Sep 2011
1 / 3
Sep 2011
Sep 2011

I have a character that I'm trying to rig for a game.

I have a high quality model and two decreasing quality LODs that I'm trying to rig. This process has always been a pain for me because of how redundant the work flow is. I have LOD 1 weighted and deforming properly but I'm trying to reuse the rigging information and apply them to my LODs without repainting weights.

I tried using the Wrap deformer, not expecting it to work correctly and achieving the expected result. The models animate great in Maya however the skeletal binding information I'm looking for douse not pass threw.

Is there a way to rig LODs without repainting all the weights?

Ryo

P.S. Maya 2K9

  • created

    Sep '11
  • last reply

    Sep '11
  • 2

    replies

  • 3.8k

    views

  • 1

    user

  Is there a way to rig LODs without repainting all the weights?

Sure, but generally easier to do the weighting on the highest level, since its most detailed one this way you dont need to hunt for problems later.

There are several approaches, one of them transfer information by baking things in a uv-map, this has been available in maya since forever i mean literally 10 years now. This of course relies on you actually having non overlapping and similar uv maps to work with. This method has some drawbacks so its not the best possible method.

Never (including 2009) mayas has this cool tool set for transferring mesh attributes (mesh -> transfer attributes). and it can pretty much transfer any attribute across meshes if used with care. Conceptually you do for cluster wights you could just dump the values of your skin weights to per vertex colors (one color set per joint), transfer and reverse. However theres a tool called Skin > Edit Smooth Skin > Substitute Geometry that does something similar.

I prefer the dump with colorsets, because the transfer is live so i can later edit it if i need to and it updates all the LODs. Besides it uvmaps things too.

If these aren't good enough then there IS a 4rd nd 5th option but it would require you to modify the way you work form scratch and is of not much use for existing models.

I always got mixed results from exporting skin weights. In the end I usually ended up repainting the weights after attempting to save working weights for similar topologies/UVs on canned riggs. The only working examples I can think of where on models that where in the same uvset, combined geometry and the texture size cranked WAY up in the options. Still, my current model is in multiple pieces and has altered UVs because of the manual reduction I did for the LODs.

I'm not sure about cluster weighting because its the same problem as using the Wrap deformer. The skin weights wouldn't be pass threw they would be an internal Maya node that wouldn't export into something a game would use... unless it was ID tech then I could just do it all with blend shapes =P but I wonder if I'm over simplifying the logic and missing out on a process I should look into.

The Vertex color sounds like it could work because it could be scalable and transparent to my work flow. If I could embed the skinning data in the vertices I could just reduce as necessary and bake from the vert data. Though its a nice theory that kind of stuff is a bit higher than my experience, until now I haven't had much more reason to play with baking vertex colors other than pre-lighting. If you know of a tutorial on the subject that would be extremely appreciated.

Hmm, a substitute geometry feature? I wonder when they threw that in there. Though I'm not surprised it could have possibly been that easy. I haven't done a refresher on my rigging process since 6.5 and I'm still using adapted techniques from the Digital Tutors kits. Though I do remember something about that coming up when I updated I probably didn't look into it because I didn't have a project in queue to test it out on.

As far as changing my work flow, I'm up for suggestions because I've been hitting this snag in the pipeline more and more as I work on game content and its something I need to change. If you have any ideas I'm more than willing to hear them, I'll be the fool if I stick with this method and not listen =3

fire away!

Suggested Topics

Topic Category Replies Views Activity
Maya 4 375 13d

Want to read more? Browse other topics in Maya or view latest topics.