On Mon, 05 Jun 2000 18:11:48 +0200, Gilles Tran <tra### [at] inapginrafr>
wrote:

>My problem with the Cornell box is that even the real-life one looks
>rendered !

Yeah, they used a "laboratory-style" CCD camera and went to a lot of
trouble to capture the image the way the *camera* saw the scene. If
this had been a more typical photograph, those white walls would have
almost certainly been color-corrected to look white or gray (depending
on how much light was actually hitting any particular spot of the
wall.) In other words, that yellow color cast would have been
neutralized. The red and green walls would probably have more
saturation also. If you viewed the physical Cornell Box with your own
eyes, it might also look much more neutral in color, because of the
tendency of our vision system to automatically adjust color balance
for us.

I think in a perfect world, we should have more than one physical
"reference scene." Each one should be something that could be
constructed by other people using fairly common materials, following a
well documented "blueprint." A modified version of the Cornell Box
could be one, but there should be others as well. When testing optical
devices, such as cameras for instance, it isn't enough to simply test
with one particular test-scene/target. Several seperate tests are
usually ran to get a more complete analysis of the device.

Later,
Glen Berry

7no### [at] ezwvcom
(Remove the "7" to reply via email.)

Post a reply to this message