Alain <aze### [at] qwertyorg> wrote:

> > Hi all,
> >
> > I was wondering if it is possible to create a photographic plate in POV-Ray? You
> > could use this plate to create a camera obscura with upside-down image on the
> > plate, or visualize the typical upside-down candle when using a lens.
> >
> > I guess not, reading on how POV-Ray works by casting rays from the origin of the
> > camera to every pixel in its viewport. What I ask requires the inverse: casting
> > rays from the light sources, and see if they happen to fall on the sensor.
> > Obviously, the former one is faster.
> >
> > If not possible out-of-the-box, are there techniques to achieve this? I was
> > thinking along the lines of putting a 'milk-glass' plane where the photographic
> > plate, or sensor, should be.
> >
> > I just read about photon mapping
> >
(http://wiki.povray.org/content/Knowledgebase:Language_Questions_and_Tips#Topic_27).
> > Would that be of any help for what I like to do?
> Absolutely!
> >
> > Keywords, to know at least what to google for, would be handy.
> >
> > Cheers!
> >
> > Bart
> >
> >
> >
> >
>
> You can use the photons feature. It uses forward ray tracing.
>
> All objects to be seen will need to have some good amount of reflection,
> and it will need to be diffuse reflection achieved using averaged
> normals. Make the objects no_reflection to prevent inter-reflection
> unless the object is to be seen in a mirror. That will improve the
> rendering time.
> All those objects will need to be made photons target with reflection on.
> The light_source will probably need to be prety strong.
> Try to use a lense as whide as possible. The more photons you capture,
> the beter the result.
> You will probably need a prety small spacing like 0.01 or less, or a
> prety large count, like over 20000000.
> Only use spacing OR count, not both.
>
> Another way is with radiosity.
> There are some stringent limitations:
> The focal lenght must be short: less that twice the diameter of the lense.
> The radiosity count must be large. No less that 400 in the best case if
> you want acceptable results. The longer the focal lenght, the more
> samples are needed. In some cases, count 1600 (the maximum value) is to
> small...
> The rendering time will be long to very long.
> The plus side is that you don't need to make the objects reflective.

Nice news! I'll give the photon thing a spin tomorrow.

Thanks!

Bart

Post a reply to this message