I have been trying to mimic an experiment where there are two boxes with
different indices of refraction one on top of the other.  I am still getting
weird artifacts in the rendering.  I have posted this problem here before and
got some advice and it looks better, but there are still problems.  I have also
posted an image on the binaries.images section to illustrate the problem.  The
bottom region should be clear and the top liquid should be a translucent
yellow.  At the top of the upper region there seems to be a weird line.  Also,
there are weird triangles formed.  I do not think these are all real.  I have
overlapped the two regions a tiny bit and used the merge operator (this seems
to work the best).  I did try separating the two regions a tiny bit instead of
the merge operator, but it still has these artifacts and I would prefer not to
do that since the real experiment does not have a flat interface, as this ideal
case does, and it seems there are other issues with a tiny separation.  I also
have the bottom region intersecting a plane (the floor) at the bottom to make
things look better (as from previous advice from this forum).  I have turned
diffuse and roughness to zero to try to mimic real life (they are supposed to
be liquids), but playing with them did not seem to fix the artifacts anyway.

Any Ideas,
Mike

Here is the code:
#include "colors.inc"
#include "textures.inc"

global_settings{max_trace_level 255}
camera {
     location  <0, 3, 260>
    look_at   <0, 3, 0>
    angle 2.5
      rotate <-30*1,-30*1,0>
   }
#declare MaterialAlcohol =
material {
  texture {
     pigment {
        color rgbf<1, 1, 1,1>
      }
 finish {
        ambient 0.0
        diffuse 0
         reflection {
          0.0, 1.0
          fresnel on
        }
        conserve_energy
        specular  0
        roughness 0
 }
  }
  }
    #declare HeavyLiquid =
material {
  texture {
     pigment {
        color rgbf<0.6, 0.5, 0,1>
      }
 finish {
        ambient 0.0
        diffuse 0
         reflection {
          0.0, 1.0
          fresnel on
        }
      conserve_energy
       specular 0
     roughness 0
 }
  }
  }
#declare AlcoholRegion=
box { <-1.899, -0.2, -1.899>, <1.899, 3, 1.899>
material { MaterialAlcohol }
  interior {
  ior 1.3800
}
 }
#declare HeavyLiquidRegion=
box { <-1.899, 2.99, -1.899>, <1.899, 5, 1.899>
material { HeavyLiquid }
 interior {
    ior 1.55
}
}
merge{
object{AlcoholRegion}
object{HeavyLiquidRegion}
}
light_source {
    <4, 8, 10>
    color White
  }
plane{y,0 texture{pigment{White}finish{ambient 0 diffuse 1}}}

Post a reply to this message

Wasn't it Mike who wrote:
>I have been trying to mimic an experiment where there are two boxes 
>with different indices of refraction one on top of the other.  I am 
>still getting weird artifacts in the rendering.  I have posted this 
>problem here before and got some advice and it looks better, but there 
>are still problems.  I have also posted an image on the binaries.images 
>section to illustrate the problem.  The bottom region should be clear 
>and the top liquid should be a translucent yellow.  At the top of the 
>upper region there seems to be a weird line.  Also, there are weird 
>triangles formed.  I do not think these are all real.

I reckon that they are real. They're just refracted views of the shadow 
and of the back of the object.

-- 
Mike Williams
Gentleman of Leisure

Post a reply to this message

I don't think you should actually be doing a merge instead of an union.
Although ior is specified in the interior block, its calculated when the
ray encounters a surface, which is removed by the merge.

Using a union, at least it doesn't produce wildly different
results between tiny overlapping and tiny separation, so I'd feel
more confident with that.

Post a reply to this message

Oh and I recommend a non-black sky because otherwise you never
know if your black areas are trace level artefacts, shadows
or some path in which the sky is visible:

sphere
{
   0, 1000 hollow pigment {color SkyBlue}
}

And before the usual suspects complain, yes, I'm perfectly
aware "hollow" has no effect here. But I do so dislike tracking
down strange issues later when media enters the scene ;)

Post a reply to this message

Christian Froeschlin <chr### [at] chrfrde> wrote:
> Oh and I recommend a non-black sky because otherwise you never
> know if your black areas are trace level artefacts, shadows
> or some path in which the sky is visible:

> sphere
> {
>    0, 1000 hollow pigment {color SkyBlue}
> }

> And before the usual suspects complain, yes, I'm perfectly
> aware "hollow" has no effect here. But I do so dislike tracking
> down strange issues later when media enters the scene ;)

  What's wrong in using sky_sphere, which exists for this exact purpose?

-- 
                                                          - Warp

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:
>   What's wrong in using sky_sphere, which exists for this exact purpose?

Various things. For instance, if you're using an image for the sky, you cannot
brighten or dim it using finish { ambient mySkyBrightness } if you map it to a
sky sphere, because - well, there's no finish on a sky sphere.

Similarly, you cannot use different sky spheres for radiosity (you may want a
strongly blurred image for that) and reflections (you'll definitely want a
high-res image there). With standard spheres, you just define two spheres of
almost same radius, and give the inner one a "no_image no_reflection" to have
it only affect radiosity. sky_sphere doesn't allow any of these statements.

Post a reply to this message

clipka <nomail@nomail> wrote:
> Warp <war### [at] tagpovrayorg> wrote:
> >   What's wrong in using sky_sphere, which exists for this exact purpose?

> Various things. For instance, if you're using an image for the sky, you cannot
> brighten or dim it using finish { ambient mySkyBrightness } if you map it to a
> sky sphere, because - well, there's no finish on a sky sphere.

  I didn't see him using an image map for his "sky sphere". His stated
purpose was to use something else than black as the background color.

-- 
                                                          - Warp

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:
>   I didn't see him using an image map for his "sky sphere". His stated
> purpose was to use something else than black as the background color.

Right, but the issues I described may be a good reason of making it a habit to
use true spheres instead of sky spheres.

In *this* particular case, you're right: A sky_sphere will do fine. Just like a
true sphere will, too.

Post a reply to this message

clipka <nomail@nomail> wrote:
> Warp <war### [at] tagpovrayorg> wrote:
> >   I didn't see him using an image map for his "sky sphere". His stated
> > purpose was to use something else than black as the background color.

> Right, but the issues I described may be a good reason of making it a habit to
> use true spheres instead of sky spheres.

  Why should it be a habit? Writing a sky_sphere is much less typing than
writing a regular sphere. If you ever need to change the sky_sphere to a
regular sphere (which probably will happen rarely), it's rather easy to do.

  If you make a habit of creating a sphere, you are only going to encounter
problems which a sky_sphere doesn't have.

-- 
                                                          - Warp

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> Warp <war### [at] tagpovrayorg> wrote:
> In *this* particular case, you're right: A sky_sphere will do fine. Just like
> a true sphere will, too.

... oh, and speaking of it: A simple background {color SkyBlue} will do the
trick, too...

Post a reply to this message