From: jpc@tauon.ph.unimelb.edu.au (John Costella)
Subject: Re: TECH: Holographic optics?
Date: Fri, 23 Apr 93 6:25:01 EET



> > >In Pimentel and Teixeira's book there's an interesting thing I never
> > >heard about before: they say it is possible to duplicate a complex set
> > >of optics on inexpensive, lightweight plastic film via an appropriate
> > >hologram (pg.109).
> 
> Someone else commented that if holographic optical elements (HOEs)
> were so easy, that everyone would be using them.
> 
> Well, they are not so easy, we are finding out the hard way by trying
> to develop a HMD using holographic optics.  While the start-up is
> hard, the benefits should be great:

A question here to the holographers that I hope is not too esoteric:

One of the messages we get here from the people in optical physics,
about holography, is an application of the `no free lunch' principle:
namely, if you encode a 3-d image on a 2-d film, then you're short a
dimension somewhere, so something has to suffer! With a hologram you
get a chunk of phase information (that gives the eye enough to infer
what the 3-d object looks like) but at the expense of fine detail,
i.e. you get speckle. (They were thinking of trying 3-d microscopy
here with holography, until they realised that if you look too
closely, the thing speckles away!)

The question then is: Do you lose anything with holographic *optics*,
and if so, what? I've always thought about real optical devices as
3-dimensional objects (for which `no free lunch' should apply), but it
is not clear that they may not in fact be two-dimensional (for which
lunch may be paid for). There are of course plenty of simple optical
elements that are obviously 2-d, but then you could imagine
complicating the optical transfer function up so badly that it
couldn't possibly be 2-d. The question then is what you must lose (and
how badly) for optics that you might want to put into a real-life HMD.

One reason this might be nasty is that, with HMD's, you would really
like the film to be small too! But this then implies a limit on the
amount of information you can pack onto the film, which is smaller
than desktop setups, for a given grade of film. If the optics were
essentially 2-d, then this would be a simple scaling (not too bad),
but if the optics were fundamentally of higher dimension, then you
might even get information-theoretical limits clamping down on you on
dimensional grounds alone, regardless of how cleverly you build the
system, i.e. the relative loss of information would scale as some
higher power of the scale change.

Apologies again if this sounds abstruse to those who are not
holographers, and childishly idiotic to those who are :-).

John

----------------------------------------------------------------------------
John P. Costella              School of Physics, The University of Melbourne
jpc@tauon.ph.unimelb.edu.au          Tel: +61 3543-7795;  Fax: +61 3347-4783
----------------------------------------------------------------------------