texture coordinate generation...

I am using tex. coord. generation for the first time. I managed to look at the results and was a great relief , especially when all the time one has been giving texcoords explicitly.

Anyway, to my question… I’m using the GL_OBJECT_LINEAR param for GL_S and GL_T generation. When I use say a 128x128 texture or a 256x256 texture, y does GL map it to the same sized region?? Actually to get my question further clarified, I’m using mipmaps which r also auto gen’d using the glu lib. Shouldn’t GL be mapping a 128x128 sized poly with a unit strobe in the same texture space too?? As in, if I gave a 256x256 texture, only 1/4 of it (0.0,0.0) to (0.5,0.5) be mapped to that 128x128 quad.

Well, can someone explain this. What I’m worried about is that, I might be creating too detailed textures for polys, and it just might end up that there r so many of those in mem, and actually never being used. What will be the ideal size of textures (if there is any thumb rule that is )??

Anyway, all said and done OGL rulez .
TIA,
Nishal.

Textures in OpenGL are resolution independent. Texture coordinates are normalized, (0,0) is always the lower left corner, and (1,1) is always the upper rigth corner. If you want the actual size of the texture to affect the generated texture coordinates, you could either set the S and T planes or scale the texture matrix to achieve the desired effect.

As for optimal texture size, well, a good guess would be no larger than needed to get a good result. There’s always a tradeoff; detail vs. resource usage. Let the user decide what he/she wants by providing variable detail level.