 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Cousin Ricky" <ric### [at] yahoo com> wrote in message
news:web.486f07c34ab32a2c85de7b680@news.povray.org...
> Kyle <hob### [at] gatenet> wrote:
>>
>> In case you hadn't noticed, there's a lack of shadow behind the guy's heel
>> too.
>
> I've had that problem when working on small scales, but not on a scale that i
> feel should challenge POV's floating point accuracy. As far as i can
> determine, it's a bug.
It's a feature! POV-Ray has to avoid self shadowing, actually I've meanwhile
checked
the sourcecode to confirm my statements made in this thread:
http://news.povray.org/47fbebb1$1@news.povray.org
Maybe this "epsilon value " should be made adjustable by the user?
Thomas, I'm quite sure this is your problem, especially as you say you scaled
your
poser figure by 0.2 and I usually scale them by a factor of 200.
-Ive
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ive" <lil### [at] hotmailcom> schreef in bericht
news:486f692a$1@news.povray.org...
> It's a feature! POV-Ray has to avoid self shadowing, actually I've
> meanwhile checked
> the sourcecode to confirm my statements made in this thread:
>
> http://news.povray.org/47fbebb1$1@news.povray.org
>
> Maybe this "epsilon value " should be made adjustable by the user?
>
>
> Thomas, I'm quite sure this is your problem, especially as you say you
> scaled your
> poser figure by 0.2 and I usually scale them by a factor of 200.
>
Ive, thank you for reminding me of that thread!! I forgot about it. Maybe
this will help also for the height_field? I shall experiment by upscaling
the whole scene.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ive" <lil### [at] hotmailcom> wrote:
> "Cousin Ricky" <ric### [at] yahoocom> wrote in message
> news:web.486f07c34ab32a2c85de7b680@news.povray.org...
> > Kyle <hob### [at] gatenet> wrote:
> >>
> >> In case you hadn't noticed, there's a lack of shadow behind the guy's heel
> >> too.
> >
> > I've had that problem when working on small scales, but not on a scale that i
> > feel should challenge POV's floating point accuracy. As far as i can
> > determine, it's a bug.
>
> It's a feature! POV-Ray has to avoid self shadowing, [...]
That makes sense. (Moves from "bug" to "feature" cubby hole.)
> Maybe this "epsilon value " should be made adjustable by the user?
I have an old book, _The Elements of Programming Style_, and it says not to use
absolute epsilons. Never having written a ray tracing program myself (always
on my to-do list since i'd first heard the term, if not before), i don't feel
right telling the POV team what to do. Not that i could demand my money back
or anything. ;-)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ive" <lil### [at] hotmailcom> schreef in bericht
news:486f692a$1@news.povray.org...
>
>
> It's a feature! POV-Ray has to avoid self shadowing, actually I've
> meanwhile checked
> the sourcecode to confirm my statements made in this thread:
>
> http://news.povray.org/47fbebb1$1@news.povray.org
>
> Maybe this "epsilon value " should be made adjustable by the user?
>
>
> Thomas, I'm quite sure this is your problem, especially as you say you
> scaled your
> poser figure by 0.2 and I usually scale them by a factor of 200.
>
A rapid render with the scene scaled up by 100 shows that all artifacts
disappear.
I rmember now that the discussion about the epsilon was (amongst other I
suppose) around the "Ringworld" scenes back in October/November 2005. There
was even a POV-Ray patch provided (By David Buck:
<4365fcf9$1@news.povray.org>) that corrected the epsilon value for those
scenes combining very distant and very near objects.
I totally forgot about this while I felt something scratching at the back of
my mind :-) Thank you again for reminding me!!
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Thomas de Groot" <t.d### [at] internlDOTnet> schreef in bericht
news:48708d66$1@news.povray.org...
>
There
> was even a POV-Ray patch provided (By David Buck:
> <4365fcf9$1@news.povray.org>) that corrected the epsilon value for those
> scenes combining very distant and very near objects.
Sorry, the link is obviously not the right one! The discussion took place at
the end of october 2005 in p.b.i. (Re: Ringworld re-re-revisited). However,
the link provided by David Buck for the patch is no longer available.
