[Bf-committers] static particles discussion

Ed Halley ed at halley.cc
Sat Dec 30 17:59:09 CET 2006


I've heard there's someone working/reworking the particle system.   
Here are some
observations I've made over the past couple weeks, trying to make a  
reasonable
wig setup for a character.  I've broken it down to discrete issues  
and invite
anyone to comment on relative merits or possible solutions.  In each  
case, maybe
I'm doing something wrong, but if so, could someone share an  
explicit .blend file
that shows how to solve the issue?

As background, the best dynamic hair I've gotten so far is by the use  
of multiple
soft-bodied guide curves attached to a scalp mesh that emits static  
particles.

* guide curves are too promiscuous
   Currently, a guide curve (a curve object with the "guide" physics  
field type)
   will control all particles on that layer.  I'd like to add a flag  
to the curve
   which says that it will only control particles from the curve's  
parent object.
   With this, I can make several scalp patches on one layer, each  
patch guided
   independently:  bangs, pate, sideburns, strays, ponytails, etc.

* guide curves disallow material displacement
   With guide curves, I could not get the "texture channel 8"  
displacement to
   work, nor any material-based displacement.  Maybe I'm doing  
something wrong
   but it looks like there are distinct ways these are handled in  
today's code.
   Without displacement along the lifetime, you can't achieve things  
like frizzy
   or wavy hair without modeling each kink in each curve.

* guide curves are immune to wind force
   I added softbody capability to a curve object, but could not get  
it to react
   to a wind force.  LetterRip was joking when he said maybe it  
didn't work
   because a curve has no surface area, but in actuality, I think  
it's very
   similar:  no force-application happens on a curve's nodes, just on  
meshes.

* guide patches would be great
   To model and style hair, supporting a "guide patch" would be a  
boon.  Strands
   would follow the warp (U direction) only, while the weave (V  
direction) would
   be rooted along the scalp at one end, and hanging freely at the  
other end.
   Currently, to emulate this, one needs several guide curves which  
have to be
   reshaped independently to style.

* curves cannot be weight-painted visually
   I suggest a weight-painting mode for guide curves be supported;  
each control
   point would get a small colored sphere or dot that shows the  
weight in the
   familiar red-green-blue false color scale.  Perhaps the  
perpendicular tilt
   lines could be similarly colored/interpolated to show the effect  
more clearly.

That's enough for now, let me hear your thoughts.

--
[ e d @ h a l l e y . c c ]






More information about the Bf-committers mailing list