Re: Ortho Cameras

Jan Hardenbergh (jch@nell.oki.com)
Fri, 02 Dec 94 13:36:00 E


> that just leaves two ways of
>specifying the 'zoom' for the camera:
>
>-- width and/or height at the "screen" (focalDistance, to me) plane
>-- horizontal and/or vertical field of view
>
>It is trivial to convert between these; a VRML implementation could
>certainly store the information in the camera either way. So the
>question is really which is easier for a person writing VRML files to
>understand.

A note on terminology: I beleive one of the above is a "dolly" which is what
happens when you roll the camera forward on a dolly (thing with wheels).
The other is a zoom. Of course, even is I'm wrong about this example (I'm
still learning the GL way) the terminology is useful.

As for how to specify cameras - PHIGS used 2 4x4s (orientation & mapping)
and
a clipping volume. There were utilities to create the matrices. This meant
that
you had to guess the eye-position if you needed to do specularity.

I believe that most mortals will find the "look at" model easier than the
PHIGS
model. In PEXlib, we just provided both. So, adding another camera type is
possible. (I don't really care, I just want people to use "dolly" when they
mean it)

-Jan
P.S. is anyone else bugged by the anonynimity of this list? Or should I just
get
a real mail reader? (all mail comes from vrml-list-owner)(www-buyinfo works)