 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Le 21/12/2010 05:07, [GDS|Entropy] a écrit :
> Link to post: http://forums.cgsociety.org/archive/index.php/t-513458.html
>
> Not vouching for accuracy, but this may be useful to someone. Maybe it
> would make a nice include or something.
>
Calling on accuracy: at least two different values for aluminum
(that's a US things too that is called aluminium in the other parts of
the world)
And refracting aluminium... just wondering.
--
Those who would give up Essential Liberty to purchase
a little Temporary Safety, deserve neither Liberty nor Safety.
Benjamin Franklin
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 21.12.2010 10:54, schrieb Le_Forgeron:
> And refracting aluminium... just wondering.
/All/ materials have an index of refraction, even metallic ones.
The value for alumin(i)um is definitely wrong though: Refractive indices
of metals are actually complex numbers.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Le 21/12/2010 13:31, clipka a écrit :
> Am 21.12.2010 10:54, schrieb Le_Forgeron:
>
>> And refracting aluminium... just wondering.
>
> /All/ materials have an index of refraction, even metallic ones.
>
> The value for alumin(i)um is definitely wrong though: Refractive indices
> of metals are actually complex numbers.
I agree if we go with the electromagnetic model (including the various
polarisations) of propagation. (and while we are at it, let's handle
chromatics too, it's just the easy part before handling birefringence
and other natural phenomenons.)
But it's a bit hard to just use it with the ray model.
Moreover, the original list came from a computer graphic site, I'm
afraid it's yet some kind of magic numbers in a hidden formula to make
it look good as the material with the usual set of light. A bit like the
ratio of ambient/diffuse or similar properties of Povray.
We could probably make a list of
phong/phong_size/specular/roughness/crand settings for each material...
it would have the same kind of magic numbers.
Just like we have the "metals.inc" with many chrome / silver / copper.
<joke>
I'm volunteering to test the precious gems rendering... please provide
me with many real big samples of diamond, rubies, sapphires, emerald,
... for comparisons. (samples won't be returned).
I can do precious metal too: gold, silver, copper,... need at least
100kg of each, thanks.
--
A good Manager will take you
through the forest, no mater what.
A Leader will take time to climb on a
Tree and say 'This is the wrong forest'.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
great! Now we need the colors. ;)
IOR is also used in povray for fresnel reflection.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
> great! Now we need the colors. ;)
>
> IOR is also used in povray for fresnel reflection.
>
>
>
We also need the dispersion values.
At first glance, the values /looks/ sane. Can't say for things like
asphalt (this one should be variable acording to the exct recipie used),
rubber and metals.
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 22.12.2010 00:53, schrieb Alain:
> We also need the dispersion values.
You might want to have a look at the include/ior.inc that should ship
with 3.7 beta 41. It's by far not as exhaustive (being focused on gems
and minerals), but does include dispersion information for most of the
materials in POV-Ray-palatable format.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Ok...I will respond to myself due to the sheer volume of comments. :)
I like what I have seen from the community...especially the more advanced
concepts; however one must have a starting point before going that
far...and this looks like it. I have lusted after many of these values for
a long time...shame it never occurred to me to just Google it before
now... >__<
1) Where would we get the missing parameters accompanying these IOR values?
1a) It seems the povray params for these missing values should be an
f(x); we have some data points already...is there a decent way to produce
the governing function, so we could just write a program and dump the
output? Given the pov model exists and is not randomly variable,
determining the function should be possible...though I do not claim
to be able to do so, at least without more research than my current
project load would allow.
2) How do we reach a consensus on the dubious IOR values?
3) Regarding magic numbers....being a programmer I have an inherent
dislike for such things...however, having a number of reference points
extrinsic to any particular gfx package, with which my current known
list of IOR values match up to, lends some credence to at least those
values which are easily confirmed by multiple sources. Margins of error
are a fact of life in any business unit, and if a software engineer
should spend the time required to eliminate all of it, I seriously doubt
he would get much programming done. Sometimes it is best to take what
you have, make something with it, and correct the errors later. I don't
like it either...but it is the way of life... :|
Ian
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
For 1a):
Perhaps the best place to start is a standard list of wide ranging (ior
material 1=a, ior material g=f) values which can be validated
independently. Then when we agree upon a list<T> where T=IOR values, we
can plug those into the pov model to determine the phase space of the
f(x)...and so as such may be able to reverse engineer the requisite
diffuse, etc.. parameters. If such a model can be completed, where the
f(x) can be defined, then any IOR x (and some parameters ??) should result
in the corresponding missing values. I have used this tactic in the
financial sector with great success...so...maybe a good strategy for this
problem as well?
