Virtual Reality in Art, Design and the Performing Arts

wafflejourneyΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 4 χρόνια)

180 εμφανίσεις

1

Virtual Reality in Art, Design and the Performing Arts



Colin Beardon
1
, Liam Birtles
2
, Rachael Beach
3

& Ben Salem
1


1

Exeter School of Arts & Design, University of Plymouth

2

Coventry School of Art & Design, Coventry University

3

Institute of Design, Univ
ersity of Teesside; (from Oct 1999: Nottingham Trent
University)


Contact information:

Professor Colin Beardon

Exeter School of Arts & Design

University of Plymouth

Earl Richards Road North

Exeter

Devon EX2 6AS

c.beardon@plym.ac.uk



© 2000

All materials
delivered by this project may be freely used within UK Higher Education,
provided that the original authors are acknowledged. Outside this community, anyone
wishing to utilise these materials should contact the authors.


Acknowledgements

The project was sp
ecified and managed by Professor Colin Beardon of Exeter School of Arts
& Design, University of Plymouth. The project was in all other respects a joint collaboration
between the Universities of Coventry, Plymouth and Teesside. Thanks are due to all those

who contributed material, read and made suggestions regarding the report, in particular
Professor Clive Richards, Dr. Bob Newman, John Owen, Janice Webster, and Nic Earle.
This project was supported by JISC, under the JTAP programme, and we are grateful
for their
assistance. Thanks especially to Tom Franklin and Tish Roberts of JTAP for their support
during this project.


The JISC Technology Applications Programme is an initiative of the Joint Information
Systems Committee of the Higher Education Funding

Councils.


For more information contact:

Tom Franklin

JTAP Programme Manager

Computer Building

University of Manchester

Manchester

M13 9PL

email: t.franklin@jtap.ac.uk

URL: http://www.jtap.ac.uk


2

0

Overview



This report aims to pull together various ma
terial relating to the current and future use of
virtual reality technologies within art, design and performing arts with particular relevance to
courses within UK HEIs.


As the field is growing daily and the boundaries of the field are by no means clear
, it cannot
aim to be comprehensive. Rather it is based around the experiences of the Networked Virtual
Reality Resource Centres for Art & Design (NVRCAD) project. It attempts to document
what we discovered, and how we reacted to the situation we found.


The principle contents are:

a)

A description of the background situation, in terms of the particular experiences of Art &
Design with respect to computing and communications technologies, especially three
-
dimensional software.

b)

An indication of the exte
nt to which staff within such units are knowledgeable and
involved with VR technologies.

c)

An indication of some of the major factors that might be limiting their further
involvement with these technologies.

d)

A description of kinds of VR work that is be
ing undertaken in Art & Design, a finding
that the most frequent type of use is the abstract exploration of virtuality, rather than real
-
world modelling as often assumed.

e)

A description of various experiences of teaching VR technologies to students of ar
t,
design and performing arts.

f)

An account of some of the technologies that may have an impact on VR in Art & Design
in the near future (collaborative virtual environments, avatar design, gesture
-
input).

g)

A description of various projects and workshops

undertaken in order to disseminate
knowledge about these technologies within Art & Design.

h)

Details of new software for visualisation in the field of theatre & performance developed
within the project to make VRML more accessible for novice computer use
rs.

i)

A summary of the NVRCAD project, its activities and outcomes.


3

Contents


0

Overview

2


1

Introduction

4

1.1

The disciplines of Art, Design & the Performing Arts

4

1.2

The NVRCAD project

5

1.3

Computers in Art & Design HE

5

1.4

Technologies of VR fo
r Art & Design

7

1.5

Teaching VRML within Art & Design

9


2

Emerging technologies of virtuality

11

2.1

Collaborative multi
-
user spaces

11

2.1.1

Using CVEs for Art & Design

11

2.1.2

Building Babel II: Construction in Virtual Environments

12

2.1.3

Virtual C
arnival

13

2.1.4

Avatars

14

2.2

G
estural interfaces

15


3

The Visual Assistant (VA): a VRML
-
based application for visualisation of
performance

18


4

Conclusions of the NVRCAD Project

20


Appendices:

A.1

NVRCAD Workshop '3D on the Internet'

