 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi all!
I wanted to create a rgb-cube by putting together 10^3 smaller cubes
filled with an emission media. But there are some strange black
squares between the smaller cubes! Changing density, transmit and/or
filter values didn't help.
Does anyone know where the problem is??
Thanx in advance,
Daniel
Post a reply to this message
Attachments:
Download 'rgb_cube.jpg' (30 KB)
Download 'rgb_cube.pov.txt' (1 KB)
Preview of image 'rgb_cube.jpg'
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
> Hi all!
>
> I wanted to create a rgb-cube by putting together 10^3 smaller cubes
> filled with an emission media. But there are some strange black
> squares between the smaller cubes! Changing density, transmit and/or
> filter values didn't help.
>
> Does anyone know where the problem is??
>
> Thanx in advance,
> Daniel
I don't have time to look at your source now but try adding a
nax_trace_level in your global settings. I would start with
something like 30 or more with this many objects in the scene.
--
Ken Tyler
mailto://tylereng@pacbell.net
http://home.pacbell.net/tylereng/links.htm
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Ken wrote:
>
> >
> > Hi all!
> >
> > I wanted to create a rgb-cube by putting together 10^3 smaller cubes
> > filled with an emission media. But there are some strange black
> > squares between the smaller cubes! Changing density, transmit and/or
> > filter values didn't help.
> >
> > Does anyone know where the problem is??
> >
> > Thanx in advance,
> > Daniel
>
> I don't have time to look at your source now but try adding a
> nax_trace_level in your global settings. I would start with
> something like 30 or more with this many objects in the scene.
I didn't see the image either until now. Um there is a problem in
pov called coincident surfaces where if you have transparent objects
whose faces occupy the same space like two boxes abutting each other
the raytracer cannot decide which surface belongs to which object.
The easiest way to avoid this is to seperate them by a tiny amount
or have them overlap by a tiny amount. If you overlap them you can
also use a merge operation to remove all of the inner surfaces.
Read up in the docs about csgs and coincident surfaces.
max_trace_level can still be an issue so don't hesitate adding
that to your scene too.
--
Ken Tyler
mailto://tylereng@pacbell.net
http://home.pacbell.net/tylereng/links.htm
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Yes. there is a need for a higher max_trace_level. About 15 seems okay,
but 150 was better, somewhere inbetween perhaps then. I was going to say
it probably is coincident surfaces but the effect goes away, or appears
to anyway. The container objects being invisible to start with might not
make it obvious but still cause trouble, not sure. At the default trace
level it didn't show any change when I rescaled the boxes down slightly.
Ken wrote:
>
> Ken wrote:
> >
> > >
> > > Hi all!
> > >
> > > I wanted to create a rgb-cube by putting together 10^3 smaller cubes
> > > filled with an emission media. But there are some strange black
> > > squares between the smaller cubes! Changing density, transmit and/or
> > > filter values didn't help.
> > >
> > > Does anyone know where the problem is??
> > >
> > > Thanx in advance,
> > > Daniel
> >
> > I don't have time to look at your source now but try adding a
> > nax_trace_level in your global settings. I would start with
> > something like 30 or more with this many objects in the scene.
>
> I didn't see the image either until now. Um there is a problem in
> pov called coincident surfaces where if you have transparent objects
> whose faces occupy the same space like two boxes abutting each other
> the raytracer cannot decide which surface belongs to which object.
> The easiest way to avoid this is to seperate them by a tiny amount
> or have them overlap by a tiny amount. If you overlap them you can
> also use a merge operation to remove all of the inner surfaces.
> Read up in the docs about csgs and coincident surfaces.
> max_trace_level can still be an issue so don't hesitate adding
> that to your scene too.
