You can export terrain data in LWO format using 2 methods, each of which is more appropriate for particular needs. The one Frank mentions is best for strict perspective matched rendering, where you mainly want to use TG2 for rendering terrain and then render objects in another application, using the exported terrain for matching and shadows, for example. The main advantages of this approach are that geometry detail is relative to distance from camera, meaning there is less geometry relative to area further away from the view and you save geometry complexity this way; the geometry exported can also account for overhangs and other non-planar features. The other type of export is heightfield-based, so overhangs don't get exported properly, instead they are cut off as a sheer surface.
Note that in either case, texture maps are not exported. That functionality is not currently available.
In any case Frank is correct that TG2's main strength is in procedural detail and rendering of terrain. While export of terrain for use in other application is a common need for various workflows, it's generally a good idea to try to do the majority of terrain-specific rendering in TG2 itself, where possible. Combining this with objects rendered in other applications for e.g. complex animations is one possible approach. This would require the aforementioned camera data exchange. TG2 reads Nuke .CHAN format currently, so you'd animate the camera in Carrara and then you'd need to find an exporter to that format for the camera motion data.
I hope that gives you a general idea of workflow to start with.
- Oshyan
