A group asked these 3 (really good) questions, and I thought everyone would want to know.
1. For splicing, do we assume skeletons in the different files have the same structure (same number of joints, effectors)
(BTW: by splicing here, I assume you mean temporal – clip after clip – splicing). You only need to make splicing work for skeletons with the same topology. Making it work when things don’t match is hard. Your program should probably notice and give a warning, rather than just doing the wrong thing.
What would be nice is if you were able to realize that files were effectively the same, but only differed in the order the variables were in, or the names of the joints, or inconsequential stuff like that.
Being able to splice different topologies (and do decent transitions between them – which require blending) is hard since it requires converting from one topology to another. Not impossible – but not something you want to do for the first phase of the project.
2. What do we need to do with marker data beyond just reading it in? Such as convert markers to a skeleton (BVH format)?
No, you don’t need to convert marker->BVH (if you want to try it, it could be something for phase 2 – students have done it in the past).
Really, the answer is “not much”. Visualize it. You might consider performing operations on it (in part 2) to see how it compares with performing them on the skeleton.
3. Do we need an automatic way to convert marker data to skeletons, or a simple point / click method to insert limbs into marker data?
No, you don’t need to convert marker data to skeletons. (see above). You might want to have some interface for saying “draw a line between RELB and RWRI” (where those are two marker names). Having a text file for that is fine (especially since the naming isn’t always consistent).
{ 1 trackback }