> > PPS: also, can you post a link to your site about normals?
> > I think would like to reread it, but cannot find the link
> > anymore


> Well, i've started to do some tests right after I realized yesterday that, thanks to Olaf's efforts, I could use Excell to make voxels...amazing huh? :)

> Anyway here are some screenshots :

> http://www.eclipsemods.com/flyby/testvoxel01.jpg

> http://www.eclipsemods.com/flyby/testvoxel02.jpg

> The test voxel has a raster of 16x16 normal colors (total 256). One cel has an 8x8 voxel size to make observation easier

> A few quick observations:
> 1)
> rotating does not change the behavior of normals, except for 1 direction that eliminates almost all normals. This was clearly not what I anticipated. I expected the normals t change while rotating.
> I think it is because they are all in the same plane. When the angle between the lightsource and groundplane (the voxel is parrallel to it) stays the same, the normal values do not change.

> 2)
> they DO change when the the voxel is inclined.


ok i don't need to release my test slab voxels again; yours are the same. anyway i think i recall from my tests that if you use some normal numbers that are never used by the official voxels then none of the normals will react; they will all stay the same shade as you rotate.

btw how do you make your test slabs? i wrote an algorithm to make them quickly. i'll release the source but it needs to be integrated into your programs. it's not self sufficient, well maybe it will be if i work a little more on it.