I thought I read somewhere along the threads that
the dispersion patch has something like
ior_min and ior_max implemented. I just downloadad
it and can only find dispersion float and
disp_nelems int in the "documentation".

How is the ior dispersed across the
wavelengths?

Any tips? Or am I completely wrong here?

Markus

-- 

 Ich nicht eine Sekunde!!!" H. Heinol in Val Thorens

Post a reply to this message

It's done by wavelength.  The dispersion controls how far the ior
differs from the shortest to the highest wavelength.  The center
wavelength is always yellow, since that's what ior in the real world is
measured against (actually a thin line of sodium, which is yellow).

-Mike

Markus Becker wrote:
> 
> I thought I read somewhere along the threads that
> the dispersion patch has something like
> ior_min and ior_max implemented. I just downloadad
> it and can only find dispersion float and
> disp_nelems int in the "documentation".
> 
> How is the ior dispersed across the
> wavelengths?
> 
> Any tips? Or am I completely wrong here?
> 
> Markus
> 
> --

>  Ich nicht eine Sekunde!!!" H. Heinol in Val Thorens

Post a reply to this message

You know a patch is big hit, when other people give the correct answer
long before the patch author gets around to reading the question  :-)


Just want to clarify some details.   The IOR in the dispersion patch is
an exponential function of color number.  By color number, i mean i'm
counting subrays that are traced, from 1 to disp_nelems.   In other
words, each subray uses an IOR a certain ratio higher than the previous
subray.  This is not based on physics, but ease of implementation and
good looks.  IMHO, a linear function did not look good enough.   Real
physical materials have a variety of dipserson curves, which would be
fun to model, but I use povray for art not physics.

Each subray is traced, and contributes some amount of a pure hue to the
sum.  The formula for converting the nth ray into a color is purely
ad-hoc, chosen to look good, again no real physical laws were refered
to, and to match some mathematical criterea, mainly that it had to sum
all-equal subrays into pure white.  For looks, I want a nice yellow, and
violet at the low-wavelength end.

At no point in the process do i calculate or use wavelengths.  

With the  source code out there, it's just a matter of time before
someone tweaks it to behave more in accordance with real optics.    I'm
too lazy, and have a totally different ray tracer for good physics
rendering.

-- 
Daren Scot Wilson
dar### [at] pipelinecom 
www.newcolor.com
----
"A ship in a harbor is safe, but that is not what ships are built for"
                                            -- William Shedd

Post a reply to this message

I was giving the technical description of what it's supposed to be
doing. :)

You did seem to 'fudge' the colors towards making yellow in the middle,
which seems to be the accuarate way of doing it.  I don't know if anyone
will ever bother for a physically accurate representation.  I know I
sure won't.

I doubt most people will notice small inaccuracies unless they are
getting a PHD in optical physics, and those people shouldn't be turning
to POV to do their dissertation anyway!  ;-)

-Mike

Daren Scot Wilson wrote:
> 
> You know a patch is big hit, when other people give the correct answer
> long before the patch author gets around to reading the question  :-)
> 
> Just want to clarify some details.   The IOR in the dispersion patch is
> an exponential function of color number.  By color number, i mean i'm
> counting subrays that are traced, from 1 to disp_nelems.   In other
> words, each subray uses an IOR a certain ratio higher than the previous
> subray.  This is not based on physics, but ease of implementation and
> good looks.  IMHO, a linear function did not look good enough.   Real
> physical materials have a variety of dipserson curves, which would be
> fun to model, but I use povray for art not physics.
> 
> Each subray is traced, and contributes some amount of a pure hue to the
> sum.  The formula for converting the nth ray into a color is purely
> ad-hoc, chosen to look good, again no real physical laws were refered
> to, and to match some mathematical criterea, mainly that it had to sum
> all-equal subrays into pure white.  For looks, I want a nice yellow, and
> violet at the low-wavelength end.
> 
> At no point in the process do i calculate or use wavelengths.
> 
> With the  source code out there, it's just a matter of time before
> someone tweaks it to behave more in accordance with real optics.    I'm
> too lazy, and have a totally different ray tracer for good physics
> rendering.
> 
> --
> Daren Scot Wilson
> dar### [at] pipelinecom
> www.newcolor.com
> ----
> "A ship in a harbor is safe, but that is not what ships are built for"
>                                             -- William Shedd

Post a reply to this message

Daren Scot Wilson wrote:
> 
> Just want to clarify some details.   The IOR in the dispersion patch is
> an exponential function of color number.  By color number, i mean i'm
> counting subrays that are traced, from 1 to disp_nelems.   In other
> words, each subray uses an IOR a certain ratio higher than the previous
> subray.

