Spherical Mapping

[Markus]

Thats odd, I recompiled it because I had some modifications myself. This is the actual result of the new code:

[Julio Jerez]

Is this the latest code?

[Markus]

Yes, I made a new checkout.

[Julio Jerez]

in teh last line do this

sphere[i].m_x = u;
sphere[i].m_y = v;


change to

sphere[i].m_x = v;
sphere[i].m_y = u;

It will align the pattern, let me see hwo it looks

[Julio Jerez]

Ok I made that change, it should be perfect now,

Let us try once again, Sync to SVN and let me knwo how it looks.

[Markus]

I think that did the trick. It is not perfect but it is better than before. Thanks for your great work!

(The rest of the sphere looks perfect, only on the top and bottom there is still a bit distortion)

[Julio Jerez]

that will never be better because the mesh topology uses a geodesic where every face has equal Area.
Open GL uses a latitude longitud to make the sphere so each face is not equal area and threfore teh UVs will be ditructed better.
the difference shows more on the pole, but normally that is not a problem because spherically mapped meshes just do not put detail on the part of the map tha cover the poles.

The only way to minimize that is to make a higher resulution mesh, and I believes that is what your OpenGl mesh shows.
My guess is that if you make the OGL lower res it will show similar distortion on the poles.