Hi all!

This is a project, i started some year ago. Now it's maybe worth posting.

<talking (too) much>
The oldest part is the thermometer itself. I once stopped working on it,
when the lathe-bug crossed my way (the dark line at the green capsule - it's
not jpeg...): I'm not that sure, what it is. In march '99 Ken discovered a
so called "media bug". But it was the "well known" lathe bug (thomas
willhalm answered). It's in povray.bugreports.
I don't know much about this bug, i guess it's a precision problem. Will
this bug be fixed in a next version? Or does a workaround exist? I realized
that the problem gets bigger when you decrease the camera-angle. It is
something with orthoganal hits on the lathe.

Than there was another problem. The problem of coincident surfaces. Try to
modell a hollow glassphere half-filled with water. I didn't know a way to
avoid the problem with coincident surfaces but a very difficult one (a union
of{ the glass and a difference of {a union of { a bigger water and
airpart }}cut out to fill the sphere} I know i can't describe it well, but I
can send you the source. (Thinking of it now, clipped_by would improve it)

The sky is the first skysphere ever made by my girlfriend.

On the floor (that is a table with (I know ) a bad wood texture) there is a
crochet work out of 11000 sphere - try to meet with reality. Sorry for the
'real' floor (simple hexagon) but i hoped jpg (compression 60%) would love
it.

And - yes, I know it's a huge format (somewhat 768*2304) but now you can see
the inscription on the weights.
Can someone of you tell me the physics of the thermometer - i once knew, but
having finished school a half year ago, I forgot all physics. :-(
What happens is, if it's getting warmer, the capsules sink onto the floor.
the lowest of the leaving swimming capsules tells the temperature (that is
the black one - 26°C in this case)
</talking (too) much>

What do you think of the scene?

cukk

Post a reply to this message

Attachments:
Download 'thermo.jpg' (98 KB)

Preview of image 'thermo.jpg'

I like this a lot, although a smaller image in a *.png might have preserved the
caustics better, but nice.

The Gallilean (sp?) themometer work on density changes of the bulk fluid verses
the fludis in the bulbs. As the temperature rises, the density of a liquid
ideally is reduced. So, as the temperature rise the density of all the liquids
in the thermometer goes down (bulk liquid as well as each bulb), but what is
important is that the rate of change, dp/dT  (p = rho = density g*ml^-1), is
different for each liquid. I forget what the composition of each liquid is (each
bulb has a slightly different composition), but that is how they work.


-paul
-------------------------------------------#
Paul Daniel Jones
120 Chandlee Laboratory
Penn State University
814-865-2090
pdj### [at] psuedu

http://research.chem.psu.edu/glassgrp/paul
--------------------------------------------#

Post a reply to this message

The dark line at the horizon? Sturm?

Grim

Post a reply to this message

Arranging the floor backgroud and the table should give better looking
results...
-- 
+-------------------------+----------------------------------+
| Simon Lemieux           | Website : http://www.666Mhz.net  |
| Email : Sin### [at] 666Mhznet | POV-Ray, OpenGL, C++ and more... |
+-------------------------+----------------------------------+

Post a reply to this message

Hi Paul!

> I like this a lot, although a smaller image in a *.png might have
preserved the
> caustics better, but nice.

Could you point me to a good png-programm. I don't know how to convert to
png.

Thinking of the Gallilean thermometer: The main trick is, that the volume of
the water increases, if it's getting warmer. The density decreases, and now
the capsules or too heavy to swim. Their density doesn't count. The mass is
constant, the volumechange is to neglect.

cukk

Post a reply to this message

"KalleK" <kal### [at] gmxde> wrote in message news:3a003d96@news.povray.org...
>
> Than there was another problem. The problem of coincident surfaces. Try to
> modell a hollow glassphere half-filled with water. I didn't know a way to
> avoid the problem with coincident surfaces but a very difficult one (a union
> of{ the glass and a difference of {a union of { a bigger water and
> airpart }}cut out to fill the sphere} I know i can't describe it well, but I
> can send you the source. (Thinking of it now, clipped_by would improve it)

First of all, got to say that now I know someone made this before I ever got
around to starting on one.  But I won't let that stop me  :-)
About using clipped_by.  Big mistake, if I understood you, because I think the
unioned difference is best to get a correct CSG.  Since clipping leaves object
open it's not very conducive to refraction (ior), imo.  At least not for all
things.
I'd just scale a copy of the object being used in the final union as the
cutaway object a fraction larger or smaller (depending on the desired result).
The laced cloth is a nice touch.  You're own creation, or from a macro or
include file someplace?

