So apparently Sketchfab has these weekend contests and I thought I’d give it a try and squeeze a quick sketch in today, but once again miscalculated the whole endeavour. Initially, I didn’t want to go beyond Dynamesh, but after I was done with the base, it just made more sense to do it that way, and once I was there, I felt like halfway decent UVs might be a good idea, so I lost a lot of time for the texture and didn’t manage to get him posed. Sad panda :(
Today I tried out the Sketchfab exporter for Maya. It’s an old model, that I made long before I had a clue how nerve wrecking bad topology can be for skinning. I hadn’t rigged the character until now, and holy shit, without ngSkinTools, this would have been a nightmare.
The base is borderline embarrassing too, but it’s not really a full blown presentation shot, so I didn’t want to spend too much time on it. I probably should have used panel loops or insert mesh brushes. Well, next time. Until then, no more radial symmetry for you, mister.
A while ago, I started working on a (sort of) joint based cloth simulation workflow, for those cases where you would like to have realistic cloth animations but your platform doesn’t support actual cloth simulation.
I used the well known Maya nHair Spline IK trick, with an additional component constraint on the CVs of the splines, so they’d behave like a closed surface. The whole thing works surprisingly well, but it’s not very animator friendly, because nDynamics.
I have a couple ideas for how to work around it, but in the meantime, I figured, I could try to make a more animator friendly Rig that is more lean than the usual FK Ctrl mayhem and can replicate the look of simulated cloth.
This is my first approach. I use Spline IKs to drive the joint chains, every CV ring is attached to a cluster, which is driven by the circular control outside (parent + scale constraints).
The animation is hand-keyed (pretty roughly), but I’m sort of satisfied with it. The problem I ran into however, is that I don’t have control over the silhouette of the skirt segments – that is, I can only drive one segment as a whole, not the individual joint chains.
I think I found a solution for that already, but I’m still dealing with some double transformation issues there, but hopefully I’ll have them figured out soon.
Earlier this year I started a new job, and so, once again, five months passed without any update to my blog.
I didn’t have too much time to work on my own stuff, hence the lack of updates. I’ve been trying to adjust my scope a little, so here is something I started working on a while ago (but haven’t finished yet). It started out as a lunchbreak ZBrush sketch, without any clear goal in mind, besides “well, make a zombie, maybe?”
Because of an overabundance of zombies, I dropped the idea and moved on to “I don’t know, what if he’s some kind of jungle creep”. Still not finished, obviously, and it’s been a couple weeks since I worked on him, but it’s better than nothing, right?
Back to that (in lack for a better word) new job of mine. I started out making character models, but, due to necessity, had to do some more rigging and technical work as well. I finally got around to start learning Python, which is something I had on my agenda for quite a while now, and the circumstances rekindled my interest in rigging.
I never really got into it before, in a professional scope, because I have the tendency to be a bit too, uh, exotic with my rigs. This, however, is actually a good thing for once, because the project I’m working on requires some exotic workarounds here and there.
Something I can share is this system I’m currently working on.
I’ll be lazy and just copy&paste the description I’ve put on the Vimeo page:
I’m currently working on a script that adds a procedural wave motion based on a sine deformer on top of hand animated FK controls. The goal is to help animators to spice up their animations with a bit of secondary overlaps on things like tentacles, hair or any kind of dangly bits, while giving them enough freedom and control over how they want to animate the general motion of the joint chain.
The sine deformer has an expression that ensures that the offset value automatically loops, and a point on poly constraint to modify the Y-translation of a set of dummy objects. Another set of expressions converts the translation values to degrees, which I feed through a double multiplyDivide chain to locally modify the amount of the rotation with a multiplier value, and another weight value to blend how much of the procedural rotation is added to the FK controls.