From TDN

Two question:

Fist of all I think the line

OUT.bumpCoord = OUT.outTexCoord;

should read:

OUT.bumpCoord = IN.texCoord;

The next thing is the line: float3 objectspace_view_vector = mul( texMat, eyePos ) - IN.position;

This is supposed produce an object space view vector. Why is the eyePos transformed by the texture matrix? Assuming that the eyePos is given in world coordinates, it should be transformed by the inverse modelview matrix, to get it into the same space (object space) as the input position.

Or am I wrong on this?