Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!husc6!mit-eddie!mit-amt!halazar
From: halazar@mit-amt.MEDIA.MIT.EDU (Michael Halle)
Newsgroups: comp.graphics
Subject: Re: 3d Computer Generated Holography
Message-ID: <629@mit-amt.MEDIA.MIT.EDU>
Date: 30 Aug 89 02:51:04 GMT
References: <441@ctycal.UUCP> <46900037@uxe.cso.uiuc.edu>
Organization: MIT Media Lab, Cambridge, MA
Lines: 38
In-reply-to: mcdonald@uxe.cso.uiuc.edu's message of 29 Aug 89 16:18:00 GMT


If the holograms in your museum are white-light viewable transmission
holograms, they almost definitely have no vertical parallax.  That's
because (see my earlier posting) different vertical perspectives would
blur together in different colors, creating an unsightly mess.  (A
full parallax white light transmission hologram is called a full
aperture hologram.  The horizontal parallax only (HPO) transmission
hologram is the Benton or rainbow hologram.)  Try tipping your head
sideways and see if the image goes flat.  White light reflection
holograms are another story; they can retain vertical parallax and not
blur out (too much) in white light.

There are many, many ways to ruin a hologram, especially a computer
generated stereogram.  Eliminating vertical parallax has not, in our
experience, proven to be one of them.  Really, most people don't
notice.  Some of these ways do, however, include: moving the object
instead of the viewer (light sources don't look right); rotating the
object (object swings around, exhibits keystone distortion); not
matching rendering to holographic setup (assorted terrible things);
and not carefully registering the frame sequence (jumpy image).  And
there are plenty of publicly displayed examples of them all...that's
in part because many people don't know or can't do the process exactly
the right way.  That doesn't mean their images don't "sort-of" work,
or that the entire idea is flawed.

I don't know of anyone who has done studies on the effect of
resolution in synthetic holography. We can't really compete with the
full bandwidth of a true hologram, so a computer image is bound to be
quantitatively inferior (in information content) to the "real thing".
Is it good enough for real-life people?  We'll see.  Also, field of
view is usually limited to about 30 degrees for the most common
stereograms.  We've got work in the pipeline that pushes that past 90
degrees.  The images are extremely good, quite convincing even with
medium resolution (640x480) input images. And no vertical parallax.

						--Michael Halle
						  Spatial Imaging Group
						  MIT Media Laboratory