Optimizing Lightmapping/mr Bake For Game Export

[skullmunky]

I'm using mental ray texture baking to create lightmaps, and have a couple questions. Since I just sort of stumbled on this method, I have no idea if I'm doing it, ah, "correctly", as in "best practices."

Is there a way to make it adaptively adjust the size of textures per-object or even per-face? For example - I'd like large objects, like a huge open floor, to have a higher resolution lightmap than small objects, like parts of a window frame.

I did some work with Radiant and another project with CrystalSpace, both of which have a built-in lightmapper. I'm hoping to get the same kinds of results just out of Maya, for use with engines/editors that don't have this. Suppose I could just read the source code for either of those but there's only so many hours in the day

Is the usual workflow to actually divide up objects into bakesets by size/resolution?

Is there a good approach for quickly uv-mapping things for lightmapping, or is it basically the same approach as regular mapping?

My hunch was that with current cards, texture lightmapping's more common than vertex prelighting ... is that kind of true or too simplistic?

thanks for any tips ...

[andy-davis]

I'm reading your post now with exactly the same sentiments, but am disappointed that no one got back to you!
I've been using an in-house light-mapping tool on the command of my bosses, simply because we need to optimise lightmaps in the way you are talking about.
However, mental ray does a much better job at doing vertex lighting tasks. But because the lightmapping optimisation sucks, we can't use it.
How frustrating!
My guess is that there's a way to script the optimisation of lightmaps, but I haven't gotten round to looking into that yet.
Another to add to my list!

Unless anyone knows of any solutions....?