Here is my current workflow for bringing characters into a DirectX9/C++ app, which seems very painful to continue with long term
1) make a model
2) collapse the mesh and delete history
3) add rigging and animation
4) export to .x file
5) since the maya exporter doesn't work with the directx samples for some reason (static meshes do, ones with animation do not), use deep exploration to open and resave the .x file (since it's a trial i only get the first 5 frames of animation right now anyway).
6) take the model to my app (basically a modified MultiAnimation SDK sample) and test it there
So here are my issues with this:
-How can i remove step 2, or make it happen after the model is complete, for some reason if i don't do it, or if i do it after the rig is added, the model just shows as a couple polygons (i'm assuming it's just the first polygon set it hits). This means i can't change the model without re-doing the skeleton, and it means i need to keep 2 sets of the model (pre-skeleton and post).
-I wouldn't mind using some other non-proprietary format, but I don't see a complete end to end example for loading and animating a character in those formats like i do with .x files, and i'm not familiar with the features supported by the other formats over the .x file.
-Overall i'm just looking for a better workflow after the character is modeled to make sure it shows in a DX app, and to make sure that i'm getting a good file from maya.
Any tips or advice would be greatly appreciated