I realise that dispersion support is still in the 'alpha' stage as stated in
the release notes, but I've been using the following simple scene to play
around and spot differences between 3.7beta11c-ss2 and 3.6:

camera {
  location <2,1,-2>
  look_at <0,0,0>
}

light_source { <0,20,0> color rgb <1,1,1> }

plane {
  y, -1
  pigment { checker color rgb x color rgb z }
}

sphere {
  0,1
  pigment { color rgbt <1,0,0,0.9> }
  interior {
    ior 1.5
    dispersion 1.1
    dispersion_samples 25
  }
}

There are some significant differences between the pics produced by both POV
versions on rendering this scene:
1. In 3.7 there seems to be no dispersion-effect at all at the boundaries of
the red and blue tiles in the middle of the sphere, whereas in 3.6 there is.
2. In 3.6 there seems to be some sort of circular interference pattern in
the middle of the sphere which smoothes out on increasing the
'dispersion_samples' paramater to just a dark shadow.
3. The effect caused by dispersion manifests itself in this scene as
pinkish/blackish edges as the checker pattern is projected through the
sphere. In 3.7 these colour changes are completely inverted compared to
3.6.

Jeroen.

Post a reply to this message

In a related test-scene employing a hexagonal prism:

camera {
  location <2,1,-2>
  look_at <0,0,0>
}

light_source { <0,20,0> color rgb 1 }

plane {
  y, -1
  pigment { checker color rgb 0 color rgb 1 }
}

prism {
  linear_sweep
  -0.5,0.5,7
  <-1,-0.5>,<-1,0.5>,<0,1>,<1,0.5>,<1,-0.5>,<0,-1>,<-1,-0.5>
  pigment { color rgbt <1,0,0,0.9> }
  interior {
    ior 1.5
    dispersion 1.1
    dispersion_samples 25
  }
}

There are again marked differences between the output of pov 3.6 and pov
3.7beta11c-ss2:
1. In 3.7 'dispersion' has an effect on all facets of the prism, which is
probably a good thing; 3.6 shows no dipersion-effect when looking straight
through the prism.
2. However, the top of the prism in 3.7 produces a highly complex and pretty
awful looking pattern wheras 3.6 produces gives slight colour attenuation at
the top facet which seems much more consistent with 'the real thing'.

Jeroen.

Post a reply to this message

In fact, these differences seem to be - at least to a certain extent - due
to 'interpretation' of ior in general. In the following scene there appears
to be a marked difference between the results for 3.6 and 3.7beta11c,
visible mainly in the centre of the sphere.

Jeroen.

camera {
  location <2,1,-2>
  look_at <0,0,0>
}

light_source { <0,5,0> color rgb 1 }

plane {
  y, -1
  pigment { checker color rgb 0 color rgb 1 }
}

sphere
{
  0,1
  pigment { rgbt 0.8 }
  interior { ior 1.5 }
}

Post a reply to this message