21

A.2

Examples
of VR in Art, Design and the Performing Arts

22

A.3

NVRCAD Case Study: VRML modelling of real world objects

25

A.4

NVRCAD Workshop 'Teaching VRML in Art & Design'

27

A.5

Multi
-
user environment software

29

A.6

NVRCAD Workshop 'Building Babel II'

31

A.7

User

comments from the NVRCAD Virtual Carnival

33

A.8

Survey of User Interfaces with Gesture Input

35

A.9

User Notes for the Visual Assistant

37



References

39

4

1

Introduction


1.1

The disciplines of Art, Design & the Performing Arts (
or

'Art & Design')


Ever
y discipline is unique and we do not normally draw attention to this fact. However, from
time to time, the specificity of a discipline, the problems it faces and the opportunities that
are open to it, do deserve special attention. The disciplines of Art
& Design
1

are currently in
that situation, emerging into the HE sector relatively recently and at the same time being
recognised as a major sector of the economy. It has been estimated, for example, that the
'new creative industries' are currently of equa
l economic importance to this country as the
coal, steel and automobile industries put together. Both Art & Design and emerging digital
technologies are at the heart of this economic and cultural revival. Our educational
institutions therefore need to pa
y particular attention to their response.


There are currently about 100,000 students studying courses within our definition of Art &
Design
2
. It is clear that this is a very large sector of Higher Education, second in size in the
UK only to Management an
d Accounting. Until the 1980s, the bulk of Art & Design
education took place in a large number of colleges under local authority control, but there
were a handful of specialist departments within the universities. During the 1980s one group
of these loca
l authority colleges joined polytechnics and later became part of 'new'
universities. A second group remained as independent specialist colleges and were
transferred as such to HE Funding Council control. The third group remained within the FE
sector. T
his diversity of organisational forms still typifies the subject area
3
.


Some of the more important features that, in the context of this report, distinguish Art &
Design educational practices from those of other disciplines have been described as follo
ws
(Beardon 1997).

a.

Learning in Art & Design is principally about developing understanding through activity
and the development of skills, unlike many other disciplines in which understanding is
achieved through the acquisition of knowledge.

b.

Art & Des
ign students learn predominantly through an iterative process of making
products and critically reviewing them.




1

To avoid lengthy terminology, the phrase 'Art & Design' will be used t
hroughout this report to refer to a
range of disciplines and sub
-
disciplines which includes: Painting, Public Art, Sculpture, Performance,
Installation, Time
-
based Art, Printmaking, Photography, Film & Video, Animation, Illustration, Graphic
and Communicat
ion Design, Environmental and Interior Design, Theatre Design, Exhibition and Events
Design, Fashion, Textiles, Jewellery and Metalwork, Ceramics, Glass, Automotive Design, Product and
Furniture Design. This list is very similar that adopted by Panel 64 (
Art & Design) in the 2001 Research
Assessment Exercise (see circular RAE 5/99). We can also identify near neighbours in such related
disciplines as architecture, interior design, engineering design, dance and music studies, cultural studies and
human
-
comp
uter interface design.

2

CHEAD (1995) reported that in 1994 there were 69,906 full
-
time equivalent students in UK ASC 10 (Art,
Design & Performing Arts). UCAS figures for 1996

(UCAS 1997), showed that there were 68,952 art and
design students undertaking

degree courses in 154 institutions, and a further 11,619 students undertaking
HND courses in 82 institutions. To this we should add about 19,000 students studying drama, theatre and
performing arts courses, making a total of around 100,000 students.

3

In

the 1996 Research Assessment Exercise, for example, eighty
-
two institutions made submissions to the
Art & Design panel and, of these 17% were 'old' universities; 44% were 'new' universities; 18% were
special purpose colleges; and 21% were HE colleges.

5

c.

Art & Design education accepts a diversity of individual performances (i.e. different
students can produce radically different solutions). I
n many other disciplines there is an
expectation of a convergence of student performance towards a single solution.

d.

The control of tools is integral to the practice of artists and designers. The educational
process provides a thorough practical knowled
ge of tools, not least because many new
products are the result of an innovative use of tools.

e.

Most working and learning takes place through activity in a studio or workshop, rather
than in a lecture theatre.

f.

Research in Art & Design is an integral
part of creative practice and embodies appropriate
methodologies which are not necessarily shared by any other discipline.



