From: rych@hagg.psy.ed.ac.uk (Rycharde Hawkes)
Subject: Re: TECH: Networked VR under 100ms?
Date: 10 Sep 1996 13:48:49 GMT
Message-ID: <513rk1$183@scotsman.ed.ac.uk>
Organization: Edinburgh University



Jason Leigh (jleigh@bert.eecs.uic.edu) wrote:
: Looking through a number of papers on networked VR (Macedonia for eg)
: I find the recurring quote that says the minimum latency for networking
: in VR should be less than 100ms. Of course no body explains why or from
: what results they gathered this statistic. 
: Does anyone have a definitive reference with an explanation.

I think this figure comes from the DIS spec.  The latency that can be
tolerated within a VR system is dependent upon the techniques used to
manage simulation computation and data and, of course, the target
application.  For example, an "interactive" visualisation may be able
to tolerate greater latency than a flight simulator.

DIS uses an approximation model (dead-reckoning) to guestimate the
position, etc. of remotely simulated entities.  I suspect the 100ms
has more to do with this than human perception.

Rych

rych@hagg.psy.ed.ac.uk (Rycharde Hawkes)