No tags yet.
Hi,
I've been looking into the idea of skinning a skeleton to a low poly mesh and then attaching this to a high poly mesh via a wrap deformer, to save time on weight painting and improve on realtime performance.
Is this still a viable workflow and still used today?
Also, do people still use influence objects now we have Maya muscle?
Thank you
We kinda do that here at DreamWorks (Wrap Deformers) but ours are a lot more fancy wrap deformers that can do tons of other things. (We have our own system and don't use Maya)
I like the workflow.
Javier
Right,
I assume all your other deformers are integrated?
I've just been taking a look at Moom's architecture, and there's an attribute on the main controller which allows you to toggle between different resolution's so I assume they've connected the visibility via set driven keys?
So a good workflow would be to paint the weights on a relatively low model and not bother skin the high res to the skeleton, but connect it via wrap deformer.
I've been told by my lecturer that the way polygons have come on (with smooth mesh) they have pretty much made nurbs and sub-d's obsolete for modeling and animation purposes. Is this true?
Cheers
GB
True, Polys and Sub-Ds are a lot easier to work with these days. Specially with all the unwrapping tools.
We still use NURBS though (or something similar to NURBS anyway) But we have some fancy tools to connect between them. Though we are starting to use more and more Sub-Ds these days.
What I like about NURBS is how easy it is to move along U or V. Which can come in handy.
J
greatbear,
you can actually skin the lores-geometry and "copy" the weights to the highres-mesh . works pretty well in most cases .
and once you need to create corrective-shapes for the highres-body , you can also trick around a bit ... 1st create corrective shapes for the lores-version , and once that is done , you can wrap the lores-blendshapes , to a highres-mesh , in order to get a highres-blendshape , after applying the highres-blendshapes it would be ideal to do some vertex-correction , and finally create the highres-corrective-blends .
that way it is much faster and easier , because in the first instance , you just need to move some lores-vertecies in order to get a nice approximation highres-shape .
but just one thing:
mayas wrap-deformer is really bad in my eyes - inbetween vertecies are usually not interpolated very well - and i am hoping on a better wrap deformer in the next version .
You must log in to post.