Re: trouble with mipmap minification filtering
Brian Paul <[email protected]>
| Newsgroups | gmane.comp.video.mesa3d.user |
|---|---|
| Message-ID | <[email protected]> |
Can you create an apitrace of the problem? -Brian On 09/28/2014 11:33 PM, David Hart wrote: > > So, I'm still trying to get mipmapping to work. I've spent countless > hours now trying to build and link Mesa the right way on mac and linux > and realizing I may be in over my head trying to reproduce this problem > in a C program. > > But I have discovered one very important detail: Mip-mapping is working > fine for me when I use the old fixed-function pipeline in Mesa, it only > breaks as soon as I start using shaders. > > Also, I can nudge the LOD level manually even when using shaders, but > the default LOD with Mesa & shaders is always sampling the base level > mipmap for me. > > Here is a pyopengl sketch of all the combinations of Mesa vs native > OpenGL, fixed-function vs shaders, and using the default LOD vs > modifying the LOD level manually. I'm drawing two polygons, one is > distorted by perspective, the other is drawn with an ortho projection > and distorted by moving two verts close together. I have dyed each > mipmap level with a different color, so the only thing I'm looking for > is multiple colors in both polygons. As you can see, there is only one > case where I don't get any colors other than the blue base mipmap level, > which is Mesa+shaders without manually setting LOD. > > https://dl.dropboxusercontent.com/u/364079/mesa-mipmap/version3/mipmap.html > <https://urldefense.proofpoint.com/v1/url?u=https://dl.dropboxusercontent.com/u/364079/mesa-mipmap/version3/mipmap.html&k=oIvRg1%2BdGAgOoM1BIlLLqw%3D%3D%0A&r=lGQMzzTgII0I7jefp2FHq7WtZ%2BTLs8wadB%2BiIj9xpBY%3D%0A&m=YWDgaf5wt4vbH78LYD%2BMiwdBEZIo0l8AuOweqeck6Yk%3D%0A&s=7565e0f1a3c57faeb1f859fd73ca22479517783cc484d8d92f6f77c32c8acee3> > > I'm still running the exact same code through native OpenGL vs Mesa and > getting wildly different results when using shaders. Am I doing > something wrong? > > -- > David. >