Anim8or Community

Please login or register.

Login with username, password and session length
Advanced search  

News:

Ian Ross has just released a book on Anim8or. It's perect for a beginner and a good reference for experienced users. It contains detailed chapters on every aspect, with many examples. Get your own copy here: "Anim8or Tutorial Book"

Author Topic: art tracer dielectric shadow and shading bug?  (Read 8127 times)

nobody_b

  • Newbie
  • *
  • Posts: 1
    • View Profile
art tracer dielectric shadow and shading bug?
« on: December 17, 2008, 03:02:09 am »

below is a render of two spheres that doesn't really make sense to me
each sphere has a dielectric material
and inside each sphere is another sphere
the difference is the left sphere has a larger inner sphere

now, why does the left sphere have a visible inner sphere
at first i thought i was the unit distance attribute of the dielectric material
but i experimented with that and it didn't seem to do anything
but even if it is, it would still not make sense;
the shadow of the dielectric sphere would have rendered the inner sphere completely black, like the right sphere
Logged

Dreadkb

  • Full Member
  • ***
  • Posts: 108
    • View Profile
Re: art tracer dielectric shadow and shading bug?
« Reply #1 on: December 17, 2008, 09:18:38 am »

I've had a similar problem in .97 but I'm just using transparency, not dielectric. If a shadow falls on a trans material, it goes dark like oil all over. idk what can be done.
Logged

Steve

  • Administrator
  • Hero Member
  • *****
  • Posts: 2126
    • View Profile
Re: art tracer dielectric shadow and shading bug?
« Reply #2 on: December 18, 2008, 03:05:42 am »

nobody_b, Anim8or doesn't handle transparent shadows, caustics and related issues.  All surfaces cast shadows.  The reason that the closer one doesn't is due to the "shadow bias" that is used to keep an object from self shadowing inappropriately.  This is a trick that ignores thing that are very close to a surface when computing shadows.  Without it the errors in floating point calculations would make all surfaces look really bad and splotchy.

Even worse, it is physically incorrect to model polygons as smoothly curved surfaces.  The jagged faces that are in the left sphere are due to this inconsistancy.  The shadow bias value has to be increased to a larger value than otherwise would be needed or this artifact would appear on many "curved" polygons.
Logged