We now only have the bindings for (u)int{8,16,32}_t and float. This will hopefully simplify things.
Like with VOLUMETYPES it is centralising the definitions of the volume and voxel types.