If any gfx people following me happen to know, a question I've had for a long time: why is the native MIP UV comp still the way it is? It seems like it's a bad way to do MIPs, because it doesn't work with borders or tiling. Why isn't the default like https://pages.jh.edu/dighamm/research/2004_01_sta.pdf …?
-
-
@cem_yuksel do you have an idea why MIPs don't support borderless filtering similar to the algorithm described in that paper? -
I can't say that I fully understand the question, but I can suggest two papers of mine that could help: Mesh Color Textures: http://www.cemyuksel.com/research/meshcolors/mesh_color_textures.pdf … Patch Textures: https://graphics.geometrian.com/research/patch-textures.html …
End of conversation
New conversation -
-
-
Texels are processed by 2x2 quads, so my guess is that using uv derivatives to compute mip level per-quad this way is very cheap if not free and just works in most cases.
-
The MIP UV specification I referenced does not change that. It is strictly a scale+bias of the UVs per MIP level. The derivatives are calculated precisely as before.
- Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.