Bob

Post a reply to this message

Hi Bob!

> About using clipped_by.  Big mistake, if I understood you, because I think
the
> unioned difference is best to get a correct CSG.  Since clipping leaves
object
> open it's not very conducive to refraction (ior), imo.  At least not for
all
> things.
Yes. When I thought about finally improving my source I realized it's not
that easy. Exactly that problem with refraction...

> I'd just scale a copy of the object being used in the final union as the
> cutaway object a fraction larger or smaller (depending on the desired
result).
But I think that this can cause some unwanted results with complex objects
(like the thermometer-lathe). If you scale the letter "H" his new 'volume'
will not cover the old one. The result would be surfaces of the cutaway
object in the final union. Therefore I used a much bigger cylinder for the
cutaway object.

> The laced cloth is a nice touch.  You're own creation, or from a macro or
> include file someplace?
It's my own creation. Some loops, no macros. It should match with reality
and represent the real crochet work the thermometer is placed on in reality.
In fact this detail encouraged me to finally post the scene.

As far as the lathe bug is concerned - i got no answers. Shell I ask in
povray.general? I want to avoid these artefacts and don't know how.

cukk

Post a reply to this message

"KalleK" <kal### [at] gmxde> wrote in message news:3a05920f@news.povray.org...
>
> As far as the lathe bug is concerned - i got no answers. Shell I ask in
> povray.general? I want to avoid these artefacts and don't know how.

I'm not so sure it can really be classified as a true "bug" since it's probably
more of a mathematics problem in the spline.  They tend to wrap back on
themselves if the points which are set up for the profile aren't in good
relation to one another.  sturm helps but can't seem to go beyond a certain
limit to clear it up.  Numerical precision and all that, you know.
Anyhow, what I'm actually getting at is that the spline calculations are either
faulty or just can't reach correct accuracy.  But if you place the points well
enough there wouldn't be artifacts.  What it needs is a over-ride to prevent
bad points and thus bad splines.  I guess in that respect it would be a bug
once the answer to that were found.
Didn't make perfect sense, did I ?

Bob

Post a reply to this message

Hi Bob!

> I'm not so sure it can really be classified as a true "bug" since it's
probably
> more of a mathematics problem in the spline.  They tend to wrap back on
> themselves if the points which are set up for the profile aren't in good
> relation to one another.  sturm helps but can't seem to go beyond a
certain
> limit to clear it up.  Numerical precision and all that, you know.
Now I get what Grimdude answered. And I thought it was a joke. Because sturm
is a german word for storm. And If there was a dark line at the horizon, it
could be a storm. but it's that keyword sturm, he ment. I see and will try
sturm now.

> Anyhow, what I'm actually getting at is that the spline calculations are
either
> faulty or just can't reach correct accuracy.  But if you place the points
well
> enough there wouldn't be artifacts.  What it needs is a over-ride to
prevent
> bad points and thus bad splines.  I guess in that respect it would be a
bug
> once the answer to that were found.
> Didn't make perfect sense, did I ?
Reading twice helped (but that are maybe language problems to be solved...)

Now I'm disappointed - I thought I prevented the spline from bad points,
therefore i did a cubicspline-prism first.
And having tried to find out what was wrong, I realized the error on the
lathe was exactly the height of the camera (those orthogonal rays to the
lathe, Thomas Willhelm mentioned in his answer to kens so called "media bug"
in march '99).
You answer is another direction, and i thought i had no "bad points" in the
lathe spline. Have to look at it.

Anyway, many thanks to you(r help)!

cukk

Post a reply to this message

I don't generally reply in German (my Deutsche being so poor).

Grim

Post a reply to this message

Hi GrimDude!

