Everyone was so great in replying to my previous question, that I'd
like to post this topic. If this was an essay question, I would phrase
it so: To what extent does Povray simulate reality and to what extent
are its images simply illusions? The point is that to allow the user to
create all sorts of images, Povray gives incredible control over
reality. Consider shadowless objects. Or, say I want to make a scene
with a mirror that doesn't reflect a vampire. In reality, mirrors
reflect and there are no vampires, but that can be changed in Povray.
You can make a ceiling out of water, or a seemingly impossible m.c.
echer model, or abstract fractals, or whatever.
  It seems obvious. Of course Povray should allow you to do all those
things. The only thing that should be real is the light and its
behavior. But consider this. When I was starting, I wanted to simulate
the effect of an oil slick on water. So, I made two thin surfaces with
different indexes of refraction, hoping to get the effect. Boy, was I
surprised that the irridescence property can achieve the interference
effect on a single surface. In reality, that's impossible.
   So here are several technical problems I'm facing now. First, I want
to make a realistic looking CD. Now, it wouldn't be wise to make a disk
with millions of dents in it, would it. So what do I use to fool the
viewer? A rainbow, a halo, what? Or how about this: as you may know, if
you shine a strong, monochromatic, coherent light.right at a penny, you
should get a bright spot in the center of its shadow as a result of
light diffraction. I can certainly achieve the effect, but how much of
it would be attributed to properties of light? Or, how about Young's
double-slit experiments and light fringes, diffraction gratings, or
polarized light. And what about "materials" with unusual optical
properties. Gypsum, for example, is trasparent but it creates a double
image of anything viewed through it. How much can you assign and let
Povray take care of it, and how much do you have to take upon yourself
to model?

Post a reply to this message

In article <399467BF.48D6C480@yahoo.com>, Adam <bel### [at] yahoocom> 
wrote:

>   It seems obvious. Of course Povray should allow you to do all those
> things. The only thing that should be real is the light and its
> behavior. But consider this. When I was starting, I wanted to simulate
> the effect of an oil slick on water. So, I made two thin surfaces with
> different indexes of refraction, hoping to get the effect. Boy, was I
> surprised that the irridescence property can achieve the interference
> effect on a single surface. In reality, that's impossible.

Effects of light wavelength are not calculated...some of these these are 
very difficult to write and slow to compute. The iridescence feature 
fakes thickness of the surface for it's calculations.


>    So here are several technical problems I'm facing now. First, I want
> to make a realistic looking CD. Now, it wouldn't be wise to make a disk
> with millions of dents in it, would it. So what do I use to fool the
> viewer? A rainbow, a halo, what?

One method is to use a very fine spiral, onion, or wood normal, 
iridescence, and high anti-aliasing settings. Rainbow isn't useful for 
this, and if your version of POV uses "halo", you need to upgrade.(it 
has been replaced by "media", which, BTW, won't help with this)
There is another method of using averaged textures being discussed in 
povray.binaries.images right now...


> Or how about this: as you may know, if
> you shine a strong, monochromatic, coherent light.right at a penny, you
> should get a bright spot in the center of its shadow as a result of
> light diffraction. I can certainly achieve the effect, but how much of
> it would be attributed to properties of light? Or, how about Young's
> double-slit experiments and light fringes, diffraction gratings, or
> polarized light. And what about "materials" with unusual optical
> properties. Gypsum, for example, is trasparent but it creates a double
> image of anything viewed through it. How much can you assign and let
> Povray take care of it, and how much do you have to take upon yourself
> to model?

Again, these optical effects(diffraction, double refraction, 
polarization...) are either too application-specific or too difficult to 
code/computationally expensive to compute. POV-Ray isn't a complete 
optics simulation, no program is(though some come close).

BTW, I don't think gypsum is double refracting, I think you mean calcite.

POV can calculate direct light, diffuse light(radiosity), refraction of 
light going to the camera, and with photon mapping(in MegaPOV), the 
refraction of light from light sources to surfaces in the scene. There 
is another MegaPOV patch which simulates dispersion(wavelength dependant 
refraction). You shouldn't rely on it simulating more.

-- 
Christopher James Huff - Personal e-mail: chr### [at] maccom
TAG(Technical Assistance Group) e-mail: chr### [at] tagpovrayorg
Personal Web page: http://homepage.mac.com/chrishuff/
TAG Web page: http://tag.povray.org/

Post a reply to this message

Povray implements a raytracing paradigm. As such its endpoint is depiction
not simulation.  Simulation is integral to, but only part of, its means ( a
raytracer's that is ).  As your question sketches out, there are some
interesting ways that the raytracing paradigm, (and by extension, povray's
excellent implementation of it ), causes us to revisite some intriguing
questions from the history of pictorial art and the world of phenonmena.
What achieves the greatest sense of "reality" in a depiction.  Does it
proceed from knowledge, is it synthesized from a depictional language, or
is it bound by the reception of pure phenonmena?  Do you present "reality"
by recording its associated phenonmena?  Or does its accurate depiction
extend from it's accurate understanding.  Is perception a matter of
sensation or concept.  These questions have stood behind the work of
imagemakers and their commentators to numerous to mention.

Post a reply to this message

>
> Again, these optical effects(diffraction, double refraction,
> polarization...) are either too application-specific or too difficult to
> code/computationally expensive to compute. POV-Ray isn't a complete
> optics simulation, no program is(though some come close).

