Here's some new software in development at Keio University, which aims to turn a standard webcam into a real-time motion capture interface for avatars -- watch, watch:
Read more about it here: "[T]he researchers," reports Diginfo TV, "aim to develop motion generation software for standard PCs. That should enable various ways to apply and implement this system." In other words, Sony Online Entertainment's upcoming motion capture system for Everquest 2 isn't the only one out there. Which probably means I should repeat what Scott "Lum the Mad" Jennings told me just a couple weeks ago:
"Hey Wagner James Au, tell Rod Humble SL needs this tech, like, yesterday." Or in this case, yesterday and two weeks ago.
Hat tip: Connie Arida.
Tweet
Want this for SL please thanks you're welcome.
Posted by: Adeon Writer | Wednesday, June 13, 2012 at 03:47 PM
Let's imagine for a while that instead of some news coming from Japan, it was a blog from Rodvik announcing the implementation of such a technology into the viewer.
(10 minutes of hysterical laughter. "Blog" and "Rodvik" in the same sentence always make me this effect... Sorry.)
Now, the big question is how do we animate those dead mesh faces? As far as I know, the current system used by mesh avatars to simply smile is to replace the whole head with some alpha trick. Excuse me but I'm having another hysterical laughter break. Can you imagine the sudden explosion of Avatar Rendering Cost if people started carrying a dozen of heads... or more?
What we need first is a system to animate custom bones in any mesh. Next we would be able to simply dream of any kind of real time motion capture.
Right now, mesh import in SL is a joke. Arbitrary nonsensical formulas computed by a psychorigid uploader, animation and texturing possibilities close to zero, imaginary costs doubling to be able to use a script. (The so-called future improvement to the mesh costs calculation in relation to scripts is a very sad joke. Seriously, what difference does it really makes for the sim where the script is? Dear Rod, if you want to change us into cash cows, just make us pay to rez scripts in-world. Don't hide behind cryptic pseudo-logical formulas.)
I really don't want motion capture in SL. Not now. Not as long as all that LL is able to do is to drop some unfinished half-*ssed blobs of technology on our heads.
(Aaaaahhhh! It feels good to vent...)
Posted by: Riisu | Wednesday, June 13, 2012 at 11:44 PM
WANT!!!!
Posted by: satine | Thursday, June 14, 2012 at 01:00 AM
Technical hurtles aside, this is awesome tech and would be one of those things that LL should work to figure out. There are many challenges to opening up virtual worlds to a wider audience and this would help to bridge a major gap.
Soooo I guess what I'm saying is..... WANT!
Posted by: Commizar Janick | Thursday, June 14, 2012 at 06:37 AM
can it be programmed to ignore a yawn?
Posted by: Graham Mills | Thursday, June 14, 2012 at 12:52 PM
The beta group asked for morph targets over and over for precisely this reason, animating faces, but got a resounding "No".
Custom bones were claimed to be the solution, but who knows when it'll happen, and I'm not sure they'll be happy when someone tries to do a detailed face rig with ~50 bones.
Posted by: Raz Welles | Friday, June 15, 2012 at 01:32 PM
The beta group asked for morph targets over and over for precisely this reason, animating faces, but got a resounding "No".
Custom bones were claimed to be the solution, but who knows when it'll happen, and I'm not sure they'll be happy when someone tries to do a detailed face rig with ~50 bones.
Posted by: Raz Welles | Friday, June 15, 2012 at 01:32 PM