It's a tricky one, this. I'm also having this problem with emissive media.
However, these white artifacts are occurring on container objects which do
not overlap, and even in reflections of the media in other objects.
Furthermore, the artifacts are different with every render, even when
nothing is changed in between. Antialiasing etc does not help. Since my
container objects are meshes, I originally thought that maybe polygon edges
were behaving like coincident surfaces or something, but I'm not so sure any
more.

I was going to deal with the problem by rendering the image at least double
the size necessary and then scaling it down in the GIMP, effectively
antialiasing them out externally, but the single spots I get on reflective
surfaces will not be ironed out so easily. My problem is that this is for
an IRTC entry, so editing the artifacts out individually (trivially easy)
is against the rules...

I am playing with media settings to see what I can do.

Bill

Post a reply to this message

"Random Pete" <pet### [at] gmailcom> wrote in message 
news:web.45200051e7045ab39cf609e10@news.povray.org...
> I'm working on a design which has a number of media-filled objects (using
> emission to produce glowing effects around other solid objects). Very 
> often
> where 2 objects overlap in the camera view, occasonal bright white pixels
> appear on the objects. I've tried rendering with anti-aliasing and a 
> higher
> max_trace_level and neither of these fix it. I've seen quite a few
> references to this kind of problem with media, but have been struggling to
> find any tips to solve it, or any discussion of the cause of this issue!
> Has anyone got any good advice?

I'm working on an emission media problem as well with some clouds .... 
(concentric spheres) and no_shadow keyword on the object nearest the camera 
appeared to have helped with my problem.

Jim

Post a reply to this message

Thanks for your suggestions, I'll let you know if I have any luck!

The media objects in question are
 - isosurface {f_poly4(x,y,z,0,1,-1,0,0)}  (a double-ended curved spike)
 - torus
 - sphere

There are many isosurfaces and torii in the scene, and one central sphere.
Artifacts are most prevalent where the isosurfaces are in front of the
sphere; however there are many elsewhere and (having looked at it again)
not necessarily where objects are in front of one another.

I'm rendering in MegaPOV since a bug in POV 3.6 (relating to spline
functions being passed to macros) often causes the scene to fail to parse.

Post a reply to this message

If your container objects are meshs, you
might try adding inside_vector, not sure, but that
might help.

Post a reply to this message

Bill Pragnell nous apporta ses lumieres en ce 02/10/2006 15:55:
> It's a tricky one, this. I'm also having this problem with emissive media.
> However, these white artifacts are occurring on container objects which do
> not overlap, and even in reflections of the media in other objects.
> Furthermore, the artifacts are different with every render, even when
> nothing is changed in between. Antialiasing etc does not help. Since my
> container objects are meshes, I originally thought that maybe polygon edges
> were behaving like coincident surfaces or something, but I'm not so sure any
> more.
> 
> I was going to deal with the problem by rendering the image at least double
> the size necessary and then scaling it down in the GIMP, effectively
> antialiasing them out externally, but the single spots I get on reflective
> surfaces will not be ironed out so easily. My problem is that this is for
> an IRTC entry, so editing the artifacts out individually (trivially easy)
> is against the rules...
> 
> I am playing with media settings to see what I can do.
> 
> Bill
> 
> 
Any jitter anywhere? jitter changes randomly from render to render. Try adding 
"jitter 0" in the media deffinition, it may help, at least, the problem pixels 
won't change.
You can also try changing the method used from the default (method 3).
Try increasing the number of samples. If using the default, keep intervals to 1, 
the default value.

-- 
Alain
-------------------------------------------------
Please hassle me, I thrive on stress.

