"MichaelJF" <mi-### [at] t-onlinede> wrote:
> "Cousin Ricky" <rickysttATyahooDOTcom> wrote:
> > "Cousin Ricky" <rickysttATyahooDOTcom> wrote:
> > > Spectrally rendered.  Nightmarish render times.  Attached is an early draft.
> >
> > I must add that this was inspired by Richard Roseman's "Glass Panels":
> >
> >   http://www.cgsphere.com/gallery/details/?submission_id=85
>
> I know that there are restrictions to the background, but one has some freedom
> to design this grid. As all I understood so far, the grid must be recognisable.
> I think a white marble texture or something alike should improve the image
> without destroying your effort with the spectral rendering. But marble invites
> to subsurface scattering prolonging the rendering time to eternity...

According to the official rules, the grid must not be modified, not even the
texture.  The density and positioning of the grid lines are strictly enforced,
from what I have seen.  Obviously the rule not to modify the texture is
completely unenforced, but, to my everlasting self-imposed frustration, I tend
to be a compulsive rules follower.

In the face of strong temptation (akin to that of driving through a 10 mph
school zone at 20 mph, instead of 30 mph like everyone else), I looked at
Richard Roseman's submissions for guidance.  (Roseman is the one running the
project.)  The prima facie ruling is extremely conservative.  However, closer
examination of "Glass Panels" reveals that the grid lines are affected by
caustics.  This is not possible with the hard-wired texture in Trevor Quayle's
template adaptation.  I am taking conservative liberties accordingly.  (The grid
lines in the present WIP have a 2% diffuse reflectivity.)

I sure as hell will not use SSLT when I am dealing with dispersion through
multiple glass surfaces.

Post a reply to this message