Ok, I understand that. So no _real_ dispersion :( (;-))

Another thing I noticed (sh*t that I didn't take the
POV source with me to demonstrate it):

Make a scene with a perfectly clear, non reflecting but
refracting sphere sitting on, say, a checker floor.
Render it once with and once without dispersion and
you'll notice that the dispersion version is slightly
brighter...
Could this be the common error made in the gamma-correction
code?

Markus
-- 

 Ich nicht eine Sekunde!!!" H. Heinol in Val Thorens

Post a reply to this message

> Make a scene with a perfectly clear, non reflecting but
> refracting sphere sitting on, say, a checker floor.
> Render it once with and once without dispersion and
> you'll notice that the dispersion version is slightly
> brighter...
> Could this be the common error made in the gamma-correction
> code?

No, actually, it's a sign that all the fundamental laws of logic, space and
time are falling apart. The physical universe as we know will cease to exist...

It is a bad thing that the dispersion patch has a hard-coded spectrum.  What we
need is user-defined spectra, and then I won't ever be troubled by such
questions.   And clever artists could then do some weird things.

Consider the average refractor telescope's objective lens, or lenses in any
optical instrument.  Normally these are made of two kinds of glass, for example
crown glass and flint glass, with different dispersions.  How to model this in
povray?  

Lemme work on my IRTC entry, and in march I'll figure out how to allow
user-defined dispersions.   In the meantime I'll accept ideas, suggestions.  

Later, for an encore, maybe I'll do birefringent materials and polarization...


-- 
Daren Scot Wilson
dar### [at] pipelinecom 
www.newcolor.com
----
"A ship in a harbor is safe, but that is not what ships are built for"
                                            -- William Shedd

Post a reply to this message

Daren Scot Wilson wrote:
> 
> It is a bad thing that the dispersion patch has a hard-coded spectrum.  What we
> need is user-defined spectra, and then I won't ever be troubled by such
> questions.   And clever artists could then do some weird things.
> 
> Consider the average refractor telescope's objective lens, or lenses in any
> optical instrument.  Normally these are made of two kinds of glass, for example
> crown glass and flint glass, with different dispersions.  How to model this in
> povray?
> 
> Lemme work on my IRTC entry, and in march I'll figure out how to allow
> user-defined dispersions.   In the meantime I'll accept ideas, suggestions.
> 
> Later, for an encore, maybe I'll do birefringent materials and polarization...
> 

I'm implementing dispersion with photon-mapping (using your code), and I will
soon switch to some form of user-defined spectra (at least for light sources...
so you can define monochromatic yellow or green+blue yellow, for example).

-Nathan

Post a reply to this message

Daren Scot Wilson wrote:
> 
> No, actually, it's a sign that all the fundamental laws of logic, space and
> time are falling apart. The physical universe as we know will cease to exist...

Read too much of Douglas Adams, I guess? ;-)

> It is a bad thing that the dispersion patch has a hard-coded spectrum.  What we
> need is user-defined spectra, and then I won't ever be troubled by such
> questions.   And clever artists could then do some weird things.

Definitely. I'll give you some time...

> Consider the average refractor telescope's objective lens, or lenses in any
> optical instrument.  Normally these are made of two kinds of glass, for example
> crown glass and flint glass, with different dispersions.  How to model this in
> povray?

No problem, if every object can be assigned different dispersion
spectra.

> Lemme work on my IRTC entry, and in march I'll figure out how to allow
> user-defined dispersions.   In the meantime I'll accept ideas, suggestions.

Ok, ok,

> Later, for an encore, maybe I'll do birefringent materials and polarization...

Oh yeah, and please, please do coherent light, so that I can
finally model interference and laser speckle and the like,
hang on, man!!!

Markus
-- 

 Ich nicht eine Sekunde!!!" H. Heinol in Val Thorens

Post a reply to this message

Markus Becker wrote:
> 
> > Later, for an encore, maybe I'll do birefringent materials and polarization...
> 
> Oh yeah, and please, please do coherent light, so that I can
> finally model interference and laser speckle and the like,
> hang on, man!!!
> 
> Markus

Yeah ! And beam splitting, and diffraction grating experiments,
and..., and...

-- 
Ken Tyler

mailto://tylereng@pacbell.net

Post a reply to this message

Markus Becker wrote:
> 
> Daren Scot Wilson wrote:
> >
> > No, actually, it's a sign that all the fundamental laws of logic, space and
> > time are falling apart. The physical universe as we know will cease to exist...
> 
> Read too much of Douglas Adams, I guess? ;-)

No, probably just went out of the door, or looked in a newspaper... :-)

//Spider

Post a reply to this message