Post a reply to this message

Random Pete nous apporta ses lumieres en ce 03/10/2006 09:52:
> Thanks for your suggestions, I'll let you know if I have any luck!
> 
> The media objects in question are
>  - isosurface {f_poly4(x,y,z,0,1,-1,0,0)}  (a double-ended curved spike)
>  - torus
>  - sphere
> 
> There are many isosurfaces and torii in the scene, and one central sphere.
> Artifacts are most prevalent where the isosurfaces are in front of the
> sphere; however there are many elsewhere and (having looked at it again)
> not necessarily where objects are in front of one another.
> 
> I'm rendering in MegaPOV since a bug in POV 3.6 (relating to spline
> functions being passed to macros) often causes the scene to fail to parse.
> 
> 
> 
> 
Try adding "all_intersection" to the isosurface.
Try adding no_shadow to the containers.
If a container is an union, try changing it to a merge.
Try adding "jitter 0" or "jitter off" to the media.
Try changing the number of samples.

-- 
Alain
-------------------------------------------------
"Sure, everyone's in favor of saving Hitler's brain, but when you put it into 
the body of a great white shark, suddenly you're a madman." 
             --Futurama

Did you know that Al Capone's business card said he was a used furniture dealer.

Post a reply to this message

"Tim Attwood" <tim### [at] comcastnet> wrote:
> If your container objects are meshs, you
> might try adding inside_vector, not sure, but that
> might help.

No .. I was using meshes in the scene but they weren't media objects, plus I
already gave them an inside_vector!

I've got rid of most of the artifacts now in any case - I've removed the
central sphere entirely from the image (it was an idea I was trying that
didn't really work out) I also (stupidly) was using method 3 with more than
1 interval. Finally I played around with the CSG to avoid coincident
surfaces. There is still the occasional artifact but since tonight was the
deadline, they'll have to stay or be manually edited out!

I'll post in binaries.images once the final renders are complete...

Thanks again for all the advice.

Pete

Post a reply to this message

Alain <ele### [at] netscapenet> wrote:
> Try adding "all_intersection" to the isosurface.
> Try adding no_shadow to the containers.
> If a container is an union, try changing it to a merge.
> Try adding "jitter 0" or "jitter off" to the media.
> Try changing the number of samples.

I'll remember that for next time I encounter problems! (Which I'm sure I
will - I love using media since it produces such wonderful effects). I've
posted my final renders in p.b.images.

Thanks again
Pete

Post a reply to this message

"Random Pete" <pet### [at] gmailcom> wrote:
> I'm working on a design which has a number of media-filled objects (using
> emission to produce glowing effects around other solid objects). Very often
> where 2 objects overlap in the camera view, occasonal bright white pixels
> appear on the objects...

If you're using spotlights in your scene, there seems to be a bug with
media-filled objects when using spotlights (and cylinder lights).  I only
recently came across this problem, and posted a discussion about it...

http://news.povray.org/povray.general/thread/%3Cweb.45125ba9d84b4b77cc29cf380%40news.povray.org%3E/

The only solution that I found was to turn off media_interaction in the
lights (which in your case would be OK, since you're using emission media.)
Perhaps this will help?

Ken W.

Post a reply to this message

Kenneth nous apporta ses lumieres en ce 26/10/2006 01:27:
> "Random Pete" <pet### [at] gmailcom> wrote:
>> I'm working on a design which has a number of media-filled objects (using
>> emission to produce glowing effects around other solid objects). Very often
>> where 2 objects overlap in the camera view, occasonal bright white pixels
>> appear on the objects...

> If you're using spotlights in your scene, there seems to be a bug with
> media-filled objects when using spotlights (and cylinder lights).  I only
> recently came across this problem, and posted a discussion about it...

>
http://news.povray.org/povray.general/thread/%3Cweb.45125ba9d84b4b77cc29cf380%40news.povray.org%3E/

> The only solution that I found was to turn off media_interaction in the
> lights (which in your case would be OK, since you're using emission media.)
> Perhaps this will help?

> Ken W.


Another solution: use a regular or parallel light and a mask. The mask is a 
simple box with a hole punched trough with a cylinder.

-- 
Alain
-------------------------------------------------
"Sure, everyone's in favor of saving Hitler's brain, but when you put it into 
the body of a great white shark, suddenly you're a madman." 
             --Futurama

Post a reply to this message