SCI: Independent on Sunday report on VR
J Hill
Independent on Sunday, accompanied by a 6"x9" photo of someone using a
VIRTUALITY hmd.
"Sega game could cause eye damage - By Steve Connor and Susan Watts
A new toy that allows children to play computer video games in
'virtual reality' could permanently damage their eyesight.
The equipment - a headset which beams stereoscopic images on to both
eyes - is already in use in such hi-tech amusement arcades as London's
Trocadero. Sega, the Japanese computer games company, intends to
launch a home version in the United States later this year, and in
Europe next year.
Tests of virtual reality headsets on adults produced visual problems
which scientists believe could be far more serious in young children.
One fear is that the toys could lead to permanent squints.
Two groups of researchers, one at Edinburgh University and one in the
Ministry of Defense, have detected side-effects in adult volunteers
who wear the 'head-mounted displays', which produce an illusion of
reality by giving all-round, three-dimensional vision of moving
objects. Such equipment is already used by the military and by
commercial designers who want to see their work in three dimensions.
The Health and Safety Executive has set up an investigation of the
headsets because of fears raised by a research team led by John Wann,
a lecturer in human perception at Edinburgh. 'Our results suggests it
seems particularly unwise to introduce them as a toy for children,' Dr
Wann said. 'If they are spending more than a few minutes with these
headsets, there are serious considerations for their eyesight.'
Mark Mon-Williams, an optometrist, said that people who used the
headsets for 10 minutes showed similar visual disturbances to those
who spend eight hours at a computer screen. 'It's amazing what you are
asking your eyes to do inside the headset,' he said.
Of 20 young adults who took part in a 10-minutes test, 12 experienced
side-effects such as headaches, nausea and blurred vision. Mr
Mon-Williams said that a particular concern is that the headset puts a
lot of strain on binocular vision, which is fully developed in adults
but is more liable to break down under stress in children under 12
years, causing squints.
The Edinburgh findings are supported by researchers at the Army
Personnel Research Establishment at Farnborough, Hampshire. In a test,
61 per cent of 150 volunteers reported symptoms such as dizziness,
headaches, eyestrain, light-headedness and severe nausea.
Mr Mon-Williams said the main problem stems from the headsets severely
straining the eye muscles, leading to slightly cross-eyed vision. A
slight misalignment of the two images in each eye produces a visual
disparity that the muscles try to correct.
Andrew Wright, software product manager for Sega in Britain, said that
the new product would be tested extensively before coming on the
market.
Other health problems associated with virtual reality are beginning to
emerge: a form of travel sickness is affecting people who spend too
long in virtual environments. Symptoms such as nausea and
disorientation are brought on by the slight time-lag between people
moving their head and the scene they are immersed in 'catching up'."
_______________________________
I would say that Sega and the rest of the VR games manufacturers have
some PR work to do. I will watch the Independent for any responses to
this article.
J. Hill
Note: My affiliations with the University of Edinburgh are totally
unrelated to those of the research team cited in the article.
D Nicoll
responsible for much of the hype of VR in the Independent over the
last couple of years. It might not be so much the PR lacking from the
companies developing these products, as the sensation seeking media
again showing lack of proportion in reporting on scientific and
technical acheivements (remember cold fusion).
Derek Nicoll (also not a member of that research team)
Robert Jacobson
between the effects of virtual environments generally and the specific
consequences associated with particular pieces of gear.
Bob Jacobson
John Wann
Bob,
Don't you think the real problem is the number of VR "experts"
who play around at developments, make comments and fiercely
protect the VR industry, but don't seem to get in there to collect
and publish hard empirical data on how the human perceptual
motor system responds to VR systems ?
We research, publish and disseminate in that order. Our original
research was on the VPL LX with fresnel lens system, with no
individual tuning of IOD, - a very crude system. We then
introduced calibration procedures that are very sensitive to
individual visual states. Then cross checked our findings on an
HMD with LEEP optics and finally with desktop stereo-displays.
Across all these experimental situations we found related
problems that are cannot be account for by fresnel lenses or
tracker lags any other model specific feature. We do have a
general hypothesis pertaining to these problems, but we go
through the critical academic peer review process first and this
isn't ready for release yet. Some indication of our proposals and
suggestions for HMD modifications are forthcoming in CyberEdge
Journal.
