CompuServe Messages

#All

    25-Mar-94 22:01:01
Sb: #93527-#All
Fm: John Tissavary 71054,15
To: dave aguilar 76260,453
You got me pegged. I was the one taking a rest in the Xaos demonstrator's chair <g>. For text I never bother with morphing on the object level, I morph on the image level using a 2D morph package. You'd be surprised how much 3D looking morphing is done that way. I don't morph much of anything, though, I find it a royal pain in the ***, and not usually worth the trouble (though it can be a great effect sometimes). For character stuff, I've seen plenty of models that were created so well the seams were virtually un-noticeable, proving that morphs can be avoided for all but the most close up scenes. And skin.axp or spline based packages can take care of that just fine, too. There are a few ways morphing between dissimilar objects can be acheived, and while I'm not real familiar with how they work, I think I've got an inkling of two methods. As a matter of fact, there are a couple of packages out now that allow you to morph between objects with dissimilar vertex counts and orientations. One way is to sort of turn both objects to "dust" and build one from the other from scratch (I don't know if it's a particle system, but I imagine that would be one way to do it). Another way is to intelligently analyze (I know that's a vague statement, but I'm not familiar with specifics of algorithms that can acheive this sort of thing) the two objects and interpolate "tween" objects for the morph. I guess this would be the more difficult of the two to implement, but I know it can be done. There are some sophisticated polygonal reduction routines (I think the Yost group plans on releasing a couple) that do something like this, and I'm sure it's not too much of a stretch to apply it to multiple objects and morphs. As I said, there's a group or two out there that are working on this for IPAS3. The way I see it is there are varying levels of "intelligence" in effects routines, and the more intelligent the routine the less the operator has to set the objects up before pressing the button, he just has to "tell" the routine what he wants and it'll figure it out. Considering how braindead I can be sometimes, I vote for people developing "smart" routines that take advantage of todays powerful software and hardware. John Tissavary (La Luna cie)