> --
> Ken Tyler
>
> mailto://tylereng@pacbell.net
> http://home.pacbell.net/tylereng/links.htm
--
omniVERSE: beyond the universe
http://members.aol.com/inversez/homepage.htm
mailto://inversez@aol.com?Subject=PoV-News
Post a reply to this message
|
|
| |
| |
|
|
From: Dave Kreskowiak
Subject: Re: rgb-cube and problems with media
Date: 25 Jul 1999 09:49:27
Message: <379b15e7@news.povray.org>
|
|
|
| |
| |
|
|
I ran into a problem probably similar to yours. I was trying to render
billowing smoke from something like a gasoline or oil fire. There spheres I
was using for containers intersected (overlapped) each other without any
coincident surfaces. The black areas I got were intermittent, but were
always confined to the INTERSECTION of the spheres, not coincident surfaces.
Moving the spheres around slightly got around the problem but it appears to
be a bug in POV.
In my experience, coincident surfaces tend to render a jittered imaged of
both textures, not a black spot. How about this... try it out 2 images
using just 2 boxes, side by side, and the texture that your using. In one
image put one box about half way into the other and render it and in the
other, make the boxes just touch each other (coincident surfaces) and render
it. Just to see what happens.
I'd try it and show you what I mean right now but I have a 6 day render
running right now with 4 days to go...
news:379a139b@news.povray.org...
>
> Hi all!
>
> I wanted to create a rgb-cube by putting together 10^3 smaller cubes
> filled with an emission media. But there are some strange black
> squares between the smaller cubes! Changing density, transmit and/or
> filter values didn't help.
>
> Does anyone know where the problem is??
>
> Thanx in advance,
> Daniel
>
>
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
yep, a wrong value for max_trace_level was the reason,
but... this is still a mystery to me, because I know I didn't specify
ANY global_settings in the scene file yesterday, and POV-Ray still
produced garbage; today I tried again (I didn't make any changes) and
these black squares were gone! The only explanation I can imagine is
that POV-Ray crashed (which happens under Windoze95 quite often) and
after restarting POV-Ray it got a bit confused and somehow used a
wrong value (6 or 7?) for max_trace_level!?
Anyway, there is left a strange white vertical line in the middle of
the rendered picture. Does anyone know why?
btw I'm using version 3.1a.watcom.win32 [Pentium II optimized]
would it be possible to achieve the same result (without that white
line of course *g*) by using density_map / color_map on a _single_
cube instead?
Daniel
Post a reply to this message
Attachments:
Download 'rgb_cube_2nd_try.jpg' (4 KB)
Preview of image 'rgb_cube_2nd_try.jpg'
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Fact is there are actually no coincident surfaces anyway in Daniels
script. The individual boxes are reduced to nine tenths their original
size (which if left unscaled would have been a potential problem,
maybe?). Being that much smaller however might be leading to some
trouble though, leaving gaps between each cube.
Dave Kreskowiak wrote:
>
> I ran into a problem probably similar to yours. I was trying to render
> billowing smoke from something like a gasoline or oil fire. There spheres I
> was using for containers intersected (overlapped) each other without any
> coincident surfaces. The black areas I got were intermittent, but were
> always confined to the INTERSECTION of the spheres, not coincident surfaces.
> Moving the spheres around slightly got around the problem but it appears to
> be a bug in POV.
>
> In my experience, coincident surfaces tend to render a jittered imaged of
> both textures, not a black spot. How about this... try it out 2 images
> using just 2 boxes, side by side, and the texture that your using. In one
> image put one box about half way into the other and render it and in the
> other, make the boxes just touch each other (coincident surfaces) and render
> it. Just to see what happens.
>
> I'd try it and show you what I mean right now but I have a 6 day render
> running right now with 4 days to go...
>
> news:379a139b@news.povray.org...
> >
> > Hi all!
> >
> > I wanted to create a rgb-cube by putting together 10^3 smaller cubes
> > filled with an emission media. But there are some strange black
> > squares between the smaller cubes! Changing density, transmit and/or
> > filter values didn't help.
> >
> > Does anyone know where the problem is??
