In article <3f68c78a$1@news.povray.org>, z99### [at] bellsouthnet says...
> If I knew more about programming than shuffling string arrays around
> in C, here's what *I'D* try to make:
> 
> a program where you set up a rigid skeleton whose joints have a type
> parameter, and attach blob or nurbs muscles to points on the bones and
> give the muscles properties such as amount they can contract lengthwise
> and expand laterally when doing so and a variable that specifies the
> amount of contraction for each muscle, string a bunch of these together,
> and dip the lot of it in a trace function that follows the normal of
> the combined muscles/skeleton and wraps upon that a 'skin'...oh yeah,
> and fatty regions and hair.  someone else can do the IK thing, i'm fine
> with just manipulating the muscles directly.
> 
> of course, the whole thing would export native POV SDL.  maybe a macro?
> 
> 

Yeah. In some ways that would be even better. Some existing modellers do 
this, but the limitation is, like with Poser, it is easier to manipulate 
existing models than build them. However, there are known measurements 
and correlations in the human body. With these they should be no reason 
to have to build entirely new models for every different human model. The 
problem Poser has is that it lacks sufficient adaptability in this 
respect, so 99.9% of the 'models' all look like twins with different 
colored hair. lol

You still have to use a form of IK though in something that supports 
skeleton and muscle systems, it is just more complex than usual. If you 
want a good one, they also appear to more or less do only one thing well, 
tend to be hard to make decent models in or are insanely expensive. 
Though admitted the ones that bug me the most are ones that do exactly 
what I would like, but are "Maya plugins" of the like that can't function 
separate from the program they are designed for. :(

-- 
void main () {

    call functional_code()
  else
    call crash_windows();
}

Post a reply to this message