P3D Re: Keystoning - mea culpa

[Gabriel Jacob <jacob@xxxxxxxxxxxxxx>]

>>Boris uses more toe in than that, and when challenged we all found it
>>difficult to see any keystoning reliably.  It could still be OK.

This is not strictly true. One person correctly identified all the
images. I had noticed it in quite a few of the images (and I imagine
others might have as well) BUT Boris (the trickster ;-) ) said only
half (or so I recall) were toed-in while the other half were not.
He challenged us to identfiy those. I found most looked like they
were, but since he said only half, I, being the trusting soul that
I am, ;-) had a difficult time deciding which half to include. I've
seen quite a few images from others and the toeing-in is quite
noticable.

As LeRoy Barco mentions another penalty is too much deviation. I've
seen this plenty of times! The subject of interest looks okay but the
background has toooo much deviation. I've often wondered if another
work around solution to this (have your cake and eat it too!), is
instead of converging the cameras at the film plane, to instead
pivot the lenses plane. I forget the exact plane to pivot but it
would be the same pivot plane as when they shoot panoramic pictures.
There is a plane where one can pivot and the perspective remains
unchanged. I wonder if that would help. Boris? Anyone! :-)

I'm constantly surprised that some people would rather toe-in
(when they do it too much) than loose some of the sides (of the
image). The way I see it, with too much toe-in, the whole image
is lost! Sure, there are times when you would loose too much image
and is unavoidable to toe-in but I've seen enough toe-in images
where the lesser of toe (I mean two) evils was loosing some
seeming precious film sides. And if you can't mask the double 
edges (don't have the appropriate masks or whatever), I'd 
still rather see that, than toe-in effects.

>It could still be OK.  Certainly worth trying it, and see what the
>results are like.

For sure! I have some scenic shots taken with 28mm lenses and
inadvertanly converged (toed-in) that show very little keystone
distortion (as for other distortions due to the mismatching of the 
lenses, that is another story!). The key is, to know how much to
converge the cameras! As Dr.T mentions, Ferwerda's book is an
excellent source for more information.

P.S. It's interesting to note, that the philosphy in 3-D movies
is to converge the cameras (I'm not sure about Imax though). The
effects are supposedly minimal, which might be true for movies but
not for still photography. 

Gabriel (don't look at me cross-eyed) Jacob, with flame shield on! ;-)


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