Am 21.08.2011 03:14, schrieb jmichae3:
> diffraction grating can be studied at
> http://en.wikipedia.org/wiki/Diffraction_grating
> it can be either transmissive or reflective.
...
> I should think it could be implemented as a plane of VERY TINY [clear or highly
> reflective] bars. I don't know whether to use square rods or triangle rods or
> circle rods.  I should think square would make a more plane-like reflective
> surface, while the circle rod would reduce the light, but work in any direction.
>   triangle rods would switch images depending upon point of view (they use larger
> ones in plastic kind of layer with an image on the back on kids toys).
>
> so maybe having a choice is a good thing.  with the diffraction grating, the
> rods are so small you cannot see them with the eye.  so working with floating
> point may or may not present a problem - so use exponents for the
> diameter/radius?

No, the only feasible approach would be to model the effects at the 
macroscopic level; they're quite easy to put into a formula though (and 
other people have done that already), and should also be pretty easy to 
integrate into POV-Ray.

Proper implementation should also allow to realistically simulate 
materials such as CD/DVD surfaces. However, this will require a 
mechanism to control the orientation of the grating, which may also 
require adding more pattern types; that would need to be evaluated.

Diffraction would also eat a lot of render time, comparable to 
dispersion (if you limit yourself to grates made of all-parallel 
elements) or worse (if you also want to do "criss-cross" grates with two 
or more optical axes).

Post a reply to this message