I guess with these programs it is customary to cut corners, like Povray does
with caustics. I mean, no one's going to model water droplets to make a
cloud, right. I wonder if you can make distant trees or plants using noise
instead of having individual leafs.
But still, say I want to model polarized sunglasses. As you know, they make
it easier to see underwater. Or, say I want to test an anti-glare coating and
I want to choose the coefficients that will cause the right frequencies to
interfere destructively. Does that mean I should get (or make) a specific
program?

> BTW, I don't think gypsum is double refracting, I think you mean calcite.

Huh, was it calcite? You may be right. I had my doubts. I used to know all
the minerals way back when.

> POV can calculate direct light, diffuse light (radiosity), refraction of
> light going to the camera, and with photon mapping (in MegaPOV), the
> refraction of light from light sources to surfaces in the scene. There
> is another MegaPOV patch which simulates dispersion (wavelength dependant
> refraction). You shouldn't rely on it simulating more.

You know, maybe I should try this Megapov, especially the
wavelength-dependent refraction. It's just that I'm a bit disappointed when I
see people who claim to make beautiful sparkling diamonds, but with lack of
color. Even the simplest things in life, like bevelling on the sides of
mirrors, or a lens next to a mirror make some spectacular effects. Sadly, if
you're not looking hard enough, you don't even know what to fake.

Post a reply to this message

Adam <bel### [at] yahoocom> wrote:
: You can make a ceiling out of water, or a seemingly impossible m.c.
: echer model

  Actually you can't. Yes, you can make an image that looks like one of those
impossible models, but if you move the camera even a bit you will immediately
see that it's only a fake that can be seen _only_ from that point of view.
This is because povray models 3D quite accurately. If an object is at a
certain location, then it is at that location, period.

-- 
main(i,_){for(_?--i,main(i+2,"FhhQHFIJD|FQTITFN]zRFHhhTBFHhhTBFysdB"[i]
):_;i&&_>1;printf("%s",_-70?_&1?"[]":" ":(_=0,"\n")),_/=2);} /*- Warp -*/

Post a reply to this message

In article <3994EB2C.844F14BE@yahoo.com>, Adam <bel### [at] yahoocom> 
wrote:

> I guess with these programs it is customary to cut corners, like 
> Povray does with caustics. I mean, no one's going to model water 
> droplets to make a cloud, right. I wonder if you can make distant 
> trees or plants using noise instead of having individual leafs.

That is something up to the user...it would sometimes be more efficient 
in memory to use the same level of detail for all trees, for example, 
when you only have a few tree meshes which are copied over the whole 
scene.(a copy of a mesh takes up much less memory than a new one)
In other cases, various methods like textured flat objects or simple 
isosurfaces have been used.


> But still, say I want to model polarized sunglasses. As you know, 
> they make it easier to see underwater. Or, say I want to test an 
> anti-glare coating and I want to choose the coefficients that will 
> cause the right frequencies to interfere destructively. Does that 
> mean I should get (or make) a specific program?

You would probably have to write the program yourself...these things are 
pretty specific. You might be able to find such a program, but I highly 
doubt it would be free(or even cheap).

-- 
Christopher James Huff - Personal e-mail: chr### [at] maccom
TAG(Technical Assistance Group) e-mail: chr### [at] tagpovrayorg
Personal Web page: http://homepage.mac.com/chrishuff/
TAG Web page: http://tag.povray.org/

Post a reply to this message

Warp wrote:

>   Actually you can't. Yes, you can make an image that looks like one of those
> impossible models, but if you move the camera even a bit you will immediately
> see that it's only a fake that can be seen _only_ from that point of view.
> This is because povray models 3D quite accurately. If an object is at a
> certain location, then it is at that location, period.

Yeah, true. In fact, I just saw that 2nd place animation by Joe Wise on IRTC for
the last Inner Workings round. Very clever, I must say. But, as he said, he could
have used orthogonal cameras to get rid of perspective, which tends to spoil the
illusion.

Post a reply to this message

>
> > BTW, I don't think gypsum is double refracting, I think you mean calcite.

Wait, of course Iceland spar, that is, calcite, exhibits birefringence. I don't
know what I was thinking with the dihydrate of calcium sulfate, I mean gypsum. I
didn't even mean gypsum when I said it. Gypsum is what causes noncarbonate
hardness of water.

Come to think of it, how would I do double refraction anyway?

Post a reply to this message

In article <3997786E.B6B66551@yahoo.com>, Adam <bel### [at] yahoocom> 
wrote:

> Wait, of course Iceland spar, that is, calcite, exhibits 
> birefringence. I don't know what I was thinking with the dihydrate of 
> calcium sulfate, I mean gypsum. I didn't even mean gypsum when I said 
> it.

I don't think crystalline calcium carbonate is the only mineral to 
exhibit this behavior, though...I just can't think of another.


> Gypsum is what causes noncarbonate hardness of water.

I didn't think calcium sulfate was soluble enough to really affect water 
hardness.


> Come to think of it, how would I do double refraction anyway?

By averaging the results of calculating refraction twice, with different 
iors?

-- 
Christopher James Huff - Personal e-mail: chr### [at] maccom
TAG(Technical Assistance Group) e-mail: chr### [at] tagpovrayorg
Personal Web page: http://homepage.mac.com/chrishuff/
TAG Web page: http://tag.povray.org/

Post a reply to this message

> I want to make a realistic looking CD.

check out the excellent cd made by Rune S. Johansen
http://rsj.mobilixnet.dk/3d/cd/cd.html

> should get a bright spot in the center of its shadow as a result of
> light diffraction

I belive I've heard Megapov (a povray patch) can do that?
http://nathan.kopp.com/patched.htm
Correct me If I'm wrong.

--
Peter
http://hertel.no/bigone

Post a reply to this message