But lets be candid, most of those involved in VR research know
there's a number of problems with current HMD's (see Bob Stone's
report to the UK Industry Conference in Coventry 1992; or
Robinett & Rollands "Presence" comments about "decoupling
accommodation and convergence"). We have been banging on a
number of HMD manufacturers doors, trying to alert them about
what they need to consider (e.g as outlined in CyberEdge Journal)
for the last 18 months without any success.
So who's going to stand up and be counted when there's the
imminent release of a home-HMD, with the potential for
unrestricted use ? Is there a single reader of this bulletin board
who seriously believes it would be wise for an 8yr old to sit at
home playing a video game for an hour or so with something
typical of our current developments in HMDs ? And if so what
possible grounds or data do they have for making this assumption
? My particular specialism is perceptual development and our
research team also contains specialists in child optometry and we
don't know of any evidence that would get you close to this
conclusion. I'm sure that if reports started coming in from
optometrists about "VR-overuse-syndrome" there would be
plenty of "experts" to say "hmmm, yes there are some device
specific problems", but is that a responsible way for VR to
progress ?
So when are all the people sitting reading bulletin boards, and
taking odd pot shots, going to get off the fence and either support
our findings or vigorously attack them with hard empirical data
of their own. I don't mind either way, but if the current UK
media fluster shakes the VR community out of their complacency
about health and safety guidelines, which are woefully lacking,
then I have no regrets.
John Wann (at the eye of the storm)
Edinburgh VR project
B...@advanced-robotics-research-centre.salford.ac.uk
Isn't it amazing! The Independent must have been sadly short of news
to publish this! At 2 UK conferences after the event, the atmosphere
amongst the practitioners and delegates was one of "so what?!" This
is a topic which has featured at all VR conferences (including those
attended by journalists from the Independent) since Computer Graphics
'90 and VR '91 in London! A few notes:
(1) Wann et al did not prompt the UK's Health & Safety Executive to
set the VR study up. We did! Colin Mackay of the HSE and I were
involved 2 years ago in a radio interview on this subject, and have
been talking ever since. Sadly ARRL did not win the contract, but I
was under the impression that it was awarded to Nottingham University
(at least I hope it was; they have an excellent reputation for
thorough human factors research).
(2) The studies referred to have been criticised at a number of
European VR functions, mainly on the basis that the experimental
designs were not up to scratch and that scant attention had been paid
to the myriad of factors contributing to "Cybersickness". Although
everyone blames the HMD, there are other cognitive, system, task-
related features which come into play and must be taken into account.
Designing a thorough and well-controlled VR study is difficult but,
given access to the right range of equipment, quite possible.
(3) For information, the Edinburgh study was, I believe, based on a
Real World Graphics board-based VR system with a VPL EyePhone LX, the
MOD Study (carried out by Clare Regan at the Army Personnel Research
Establishment in Farnborough whilst they were waiting for funds to
come through for their main line of work) used a Division ProVision
System with, I think, a Flight Helmet.
(4) The Independent is one of the UK's best-known "over-glamifiers" of
VR - front page status for these studies surprised us all!
(5) There is little doubt that there are problems, but these studies
have only considered the tip of the iceberg and may be extrapolating
too far on the basis of quite specific results. I only hope that some
of the proposals we put into our 102-page proposal to the HSE get
taken on-board. Whilst I don't want to appear jaundiced or blow our
trumpet too heavily, unless attention is given to the experimental
design, the virtual environment used, the selection and even invention
of appropriate objective and subjective measures, then a lot of the
work will be scientifically meaningless.
Prof. Bob Stone
Technical Manager (Advanced Robotics Research Limited)
Robert Jacobson
John Wann <jo...@HMS01.HMS.UQ.OZ.AU> wrote (in response to my
remarks about differentiating between systems and virtual worlds):
>Don't you think the real problem is the number of VR "experts"
>who play around at developments, make comments and fiercely
>protect the VR industry, but don't seem to get in there to collect
>and publish hard empirical data on how the human perceptual
>motor system responds to VR systems ?