1.2

The NVRCAD Project


The Networked Virtual Reality Resource Centres for Art & Design (NVRCAD) Project is a
collaboration betwee
n Coventry University (lead institution and JTAP fund holder), the
University of Plymouth (co
-
ordinator of the project) and the University of Teesside. It aims
to provide resources to help promote digital 3D modelling and networking within Art &
Design in

UK higher education institutions. It received three years' funding from JTAP
which commenced in 1996, leading to the Centres becoming operational in early 1997.



It was recognised that 3D modelling and networking had attractions to the Art & Design
comm
unity yet the sector then required more access to resources, experience, technical
infrastructure and support. The original JTAP proposal stated our belief that the only
effective way to do this was to have staff who were both knowledgeable of the subject
(s) and
the technology, and who could demonstrate the software and advise academics. A research
assistant was therefore appointed to each of the three Centres (Birtles, Bensalem & Beach).


We aimed to use various methods to ensure that our activities were

effective and our
expertise was disseminated with respect to our target community:



establishment of web sites promoting VR networked services for the Art & Design
community;



provision of access to the technology, where this was a problem;



encourage
ment of greater collaboration between institutions;



dissemination of information about the technologies and the curriculum possibilities of
3D modelling;



organisation of workshops to share experiences and provide training;



operation of an advice serv
ice, and the offer of site visits.



1.3

Computers in Art & Design HE


In September 1995, the Conference for Higher Education in Art & Design published
Information Technology Survey

(CHEAD 1995)
4

which highlighted a serious under
-
provision
of computing eq
uipment in Art & Design disciplines. It was concerned with a perceived



4

T
he report, based upon 48 questionnaire responses and 12 management interviews, focused mainly on the
deficit of IT equipment for staff and students in art and design.

6

anomaly in the equipment funding model for Art & Design and estimated a total shortfall
across the sector of 1,972 'standard Art & Design computers' and 527 'high performance Art
& Des
ign computers'. It costed the value of this shortfall at £52.8 million.


An AGOCG report into IT strategies for Art & Design, published in 1997, found the amount
and quality of equipment within Art & Design departments varied greatly and observed that
very

few Art & Design lecturers had a working environment with access to networking
(Beardon 1997). Despite this lack of access through their institutions, the CTI for Art &
Design found that, in the same year, staff were finding ways of accessing networks.
It
presented its findings from a questionnaire of computer users in Art & Design (Reast 1997).
Among the users surveyed, 85% had access to Internet, 81% used email, and 61% already had
or were proposing to create web resources of their own.


Most of this
use, however, involved only text (i.e. HTML) and 2D images. In general, three
-
dimensional work had been slower to take off within Art & Design (for reasons we will
discuss later). From external sources, however, the potential of the Internet for carrying

material with visual, audio, interactive and time
-
based content was becoming apparent. The
first guide to VRML in Art & Design had been produced by AGOCG (Ashdown 1996). It
described the main elements of the VRML language and saw its role almost exclusi
vely in
terms of the ability to support modelling. It argued that VRML presented interesting
possibilities, especially for the disciplines of architecture, interior design, landscape and 3D
design, but it did not find any major users or uses within fine a
rt or performing arts. The
report presented five examples from architecture, interior design, industrial design, and
ceramics and one gallery site.


The design of virtual worlds is, of course, a matter that requires creative design skills as well
as techn
ical skills and these were well covered in a SIMA report by Stephen Boyd Davis
(1996). The availability of archives of models is also covered by an AGOCG Report by
Costello and Bee (1997). Health and Safety issues with respect to VR in general were
handl
ed in a further report by Costello (1997).


At the commencement of the NVRCAD project in January 1997, a survey of a representative
sample of institutions was conducted. We wished to find out how much knowledge and
experience existed of technologies such
as VRML and we had been advised that a postal
survey was unlikely to produce this knowledge. We therefore approached eleven colleges
and sought the people most likely to know about computer
-
based 3D applications
5
. If this
sample survey was representative

then there were about 100 Art & Design staff in UK HEIs
who knew anything about VRML, just over 50 of whom had used a VRML browser and
about 30 of whom had created a VRML file (directly or indirectly). Our subsequent
experience would tend to support thes
e estimates. An implication of this finding is that, in
about one half of all HEIs that teach Art & Design, there was no knowledge of these
technologies at all in 1997.


