I wish I could say I know what you mean, but I don't have the patience to try my hand at 3D... i have too short a fuse, one too many guns, a lot of ammo and no love for technology...

Well... you'd need a bulletproof computer, I'm sure.

y'know, i was thinking the same thing.. Back when I was big one sketching and art I loved working with Adobe Photoshop CS2... or The GIMP.. but I got more into working... started hating life.. and sadly, forgot what I had a passion for, art.. :(

Wooo! So what was the problem, not pulling the levers in the right order or something like that?

(Trust you to correctly factor the relative time rather than just say you halved it. )

It's a bizarre workaround. The target is copied, mirrored, and then given a morpher (vertex manipulation) modifier. The modifier is fed a copy of the object that's in the inital state. That object then affects another copy of the intial-state target which has a skin wrap modifier applied. The falloff and distance influence values on the skin wrap modifier were out of whack leading to blah blah blah widgets blah blah reticulated splines blah blah...

...all of that to get a character to smile on both sides of the face.

(Fuck. Now that you've mentioned it, I've only just realised.)

Not that I completely understand (though I fully understand spline reticulation), I'm curious, were the influence values very out of whack (as in, actually completely nonsensical), or out of whack only just enough to look okay but actually be a problem?

Well, they were out of whack relative to the application of the modifier. I understood their function, but not their dependency on each other to get the job done. I had to blast the falloff upwards, not downwards. Which is very unintuitive. Falloff is supposed to reduce the amount of influence - and it does in most cases. But for the particular model I was using, for some reason it was living in opposite world.

Funny enough, for most morph targets in 3ds max I just use soft selection and select the two exact points I want and all, no futzing about with all this tomfoolery.

Kind of a shame they haven't implemented the Maya method of soft selection, where changes to one side are automatically done to the other side and all. Ah well, maybe in a new version.

Yeah, I imagine that works out okay. The real trick is in figuring out how low poly to go - too low poly and the expressions you create won't be very meaningful, but too high poly and you wind up with a whole mess of troubles.

The thing is, if you're working on two symmetrical morph targets, you're using two separate models. So jumping back and forth between each model, making sure that each vertex is selected, the soft selection falloff is at the right level, and the vertices have travelled the correct distance, in three axes, really shits up my rhythm. Being able to just not care and make a morph target is so conductive to the artistic process.

My faces clock in at around 3,000 polys before smoothing. I'm a control freak.

Or, conversely, come up with a symmetrical facial expression on an assymetrical mesh.

Well, consider a character with one eye bigger than the other, for instance. An assymetrical mesh that might need some symmetrical expressions - like raising the eyebrows!

Sure, this is why I create morph targets that affect only one side of the face, then mirror them over (now that I've figured it out). Each eyebrow has its own range of targets for example. I create targets on the muscular level for the greatest degree of control.

But pure symmetry in animation should be avoided - the character often ends up looking robotic. If each side of the face has its own range of targets, they can be dialled up to be vaguely similar - say 76 and 83.

I usually end up with around 40 targets - 34-36 of which would be symmetrical, the rest operating on the centre of the face (like nose twitching and nostril flare).

Ah, fair point there, fair point. I can see why you'd want to work on only one side in that case, to make sure that, say, one eyebrow's maximum height is the same as the other. XD

Right. If you start with the base mesh in a completely relaxed position (which makes the character eerily 'dead' looking), and each morph target describes the extreme tension of each muscle, theoretically it's impossible to go wrong and have the character create an expression that can't be achieved. Then again, if you're making cartoon characters, go nuts on the expressions!

Yeah, I know exactly how you feel. I don't use 3ds but I've used Blender for years and until they added raytracing in, getting accurate reflections was harder than pulling one's own teeth.

Hey,

Great news indeed. I recall making the suggestion for working more with facial expressions a little over two weeks ago. I'd like to think that one of my suggestions helped you pick a direction, but I get the feeling you've been beating your head over this problem for months.

I had the plan to mirror them months ago, but I only started tackling the problem about a week ago.

Ahh, cool. Well, hopefully the next time you start running into problems like this, you don't have a large blunt object nearby to vent your frustration.