Duh - from the looks of it, I'd say your glass globe is *solid*!

Aside from this giving totally unrealistic results anyway, this will result in
the media containers cutting "bubbles" and a chimney into it, which perfectly
explains the effect.

So your objects *do* intersect - not their surfaces, but their volumes.

Post a reply to this message

Just a quick question: How large is the whole smash, in POV dimensions?

Try scaling everything up by a factor of 10 or even 100. I have the impression
that POV rejects the inner wall of your glass globe (there are hints that there
*is* such an inner wall) due to the glass being too thin.

IIRC the limit is 0.001 units (may be 0.0001, need to look at the code).
Surfaces closer than that will give you trouble.

Post a reply to this message

"clipka" <nomail@nomail> wrote in message
news:web.49d10dc330aab812f708085d0@news.povray.org...
> Duh - from the looks of it, I'd say your glass globe is *solid*!

well double duh .... that's what appears to be happening but I've proven
that the globe is a tube. I gave it just a pigment and rotated it so I could
see ALL the way throught it!

I even used just a plain cone object with open and hollow keywords and the
effect is still there.

> So your objects *do* intersect - not their surfaces, but their volumes.

doesn't open (if applicable) and hollow keywords overcome that problem?

Cheers

Post a reply to this message

"clipka" <nomail@nomail> wrote in message
news:web.49d10e9230aab812f708085d0@news.povray.org...
>  I have the impression
> that POV rejects the inner wall of your glass globe (there are hints that
there
> *is* such an inner wall) due to the glass being too thin.

then a differenced cone (with some thickness) should overcome that right?

> IIRC the limit is 0.001 units (may be 0.0001, need to look at the code).
> Surfaces closer than that will give you trouble.

if that's true then perhaps that's what I'm battling (infinite thinness)
haha ... that sounds odd!

Exit stage left!

Post a reply to this message

"Jim Holsenback" <jho### [at] hotmailcom> wrote:
> > So your objects *do* intersect - not their surfaces, but their volumes.
>
> doesn't open (if applicable) and hollow keywords overcome that problem?

Uhm - no.

"Open" basically creates an object with an infinitely thin wall - but such an
object has no interior that an IOR would be applicable to. Or, seen from a
different perspective (and that's also POV's perspective), the interior does
*not* pertain to the wall (which is just an infinitely thin surface), but
rather to whatever is enclosed by it.

But obviously, with some of the surface removed, it is no longer possible to
clearly specify where the inside begins and the outside ends. So combining
"open" (or "clipped_by", for that matter) with an IOR other than 1.0 (or any
other different interior property than the air outside) messes with the object
in a fashion that has no real-world equivalent whatsoever. Unless someone
discovers a material in which the speed of light and the media density depend
on where the light entered the object.


All the while, "hollow" just specifies that the object itself is fit to contain
media. Which, as I just mentioned, doesn't mix well with "open"/"clipped_by"
either.


So what you need is either an "open" shape with a surface that models glass
sufficiently well without relying on IOR (which is difficult), or a CSG
difference of two sor objects defining a finitely thin wall. Or a single lathe
object comprised of a closed spline, or something alike.

Post a reply to this message

Jim Holsenback nous illumina en ce 2009-03-30 14:21 -->
> "Alain" <ele### [at] netscapenet> wrote in message
> news:49d0df24$1@news.povray.org...
>> Does your glass have a thickness, or is it solid trough?
> 
> It's a lathe object, and I rotated it slightly just to make sure I could see
> all the way through it. Without a texture it looks like a hollow tube ...
> btw: when it's textured hollow on or off appears to make no difference. It
> indeed appears to be acting like a solid object.
> 
> Just to be clear about it I'm running 3.6.1 on linux ... I'm not on that
> machine now but planned on doing some more testing (purely trial and error
> at this point) this evening .... perhaps I should post a couple of relevent
> code segments to see if anyone can spot what's going on here.
> 
>> Now, there is a magnifier like effect that should NOT be there.
> 
> yes that describes what we're seeing perfectly ....  the effect gets more
> pronounced as ior is increased.
> 
>> Then, the flame's container apears to have an ior of 1.000.
> 
> I don't recall (it's someone elses macro) but I'll check
> 
> Thanks
> Jim
> 
> 
> 

Your lathe spline should go back to form an inner surface and form a closed 
shape that never reatch the axis.

-- 
Alain
-------------------------------------------------
Keep your eyes wide open before marriage, half shut afterwards.
Benjamin Franklin

Post a reply to this message

Jim Holsenback nous illumina en ce 2009-03-30 16:16 -->
> "clipka" <nomail@nomail> wrote in message
> news:web.49d10dc330aab812f708085d0@news.povray.org...
>> Duh - from the looks of it, I'd say your glass globe is *solid*!
> 
> well double duh .... that's what appears to be happening but I've proven
> that the globe is a tube. I gave it just a pigment and rotated it so I could
> see ALL the way throught it!
> 
> I even used just a plain cone object with open and hollow keywords and the
> effect is still there.
> 
>> So your objects *do* intersect - not their surfaces, but their volumes.
> 
> doesn't open (if applicable) and hollow keywords overcome that problem?
> 
> Cheers
> 
> 
This is a pair of open cylinder with an ior of 1.5.
The left one is rotated so that you can see trough both open ends, the other is 
seen edge on.
There is also a small spgere with only a pigment of rgbt 1 and no interior at 
the center of both cylinders.

As you can see, the ior does affect the background even if you only pass trough 
one surface. The sphere is also visible on the right where you see it trough the 
side of the cylinder.
On the left, the sphere is completely invisible, but is very visible on the right.

-- 
Alain
-------------------------------------------------
'Honesty is the best policy, but insanity is a better defense'

Post a reply to this message

Attachments:
Download 't2.jpg' (36 KB)

Preview of image 't2.jpg'

"Alain" <ele### [at] netscapenet> wrote in message
news:49d17ad0@news.povray.org...
>
> As you can see, the ior does affect the background even if you only pass
trough
> one surface. The sphere is also visible on the right where you see it
trough the
> side of the cylinder.
> On the left, the sphere is completely invisible, but is very visible on
the right

ah yes simular to some tests I did ....

a differenced cone fixed the problem I was having with the infinately thin
walls of the lathe object. Before I did that scaling the entrie object by
100 didn't give any joy, just like the open/hollow cone .....

I wanted the globe to have a curved profile not straight, so maybe a merge
of a cone and an irregularly scaled sphere, then difference to trim out the
inside might suit .... maybe an isosurface with thickness.

thanks for your insights.

Jim

Post a reply to this message

"clipka" <nomail@nomail> wrote in message
news:web.49d146c530aab8122f4aa4190@news.povray.org...
> So what you need is either an "open" shape with a surface that models
glass
> sufficiently well without relying on IOR (which is difficult)

well this excercise has proven that ....

> Or a single lathe
> object comprised of a closed spline, or something alike.

well it's back to school for me on this subject .... I thought that's what I
had with the lathe object. I even set up a "Spline Lab" environment simular
to that outlined in the documentation. I'll be reading between the lines to
see what I've missed. I may be slow but I eventually get there.

Ciao

Post a reply to this message

"Jim Holsenback" <jho### [at] hotmailcom> wrote:
> a differenced cone fixed the problem I was having with the infinately thin
> walls of the lathe object. Before I did that scaling the entrie object by
> 100 didn't give any joy, just like the open/hollow cone .....

Yeah, that scaling idea was based on the assumption that you already *did* have
a differenced object ;)

Scaling a thickness of 0 by any value doesn't help, of course :P

Post a reply to this message