|
|
Reactor schrieb:
> That doesn't sound good, but I made an image that shows more of what I had in
> mind. I posted it in p.b.i here:
>
>
http://news.povray.org/povray.binaries.images/thread/%3Cweb.4a7ca1d884f249fd4646012a0@news.povray.org%3E
Hum... now, that's a *totally* different thing. And I guess I can
confirm that indeed it appears to be based on a poor understanding of
lighting computations ;-)
It would theoretically be possible to *misuse* a transmissive or filter
component of the ambient to control such an effect, but that's about it.
The math behind it would have not much similarity with what is done on
the RGB channels. (And it would pretty much mess up the math, which is
quirky enough as it is.)
If I were to hard-code anything in POV-Ray to model such an effect, I'd
go for a kind of "light map" pattern: Taking a #define'd light source
(or, alternatively, the actual illumination, including radiosity and
photons and all), possibly multiplying it by some RGB colour, converting
to greyscale, and using the resulting value to index a pigment or
texture map. From there, you could easily model precisely the effect you
want - plus some.
Post a reply to this message
|
|
|
|
clipka <ano### [at] anonymousorg> wrote:
> Reactor schrieb:
> > That doesn't sound good, but I made an image that shows more of what I had in
> > mind. I posted it in p.b.i here:
> >
> >
http://news.povray.org/povray.binaries.images/thread/%3Cweb.4a7ca1d884f249fd4646012a0@news.povray.org%3E
>
> Hum... now, that's a *totally* different thing. And I guess I can
> confirm that indeed it appears to be based on a poor understanding of
> lighting computations ;-)
>
> It would theoretically be possible to *misuse* a transmissive or filter
> component of the ambient to control such an effect, but that's about it.
> The math behind it would have not much similarity with what is done on
> the RGB channels. (And it would pretty much mess up the math, which is
> quirky enough as it is.)
>
> If I were to hard-code anything in POV-Ray to model such an effect, I'd
> go for a kind of "light map" pattern: Taking a #define'd light source
> (or, alternatively, the actual illumination, including radiosity and
> photons and all), possibly multiplying it by some RGB colour, converting
> to greyscale, and using the resulting value to index a pigment or
> texture map. From there, you could easily model precisely the effect you
> want - plus some.
I think I just might have to start a new feature request thread, then. That
description suddenly reminded me of the projection pattern in MegaPov, so I'm
going to go try abusing that some more first. First experiments look sort of
promising.
-Reactor
Post a reply to this message
|
|