It seemed clear that the level of awareness and experience of VR technologies within
UK
colleges was low. Further questioning elicited some of the reasons for this.

a)

There was not a strong existing tradition of 3D computing within Art & Design. Such
applications were seen as complicated and expensive (particular within small budgets),
and they did not always address the topic from the creative designer's point of view.



5

The NVRCAD Survey Report (1998) can be found at http://vr.tees.ac.uk/rachael/article
s/survey/surv.html

7

Furthermore, there were no potential cost savings involved, as staff would not anticipate
giving up any previous technologies as a result of such computer use.

b)

Many A
rt & Design departments use Apple Macintosh equipment, in some cases
exclusively
6
. VRML software has not been readily available for the Apple Macintosh;
there has only been one beta
-
version VRML 2.0 browser available and there is no
purpose
-
built editing
software. The traditional loyalty of Art & Design users towards
Apple platforms was a major limiting factor in the spread of VRML in this community.

c)

The slow spread of networking (especially to workshop areas) within Art & Design in
HEIs was an initi
al problem. In some colleges, networking has only recently arrived in
Art & Design departments and many staff do not have their own computer to access the
network. Any additional technology, requiring plug
-
ins, extensions, extra cards, etc.
requires extr
a finance and can be difficult to get approval for in these circumstances.

d)

The low quality of output images within VRML browsers has meant that work done using
VRML format is much poorer visually than work done using other applications software
or non
-
d
igital means. In Art & Design the quality of the finished product has always been
very important and this is another major drawback to its acceptance.

e)

The instability of the VRML standard and its implementations has meant that VRML files
will often onl
y run successfully on the browser under which they were developed.
Attempting to browse the same object on different hardware, a different browser, or even
different versions of the same browser can produce radically different output. Any
supposed advant
ages of using a net
-
based standard are therefore not apparent.


Nevertheless, there were developments which led to a certain optimism that VRML might
become accessible and useful for Art & Design.

f)

Most HEIs either had extended, or had plans to extend, t
heir networking to include all
staff offices in Art & Design and students would also have reasonable access; in addition,
the speed and bandwidth of SuperJANET would make graphics and similar high capacity
files more accessible;

g)

increases in computing p
ower and commercial developments would make full multimedia
workstations not only affordable, but the entry
-
level machine;

h)

developments in operating systems would make systems easier to learn and use, and the
development of Internet browsers in general
would increase user sophistication of the
possibilities;

i)

hopefully, new applications software would be addressed to the creative arts market.



1.4

Technologies of VR for Art & Design


When most people think of Virtual Reality they think of technologies

such as head mounted
displays, data gloves and immersive environments such as CAVEs


(Cruz
-
Neira et al 1993).
While there are examples of artists and designers using such technologies, there are very few
examples of such equipment within Art & Design dep
artments in UK HEIs. Virtual reality,
within this context, tends to mean something other than a discipline defined by the use of



6

The CTI Art & Design questionnaire indicated that 64% of respondents used Apple Macs, 52% PCs and
only 4% Unix machines (Reast, 1997).

8

certain pieces of high
-
technology. Artists and designers tend to concentrate more on
practice
-
based explorations which are ba
sed upon the concept of virtuality (Levy 1997).


As indicated in the AGOCG Report on VRML in Art & Design (Ashdown 1996), CAD
-
type
3D
-
modelling was initially seen as the main application of VR within Art & Design. Yet
experience had shown that full CAD sy
stems were both expensive and appeared to be
difficult to learn for the majority of students. Apart from a handful of staff and students who
were prepared to undertake a lengthy period of training, the use of such systems in teaching
and learning was prob
lematic. They seemed technically difficult (and 'unnatural') and so the
learning exercise was in danger of being diverted into a struggle to control the technology
rather than to develop subject
-
relevant skills. In short, CAD technology often got in the
way
of teaching and learning rather than supporting it.


In order to explore and learn from the positive experiences of those artists and designers who
we knew had used VRML for creative work, we organised the first NVRCAD Workshop
entitled '3D on the Inte
rnet'. It was held at Coventry University on 23 September 1997 and
attracted about 100 participants. The workshop featured the work of six artists and designers
who had used VRML and related technologies (for a fuller report see Appendix A.1).


