Yes have done that also. But it was a while ago.

But you're on the right track. Each phoneme and expression has to be created from the Base obj as previously discussed. Pick a figure (let's say Sydney since that's who we are working with now) that has the phonemes and expressions that you want to include in the Talk Designer. Name the morphs EXACTLY the same in your new figure. Make sure you zero out the mesh to the default shape before you create the next phoneme. Then you can just use the Sydney G2 Viseme Map in the Talk Designer rather than create an entirely new one.. 

As far as the OBZ format. If you have Winzip or something similar, open that OBZ file in that software. Then export the OBJ out from it.  The files ending in Z are nothing more than compressed versions of the library content.