Attribute format specifiers for MDL0 vertex attribute groups are now detected automatically rather than read & stored from BRRES files, and some basic quantization (to the nearest 2^-14) has been added during Blender export. This all means that mesh data exported from Blender is now significantly more compressed on average (the W1 map model dropping from 3.53 MB with the previous method to just 2.23 now, compared to its regular retail size of 1.55).
Attribute format specifiers for MDL0 vertex attribute groups are now detected automatically rather than read & stored from BRRES files, and some basic quantization (to the nearest 2^-14) has been added during Blender export. This all means that mesh data exported from Blender is now significantly more compressed on average (the W1 map model dropping from 3.53 MB with the previous method to just 2.23 now, compared to its regular retail size of 1.55).