Hi all !


A Mayan stela with the date glyphs of Dec-21-2005.



Any idea how to get rid of the silly dots on top of the stela?
It is an isosurface with function f_superellipsoid. But these dots also turn
up when using a real superelli of the same size and material.
Find the code at povray.binaries.scene-files with the same subject name.

Comments welcome!


Yours,
Bu.

Post a reply to this message

Attachments:
Download 'stela.jpg' (71 KB)

Preview of image 'stela.jpg'

I have not checked your code, but perhaps you could try moving the glass
block upwards very slightly, to make sure there is not a coincident surface
at the bottom. This could cause the strange dots you see at the top.

But if there is no coincident surface down there, then I am just rambling
for no apparent reason. =P

I like your picture though, very neat. Only thing I could suggest is perhaps
to get the glyphs to stand out a bit more.

"Burki" <nomail@nomail> wrote:
> Hi all !
>
>
> A Mayan stela with the date glyphs of Dec-21-2005.
>
>
>
> Any idea how to get rid of the silly dots on top of the stela?
> It is an isosurface with function f_superellipsoid. But these dots also turn
> up when using a real superelli of the same size and material.
> Find the code at povray.binaries.scene-files with the same subject name.
>
> Comments welcome!
>
>
> Yours,
> Bu.

Post a reply to this message

"etrask" <et7### [at] yahoocom> wrote:
> I have not checked your code, but perhaps you could try moving the glass
> block upwards very slightly, to make sure there is not a coincident surface
> at the bottom. This could cause the strange dots you see at the top.
>
> But if there is no coincident surface down there, then I am just rambling
> for no apparent reason. =P
>
> I like your picture though, very neat. Only thing I could suggest is perhaps
> to get the glyphs to stand out a bit more.
>
> "Burki" <nomail@nomail> wrote:
> > Hi all !
> >
> >
> > A Mayan stela with the date glyphs of Dec-21-2005.
> >
> >
> >
> > Any idea how to get rid of the silly dots on top of the stela?
> > It is an isosurface with function f_superellipsoid. But these dots also turn
> > up when using a real superelli of the same size and material.
> > Find the code at povray.binaries.scene-files with the same subject name.
> >
> > Comments welcome!
> >
> >
> > Yours,
> > Bu.

I tried to follow etrask's idea and moved the stella a little upward, but
it still persist. I also tried to reduce the accuracy of the
isosurface down to 0.00001, but still the same. I also tried to put a white
plane above the whole scene in case it would come from reflection of the
black sky ...

I just notice that the bottom of the stella renders quite dark.

Maybe someone else can help you. Could it be a numerical artifact (things go
better with a box instead of superllipsoid-like shape)? Anyway, I wonder
why this undesirable effect...

Regards.

    Bruno.

Post a reply to this message

Could this be down to max_trace_level?

I have read before that black dots in a translucent/reflective object 
could be caused by too low max_trace_level.

Dave.

Bruno Cabasson wrote:
> "etrask" <et7### [at] yahoocom> wrote:
> 
>>I have not checked your code, but perhaps you could try moving the glass
>>block upwards very slightly, to make sure there is not a coincident surface
>>at the bottom. This could cause the strange dots you see at the top.
>>
>>But if there is no coincident surface down there, then I am just rambling
>>for no apparent reason. =P
>>
>>I like your picture though, very neat. Only thing I could suggest is perhaps
>>to get the glyphs to stand out a bit more.
>>
>>"Burki" <nomail@nomail> wrote:
>>
>>>Hi all !
>>>
>>>
>>>A Mayan stela with the date glyphs of Dec-21-2005.
>>>
>>>
>>>
>>>Any idea how to get rid of the silly dots on top of the stela?
>>>It is an isosurface with function f_superellipsoid. But these dots also turn
>>>up when using a real superelli of the same size and material.
>>>Find the code at povray.binaries.scene-files with the same subject name.
>>>
>>>Comments welcome!
>>>
>>>
>>>Yours,
>>>Bu.
> 
> 
> I tried to follow etrask's idea and moved the stella a little upward, but
> it still persist. I also tried to reduce the accuracy of the
> isosurface down to 0.00001, but still the same. I also tried to put a white
> plane above the whole scene in case it would come from reflection of the
> black sky ...
> 
> I just notice that the bottom of the stella renders quite dark.
> 
> Maybe someone else can help you. Could it be a numerical artifact (things go
> better with a box instead of superllipsoid-like shape)? Anyway, I wonder
> why this undesirable effect...
> 
> Regards.
> 
>     Bruno.
> 
>

Post a reply to this message

I doubt raising the max_trace_level would help but you can give it a try. i
would recommend raising the max_gradient of the isosurface instead (a value
of 4 should work i guess). at least this always helped with my isosurfaces
;-)

Regards Roman

Post a reply to this message

"Roman Reiner" <lim### [at] gmxde> wrote:
> I doubt raising the max_trace_level would help but you can give it a try. i
> would recommend raising the max_gradient of the isosurface instead (a value
> of 4 should work i guess). at least this always helped with my isosurfaces
> ;-)
>
> Regards Roman

"Roman Reiner" <lim### [at] gmxde> wrote:
> I doubt raising the max_trace_level would help but you can give it a try. i
> would recommend raising the max_gradient of the isosurface instead (a value
> of 4 should work i guess). at least this always helped with my isosurfaces
> ;-)
>
> Regards Roman

I did some tests and found that max_gradient, adc_bailout, max_trace_level
all have no effect on the outcome. Neither does moving the superellipsoid
(except if you move it down so that it DOES intersect the plane).  This is
true for both the superellipsoid and isosurface versions.  The only thing
that I could change to remove the dots was reducing the ior value. In the
end, I'm not sure what causes the dots, but perhaps this will help someone
else identify the problem...

