> What we did was take a copy of the APC from TS and change the normals flag to '4' (it is the last byte of the file, making it easy to find in a hex editor).

> Then flyby put both TS and RA2 voxel versions into the game and got them side by side.

> A few screenshots would be nice, flyby :)

> To make this experiment meaningful, maybe more people should try it, and with other voxels?

so maybe why my cube test failed is because i used a vxl with the header of 4?
yea, i think that's it, i made the BMP-1 from a rhino whcih uses the TS normals, whereas the cube was i think the apoc or something which i cleared and made it

btw, is the 2 or 4 changable?
it'll be helpful for using the TS perfect normals on a previously RA2 normalised vxl