Hi again,

I wonder, if others have experienced the same: Sometimes, when I just change
my positioning in a scene a little bit (changed a "translate") and render it
with megapov, one isosurface occasionally disappears. I can't reproduce the
problem, because when I render the scene again, the isosurface is there
again. Everything I do between the two renders is undo my changes and redo
them (with ctrl-z, ctrl-a). So I end up with the same scene after that.
Just wondering ...

Marc-Hendrik

Post a reply to this message

Marc-Hendrik Bremer wrote:

> I wonder, if others have experienced the same: Sometimes, when I just change
> my positioning in a scene a little bit (changed a "translate") and render it
> with megapov, one isosurface occasionally disappears.

Possibly a max_gradient related problem.
The usual solution is to look for the right max_gradient in the stats (it
appears when using the "eval" keyword) and write it in the isosurface
definition.
Have a look at Mike Williams' tutorial here for a better explanation :
http://www.econym.demon.co.uk/isotut/keywords.htm

G.

Post a reply to this message

Thanks for the link to the tutorial, but I think that is another case. I
know those holes in Isosurfaces and yes sometimes they even disappear.
But I did not change anything in the Isosurface settings, but I translated
it only one tenth of a pov unit upward. In my case the same scene source
produces to different pictures - first, one isosurface misses, the second
time it doesn't. With this max_gradient-problem, the surface will always be
the same?

Marc-Hendrik


Gilles Tran schrieb in Nachricht <395C8D22.35D9BE2B@inapg.inra.fr>...
>Marc-Hendrik Bremer wrote:
>
>> I wonder, if others have experienced the same: Sometimes, when I just
change
>> my positioning in a scene a little bit (changed a "translate") and render
it
>> with megapov, one isosurface occasionally disappears.
>
>Possibly a max_gradient related problem.
>The usual solution is to look for the right max_gradient in the stats (it
>appears when using the "eval" keyword) and write it in the isosurface
>definition.
>Have a look at Mike Williams' tutorial here for a better explanation :
>http://www.econym.demon.co.uk/isotut/keywords.htm
>
>G.
>

Post a reply to this message

Nope, translations can have an effect - at least I ran into similiar
problems just moving the camera slightly. In one position a simple sphere
iso rendered without problems, moved the camera and someone had been
munching on it.

Sorry, I would have suggested this before, but I misread your original
posting and thought you were saying that you were having odd results that
were corrected by just re-typing the same co-ords.

"Marc-Hendrik Bremer" <Mar### [at] t-onlinede> wrote in message
news:395c9194@news.povray.org...
> Thanks for the link to the tutorial, but I think that is another case. I
> know those holes in Isosurfaces and yes sometimes they even disappear.
> But I did not change anything in the Isosurface settings, but I translated
> it only one tenth of a pov unit upward. In my case the same scene source
> produces to different pictures - first, one isosurface misses, the second
> time it doesn't. With this max_gradient-problem, the surface will always
be
> the same?
>
> Marc-Hendrik
>
>
> Gilles Tran schrieb in Nachricht <395C8D22.35D9BE2B@inapg.inra.fr>...
> >Marc-Hendrik Bremer wrote:
> >
> >> I wonder, if others have experienced the same: Sometimes, when I just
> change
> >> my positioning in a scene a little bit (changed a "translate") and
render
> it
> >> with megapov, one isosurface occasionally disappears.
> >
> >Possibly a max_gradient related problem.
> >The usual solution is to look for the right max_gradient in the stats (it
> >appears when using the "eval" keyword) and write it in the isosurface
> >definition.
> >Have a look at Mike Williams' tutorial here for a better explanation :
> >http://www.econym.demon.co.uk/isotut/keywords.htm
> >
> >G.
> >
>

Post a reply to this message

Tom Melly schrieb in Nachricht <395c925e@news.povray.org>...
>Nope, translations can have an effect - at least I ran into similiar
>problems just moving the camera slightly. In one position a simple sphere
>iso rendered without problems, moved the camera and someone had been
>munching on it.
>
>Sorry, I would have suggested this before, but I misread your original
>posting and thought you were saying that you were having odd results that
>were corrected by just re-typing the same co-ords.


Well, actually it was not so much misread!
This were my steps:
1. render a scene with 3 isosurfaces and a plane
2. move all three isos and the plane 1/10 unit up by changing the translate
3. rendering the scene -> first iso is missing
4. (Wondering and) undo the changes by pressing ctrl-z several times,
rendering. Isosurface is again there.
5. Hm, redo the changes by pressing ctrl-a several times, rendering.
Isosurface remains there.

This problem occurred now 2 times for me. As I said I can't reproduce it
though.

<Shrug> ... if no one else had this problems it might be my computer, but I
don't think so. I can definitely life with it, as long as my scene renders
in some 3 minutes, but it would be very annoying, when rendertimes go up.

Marc-Hendrik

Post a reply to this message

"Marc-Hendrik Bremer" <Mar### [at] t-onlinede> wrote in message
news:395cb4fa$1@news.povray.org...
> Well, actually it was not so much misread!
> This were my steps:
> 1. render a scene with 3 isosurfaces and a plane
> 2. move all three isos and the plane 1/10 unit up by changing the
translate
> 3. rendering the scene -> first iso is missing
> 4. (Wondering and) undo the changes by pressing ctrl-z several times,
> rendering. Isosurface is again there.
> 5. Hm, redo the changes by pressing ctrl-a several times, rendering.
> Isosurface remains there.
>
> This problem occurred now 2 times for me. As I said I can't reproduce it
> though.

Hmm, IMHO it still sounds like a problem with max_gradient. I convinced
myself the other day that my PC at work was rendering an iso very
differently than my home PC. Nope, just blindness on my part. Maybe you
undid one less or one more change than you thought?

Have you tried eval? It's interesting to run eval without a max_gradient,
use the eval output to set a gradient, then run the scene again. Eval quite
often then rises above it's first estimate. For a recent scene, I would use
the output of eval to set max_gradient, only to have the scene give me a
higher eval on the next render. I had to increase max_gradient about three
times until I reached a stage when eval didn't go up any further.

Post a reply to this message