Three main

observations emerged from the day. Firstly, there was a sharp contrast between
the formal approaches to learning about VRML which seemed to be promoted by most
technical developers, and the more experimental methods that were frequently employed by
artis
ts and designers. Creative users avoided using tutorials or other such methods and
preferred instead to find working examples and to learn by adapting them to their needs.


Secondly, there was a growing interest in the potential of shared virtual worlds f
or Art &
Design. Simply seeing the Internet as a means of distributing 3D models did not, in itself,
suggest many interesting applications, but the idea of communicating with others through a
shared 3D environment did excite interest.


Thirdly, for many f
ine artists and performing artists, seeing VRML as a variant of CAD
systems was not the obvious way to explore virtuality. Artists want to create alternate
realities and to play with concepts such as representation and interactivity. For some
designers,
too, complex modelling of objects was not necessarily the most relevant approach
and they wished to explore new ways to explain their products to a new audience.


Following this original Workshop, we have become aware of many instances of interesting
VR wo
rk being undertaken by artists and designers. Appendix A.2 contains a list of some of
the more interesting examples we found, intended to be of benefit to tutors or students who
wish to see different types of usage. The field is now increasing in size at

such a rate that we
cannot hope to be comprehensive and we apologise in advance if we have missed any
particular site.


Our initial research and activities indicated that the majority of artists and designers who
were using VRML came to it from a creativ
e arts, rather than product design, perspective.
This raised the initial question of whether VRML was an effective teaching/learning tool for
projects that required a high degree of realism. We decided to test this in the form of a
specific case study (s
ee Appendix A.3).


The case study involved a student who was familiar with the use of 3D software, but not
VRML, attempting to model a real
-
world ceramic object within a reasonable time
-
frame (3
9

weeks). It was successful in that the VRML model was achieve
d and we learned a lot about
the limitations of the technology for this purpose for this group of users. The software
support for VRML was, at the time, not well developed and the technology failed in some
important respects that would be critical if it w
ere to be used for realistic modelling for artists
and designers. The visual quality of the output was poor and inconsistent across platforms.
The lack of sophisticated lighting effects meant that the images sometimes lacked a sense of
realism. At this
time, artists and designers were best advised to use non
-
VRML approaches if

they wished to create realistic models of physical artefacts.



1.5

Teaching VRML within Art & Design


Having established a range of uses of VRML in Art & Design, we then turned ou
r attention to

how such a topic can be taught to such students. In April 1998 a workshop was held at the
University of Teesside concerning the issues involved in introducing VRML into taught
courses within Art & Design (Edgar 1998). There were six speake
rs, all of whom had been
involved in some way in art, design or multimedia courses that involved the student using
VRML. Their various contributions are summarised in Appendix A.4.


One of the main issues to arise during the Workshop was the advantage of
running a VRML
module in conjunction with a department of technology. In these circumstances, there was
not only technical knowledge available, but also more and better quality computers and more
technicians. The necessity for students' work to be review
ed in peer group sessions was also
raised, and the need to specify a particular hardware/software configuration with which it was
the responsibility of the student to comply.


NVRCAD staff made many visits to HEIs, and they were frequently disappointed by
the level
of the problems they encountered. There was often a lack of relevant technical support for
lecturers, and basic software was frequently unavailable or outdated. Though invited to
demonstrate VRML, they often had to make presentations that introdu
ced the Internet, as
neither staff nor students had ever experienced it. Often the technology available prevented
them from demonstrating VRML 'live' and they could only show a video of its use.


We remained convinced that artists and designers would use

these technologies if only the
right way can be found to appeal to them. To this end, a CD
-
ROM is being produced which
attempts to map the experience of artists and designers onto VRML (and not the other way
around). Currently most tutorials in VRML tak
e a language
-
based approach. Artists and
designers do not naturally work in this way and will approach the possibility of VRML only
if they have clear motivation. They need to have their interests mapped onto what VRML
can offer. Only if this happens can
they make an informed decision as to whether VRML is
useful to them or not. The CD
-
ROM takes the user through a series of binary choices which
ask them to consider which aspects of the design process they are interested in and the type of
result they are i
nterested in obtaining. Only when the user has made these choices are they
pointed to concepts that they may need to understand if they wish to pursue VRML further.
Only after the user has been through these concepts, are they shown the aspects of VRML
nod
es that relate to what they want to achieve.


