If you’re referring to the colour difference between head and body, it’s just because it’s using a different material. The head and body are the one mesh.
They’re on the mesh itself. For clothes I was thinking of using the cloth sim - or failing that, sculpting them into a multires - or failing that, displacement maps. Or combination.
Someone mentioned using a deform cage earlier but I still haven’t got around to testing one… so if it all falls apart at some point I figure that will be a good time to try it out. At least then I’ll know how far I can get with the current approach.
They’re just driven by the rotation of the upper arm bone or the clavicle, and come into affect within a range of rotation that’s dictated by the curves in the driver editor.
I also added some ease-ins on the curves to smooth the transitions, but I can still see them popping, so that needs more work (possibly by overlapping the curves). Which, by the way, made me realise this might be the more direct solution I was originally looking for when I came up with that crazy eyelid rig.
Hi Chris! Awesome work!
I have a small question, how does your shape keys work? It seems that once I am turning on more than one shape key, my mesh is getting distorted.
Is there a good way to fix it?
Very impressive! It seems you’ve managed to wrestle the Fresnel node to work really well! I find the effect way too strong even with that IOR. The principled BSDF Fresnel is much more subtle but unfortunately the SSS with that node is inferior to the standalone SSS node.
BTW do you mind showing a quick screenshot of how your spec map looks?
Chris, I’m REALLY struggling with the face rig here, can you tell me what you used for the setup? what bones are there (if they were bones) and are they using corrective shapekeys?
by far your rig was the most optimal for a good result, trying to study your old lightwave rig but having some hard time distinguishing the bones
I want to use them for a game engine application, right now I’m sculpting FACS into Blenshapes but I’m not even sure of those being done right, the info about this is very rare around for some reason.
@hfilben Thanks, yeah it’s a pity the quality of the Principled SSS isn’t up to par (or so I’ve heard - I forget whether I’ve put that to the test myself).
My spec map is a bit rubbish… I need to do it again properly, but you can probably get an idea of the values from it.
@ComplexAce This might answer some of your questions:
FACS is an acronym I’ve seen before but hadn’t found reason to investigate until now… and it turns out it’s basically what I’ve been using all along! I arrived at my system by studying my own face and working out how best to group the muscles into individual shapes though.
Taking another look at the LW version I see that there are more bones visible, but they’re just hold bones to keep the jaw and parts of the face from being influenced by other bones in the rig. The shape keys are driven by controls which appear as dots or text descriptions.
There are corrective shapes for the jaw and eyelids in that version, but they were a nightmare to work with in LW. They’re much easier to do in Blender, but as mentioned I’m going to see if I can do without a jawbone this time.
NB: I corrected the above quote as I had initially thought I wasn’t using bones in the eyelids.
Damn you, Chris!! I am STRUGGLING to get my new rig to perform even close to yours, and now I see you even use about half as many polys! This is witchcraft, witchcraft I say!!
Added finger flexion/extension controls, and got a bonus posing feature for free by selecting all the controls and dragging them left/right/up/down from different viewpoints. This causes them to move at different speeds - which I normally find to be a hindrance when manipulating items - but in this particular case it turns out to be “handy”.
Also added IK damping thanks to this tutorial (which especially makes the fingers easier to control when at full extension), IK/FK switches, and general model/rig improvements.
The IK/FK switching gave me some troubles, which I ultimately reported as a bug - only to be informed by Brecht (within 2 hours of me submitting the report no less) that it was a cyclic dependency issue - for which he proceeded to prescribe directions to locate the error (which alerted me to a whole cascade of such errors), along with a snippet of code that made them all go away.
I got the video to play on Twitter without the sensitive content warning after trimming a few duplicate frames off the end and reposting it, so apparently it wasn’t about the gestures after all (or anything for that matter). So that cost me another couple of days of stress, thanks again Twitter!
Anyway obviously no offense was intended, so I’ll leave it be unless I hear of any actual objections.
At least then I’ll know how far I can get with the current approach.
