As for the media, you didn't forget the hollow, no?
I'll take a look at the scene now.

--
Tim Nikias v2.0
Homepage: http://www.digitaltwilight.de/no_lights
Email: Tim### [at] gmxde

> In this scene:
>
> global_settings { assumed_gamma 1.0 }
> camera { location  <1.5, 3, 3> look_at <1.5, -1, 0> }
> light_source { 10000, rgb 1 media_attenuation on }
> intersection { plane { y, 0 } plane { -y, 2 } pigment { transmit 1 }
>   interior { fade_power 1001 fade_distance 2 fade_color <1, 0, 0> } }
> cylinder { <3, 2, 0>, <3, -4, 0>, .3 pigment { rgb 1 } }
> cylinder { <2, 2, 0>, <2, -4, 0>, .3 pigment { rgb 1 }
>   finish { ambient 0 diffuse 1 } }
> cylinder { <1, 2, 0>, <1, -4, 0>, .3 pigment { rgb 1 }
>   finish { ambient 1 diffuse 0 } }
> cylinder { <0, 0, 0>, <0, -2, 0>, .3 pigment { blue 1 } }
>
> The camera is looking down at three white cylinders, which pass through a
> two-unit clear layer with red-colored light attenuation in its interior.
The
> one on the left uses the default finish, the middle one has only diffuse
1,
> and the right one has only ambient 1. (A fourth, blue cylinder simply
marks
> the height of the clear layer.)
>
> Notice the abrubt change in color at the bottom of the layer, in all three
> cylinders. This should not happen because the pigment is clear and the
> attenuating interior should result in only continuous changes.
>
> Also, the ambient cylinder is completely white all the way to the bottom
of
> the layer. The attenuation is having no effect on it, except where it
pokes
> all the way through. The problem is visible in the default-finished
cylinder
> too.
>
> These problems are making attenuation pretty much useless for what I'm
> trying to do (in particular, model some reasonable-looking water). So I
> tried using absorbing media instead. But when I replace the interior
> statement in the code above with:
>   interior { media { absorption rgb <1, 0, 0> } }
> absolutely nothing happens! All three cylinders are as white as ever.
>
> It doesn't seem possible to me that all these features could be so
horribly
> broken, but it seems that they are. Am I doing something wrong, or are
these
> bugs?
>
> Anders
>
>

Post a reply to this message