> >
> > Thanx in advance,
> > Daniel
> >
> >
> >
--
omniVERSE: beyond the universe
http://members.aol.com/inversez/homepage.htm
mailto://inversez@aol.com?Subject=PoV-News
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I overlooked the scaling down of each individual box before. There is no
possibility of coincident surfaces.
Bob Hughes wrote:
>
> Yes. there is a need for a higher max_trace_level. About 15 seems okay,
> but 150 was better, somewhere inbetween perhaps then. I was going to say
> it probably is coincident surfaces but the effect goes away, or appears
> to anyway. The container objects being invisible to start with might not
> make it obvious but still cause trouble, not sure. At the default trace
> level it didn't show any change when I rescaled the boxes down slightly.
>
> Ken wrote:
> >
> > Ken wrote:
> > >
> > > >
> > > > Hi all!
> > > >
> > > > I wanted to create a rgb-cube by putting together 10^3 smaller cubes
> > > > filled with an emission media. But there are some strange black
> > > > squares between the smaller cubes! Changing density, transmit and/or
> > > > filter values didn't help.
> > > >
> > > > Does anyone know where the problem is??
> > > >
> > > > Thanx in advance,
> > > > Daniel
> > >
> > > I don't have time to look at your source now but try adding a
> > > nax_trace_level in your global settings. I would start with
> > > something like 30 or more with this many objects in the scene.
> >
> > I didn't see the image either until now. Um there is a problem in
> > pov called coincident surfaces where if you have transparent objects
> > whose faces occupy the same space like two boxes abutting each other
> > the raytracer cannot decide which surface belongs to which object.
> > The easiest way to avoid this is to seperate them by a tiny amount
> > or have them overlap by a tiny amount. If you overlap them you can
> > also use a merge operation to remove all of the inner surfaces.
> > Read up in the docs about csgs and coincident surfaces.
> > max_trace_level can still be an issue so don't hesitate adding
> > that to your scene too.
> > --
> > Ken Tyler
> >
> > mailto://tylereng@pacbell.net
> > http://home.pacbell.net/tylereng/links.htm
>
> --
> omniVERSE: beyond the universe
> http://members.aol.com/inversez/homepage.htm
> mailto://inversez@aol.com?Subject=PoV-News
--
omniVERSE: beyond the universe
http://members.aol.com/inversez/homepage.htm
mailto://inversez@aol.com?Subject=PoV-News
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Shift your camera or rotate the entire cube itself (using a union) ever
so slightly and the vertical white line will disappear. It's a result of
the opposite front and back edges lining up.
>
> yep, a wrong value for max_trace_level was the reason,
> but... this is still a mystery to me, because I know I didn't specify
> ANY global_settings in the scene file yesterday, and POV-Ray still
> produced garbage; today I tried again (I didn't make any changes) and
> these black squares were gone! The only explanation I can imagine is
> that POV-Ray crashed (which happens under Windoze95 quite often) and
> after restarting POV-Ray it got a bit confused and somehow used a
> wrong value (6 or 7?) for max_trace_level!?
>
> Anyway, there is left a strange white vertical line in the middle of
> the rendered picture. Does anyone know why?
>
> btw I'm using version 3.1a.watcom.win32 [Pentium II optimized]
>
> would it be possible to achieve the same result (without that white
> line of course *g*) by using density_map / color_map on a _single_
> cube instead?
>
> Daniel
>
> [Image]
--
omniVERSE: beyond the universe
http://members.aol.com/inversez/homepage.htm
mailto://inversez@aol.com?Subject=PoV-News
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Bob Hughes wrote:
>
> Fact is there are actually no coincident surfaces anyway in Daniels
> script. The individual boxes are reduced to nine tenths their original
> size (which if left unscaled would have been a potential problem,
> maybe?). Being that much smaller however might be leading to some
> trouble though, leaving gaps between each cube.
The only problems I foresee with that would be the increased number
of intersection tests needed. This of course can be compensated for by
increasing the max trace level to an adequate amount.
--
Ken Tyler
mailto://tylereng@pacbell.net
http://home.pacbell.net/tylereng/links.htm
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
