![](/i/fill.gif) |
![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
From: Thomas de Groot
Subject: Re: any ideas what these black patches are..
Date: 15 Mar 2008 03:11:06
Message: <47db849a$1@news.povray.org>
|
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
"s.day" <s.d### [at] uel ac uk> schreef in bericht
news:web.47db059fd387c8866ba4aa160@news.povray.org...
>
> I have had similar issues to you but find rendering at 50% of the final
> width/height to usually be acceptable (although comparissons may show that
> rendering with 50% the radiosity quality at the correct resolution will be
> just
> as good..).
I think that you can try less than 50% and get good results. The only thing
probably is that some more radiosity has to be calculated in the second
pass, but that might be compensated somewhat by a higher trace resolution in
the first..
>
> I do not really understand the radiosity settings very well I have a rough
> idea
> of how most of them work but tend to resort to trial and error more than
> educated guessing..
>
Yes, I understand what you mean. I am slowly getting understanding of it,
but experimentation remains dominant.
Thomas
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
"Tim Attwood" <tim### [at] comcast net> wrote:
> > The scene code is on the povcomp site for all to use:
>
> sky_sphere ...
>
> The radiosity engine only uses rays that travel finite
> distances, sky_sphere and background are both an
> infinite distance. When a radiosity ray "leaks" thru
> a seam, and then goes infinite, that ray is considered
> pure black, which affects previous bounces of the
> ray, coloring them black as well. Higher quality
> settings means there are more rays to leak thru.
>
> You should use a large inverted sphere for sky
> in radiosity scenes, that way there will be no infinite
> rays at all. That should reduce the artifacts.
Hi Tim,
That's something I did not know, not completely sure it is affecting this image
as some of the black patches are on areas I am pretty confident do not have any
seam's but as I created this over 3 years ago (was povcomp really that long ago)
I don't remember too much about it.
I will wait to see how the current render is when it gets past the pretrace
stages and if I still have the black patches will give this a go. I think I
will try some experiments with this anyway without all the objects in the scene
to see if I can find out a bit sooner..
Sean
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
The mystery deepens.. The machine I was running this render on crashed over the
weekend (not sure of the reason for the crash), but it had (just) done enough
of the render to see that the patches were still present (and may have been
even worse).
This was with the following radiosity settings:
pretrace_start 0.08
pretrace_end 0.01
count 1000
nearest_count 20
error_bound 0.2
recursion_limit 4
low_error_factor .9
gray_threshold 0
minimum_reuse 0.015
brightness 0.8
adc_bailout 0.01/2
I am now wondering if this is a problem with the machine I am using for the
render (maybe a dodgy memory chip issue which would explain the crash), I think
I will try stripping the radiosity settings back to something really basic that
should not cause any issues (and will be much quicker) then see how that goes.
Sean
Post a reply to this message
Attachments:
Download 'temp2.jpg' (6 KB)
Preview of image 'temp2.jpg'
![temp2.jpg](/povray.binaries.images/attachment/%3Cweb.47de6242d387c8867eb3d110%40news.povray.org%3E/temp2.jpg?preview=1)
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
From: Thomas de Groot
Subject: Re: any ideas what these black patches are..
Date: 17 Mar 2008 08:10:25
Message: <47de6dc1$1@news.povray.org>
|
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
"s.day" <s.d### [at] uel ac uk> schreef in bericht
news:web.47de6242d387c8867eb3d110@news.povray.org...
>
> The mystery deepens.. The machine I was running this render on crashed
> over the
> weekend (not sure of the reason for the crash), but it had (just) done
> enough
> of the render to see that the patches were still present (and may have
> been
> even worse).
>
> This was with the following radiosity settings:
>
> pretrace_start 0.08
> pretrace_end 0.01
> count 1000
> nearest_count 20
> error_bound 0.2
> recursion_limit 4
> low_error_factor .9
> gray_threshold 0
> minimum_reuse 0.015
> brightness 0.8
> adc_bailout 0.01/2
>
> I am now wondering if this is a problem with the machine I am using for
> the
> render (maybe a dodgy memory chip issue which would explain the crash), I
> think
> I will try stripping the radiosity settings back to something really basic
> that
> should not cause any issues (and will be much quicker) then see how that
> goes.
>
This is strange indeed. I am wondering now if those artifacts are caused by
radiosity... Have you had a mention of I-Stack Overflows reported in the
statistics? Maybe max_intersections is too low?
OTOH, it looks a bit like a max_gradient problem, but if no isorfaces are
used here... Neither max_trace.... And without reflective surfaces
max_trace_level seems out of the question too...
Thomas
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
> This is strange indeed. I am wondering now if those artifacts are caused by
> radiosity... Have you had a mention of I-Stack Overflows reported in the
> statistics? Maybe max_intersections is too low?
