From: mark@cs.ualberta.ca (Mark Green)
Subject: Re: REV-PROD: VirtualIO iglasses!
Date: 	1 Sep 1995 05:47:00 GMT
Message-ID: <4266ok$gnb@scapa.cs.ualberta.ca>
Organization: Computing Science, U of Alberta, Edmonton, Canada


holupka@keith.dfci.harvard.edu (Edward J. Holupka) writes:

>	I am in the process of putting together a project for VR in
>medicine and I am looking for a suitable HMD. I have looked at
>vitrual-iO 's i-glasses! and they have faxed me some info on them.
>However, I get the impression that the glasses are really "toys" and
>not suitable for developement for a serious project. 

You seem to be really confused about what's going on.  We have been
using the i-glasses for a few weeks now.  We are currently driving the
video from an SGI workstation, and are in the process of producing a
Unix driver for the tracker.  This will be part of the MR Toolkit v2.0
release that will occur some time this fall.  For more information on
MR, see our Web page:

http://www.cs.ualberta.ca/~graphics/MRToolkit.html

  We have been impressed with the overall quality of the i-glasses.
It is one of the best HMD's in the price range (under $5000).  I will
try to clarify some of the details.

>The project I am working on will involve a large amount of image processing 
>and I am concerned about the iglasses IO interface. They use the serial 
>port of the PC. Wouldn't connecting to the ISA/EISA bus be alot faster?  

The output to the i-glasses takes one of two forms.  It can either be
field sequential NTSC video, or on the PC version you can also have a
VGA interface.  We have only worked with the field sequential NTSC
version, since we are using SGI workstations, and not PC's.

>I didn't get the impression that one could develope actual virtual
>worlds on them. Is it indeed possible to view 3D perspective images in
>the HDM? For example, each monitor of the HDM would have a slightly
>different image dependent on the value of interoccular distance. In

This has far more to do with the software being used than the actual
HMD being used.  The software generates the views for the two eye, and
takes into account the IOD, this can't be done in the HMD itself.

>addition, passing information about the head position (tracking) would
>also have to go through the serial port. Given the complexity of
>creating and interactively viewing a virtual object I find it
>difficult to believe that all this can be handled through the serial
>port.

Almost every tracker that I know of uses a serial interface, so this
shouldn't be a problem.

Dr. Mark Green                                 mark@cs.ualberta.ca
Professor and Associate Chair                  (403) 492-4584
Department of Computing Science                (403) 492-1071 (FAX)
University of Alberta, Edmonton, Alberta, T6G 2H1, Canada