Hugo wrote:
> 
> Hey hoo,
> 
> Nice with some comments, as long as it doesn't develop into a "war", and
> so far it hasn't.  :o}  I like the way you expressed your thoughts Ryan.
> Personally I am wether a fan of imagemaps or procedural textures, I'm
> sort of neutral.
> 
> BUT I try to understand why these questions arise in the first place. I
> think, the thing we must understand here, is that artists want praise
> for their "work". They don't want, or deserve, praise for something they
> didn't work for... With imagemaps, no one knows how much work has gone
> into the picture, the final raytracing, and how much praise it deserves.
> That's why some of us don't want to use imagemaps. It's frustrating,
> when people can't find out how to credit our work.
> 
> But I'm afraid it's an illusion to think, procedural textures are safer
> to use.. They don't necessarely show the truth either; the truth about
> the amount of work behind the picture. A lame POV-user can grap and use
> whatever procedural texture he finds...
> 
> Hugo

right.  one thing i think people forget is where image maps can come
from.  if you have a scanner and/or digital camera you can take pictures
and use those as image maps (this is the kind i assume aren't liked?). 
leaves are a good candidate for this.  one can also paint their own from
scratch.  if anyone has ever tried that, it's pretty hard.  i'm
currently trying to do it for one of my models.  another way to do image
maps is to make a prodedural texture, make the image, and then use it in
another image as an image map.  for those of you who remember my r2d2 a
few months back, you may recall that i originally made r2d2's head one
giant procedural texture.  it was giant because it was accurate, not
some made up blue and silver sections that some other people have made. 
the texture alone was over 2000 lines of code.  i couldn't have done it
without moray.  i then used a megapov spherical camera inside a hollow
r2d2 head to get an image map i could spherically map to my real model's
head instead of using the procedural texture.  why?  multiple layers of
specularity if i remember correctly.  my texture was many layers thick. 
most of each layer was transparent.  i thought this would mean that i
could make it so only the colored areas would be shiny on each layer. 
but that's not how it worked.  the layers added together so that the
bottom layer looked really shiny and the top layer wasn't.  i'm sure
there were posts about it during that time. in fact, i'm positive i had
asked any pov programmers to look into pattern based specualrity (i.e.
my r2d2 used patterns where most entries were clear, and only some had
color.  it would be nice to make the clear part not shiny). my only
solution was to 'flatten' my texture to one layer which was impossible
procedurally.  so i rendered a 4000 by 4000 image and then wrapped it on
r2d2's head.  then i gave him the proper shininess and even used it as a
bumpmap.  the original 2000 line texture is available at the moray site
for those who are curious.  it shows how one can get precise texture
geometry on some shapes.  organic shapes are another story, and i won't
get into that right now.

Post a reply to this message