Ian
On Tue, 21 Dec 2010 21:55:41 -0500, [GDS|Entropy]
<gdsHYentropyAThotmailDTcom> wrote:
> Ok...I will respond to myself due to the sheer volume of comments. :)
>
> I like what I have seen from the community...especially the more
> advanced concepts; however one must have a starting point before going
> that far...and this looks like it. I have lusted after many of these
> values for a long time...shame it never occurred to me to just Google it
> before now... >__<
>
> 1) Where would we get the missing parameters accompanying these IOR
> values?
> 1a) It seems the povray params for these missing values should be an
> f(x); we have some data points already...is there a decent way to
> produce the governing function, so we could just write a program and
> dump the output? Given the pov model exists and is not randomly
> variable, determining the function should be possible...though I do not
> claim to be able to do so, at least without more research than my
> current project load would allow.
> 2) How do we reach a consensus on the dubious IOR values?
> 3) Regarding magic numbers....being a programmer I have an inherent
> dislike for such things...however, having a number of reference points
> extrinsic to any particular gfx package, with which my current known
> list of IOR values match up to, lends some credence to at least those
> values which are easily confirmed by multiple sources. Margins of error
> are a fact of life in any business unit, and if a software engineer
> should spend the time required to eliminate all of it, I seriously doubt
> he would get much programming done. Sometimes it is best to take
> what you have, make something with it, and correct the errors later. I
> don't like it either...but it is the way of life... :|
>
> Ian
--
Using Opera's revolutionary email client: http://www.opera.com/mail/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 22.12.2010 03:55, schrieb [GDS|Entropy]:
> 1a) It seems the povray params for these missing values should be an
> f(x); we have some data points already...is there a decent way to
> produce the governing function, so we could just write a program and
> dump the output?
The actual surface properties depend on:
[ambient] the ambient illumination, so no material property here at all;
[specular/roughness,phong/phong_size,reflection] the surface structure,
which may vary on how the material was processed (e.g. polished, sanded,
whatever); not a material property in that sense either; and while the
parameters also depend on the IOR, this is already accounted for by the
fresnel model, at least for reflections;
[diffuse] the IOR, but also a set of wavelength dependent parameters: An
absorption coefficient, a scattering coefficient, and - if you want to
do it really precise - a scattering distribution. From these you can
even compute the material pigment (and model subsurface scattering
besides), but without these you cannot even compute the diffuse parameter.
So sorry, but no banana here.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Wed, 22 Dec 2010 02:52:59 -0500, clipka <ano### [at] anonymous org> wrote:
> Am 22.12.2010 03:55, schrieb [GDS|Entropy]:
>
>> 1a) It seems the povray params for these missing values should be an
>> f(x); we have some data points already...is there a decent way to
>> produce the governing function, so we could just write a program and
>> dump the output?
>
> The actual surface properties depend on:
>
> [ambient] the ambient illumination, so no material property here at all;
>
> [specular/roughness,phong/phong_size,reflection] the surface structure,
> which may vary on how the material was processed (e.g. polished, sanded,
> whatever); not a material property in that sense either; and while the
> parameters also depend on the IOR, this is already accounted for by the
> fresnel model, at least for reflections;
>
> [diffuse] the IOR, but also a set of wavelength dependent parameters: An
> absorption coefficient, a scattering coefficient, and - if you want to
> do it really precise - a scattering distribution. From these you can
> even compute the material pigment (and model subsurface scattering
> besides), but without these you cannot even compute the diffuse
> parameter.
Material was a poor choice for a word in relation to its meaning in SDL;
it just fit my mental picture of the data construct best at the time.
> So sorry, but no banana here.
Hmm...well, banana or no, there simply has to be a way to do this...people
either collectively or individually have solved more complicated problems
than this. I do not currently have enough understanding of all intrinsic
factors requisite to performing such a feat at this time, quite obviously.
:/
Ian
--
Using Opera's revolutionary email client: http://www.opera.com/mail/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
