Mailinglist Archives:
Infrared
Panorama
Photo-3D
Tech-3D
Sell-3D
MF3D

Notice
This mailinglist archive is frozen since May 2001, i.e. it will stay online but will not be updated.
<-- Date Index --> <-- Thread Index --> [Author Index]

P3D Re: Instant Anaglyphs


  • From: aifxtony@xxxxxxxxxxxxx (Tony Alderson)
  • Subject: P3D Re: Instant Anaglyphs
  • Date: Wed, 12 Aug 1998 15:23:01 -0700

>Dylan The Hippy Wabbit wrote (digest 2892):
I disagree with Tom Deering about that anaglyph thing. If you allow the
images to overlap, which the example on his web site did *not* do, you
would have an anaglyph that worked, but not very well. <

Yes, but with this proposed "instant anaglyph pinhole camera" the images
WON'T overlap properly--the overlap will be at "infinity" (the spacing of
the pinholes). Of course, reducing the spacing of the pinholes to "improve"
the overlap will eventually eliminate the stereo.

>Firstly I don't care about the red and blue ball thing. That's always
>going to happen with a "full colour" anaglyph, and it's one reason I don't
>think much of them.<

Well, Tom has partly missed the point--color anaglyphs ARE a color
separation gag; that's how they work, but also why they don't work very
consistently.

>Perhaps a prism behind each pinhole to move the images closer together
>would fix it. But then once you've started putting glass in you may as
>well use lenses.<

Well, exactly! But separate lenses will have the same problem two pinholes
have in aligning the anaglyph: the convergence won't ever be right...unless
you design a system with some method to converge the two channels. Then
you're getting into such engineering complexity you've defeated the whole
initial concept. You could test this with a slide bar set up: Double expose
the L & R images onto the same frame, thru appropriate Red and Cyan
filters. You will have to toe-in to get the images to overlap properly,
unless you reduce your interaxial to zero.

Using an anaglyph filter in the aperture of a single lens (such as the QDOS
lens, or the Video West (?...depending on memory here, my copy of the
American Cinematographer issue in question is at home) is somewhat
different.  The common lens elements will converge the optical paths. The
thing you have to remember is the effective "interaxial" in such a system
is the width of the aperture; which means you need to be pretty wide open
(at least f5.6) to get any stereo effect. This also means your DOF must be
pretty narrow...forget having everything in focus. This is one case where
the stereo base is literally dependent on f-stop. Also, in my opinion, such
lenses cardboard noticeably, as the effective interaxial is pretty small
for the focal length and taking distance. But they have some applications,
I guess.

Tony Alderson
aifxtony@xxxxxxxxxxxxx



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