P3D Re: wet toes winner.

[roberts@xxxxxxxxxxxxxxxxx (John W Roberts)]

>Date: Sat, 21 Feb 1998 20:54:44 -0700
>From: boris@xxxxxxxxxxxx (Boris Starosta)
>Subject: P3D wet toes winner.

>The correct answer was that ALL OF THE SHOWCASES, except the computer
>generated digital work, were shot toed-in. 

>...The architecture images were shot with reduced
>convergence (about 2 degrees), placing the stereo window (i.e. the
>intersection of the optical axes) at a distance of about seven feet.  For
>that reason, keystoning is all but undetectable in those images.  

That reminds me of an alleged "lie detector" I recently saw in a store.
I suggested that the reliability of the device be tested by stating
"3 + 3 = 6.001" - because that's "very nearly true", it could help to
quantify the lie detector's ability to differentiate between truth and
falsehood. Different numbers could be introduced - for example, all but
the most discriminating lie detectors ought to pass "3 + 3 = 6.00000001"
as true.

I might as well admit that in some of my Grand Canyon hyperstereos, shooting
with 50-100 yard interocular and with the alignment reference 15-20 miles
away, I have occasionally used the *same* reference point for both photos.
This makes the stereo pairs technically toed-in, but so far nobody has
noticed.

A note of caution: cosmological measurements (COBE, HIPPARCOS, etc.) have
not yet shown any definitive deviation, but it's still possible that the
universe we live in is non-Euclidean, in which case it could be possible for
two lines to intersect and still be parallel. So it's at least theoretically
possible that some of your "toed-in" shots are in fact also parallel.

John R


------------------------------