"Tim Nikias" <tim### [at] gmxde> wrote in message
news:3CC85619.15FEFEB2@gmx.de...
> I'll try, but I'm thinking it all has to do either
> with the incoming ray being in such a strange
> angle to the glass, that its just bouncing around
> in the glass.

Probably, AA would probably help this as well.

-tgq

Post a reply to this message

You have some cool macros. The black spots are just what was expected, not
enough max_trace_level. Just a torus primitive placed into the scene still
had black dots at max_trace_level 40. I carefully examined a wireframe of
your meshes and found no twisting or overlapping which might cause inverted
normals or other problems.

The mesh did have some odd properties which might cause a some problems. The
inside intersected with the outside and the mesh was not closed. I fit the
inside into the outside and closed the mesh and the pinholes on the bottom
and combined the whole thing into one mesh2 statement. I can post it if you
would like, but I'm not sure if you would notice the difference in your
scene.

 -Shay


Tim Nikias <tim### [at] gmxde> wrote in message
news:3CC85E72.FAB68679@gmx.de...

Post a reply to this message

The black spots could be removed by cheating your glass and lowering the ior
to 1.

 -Shay

Shay <sah### [at] simcopartscom> wrote in message news:3cc96fba@news.povray.org...

Post a reply to this message

Ah, well, then it wouldn't look like real glass anymore, would it?

I am thankful for your experiments, but rather than having you send me
your mesh2 object, I'd like to know what exactly the problems are
with the mesh.
I think I can overcome the not-closed problem with some vector-
copy-pasting, shouldn't be too hard to implement into the macros. Would
be an enhancement, I guess then the macros would be doing the job
they are supposed to (merging meshes, attaching new ones etc).

The pinholes can probably be overcome with another macro not
included into the version I gave you, which I use for creating the surface
of the water: it takes one line of the mesh and connects all those to
a single point. Making either the pinholes larger and using that, or just
keeping the small pinholes, would probably do the job.
Where did the inside intersect with the outside? I did some trials,
and found no intersection (though I may not have looked closely enough),
but that could be the possibility on the bottom, the algorithm used
there doesn't fit the exact needs of being not-intersecting there...

Shay wrote:

> The black spots could be removed by cheating your glass and lowering the ior
> to 1.
>
>  -Shay
>
> Shay <sah### [at] simcopartscom> wrote in message news:3cc96fba@news.povray.org...

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

Post a reply to this message

Thank you very much.

DZ wrote:

> Holy cats-- that's beauteous. Rocks the Casbah. Great job.
>

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

Post a reply to this message

I wrote some new macros to get rid of the intersection
of inside and outside (an Uneven_Translate-Macro).
And some macros now copy other vectors instead
of doing stuff like vrotate(Vector,y*360), so float-
inconsistencies should also be taken care of.

But I tried with max_trace_level 256, and I still get
those black edges on the top-inside of the vase. What
did you exactly do?


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

Post a reply to this message

"Tim Nikias" <tim### [at] gmxde> wrote in message
news:3CCA9878.760010E0@gmx.de...
>
> But I tried with max_trace_level 256, and I still get
> those black edges on the top-inside of the vase. What
> did you exactly do?
>

I think that some of this is natural behavior. I tried displaying just the
Lathe_Outside object with pigment {rgbt 1} interior {ior 1.56} and nothing
else. I still saw some dark spots. The spots can be minimised by specifying
fade_color at .5.

I thought that the spots might be caused by reversed normals or crossed
triangles, but there is none of that in your mesh except for the bottom/top
intersection which you already took care of.

I do know that the default max_trace_level will give a max reached warning.

I closed the mesh by connectiong all of the meshes and then connecting the
outside pinhole with the inside pinhole. This gave a torus type mesh.

I'm sure you also noticed that the defaut aa settings are not sufficient for
the vase.

 -Shay

Post a reply to this message

Okay, so max_trace_level is of no real concern.

What do you mean with aa settings? Do you mean the
blur_samples of 125 should be higher? They'll be higher for the
final version, of course, but what would you propose?

Shay wrote:

> I'm sure you also noticed that the defaut aa settings are not sufficient for
> the vase.

I also took care of the vase intersecting with the background
mesh (by writing a macro which searches the mesh for a lowest
given axis... hehe, another one for the list), the water doesn't
intersect with the vase and has a better surface normal.

Oh, another attached image showing a new change of the vase.
Notice the nice bumps near the bottom? Adding surface-normals
with these useful sine-waves...

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

Post a reply to this message

Attachments:
Download 'newvase.jpg' (158 KB)

Preview of image 'newvase.jpg'

"Tim Nikias" <tim### [at] gmxde> wrote in message
news:3CCADCD4.FBA8867F@gmx.de...

The high blur settings may solve the problem. Other than that, you may have
to use +A0.03 or something like that. I just ran a pic at 1280x1024 with
default aa and shrank it to 640x512. The results were quite nice(attached)

 -Shay

----------------------------------------------------------------------------
----

Post a reply to this message

Attachments:
Download 'vaseJPG.jpg' (3 KB)

Preview of image 'vaseJPG.jpg'

The image will be traced even larger than 1600x1200, because
I will need a 600dpi resolution to print the image with a
color-laser-printer.

Also, I cannot combine AA and focal-blur, its one or the
other. But I'll probably use very high blur-samples, because
focal-blur appears much more realistic.

Well, thanks for your help! I think I'm going through the final
steps of working on blur/radiosity/photon details, cause the
image needs to be traced in a few days time...

Shay wrote:

> The high blur settings may solve the problem. Other than that, you may have
> to use +A0.03 or something like that. I just ran a pic at 1280x1024 with
> default aa and shrank it to 640x512. The results were quite nice(attached)

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

Post a reply to this message