Tue, Nov 12, 8:00 PM CST

Renderosity Forums / Poser - OFFICIAL



Welcome to the Poser - OFFICIAL Forum

Forum Coordinators: RedPhantom

Poser - OFFICIAL F.A.Q (Last Updated: 2024 Nov 11 8:37 pm)



Subject: Emotional expression in Poser


Grammer ( ) posted Mon, 20 January 2003 at 9:57 AM · edited Fri, 08 November 2024 at 5:22 PM

Attached Link: http://evolution.anthro.univie.ac.at/institutes/urbanethology/emosym.html

No Vicky in a temple but may be of interest for some of you - The link above goes to our current work in emotion simulation and facial expression using Poser - but basically all other 3d programms are possible, we also have a version runnning in 3ds. Graphics will be higher rez soon. The python script can generate any emotion and any blend of emotion scientifically correct. You just can use sliders for the emotions and other affects like pleasure and arousal. It uses 44 morph targets which are based on single muscle contractions. A version for V3 with asymmetries will follow soon - any interest ? Karl


SamTherapy ( ) posted Mon, 20 January 2003 at 10:06 AM

Definitely interested.

Coppula eam se non posit acceptera jocularum.

My Store

My Gallery


FrankJann ( ) posted Mon, 20 January 2003 at 10:26 AM

Yes this sounds very interesting. The artistic implications are excellent. Will it be possible to animate the expressions?


Grammer ( ) posted Mon, 20 January 2003 at 10:31 AM

We are currently working on the implementation of the animation part, which will also include correct head movements, tilts, eye movements and breathing - but this will take some time of research


ockham ( ) posted Mon, 20 January 2003 at 10:36 AM

Wow! Definitely interested, especially on the scientific level. I produce interactive learning tools in speech and hearing science, and would be THRILLED to have a model that directly morphs each actual facial muscle!!!!!!!! This has always struck me as a basic limitation of the existing Poser models; the expressions are empirically determined, rather than being pulled by honestly modeled muscles.

My python page
My ShareCG freebies


ockham ( ) posted Mon, 20 January 2003 at 10:37 AM

Especially with asymmetric morphs! A dream come true!

My python page
My ShareCG freebies


Valandar ( ) posted Mon, 20 January 2003 at 11:26 AM

Yowsers!!!

Remember, kids! Napalm is Nature's Toothpaste!


xoconostle ( ) posted Mon, 20 January 2003 at 11:30 AM

This is one of the most fascinating and intelligent uses of Poser I've seen. Thank you for sharing this work, Karl. I found the "mixed" emotions especially compelling, since in real life our faces do tend to express many things at once. VERY interested to read more about your research and work.


farang ( ) posted Mon, 20 January 2003 at 12:11 PM

I'm still not quite sure exactly what it does or how it works except that its some sort of python script. I just want to know how do I get my hands on it thats all.


lululee ( ) posted Mon, 20 January 2003 at 12:51 PM

Great stuff. I'm definitely interested.


leather-guy ( ) posted Mon, 20 January 2003 at 1:03 PM

This is fascinating. Does it function by re-setting existing dials or would it create new morph-targets (new dials) in Poser?


PabloS ( ) posted Mon, 20 January 2003 at 1:04 PM

Triple Yes! Definately interested in a V3 version! And like farang, I want to know where/when it will be available.


pdxjims ( ) posted Mon, 20 January 2003 at 1:19 PM

Michael 2 Xpression Magic takes a similar approach. You select upper and lower face expressions to create a total expression. Example: Top face could be "Concern up left" while the bottom half could be "Smile open silly". The total effect is a sheepish grin. Using "Snarl tight right" instead on the bottom half creates an angry, almost wicked expression. You then copy it from Xpression Magic and past it onto your character in Poser. Doesn't work correctly in P5 with the SR2.1 release though sigh. This is a pretty neat product. There are also expression pose sets available that take the same approach. Only part of the face is posed with each pose, and using them in combination creates the whole expression. The Poser 5 figures have emotion morphs built in. They're few in number, but combining different morphs creates a broad range of expression. One of the nice things about Don and Judy (see, I can say something nice about them).


Grammer ( ) posted Mon, 20 January 2003 at 2:05 PM

If it will be a product will be decided soon - the main difference to XPression Magic and the emotion morphs in Don and Judy is that our approach is completely non-interpretative which it allows to construct facial expressions from existing scientific databases. These databases and their interpretations have been collected on a muscular level. Not everybody would accept the interpretations of the emotion names presented in the "emotion naming" approaches oming from artbooks. Our final goal is a realtime simulation (not with Poser), this is only the beginning where we tested if it is feasable or not.


gryffnn ( ) posted Mon, 20 January 2003 at 2:27 PM

Hi pdxjims, The basic copy/paste function that supports MichaelXprssnMagic (also Vicky2, Stephanie and P4Fem) works in Poser 5 the same as P4, and I've heard from customers saying they're using them in P5. Settings for the new set of expression morphs I have almost ready for Judy copy/paste as usual. So let me know what problems are you having in Poser 5 and I'll see if I can help. Thanks for the mention, too; they're not very visible in DAZ's software section - Elisa/gryffnn


EricofSD ( ) posted Tue, 21 January 2003 at 1:08 AM

Very well done. Definitely worth the effort. Keep us informed of the progress.


hauksdottir ( ) posted Tue, 21 January 2003 at 11:27 PM

As an animator with 15 years experience, including creating hundred of avatars for chat sites, I will say that NO approach using "sliders in combination" is ever going to truly simulate human expression. I've been watching and arguing against this trend for several years now as various companies have shown their products in the trade shows. It is fake in origin and fake in result. If you start with the wrong hypothesis, no matter how tidy the argument, you still have bad science.... and in this situation, bad art. Declaring that "18% surprise plus 82% happy = very happy surprise" (for example) only satisfies programmers... and deceives people who thoughtlessly accept numerical approximations for reality. Forget the sliders. Ditch the named emotions. You may be providing a crutch, but it is a disservice to artists who will never learn how to observe and recreate for themselves. Who are you to say that this expression equals sadness or that this set of morph values reveals arousal? Scientific papers and degrees notwithstanding, there is no substitute for experience. None. An artist wanting to recreate a sorrowful expression is going to have to do more than dial the mouth down at the corners or pick a slider value. An artist who has never observed the weight of sorrow in a real human face, but takes some engineer's or psychologist's word that this number = reality is more crippled than an artist who struggles to interpret and reveal what he sees. 44 morph targets is a decent number. It is not quite the number of muscles in a human face (and nowhere near the 300 muscles in a cat face), but still a sizable number. Providing a product with a lot of morph targets may be a good thing. Just don't try to delude people into thinking that your stated values reveal emotion. Carolly


Privacy Notice

This site uses cookies to deliver the best experience. Our own cookies make user accounts and other features possible. Third-party cookies are used to display relevant ads and to analyze how Renderosity is used. By using our site, you acknowledge that you have read and understood our Terms of Service, including our Cookie Policy and our Privacy Policy.