Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site cornell.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!houxm!vax135!cornell!jqj
From: jqj@cornell.UUCP (J Q Johnson)
Newsgroups: net.periphs,net.research,net.graphics
Subject: Re: volumetric displays
Message-ID: <816@cornell.UUCP>
Date: Mon, 30-Sep-85 09:28:28 EDT
Article-I.D.: cornell.816
Posted: Mon Sep 30 09:28:28 1985
Date-Received: Wed, 2-Oct-85 07:25:58 EDT
References: <2@unc.unc.UUCP> <486@olivee.UUCP> <394@bbncc5.UUCP> <306@bdaemon.UUCP>
Reply-To: jqj@cornell.UUCP (J Q Johnson)
Distribution: net
Organization: Cornell Univ. CS Dept.
Lines: 12
Xref: watmath net.periphs:875 net.research:241 net.graphics:1156

Although it might seem that a 3d display requires bandwidth o(n^3) for 
resolution n, in fact a great deal of data compression is possible.  We
don't actually see in 3d, but rather in 2-1/2 d or so (who cares about
the inside of a solid?).  Consider a representation that encodes only
range data at each point on a 2d display -- it requires bandwidth
o(n^2*log(n)).  Seems to me that most 3d images should be encodable
with a small constant multiple of the range data (e.g. as up to k
visible z values at each x,y position; by continuity, the k is not a
function of the resolution).

What representation actually achieves o(n^2*log(n)) for a 3d display?
Well, how about a hologram?