[Bf-committers] Bump mapping computation issue

[Stephane SOPPERA]

Hi,

Since I felt that bumpmapping in blender was behaving strangly, I 
decided to setup a simple test:
http://perso.wanadoo.fr/stephane.soppera/blender/bump_bug/bug_bump.blend
This test scene is made of:

    * two textured planes: these planes have a material with a texture
      for bumpmapping. The texture is a simple bitmap. The material has
      no specular, only diffuse. Diffuse function is lambert one.
    * a parallel light (sun light in blender)
    * an orthographic camera

For the animation, the light and the planes are parented to an empty 
that simply rotates around the Z axis.

Since the material diffuse function is lambert one and since the light 
is parented to the same empty as the planes, the colors of pixels of the 
planes should always be the same during the whole animation. (with 
lambert, colors of pixels only depends of point normal and light direction).

But that's not what is happening, see the result of that animation in 
the following video:
http://perso.wanadoo.fr/stephane.soppera/blender/bump_bug/videos/bug_bump_235.avi

To illustrate the result I excpected, I made an equivalent test scene in 
3DSMax:
http://perso.wanadoo.fr/stephane.soppera/blender/bump_bug/videos/bug_bump_3ds.avi

It seems to me that the problem comes from perturbated normal 
computation made by blender.
The formula for normal computation should be:
     perturbated_normal=initial_normal+dI/du*dP/du+dI/dv*dP/dv
where dI/du and dI/dv are the partial derivates of the texture color in 
texture plane (these are scalars), and where dP/du and dP/dv are the 
partial derivates of the position against texture coordinates (vector 
dP/du "points" in the direction of greater "u").

But in void do_material_tex(ShadeInput *shi) function (texture.c), 
normal is computed with:
                    shi->vn[0]+= Tnor*tex->norfac*tex->nor[0];
                    shi->vn[1]+= Tnor*tex->norfac*tex->nor[1];
                    shi->vn[2]+= Tnor*tex->norfac*tex->nor[2];
instead of dP/du and dP/dv, the x and y vectors of the space where 
shi->vn is expressed are used.



I thinks that's where the problem is coming from. Am I right?

About, shi->vn, In which space is it expressed?


Thanks,
Stephane

-- 
Stephane SOPPERA
http://perso.wanadoo.fr/stephane.soppera