If I may share some insight, the meshes are not resized with bounding boxes. They use
mad shader instructions with float constants stored in each xbg file.
Specifically, in the PMCP chunk (Position Constant Multiplier?) you will find two 32-bit floats. Each vertex coordinate is calculated like this: vp = vp*float2 + float1. Then you will get correct sizes.
It's probably the same result, more or less. but this way you'll know for sure it's accurate.
The same thing is done with texture coordinates using data from the PMCU chunk. You will see that many xbg models need a certain scaling for the UV.
You can see it all in action in my maxscript.
Good luck with your research!
Please post any requests or issues with my tools in the appropriate topics.
I'm sorry if I don't reply or if I ignore PMs. My time is very limited.
