> I could be wrong about this, i wish i had the game now so i could test it again instead of making mistaken recollections but please just consider what i'm saying for one moment. in my opinion, that faces pic should be a sphere with 6 faces in each longtitude and 6 faces in each latitude. make a square test slab voxel that is 6x6 to test my theory.

Vipr,
I agree that your idea looks more appealing from a mathematical point of view (6x6=36), but it does not match my observation from screencaps I made.Among others, I used the Gcannon turret, a TS style vxl, to study the large standing faces.

In addition, if you make a crosssection of the sphere you have in mind, you'll see that you can't have a top-bottom and a left-right with 6 faces...There definately is a top-bottom-front-back-left-right side.
So with your opinion in mind, i'll see if I can come up with something that fits both points of view.

I did make a reference tile 6x6 that has the complete normal palette on it. I use the RGB shift sequences (over 8 rotation directions) that occure on the test voxel to find the matching colorshifting sequence on real TS voxels. This is how I can determine what type of surface corresponds with a normal index...