To get a bone's world transform, you multiply a bone's local transform by its parent's world transform (if the bone doesn't have a parent, copy the local transform to the world transform). For the other way around, you multiply the bone's world transform by the inverse of the parent's worldtransform...

For a ghetto solution you could invert the axis in your vertex shader after applying the skinning, before applying the other transforms.

Well, what was it?

Can you provide the box file?

For the box you have 8 vertices, but your indices go up to 23, so that's probably not index data or there are more vertices.

Knowing which game it is would be useful.

Yeah. I've never seen it done like this before, but it certainly saves space and it doesn't take too many instruction to decode in the shader.

inline CVec2f SinCos(float f){ return CVec2f(cosf(f), sinf(f)); } void BDOGetTangentSpaceVectors(const CVec4f &input, CVec3f &Normal, CVec3f &Tangent, CVec3f &Binormal){ CVec2f bxy, bzs, txy, tzs; CVec4f scaled = input * (M_PI * 2.0f) - M_PI; //Rescale from 0.0 - 1.0 to -PI - +PI txy = SinCos(scale...

Open the exe in disassembler then look for references the the debug/log strings.

That's the older version of the format used by Planetside 2/H1Z1.

Some of your UVs are broken because you aren't scaling the SNORM ones.

As for the mirroring, that's just an engine coordinate system thing.

I haven't seen any variable length integers in MMB files.

Yeah, the Division's approach is certainly something new. But you probably won't get far without looking at the game's disassembly, because most of the per-file info is encoded in a way that's very hard to deduce from looking at it in a hex editor or something like that.

Or you could challenge yourself and try to figure it out on your own!

Those are just Flash files with compression enabled, you can look at the file spec for SWF to see how to decompress them.

You can get the PDF with the swf spec here: http://www.adobe.com/devnet/swf.html.