>
>there's a number of problems with current HMD's (see Bob Stone's
>report to the UK Industry Conference in Coventry 1992; or
>Robinett & Rollands "Presence" comments about "decoupling
>accommodation and convergence"). We have been banging on a
>number of HMD manufacturers doors, trying to alert them about
>what they need to consider (e.g as outlined in CyberEdge Journal)
>for the last 18 months without any success.
>
>
>John Wann (at the eye of the storm)
>Edinburgh VR project
John, for a long time I have not identified virtual worlds with HMDs,
anymore than I would identify computing with a VGA terminal. The two
are not synonymous. Our firm works mainly with projection technology,
for practical reasons having to do with group experiences. It is
possible to create persuasive virtual worlds without HMDs. (Some
would argue, more persuasive, given the relative maturity of projec-
tion technology versus HMD technology.) But the press makes no
distinction between the experience and the particular technology of
which you are critical. I am sure you made efforts to distinguish
between the two; unfortunately, in the retelling, these distinctions
vanish.
I appreciate the good you may be doing for young children whose
perceptual skills are at risk. But the overall damage, even if
slight, that may be done to the field at large by questionable
reporting is regrettable.
Bob Jacobson
(Running from the storm)
Benjamin...@uc.edu
ago. it seemed like both those studies used stereographic (two
different displays with slightly different angles of view,
representing left and right eye) in a head mounted display.
Monographic (one display, sent to both eyes) seemed like they weren't
tested in either study. is this correct? are there differences in
level of eye strain? it would seem that the mechanical structure of
the hmd and the positions of the displays varying (or not varying) for
each person's set of eyes would also greatly affect level of strain.
perhaps we should have prescription hmd's, just like NASA! Cool!
seriously, what about monographic displays, same problem, different,
no problem, or not tested?
John Wann
issue as I have no intention in being drawn into an electronic
argument with Bob Stone or anyone else. - But as Bob's mail crossed
over with mine, I have the following comments on his attack:
John Wann (Edinburgh VR project)
-------------------------------------
Bob (Stone) posted:
> Isn't it amazing! The Independent must have been sadly short
>of news to publish this! At 2 UK conferences after the event, the
>atmosphere amongst the practitioners and delegates was one of
>"so what?!"
>
The "so what ?" is exactly the attitude I take issue with:
"everyone" knows there are problems with HMD's yet people are
still putting them on visitors/subjects with no agreed guidelines
or precautions, and when a home-use HMD of unknown spec.
was/is due to be released in the new year, where was "everyone"
who should have been saying "Hey, hold on a minute..."
Bob states:
> (2) The studies referred to have been criticised at a number of
>European VR functions, mainly on the basis that the
>experimental designs were not up to scratch and that scant
>attention had been paid to the myriad of factors contributing to
>"Cybersickness".
>
Our studies say nothing about "Cybersickness" or the "myriad"
factors involved, we were merely concerned about visual
function. It is galling to have "researchers" like Stone repeatedly
maligning our results without ever stating what he thinks are the
problems or what should be done. These findings have been
reviewed and published in a leading vision science journal. Is it
that Stone thinks our test of fixation disparity was inappropriate
or that there are other measures of heterophoria than the ones
we used ? I didn't realise that Stone was qualified to make these
sort of judgements. Interestingly Sony's H&S testing on their
headset used similar tests to our own [1].
Bob goes on to state:
>
>unless attention is given to the experimental design, the virtual
>environment used, the selection and even invention of
>appropriate....measures, then a lot...will be scientifically
>meaningless.
>
I agree, although the "invention" of measures is unneccessary,
there is a huge history of appropriate research in prismatic
adaption, visual response to disparity etc..(see for instance
Experimental Brain Research 1950-93) with well developed
measures. We borrow from and build on these, some recent
proposals from Stone/ARRL suggest they are unaware of much of
this work [2].
Why don't centres like ARRL (Stone) collect and publish similar
data to ours to show we are wrong ? Stone (1992) has previously
listed HMD problems in a conference paper but they are of the
form:
"experienced by the author and some visitors"
"an experience reported by .....a husband, wife and son"
This such anecdotal evidence is useless. Of course if the
Independent had picked it up the message would have become
even more distorted than ours did.
