 pixels for the whole map.
I think what I need to do is create a render function that can replace texture streaming. Give it an x,y position and a zoom factor and it'll produce that texture.
2048x2048 texture for around the camera. 1024x1024 texture at quarter res for around that, 512x512 texture at 1/16 res around that. first texture has a backup texture for rendering into.
Or I could do a sparce quad tree. Render into a 2048x2048 buffer and populate the quadtree from that. Sounds better since data wouldn't overlap and merging quads as they get away from the camera would be cheap. Downside is that I'd need a custom shader to draw this quadtree on screen. Could be too slow for VR.
If you hadn't noticed, this is Forum 2.0