From: draperjv@ornl.gov (John V. Draper)
Subject: Re: DESIGN: Definitions for Virtual Reality Systems
Date: December 28, 1995


Tony,

A worthwhile endeavor and an excellent start. Some comments follow.


At 02:43 PM 12/27/95 +1100, Tony Healy wrote:
>
>Definitions for Virtual Reality Systems
>=======================================

>I distinguish three parts of a system - how we view something, what we are
>viewing and how the resulting 2D imagery is presented to us. I seek to
>include related systems, such as film, into this system of definitions to
>help differentiate between them when required.

At some point, the fourth part needs to be added: how we act in the
synthetic environment.

>Mirror World:
>
>A representation of an actual world, such as a patient undergoing surgery at
>a remote location, presented in a telepresence system for the user to take
>actions in. The mirror world will be slaved to the actual world, so that
>actions the user takes in the mirror world will be executed in the actual
>world. Mirror worlds include the worlds in telepresence systems only.

Would this also include a 3D animation of a real world, constructed
from integrating world models with real-time sensing? The previous
definition seemed to give a sense of exclusivity vis a vis graphic
animations. (On second reading, I don't think that the definitions
contain this exclusivity; I leave the question in to see what you
think about this.)

>Depth simulation using perspective views:
>
>Creating the illusion of depth by correctly incorporating perspective in the
>generated views of a scene. Most virtual reality systems do this.

I think the use of "perspective" is too limiting. There are many
mono-image cues to depth, including perspective, interposition,
surface texture, etc.  Perhaps "Depth simulation using single views"
or "Depth simulation using mono-image views" would serve better.

>Depth simulation using stereoscopic viewing:
>
>Creating the illusion of depth by correctly generating the different views
>from the left and right eyes. Most virtual reality systems do this if they
>are designed to be viewed using headsets. However if they weren't, then some
>headsets will display the same image to the left and right eyes.

I think "correctly" is too limiting because it implies matching
inter-camera separation in the virtural world to the inter-ocular
separation of the user.  Hyperstereopsis may sometimes be used to
enhance distance estimations, for example.

>Stereoscopic viewing using single screen:
>
>Presenting left and right eye views by having the user wear glasses with
>different filters in each eye, or by rapidly alternating views on the single
>screen. This is not considered a powerful way of presenting stereoscopic
>views, but needs to be defined to distinguish it from conventional
>stereoscopic viewing.

In what sense is this method less powerful than headsets? In what
sense is it different?

>Stereoscopic viewing using dual screens:
>
>Presenting left and right eye views, one each to a screen. This is the way
>stereoscopic views are normally presented. Normally the two screens are
>eyepieces in a headset.

I think the references to what is "normal" imply a judgement. Perhaps
you could say "frequently" instead.

John Draper
Oak Ridge National Laboratory