From: G. Jounghyun Kim [gkim@postech.ac.kr]
Sent: Sunday, November 07, 1999 7:23 PM
Cc: 3dui List
Subject: Re: exo- vs. ego-centric tasks

Dear Doug,

We have been developing a VR based "motion" training system using a first
person viewpoint (using an HMD).
We have actually carried out a usability test on the effect of first person
viewpoint on learning motion profiles
and compared its performance against learning them from 3rd person
viewpoint.  Even though there were few
glitches in the way we conducted our experiment (and so we are redoing the
whole thing right now), the
preliminary results tell us that there is a significant performance
difference in learning motion profiles between the
two methods (first person viewpoint outperformed the other).  Please refer
to our on-line paper, if interested.

http://www.postech.ac.kr/~gkim/vsmm99-jfm//jfm.htm

Regards,

Gerry Kim.

Doug Bowman wrote:

> Greetings all:
>
> As you may remember from previous postings, I'm thinking about
> doing some comparative studies of different VE display types.
> One of the differences that seems pretty intuitively clear is
> the distinction between an egocentric, 1st person, point of view
> (HMDs, CAVE), and an exocentric, 3rd person, point of view
> (Workbench).
>
> Of course, you can do an exocentric view in an HMD and an
> egocentric view on the workbench, but that seems to go against
> the "natural" use of these displays. So, I'd like to test both
> types of tasks using both types of displays, and try to show
> that HMDs and CAVEs excel when the user is doing an egocentric
> task from an egocentric point of view, and that workbenches excel
> when a user is doing an exocentric task from an exocentric point
> of view.
>
> The problem is finding *tasks* that are by nature exo- or egocentric.
> In fact, for almost every task I can think of ways to accomplish
> it from a first- or third-person point of view. Take driving: it
> seems clearly egocentric, but some people can develop highly accurate
> driving skills for the task of steering a vehicle while looking down
> on it from above (remember the old video game "Tank"?). The same is
> true of a task like surgery, which seems exocentric, but you could
> argue that certain delicate procedures could be done even better if
> I could see through the "eyes" of the instrument instead of my own.
>
> Rudy Darken talked about this in his VR '99 paper with relation
> to the usage of maps in virtual environments. He claimed that
> targeted search (the target is shown on the map) is best performed
> from an egocentric reference frame and primed or naive searches
> are best performed from a world reference frame. However, this is a
> bit different from doing these tasks from an ego- or exocentric
> *point of view*.
>
> So, two questions:
> 1. Are tasks inherently exocentric or egocentric?
> 2. If so, what are some examples of each?
>
> My current list for #2, which I'm not satisfied with:
> exocentric: puzzle assembly, global spatial relationships, surgery,...
> egocentric: line of sight tasks, driving, local spatial relationships...
>
> Looking forward to your comments.
>
> --
> Doug A. Bowman, Ph.D.          (540) 231-7537
> Assistant Professor             bowman@vt.edu
> Computer Science        www.cs.vt.edu/~bowman/
> Virginia Tech