I didn't think, you would reply in german - I didn't think at all, I think.
I read "sturm" and thought "storm"... The english word for "sturm" would be
the same in german... (not true!)

Back ontoppic:
The helps says "sturm" is always used in cubic splines. So this is not the
solution... It's something with the precision, I can't help.

But thanks for the replies!
cukk

Post a reply to this message

> Back ontoppic:
> The helps says "sturm" is always used in cubic splines. So this is not the
> solution... It's something with the precision, I can't help.
>
> But thanks for the replies!
> cukk
>
>

What was your max_trace_level?
If that's not it, could you post your code to p.t.s.f?

Danke, viel gluck (no umlaut available)!
Grim

Post a reply to this message

Hi Grim!
> What was your max_trace_level?
it was 250. That's not the trick...

> If that's not it, could you post your code to p.t.s.f?
I do. But since the code consist of several files, I will post a zip to
p.b.s-f,
if you don't mind.
see "thermometer (glassthingie) code 6kbbu"

> Danke, viel gluck (no umlaut available)!
> Grim

Whenever I 'm not able to use an "Umlaut" I insert a 'e' after the vowel.
For
example "Hoehe" instead of "Höhe" (that is "height" and often used for
variables
in povray...)

Dankeschön (dankeschoen:-)

See you in p.b.s-f
 Kallek

Post a reply to this message

..and that would be why I don't reply in foreign tongues. :)

Thanks, I'll check out your code.

Grim

Post a reply to this message

Tried making one myself now.
"artifact" across it too, lined up with the horizon, but I don't think it's in
error.
max_trace_level was at 15 and the camera offset from any axis.  The top of the
liquid is very prominent too but it's due to the surface tension or adhesion to
glass I put there.  Too much exaggeration and reflection possibly.
Needs work, I know.  Just showing what I got when making this same thing.  I
had put a heat wave effect near the bottom of the picture but it's an awful
attempt.

Bob

Post a reply to this message

Attachments:
Download 'therm.jpg' (45 KB)

Preview of image 'therm.jpg'

How come it doesn't cast a shadow?

"Bob H." wrote:

> Tried making one myself now.
> "artifact" across it too, lined up with the horizon, but I don't think it's in
> error.
> max_trace_level was at 15 and the camera offset from any axis.  The top of the
> liquid is very prominent too but it's due to the surface tension or adhesion to
> glass I put there.  Too much exaggeration and reflection possibly.
> Needs work, I know.  Just showing what I got when making this same thing.  I
> had put a heat wave effect near the bottom of the picture but it's an awful
> attempt.
>
> Bob
>
>  [Image]

Post a reply to this message

The problem seems to be exacerbated by the use of 'filter' versus
'transmit.' I didn't play with your scene enough to render a complete image
(it is your effort after all), but I can tell you that I believe this is the
way to go. Instead of using rgbf you might try rgbt. The fog doesn't help
either, so you may wish to add a transmit value to it, too.

Good luck!

Grim

Post a reply to this message

"Thomas Lake" <tla### [at] homecom> wrote in message
news:3A0DFEAC.AEE8D32A@home.com...
> How come it doesn't cast a shadow?

Oh, you're right.  At least it doesn't seem to be doing so.  There was one
there when the clear liquid was a red color for testing and the lights were
closer to the floor plane but all that changed.  Didn't think about it on this
render.  And I have caustics in there too, gone to waste.  Photons are next so
I'll be forced to arrange it better anyhow.

Bob

Post a reply to this message

Hi Bob!

> Tried making one myself now.
Nice! The rainbow is a good idea. And it's not a background rainbow, is it?
What is it then?

> "artifact" across it too, lined up with the horizon, but I don't think
it's in
> error.
What else? An effect with the horizon?

> max_trace_level was at 15 and the camera offset from any axis.  The top of
the
> liquid is very prominent too but it's due to the surface tension or
adhesion to
> glass I put there.  Too much exaggeration and reflection possibly.
That's the game with refracting objects. At first I used a rather high
amount of reflection, too - a white object and a rendertime going through
the roof. The former times...

