From: kn008@rrz.uni-koeln.de (Dipl.-Ing. Olaf Kelle)
Subject: SCI: Some notes on J.Wann: What's wrong with your HMD
Organization: Regional Computing Center, University of Cologne

Dipl.-Ing. Olaf H. Kelle
University of Wuppertal, Germany
Department for Traffic Security and Simulation
EMail: kelle@wrcd1.urz.uni-wuppertal.de

Some notes on John Wann: What's wrong with your HMD
(CyberEdge, August 93)

John reports in his article about visual stress with the use of the
VPL LX (with Fresnel / LEEP optics lens). Over 50% of his subjects
showed nausea/ headache and tired eyes. My experience with the VPL
system supports that.

When I do my research on depth perception with and w/o motion parallax
and motion perspective with our selfengineered HMD (two CRT-viewfinder
(200 TV-lines vertical) with a high quality Zeiss optical system and a
optical HeadTracker) I have to ascertain that only 10% of my subjects
(from 60 people) reports about visual stress.  The average viewing
duration was about 45min each.

So why this difference ?
John fixes the problem to the depth of the virtual screen image (I
think the resolution of the used LCD will be also a problem).  His
solution is to set the depth depending on the visual tasks.  So his
proposal is to use near - mid - far virtual image depths.
My opinion is: He can do this but he don't have to do it.

Why ?
We think that one fixed virtual image depth will suffice.
We decide to install the depth plane at 3m.

This value has two reasons which can be calculated:

1. Accommodation: At the distance of 3m the accommodation
has no effect on visual stress. This is in line with many
psychologists who reported no effects from 2m on. The
importance of accommodation seems also not so powerful.

2. Stereoscopic Depth Zone: This term was first introduced by
the german optical researcher H. Luescher [1] (1944) who
called it "Stereoskopische Tiefenzone" which can translated to
"stereoscopic depth zone". Another german researcher O.
Vierling [2] (1965) (he works at Zeiss, Stuttgart) do some
practical extensions on the theory from Luescher. This
stereoscopic depth zone has an important effect on the
stereoscopic presentation of a scene. To avoid binocular fusion
problems (and so binocular stress) Vierling says that you have
to consider this depth zone. Luescher speaks about the
"Tiefenbedingung" which is similar to "depth condition" or as he
states the "70' (arc min) - condition". If you like to display a
specific depth range you should take a look on this 70'-
condition.
Example: Our HMD should be able to display a depth range to
infinity without binocular fusion problems. With the formulas
Vierling gives in his book the calculation is no problem.
Assuming an eye seperation of 65mm the result for the nearest
point (the virtual plane) is 3m. In the zone between 3m - infinity
there will be no fusion problems. Depending on other eye
seperation values and other maximal depth values you can
calculate different nearest points.

With the value of 3m for the depth of the virtual screen image I
can report only good experience, see the stress-percentage
above !

Conclusion:
I think it is not necessary to install different depth planes
depending on the visual task. With one fixed value you have
the ability to look in a range from 3m to infinity without any
binocular stress. Accommodation problems are also eliminated.
For the most tasks this will suffice.

I think this is the right interpretation, always with a look on the
different image resolutions at VPL and our system,  for the low-
level percentage of stressed subjects in my experiment.

Comments on this notes are always welcome !

Bibliography:

[1] Luescher, H.
Stereoskopische Tiefenzone - Akkomodation und
Konvergenz
Photo-Industrie und Handel, 1944

[2] Vierling, O.
Die Stereoskopie in der Photographie und Kinematographie
Wissenschaftl. Verlagsgesellschaft, Stuttgart, 1965

-- 
Dipl.-Ing. Olaf H. Kelle
University of Wuppertal, Germany
Depart. for Traffic Security and Simulation
E-Mail: kelle@wrcd1.urz.uni-wuppertal.de