Now, my next hunch is that epsilon has been changed for version 3.7, can
anybody confirm this? I have not the beta installed at the moment, pending
the installer problems are ironed out.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Sun, 6 Jul 2008 11:26:43 +0200, "Thomas de Groot" <t.d### [at] internlDOTnet>
wrote:
>I have not the beta installed at the moment, pending
>the installer problems are ironed out.
You can still use Beta 26, Just "extend" it.
--
Regards
Stephen
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Thomas de Groot" <t.d### [at] internlDOTnet> wrote in message
news:48708fd4@news.povray.org...
>>
> There
>> was even a POV-Ray patch provided (By David Buck:
>> <4365fcf9$1@news.povray.org>) that corrected the epsilon value for those
>> scenes combining very distant and very near objects.
>
Interesting - I did not follow the NG at this time.
The "epsilon-problem" appears not only in scenes with large/small scales
but also - as in your "paintes desert" scene (great idea BTW) - effects
the shadow ray calculation. And is of course not limited to poser figures
or meshes imported by PoseRay.
> Now, my next hunch is that epsilon has been changed for version 3.7, can
> anybody confirm this? I have not the beta installed at the moment, pending the
> installer problems are ironed out.
>
Well, I have (not yet) installed the 3.7 beta sources but my guess is the
hard coded EPSILON remains the same. Changing it would surely help
for some scenes but meight also very well break others. I think it should
not be hard coded but part of the global_settings block.
-Ive
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Stephen" <mcavoysAT@aolDOTcom> schreef in bericht
news:km4174d28sdugubic42m1eaea72eo94b0o@4ax.com...
>
> You can still use Beta 26, Just "extend" it.
> --
>
Oh, yes. I still have 25 somewhere. Didn't install later versions....
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ive" <lil### [at] hotmailcom> schreef in bericht
news:4870c7c8$1@news.povray.org...
> "Thomas de Groot" <t.d### [at] internlDOTnet> wrote in message
> news:48708fd4@news.povray.org...
>>>
>> There
>>> was even a POV-Ray patch provided (By David Buck:
>>> <4365fcf9$1@news.povray.org>) that corrected the epsilon value for those
>>> scenes combining very distant and very near objects.
>>
>
> Interesting - I did not follow the NG at this time.
>
> The "epsilon-problem" appears not only in scenes with large/small scales
> but also - as in your "paintes desert" scene (great idea BTW) - effects
> the shadow ray calculation. And is of course not limited to poser figures
> or meshes imported by PoseRay.
Indeed. Everything is affected as we "discovered" in the Ringworld
experiments.
>
>> Now, my next hunch is that epsilon has been changed for version 3.7, can
>> anybody confirm this? I have not the beta installed at the moment,
>> pending the installer problems are ironed out.
>>
>
> Well, I have (not yet) installed the 3.7 beta sources but my guess is the
> hard coded EPSILON remains the same. Changing it would surely help
> for some scenes but meight also very well break others. I think it should
> not be hard coded but part of the global_settings block.
>
Yes, you are most probably right. I seem to remember that it was suggested
to Chris Cason at the time, but I cannot find that thread. I probably
dreamed it... :-)
Again, thanks for reminding me of the epsilon issue. I am going to note it
down seriously this time so that I shall have no excuses next time. ;-)
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Thomas de Groot nous illumina en ce 2008-07-05 03:29 -->
> "Cousin Ricky" <ric### [at] yahoocom> schreef in bericht
> news:web.486f07c34ab32a2c85de7b680@news.povray.org...
>> Kyle <hob### [at] gatenet> wrote:
>>> In case you hadn't noticed, there's a lack of shadow behind the guy's
>>> heel too.
>> I've had that problem when working on small scales, but not on a scale
>> that i
>> feel should challenge POV's floating point accuracy. As far as i can
>> determine, it's a bug. I haven't brought it up in these newsgroups
>> because i
>> haven't yet reproduced in in a simple scene.
>>
>
> That is interesting. The scale is reasonably scaled in the present scene.
> For instance, the figure is scaled to 0.2 from original, so nothing extreme
> here. However, I am not sure how the hf subdivision in John VanSickle's
> macro might influence the foreground scaling, although the final hf is
> scaled up.
>
> Thomas
>
>
Try not scaling the figure, but scale the rest of the scene up by 5. It should,
at least, reduce the problem and maybe make it fade away.
--
Alain
-------------------------------------------------
An organizer for the "Million Agoraphobics March" expressed disappointment in
the turnout for last weekend's event.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
