After a 15-Year Hiatus -
I was annoyed at her elbows, knees, and a few other bits. I knew the topology wasn’t great, and I was running into the limitations. I’d looked at a lot of other topologies, such as Hippydrome, and wasn’t that happy with them.
I did yet another search and this time found Manuel Bastioni Labs links. I dowloaded a copy, generated a girl, and looked at the resulting mesh. Although suboptimal in a lot of ways, it was better than a lot of the meshes I’d seen.
So I adapted it to her form.
It looked better in the default pose, and in various slight changes to the default pose. The extra work I did on her butt really paid off - it looks amazing. Fingers? Nice. Toes? Not bad. Knees and calves? Not bad. Face? Too many polys, and the face rig is pretty crap now.
But worst of all were her shoulders which were way worse in any slightly-deformed pose. So bad that I ended up deleting all the other couple dozen renders I did, because they were so ugly.
Her elbows and knees seemed destined to be just as lame in many of the poses I want her in.
But Subdiv Bug?
One weird thing. This mesh is probably about 30-40% more dense than my original mesh, and like my original mesh, is all-quads, but for some reason the subdivision surface modifier is about 10-20x faster on it than on my mesh. Unsure what to make of that. I can leave the modifer enabled while posing, as long as I’m not on a rendered 3D view.
So maybe I can figure out what the difference is and do that on my original mesh. But then…
I’m kind of hitting the limits of what I want to do with this mesh and rig. I want to start getting into extreme poses where also her body interacts with other parts of her body, with her costume/clothes, with Annya-Leona, give her long hair, etc, and all that is looking like a huge timesuck.
As hinted early in this series, I’m not a professional artist. I can’t devote 80-120 hours every week to this endeavour. I just used up my entire Thanksgiving weekend to vertex painting, modifying, reworking topology and morphology of the MB-Labs mesh, and what do I have to show for it? Two lousy renders. Lousy in every sense of the word. They’re not that great.
What to Do Then?
So do I pull up the Blenrig deform mesh cage, and use it? That’s a lot of time involved.
Do I try to reduce the polycount of the MB-Labs mesh so I can pose it with Rigify? Probably more time.
Do I do one or the other of the above and do driven shape keys? I really don’t like this idea, since it’s just a lot of slogging through vertex twiddling in various extreme poses, and I still don’t know if I just model (say her shoulder) it in an extreme up, down, back, forward set of poses, and then linearly interpolate between them, if Blender does that, or if I’ll have to spend six weeks with a Python script to do it.
Time for another 15-year hiatus?
Honestly the latter feels like the right answer. I can get back to music, poetry, 2D vector art, and whatever else I might fancy.
This is turning into more frustration than enjoyment.
Okay. Here’s the original mesh again, for one last dark studio render.
If I come up with any great ideas the old mesh would be suited for, I’ll give it a go. The Hearts Girls 2D vector art does need being re-created.