Hello,
I was wondering how much texture resolution of the models would affect the calculation processes and if there were any indication about it.
Thank you!
It’s complicated. Most of a pattern or texture’s effect on calculation time is from the evaluation of texture coordinates. If image-plane sampling is applied (-ps > 1 in rpict), then some pixels will be sampled and others will not. If your texture is “busy” at the pixel resolution, then the extra sampling will be expensive. If on the other hand, there is not much going on in your texture, or it is very low-resolution, then image-plane sampling will be more effective in reducing rendering time. If you are not doing image-plane sampling, then the resolution of the texture hardly matters to the calculation time. More complex coordinate calculations may cost more, however.
Thanks Greg for your reply. Mine are texture atlases (I think they are called like this), so one room is mapped with one texture containing all the details concerning 4 walls. They can reach 4096x4096 each. So I should not worry about the resolution but more about the complexity of the mapping coordinates, right? So mapping a single wall with a single planar map would be less costy in computational terms than collecting 4 textures in one?
I have not clearly understood why one should use the function you mention.
Thank you for your help.
If you are using a texture atlas and local coordinates passed through a Radiance triangle mesh using obj2mesh, then the coordinate calculations are reasonably fast. Combining everything into a single, high-resolution atlas will not slow things down, provided you do not run out of memory. On most modern machines, this is less of an issue.
Cristal clear, thank you.