"> So did you make the model using Blender and then you wrote the shading
> of it in RenderMan shading language and used POVMan to patch POVray into
> understanding the shading language.
>
> Correct?
>
That's a simply tweaked Fsss shader from POVMan distribution, and layered
image map pigment-texture with 50% filtering over, using UV mapping, coming
from Blender.

I working as a 3d modeler-rigger-animator (well, not with POVRay or Blender
:), so I know something basic about coding-scripting, enough for
understanding. If you use any commercial app., most of the common shaders
exists, so you really don't need to programming the same things again ...
This is homework, so I using freewares.

Just for now, there are no features like a SSS in Blender internal, or
external renderer called Yafray, so I using POVMan  - I'll render model in
Blender with 'env' option in material - this will leave empty places in
image where I'll add POVMan image, or I'll compose simply black-white POV
Man image with pure Fsss, over Blender rendering, using this 'env' alpha as
a mask - this can be done with Gimp.
Procedure is common in commercial 3d, because you have much more control and
ability for changes.

> How hard is it to come up with many faces like this in blender and then
> convert them to POV? Was this a premade face that took someone 100's of
> hours?
>
That's a subdividion surface (sMesh in Blender) model, where 'base' model in
Blender have about 4000 quads ( a whole body, with details like a fingers,
ears - these details are about a half of poligons) - then subdivided two
times - this is about 80000 quads (two times more triangles). 'Base' model
stay as is in Blender, I can use different resolutions at any time. When I
editing model, I work with 'base' of 4000 quads, or half of this, if I using
instanced symmetry.

Model don't need to be modeled just in Blender, Blender can import UV-mapped
mesh, among others vawefront obj. format, just this 4000 quads, not more.
Even it can be; this model is a hardly tweaked just in Blender - some parts
are from old sPatch models . Blender itself have very fast viewport OpenGl,
if you have good graphic card (nothing esspecialy, my first modeling with
this kind of models in Blender was with GeForce2GTS - 2001....).
Povanim will export Blender's  'viewport resolution' , in this case about
80000 quads -160000 triangles with UV-mapping - if you have 'texture face'
in Blender. You'll get Mesh2 in POV - with paths to original Blender
textures - in this case, I change these in POV-Ray editor.

Export time was, about, half of minute, or faster.
POVanim can, also, export Blender animation to POV, but, separate mesh for
every frame -
so I can't use Blender motion blur if I like to animate it, or to add some
motion-blur effects to a static image.
So I decided to use Blender for main rendering, POV-Man only for extra
features.

Also, I don't know for existing any of POV-Ray animation utilities with
ability to do serious animation, like commercial apps - Blender have the all
what I need, not just as a Max, but enough (function curves, bones,
constraints, morphing, motion clips, even some simulations of cloth and
rigid bodies).

Anto
http://www.matkovic.com/anto

Post a reply to this message