-tgq

Post a reply to this message

Thank you etrask and Bruno, David and Roman.

Before I read Bruno's post I already tested etrask's proposal and gave the
stela a translate 0.002 * y (1/1000 of its height). That already removed
most of the dots.

The carving should really a bit deeper. I have to balance the depth of the
carving with the refraction of the light in the shadow to the right.

Didn't try the accuracy trick as I feared the rendering time. BTW Bruno, you
must have spend a real long time with it, hope you didn't do it all with
photons...

The dark part at the bottom of the stela I think comes from the large ior.
Value is 1.765 - that of corundum (or of ruby what I just found out) ,
somewhere in between of glass and diamond. I will also try a lower value,
say that of emerald ( 1.575, ok emerald is hardly ever red but we're all
flexible aren't we).

To David and Roman: max_trace_level was 24. I'm asking myself if it can be
to high. But this is an the global block, not in the photons block.
max_gradient was 4. Wondering if higher values do anything

---
What I want to do further is to place a few tools that archaeologists use
around the stela. E.g. a brush, trowel, plumb bob, these kind of things.
Thinking about a background but still not sure what. Maybe a book shelf.


Thanks again all !
Yours,
Bu.

Post a reply to this message

Wasn't it Trevor G Quayle who wrote:
>"Roman Reiner" <lim### [at] gmxde> wrote:
>> I doubt raising the max_trace_level would help but you can give it a try. i
>> would recommend raising the max_gradient of the isosurface instead (a value
>> of 4 should work i guess). at least this always helped with my isosurfaces
>> ;-)
>>
>> Regards Roman
>
>"Roman Reiner" <lim### [at] gmxde> wrote:
>> I doubt raising the max_trace_level would help but you can give it a try. i
>> would recommend raising the max_gradient of the isosurface instead (a value
>> of 4 should work i guess). at least this always helped with my isosurfaces
>> ;-)
>>
>> Regards Roman
>
>I did some tests and found that max_gradient, adc_bailout, max_trace_level
>all have no effect on the outcome. Neither does moving the superellipsoid
>(except if you move it down so that it DOES intersect the plane).  This is
>true for both the superellipsoid and isosurface versions.  The only thing
>that I could change to remove the dots was reducing the ior value. In the
>end, I'm not sure what causes the dots, but perhaps this will help someone
>else identify the problem...

I stripped the problem down to a fairly minimal scene, to eliminate
anything to do with isosurface parameters, coincident surfaces, photon
artefacts, etc. and the dots are still there.

My guess is that they might possibly be real effects. The particular
shape of the superellipsoid causing infinite internal reflections.

If you set the max_trace_level to 100000, then there are fewer dots, and
the ones that are left turn red. That suggests to me that some rays are
getting trapped forever inside the object, becoming slightly redder with
each internal reflection.



global_settings {
     max_trace_level 24
}

background {rgb 1}
camera {
     location  <0.0, 1.5, -4.0>
     look_at   <0.5, 0.0,  0.0>
}

light_source {
     <200, 180, -200>                  
     color rgb 1
}
       

#declare gemmat=
material {
     texture {
          pigment {
               color rgbt <1,0,0,0.9>
          }
     }
     interior {
          ior 1.765                // Corundum
     }
}     


superellipsoid {
     <0.15, 0.15> 
     scale <0.6, 1, 0.25>
     rotate 15 * y
     translate z
     material {gemmat}
}

Post a reply to this message

> My guess is that they might possibly be real effects. The particular
> shape of the superellipsoid causing infinite internal reflections.
>
> If you set the max_trace_level to 100000, then there are fewer dots, and
> the ones that are left turn red. That suggests to me that some rays are
> getting trapped forever inside the object, becoming slightly redder with
> each internal reflection.
>


I also assume that this must be the case (I am unable to do any further
checks here at work, when I try to DL the source it gets corruped.  I'll
have to check later at home).

BTW I believe that max_trace_level can be maximum 256:

3.3.3.7  Max_Trace_Level
....
Values for max_trace_level can be set up to a maximum of 256. If there is no
max_trace_level set and during rendering the default value is reached, a
warning is issued

-tgq

Post a reply to this message

Hi!

I agree with Mike:


>My guess is that they might possibly be real effects. The particular
>shape of the superellipsoid causing infinite internal reflections.

When I set
max_trace_level 10
color instead if rgbt 1 to 0.96
and in the ior: fade_power 1001
then also black dots appear:

Bu.

Post a reply to this message

Attachments:
Download 'black_dots.jpg' (10 KB)

Preview of image 'black_dots.jpg'