Question

I’m currently learning about cube mapping, I found this tutorial: http://www.keithlantz.net/2011/10/lighting-and-environment-mapping-with-glsl/ He transform normal vector by multiplies with inverse of transposed Model matrix. I wonder why don’t he multiplies by inverse of transposed ModelView matrix but Model matrix?

Answer 1

cube map is defined in word space, it seems that the author wants to have texture coords in word space as well.

note that one normal is transformed to view space – for lighting, and the second is used only to lookup in cubemap.

two links for reference:

• http://en.wikibooks.org/wiki/GLSL_Programming/Unity/Reflecting_Surfaces
• http://www.ozone3d.net/tutorials/glsl_texturing_p04.php#part_4