> Needs work, I know.  Just showing what I got when making this same thing.
I
> had put a heat wave effect near the bottom of the picture but it's an
awful
> attempt.
It is a bit strange effect - the heat wave. How did you make it? a plane
with a gradient y decreasing normal-turbulence?

cukk

Post a reply to this message

Hi Grim!

> The problem seems to be exacerbated by the use of 'filter' versus
> 'transmit.' I didn't play with your scene enough to render a complete
image
> (it is your effort after all), but I can tell you that I believe this is
the
> way to go. Instead of using rgbf you might try rgbt. The fog doesn't help
> either, so you may wish to add a transmit value to it, too.
But it even appears without filter nor transmit. Somehow the normal in that
region is not right, causing the rays to refract the wrong way. But you can
see the error even just with a strange shading

> Good luck!
Thanks

cukk

Post a reply to this message

"KalleK" <kal### [at] gmxde> wrote in message news:3a12d809@news.povray.org...
> Nice! The rainbow is a good idea. And it's not a background rainbow, is it?
> What is it then?

A rainbow { }.  The POV-Ray kind.  Seems to be overlooked a lot that one exists
in POV-Ray.  Maybe because it's one of those not-so-easy to control sort of
things.  At least not right away.

> > "artifact" across it too, lined up with the horizon, but I don't think
> it's in
> > error.
> What else? An effect with the horizon?

Effect of refraction, I'd say, of the horizon.  I think I used too much
transmit in my glass and liquid too though.

> It is a bit strange effect - the heat wave. How did you make it? a plane
> with a gradient y decreasing normal-turbulence?

Good guess.  A ior 1.01 too.  I used it in the final material { } which had a
texture {texture_map{}} because of not knowing any way to do a material_map.
material_map, as you might know, is not a map made up of material { }
statements.

Bob

Post a reply to this message

I have since made another render of my attempt at a Galileo Thermometer with
the floor plane clipped to a small size and found that the lathe does indeed
continue to produce the artifact.  The camera has always been offset from any
axis and yet it still persists, probably until the camera is completely
non-parallel with any point along the lathe height, but I haven't checked.
Sorry if I sounded argumentive before.  I was going on what I experienced
before and figured it was either a case of point placements of spline (it might
still be only cubic_spline doing it, I just don't know, sturm isn't used at all
for it, only the quadratic_spline I think) or axis-aligned camera.
Looked at the bugreport group about what Thomas Willhalm had said and I guess
there was never any fix found or produced.

Bob

Post a reply to this message

Bob it may be worth posting your code so that we can see if the
problem can be reproduced on other systems, or even find a 
solution.

-- 
Cheers
Steve              email mailto:ste### [at] zeroppsuklinuxnet

%HAV-A-NICEDAY Error not enough coffee  0 pps. 

web http://www.zeropps.uklinux.net/

or  http://start.at/zero-pps

  3:12pm  up 37 days, 17:35,  2 users,  load average: 0.99, 0.98, 0.99

Post a reply to this message

"Steve" <ste### [at] zeroppsuklinuxnet> wrote in message
news:slr### [at] zero-ppslocaldomain...
> Bob it may be worth posting your code so that we can see if the
> problem can be reproduced on other systems, or even find a
> solution.

Well, I found that going with a quadratic_spline instead of cubic_spline appears to
almost fix it completely
when sturm is used.
Just now getting back here to post a reply and now I have to leave again.
The pov script was put at povray.scenes.text-files yesterday as if anyone sees this
message try replacing the
cubic_spline type with the quadratic_spline and uncomment the sturm keywords.  The
shape has to be adjusted for
but it seems to be the answer.  Almost but not quite maybe.  I think it might still be
susceptible to problems.
I haven't tried a sor yet.

Bob

Post a reply to this message

"Bob H." wrote:

> I haven't tried a sor yet.

For the record I have had this problem more times with a SOR object
than I have had with lathes. I have pretty much given up on SORs
altogether because of this.

-- 
Ken Tyler - 1400+ POV-Ray, Graphics, 3D Rendering, and Raytracing Links:
http://home.pacbell.net/tylereng/index.html http://www.povray.org/links/

Post a reply to this message