Greetings all! I am new to POVray so please bear with me if my questions
seem strange. I have checked POV documents (3.6.1 Reference and
Introduction) and some related books (Henrik Jensen's, Glassner's etc),
websites and did some testing myself, still can't find a satisfying answer.
Before I go further, I'd like to get your opinions and insights on this.
Thanks in advance.

I know POV is designed for computer graphics rather than optics. Still,
since it is largely based on precise optical principles, I am hoping it can
simulate optics accurately, for example, caustics on the bottom of a
swimming pool.

I used two perfect sine waves at 90 degrees to each other to make a fake
wavy water surface and it showed square caustics on the bottom, when photon
mapping used. Since in real life, lights will be focused and defocused in
the water body with depth, I put in a slope in the water (isosurface,
pretty much from surface to bottom) to see if indeed it'll converge and
diverge. It seems showing only divergence (focused region/line get wider
and wider with depth). I varied the water depth and wavelength of waves
enough to make me believe that POV is not doing what I hope it was doing. I
had thought when photon mapping was used, it stores actual refraction point
locations for later ray tracing so it'll be very accurate when enough
photons used. Why am I wrong here? Are there any corners being cut to
"fake" these caustics?

This is a crude test so I hope I did something wrong. If not, how difficult
is it to make POV optically accurate (in this particular case and maybe
more)? Are there any efforts out there on this? Who is the person to
contact in POV team? By the way, I am an optical guy and I have to admit I
am confused about the interior media scattering/absorption as well: why an
extinction value is used relative to scattering to describe absorption?
(P256, POV Reference 3.6.1). It seems to me it'll be much easier to
understand and use if absorption and scattering were just two parallel
properties of the interior media. I was surprised to see emission
properties though and for a second, I thought this is
something like fluorescence but I digress. (hey how about adding
fluorescence later?! it'll surely make underwater world more interesting)

Here's what I used for waves and slope (help from Christoph Hormann page):

isosurface {
  function {
    z - 4-0.01*sin(50*x)-0.01*sin(50*y)
  }
  max_gradient 2
  contained_by { box { <0, 0, 0>, <18, 18, 5> } }
  material { M_Watx }
  hollow on
  photons {
    target
    reflection on
    collect off
  }
}
isosurface {
  function {
    z+2.4-1.2*y
  }
  contained_by { box { <2, 2, 0>, <4, 4, 4.8> } }
  material {
          texture { pigment {Blue}  }
               }
}

and

#declare M_Watx =
material {
  texture {
    pigment {
      color rgbf 1
    }
    finish {
      diffuse 0.0
      ambient 0.0
      reflection {
        0.0, 1.0
        fresnel on
      }
      conserve_energy
      specular 0.4
    }
  }
  interior {
   ior 1.3
   }
}

Thanks.

- Will

Post a reply to this message

Weilin Hou wrote:
> 
> I used two perfect sine waves at 90 degrees to each other to make a fake
> wavy water surface and it showed square caustics on the bottom, when photon
> mapping used. Since in real life, lights will be focused and defocused in
> the water body with depth, I put in a slope in the water (isosurface,
> pretty much from surface to bottom) to see if indeed it'll converge and
> diverge. It seems showing only divergence (focused region/line get wider
> and wider with depth). I varied the water depth and wavelength of waves
> enough to make me believe that POV is not doing what I hope it was doing. I
> had thought when photon mapping was used, it stores actual refraction point
> locations for later ray tracing so it'll be very accurate when enough
> photons used. Why am I wrong here? Are there any corners being cut to
> "fake" these caustics?

There are examples in the POV-Ray sample scenes showing how you can use 
photon mapping for generating caustics including glass focussing light. 
  You also have to understand the way photon mapping works and realize, 
that the appearance of caustics is much influenced by the photon 
settings used and not only by the scene geometry and textures.

If you still think there is something wrong in the way POV-Ray does 
photon mapping you should post a minimal but complete test scene and 
explain what you think should be different in the result.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 03 May. 2005 _____./\/^>_*_<^\/\.______

Post a reply to this message

Christoph,

Thanks for your kind reply and suggestions. I'll look more into photon
mapping and especially Nathan's paper and see if I missed something.

Have a great day!

Will

Christoph Hormann <chr### [at] gmxde> wrote:
> There are examples in the POV-Ray sample scenes showing how you can use
> photon mapping for generating caustics including glass focussing light.
>   You also have to understand the way photon mapping works and realize,
> that the appearance of caustics is much influenced by the photon
> settings used and not only by the scene geometry and textures.
>
> If you still think there is something wrong in the way POV-Ray does
> photon mapping you should post a minimal but complete test scene and
> explain what you think should be different in the result.
>
> Christoph
>

Post a reply to this message