>
> OTOH, it looks a bit like a max_gradient problem, but if no isorfaces are
> used here... Neither max_trace.... And without reflective surfaces
> max_trace_level seems out of the question too...
>
> Thomas
Hi Thomas,
Not sure about the max_intersections as the scene has not completed yet (it
crashed and now I have restarted it with lower quality radiosity settings but I
will look into this next).
However, I think I may have found the issue (although I am not sure yet and it
will require more test renders to check..) I inserted the basic radiosity
settings and re-rendered (wow! that was much quicker) the scene looks fine.
However, I noticed that the comment shows a value for nearest count of 1..10.
Now the documentation states that 20 is the highest possible value but on
Christoph's radiosity tutorial:
http://www.imagico.de/pov/radiosity01.html
He states 10 is the highest, obviously only one of these is correct and maybe
setting this too high is the cause.
I will do some tests to find out but am pretty sure I have used nearest_count=20
before without problems..
Sean
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
From: Thomas de Groot
Subject: Re: any ideas what these black patches are..
Date: 17 Mar 2008 10:27:45
Message: <47de8df1@news.povray.org>
|
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
"s.day" <s.d### [at] uel ac uk> schreef in bericht
news:web.47de73e8d387c8867eb3d110@news.povray.org...
>
> I will do some tests to find out but am pretty sure I have used
> nearest_count=20
> before without problems..
>
Sean,
20 should work fine, according to the docs. However, it certainly might up
render time. I am not sure about that 10 on Christoph's site. Might be a
typo, but it might also be a maximum Christoph considers as reasonable and
within render time tolerance.
I am certainly curious to see the outcome of your experiments.
Thomas
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
s.day nous apporta ses lumieres en ce 2008/03/17 09:36:
> http://www.imagico.de/pov/radiosity01.html
>
> He states 10 is the highest, obviously only one of these is correct and maybe
> setting this too high is the cause.
>
> I will do some tests to find out but am pretty sure I have used nearest_count=20
> before without problems..
>
> Sean
>
In early MegaPOV radiosity implementations, nearest_count was limited to 10. The
tutorial was originaly writen for those early specifications. It was edited for
the POV-Ray radiosity in version 3.5, then updated for 3.6. Some details got
forgotten.
--
Alain
-------------------------------------------------
You know you've been raytracing too long when you think it's a failing of the
universe that the large software companies like Corel or Fractal Design do NOT
export to POV primitives.
George Erhard
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
s.day wrote:
> I forgot to add that the machine is only running XP 32 bit and 32 bit POV which
> is why I wondered if approaching the 2GB memory limit may be causing the
> problem.
>
> Sean
>
Start XP with /3GB -switch and it'll be able to allocate 3G for one process.
--
Eero "Aero" Ahonen
http://www.zbxt.net
aer### [at] removethis zbxt net invalid
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
"s.day" <s.d### [at] uel ac uk> wrote:
> Wow! thanks for all the replies.
>
> No media in the scene to be causing any issues...
>
A beautiful scene, and very complex (which makes it difficult to nail down the
problem, I'm sure.)
In looking at your scene code, I see that there are lots of spotlights that
actually DO make use of media in some way. In the current version of POV-Ray,
there are *multiple* problems associated when using media, radiosity and
spotlights, which can cause really strange visual aberrations to appear in
places you wouldn't expect. So far, I've only posted info about one such
problem (not using radiosity). But there are others.
Here's the link...
http://news.povray.org/povray.general/thread/%3Cweb.45125ba9d84b4b77cc29cf380@news.povray.org%3E/
and the link to the associated image...
http://news.povray.org/povray.binaries.images/thread/%3Cweb.45125e6ed84b4b77cc29cf380%40news.povray.org%3E/?ttop=267426
&toff=900
Another problem is that radiosity illumination from media does not show up on
any objects (or parts of objects) that are inside a media-filled spotlight
cone. Said another way, the media IN the cone doesn't generate radiosity
illumination on objects that are embedded in it. I think that's a problem with
the spotlight itself.
My suggestion would be to run your complete scene--radiosity and all-- with
UseTestTex=1 rather than 0 (which will turn off your media) and see if the
black spots still appear. A long and laborious testing process, to be sure!
Ken W.
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
From: Nicolas Alvarez
Subject: Re: any ideas what these black patches are..
Date: 18 Mar 2008 17:50:21
Message: <47e0472d$1@news.povray.org>
|
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
>
http://news.povray.org/povray.binaries.images/thread/%3Cweb.45125e6ed84b4b77cc29cf380%40news.povray.org%3E/
Wow!
Post a reply to this message
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |
| ![](/i/fill.gif) |
|
![](/i/fill.gif) |
|
![](/i/fill.gif) |
| ![](/i/fill.gif) |