Centres like the ARRL and commercial VR groups promote
systems to industry but don't get in and publish the basic
research that the science needs [3]. Sure our research isn't
perfect, but I think its as good as anything else around, and there
isn't much. It is also POSITIVE, not negative. We identified,
specific problems related to the neural control of accommodation
and ocular vergence and as a result propose design modifications
to avoid these [4] (see also forthcoming CyberEdge Journal). This
is what the Science of VR, needs - data, hypotheses, replications,
disagreements & heated debate, the science will not progress, by
people sitting around saying:
"Oh, yes the present HMD's make you 'legally blind' (sic) .... but
next year the cybercorp are releasing a fantastic HMD"
Science like nature should be red in tooth and claw. So why
doesn't Bob stop throwing Stone's and conduct controlled studies,
with suitable no.s of subjects, standardised tests and inferential
statitistics, and publish this work in a refereed vision science
journal. Then he will have a huge boulder of data to drop our
heads and I will be genuinely delighted to receive such
de/con-structive criticism.
Thank you for reading this far
______________________
Footnotes:
[1] Sony observed "tendencies" in the same direction (after 2
hours), but nothing like the results we obtained. Their testing,
however, used a non-disparate (mono-scopic) image which
circumvents all the problems we have highlighted. I was also
extremely pleased to see that Sony state "We are hoping to
establish some display usage standards......only after that will
Sony turn to the home market". - A pat on the back for Sony.
[2] In Robotica (1992), for instance, Stone proposes the use of
remote camera adjustments in stereo-HMD tele-operation, that
clearly conflict with the current understanding of human 3-D
space perception.
[3] These comments do not apply to centres such as Chapel Hill
and Nasa-Ames where Robbinett, Ellis etc.. clearly make a
substantial contribution to the science of VR rather than just to
the "industry".
[4] Just to re-post some details: Our original research was on the
VPL LX with fresnel lens system, with no individual tuning of IOD,
- a very crude system. We then introduced calibration
procedures that are very sensitive to individual visual status.
Then cross checked our findings on an HMD with LEEP optics and
finally with desktop stereo-displays. Across all these
experimental situations we found related problems that are
cannot be account for by fresnel lenses or tracker lags any other
model specific feature. These were measured with standardised
and accepted optometric measures of visual function. In all cases
the VR environment was one in which a large range of disparities
occurred giving percepts of objects in near and far space. - Herein
the problem lies, due to the nature of the neural control of
accommodation and convergence. We are presently working on
display procedures that will reduce these problems and these are
sketched out in a forthcoming CyberEdge Journal. This is not our
main area of interest, however, we merely started this testing for
our own H&S concerns, hence we didn't submit a tender for the
UK HSE contract. We will be delighted if someone else takes on
this research and tells us we can use an HMD with our subjects for
more than 10 minutes !
Chris Shaw
>tested in either study. is this correct? are there differences in
>level of eye strain?
Well, if you actually mean that the identical image is sent to both
screens of a HMD that is set up for stereopsis, you're just going to
see a double image. When we first got our VPL EyePhone 1 about 3-4
years ago, I tried this for about 5 minutes and immediately suffered a
major headache.
What you really mean is biocular display (no "n"): The hardware is set
up so that there are two views of the same image. The effect is the
same as looking at a photograph nailed to a stick projecting from your
forehead. In this configuration, there is no stereo disparity, because
both eyes are getting the same information. In this situation, there
is nominally NO conflict between accomodation and convergence, because
the eyes are always converged on the single image plane, and
accomodation is also set at that depth. The 3D graphics projection for
biocular simply uses the same centre of projection for both eyes, and
adjusts the angle between the two lines of sight so as to fuse the two
images. You must be careful to use the same projection plane for both
eyes. In essence, each eye gets a subwindow of a somewhat larger
single image.
Binocular uses two independent centres of projection with lines of
sight parallel and separated by inter-ocular distance. Conflict
between convergence and accomodation arises because accomodation has
to be at the focal distance of the displays, and convergence can occur
at many different depths.
--
Chris Shaw University of Alberta
cds...@cs.UAlberta.ca CatchPhrase: Bogus as HELL !