More information about the CD
-
ROM can be found at:
http://vr.tees.ac.uk/rachael/articles/idater/idater.html and in Beach, R., & Birtles, L. (1999b)


10

Given that the motivation of artists and designers was the s
ingle most important prerequisite
to our fulfilling our objective of promoting VR technologies in Art & Design, we set about a
series of workshops and projects based around new and emerging technologies of Internet
-
based VR that we thought might capture th
e imagination of our users, and might lead to
interesting applications within these disciplines.

11

2

Emerging Technologies


We have noted that, for artists and designers, VR frequently means the exploration of
virtuality rather than the digital modelling of

physical objects. But artists and designers also
respond to the tools and media available to them. As we noted at the beginning, the control
of tools is integral to the practice of artists and designers.


As the NVRCAD project progressed, one of the way
s in which we found we could support
artist and designers was through introducing them to particular tools that were emerging and
seemed to be of relevance to creative work.


2.1

VRML Multi
-
user Environments


The first of these was 3D multi
-
user environmen
ts which are based upon a VRML model of
the world but which contain representations of all the users logged in at any one time. The
idea of a virtual 3D environment as a meeting place received a lot of interest at our first
Workshop.


2.1.1

Multi
-
user env
ironment software

We first investigated four examples of software that make possible a shared three
dimensional interactive space in which users can meet and collaborate. A comparison of
these four examples can be found in Appendix A.5. All of the softwa
re runs only on PCs (or
Unix) and is available free via the Internet (we could find no such software that runs on
Apple Macintosh). As well as considering the usefulness of these VRML Collaborative
Virtual Environments (CVEs) it is useful to understand ho
w they work.


After installing the client
-
end software (i.e. the browser or browser plug
-
in), and opening it,
there will be an attempt to connect to a server containing a VRML world. The server is most
often maintained by the company which distributes the

server software or else a third party
which has obtained the software and built a world for visitors. Once connected there is an
initial download of the description of the constant parts of that world (in VRML). Then, as
each user moves about and communi
cates, there are short bursts of communication between
the user's machine and the server to enable updates of only those aspects of the world that
have changed. So, while there may be some delays when first entering a world, there will not
be heavy networ
k traffic in order to represent real
-
time updating.


To enter a CVE one must first choose an ‘avatar’ (see section 2.1.4). One then
communicates by navigating (i.e. moving your avatar) and participating in text
-
based 'chat' .
Most multi
-
user spaces have

a 'chat
-
box' and a 'chat
-
log' ('chat
-
log' shows the conversation of
all the users and 'chat box' is where you type your own messages). There are often several
independent conversations going on at once and keeping track of them at the same time as
watchin
g the 3D world is something of an acquired art.


For what purpose might one use a CVE? A multi
-
user world presents us with the opportunity
to communicate in a visible, computer
-
generated, space. How is this better than using VDS
(Virtual Design Studio) pa
radigm or communicating via other ‘white
-
boarding’ technologies
in which one collaborates through multiple 2D windows and shared files? NVRCAD
conducted an experiment to see whether VRML CVEs could be used for ‘constructive’
collaboration. We set two objec
tives for our collaboration. Firstly, it should make functional
use of the text feature of the CVE and not simply use it for incidental chat. Secondly, since
we were presented with a space full of objects, we should be able to construct new objects
12

and spa
ces collaboratively. This seemed the most likely way in which these spaces might be
used by those in Art & Design (Edgar and Salem 1998).


Upon investigating these CVEs, however, it soon became clear that 'social chat' was their
primary aim. Only one piec
e of software, Active Worlds, came close to our need of being
able to construct objects in real
-
time. To pursue our objective would have required us to
begin amending or maintaining server sites. Although this would have given us more control
over what is

possible, it would not be an appropriate level of technical involvement for the
typical student/tutor involved in Art & Design.


Active Worlds actively promotes itself as a social tool in which people are encouraged to
‘homestead’. Users claim their own
piece of 'cyberspace', build on it and partake in
socialising with other 'homesteaders'. The user is encouraged to become a full ‘citizen’ of
Active Worlds but it is possible to build in Active Worlds while remaining a 'tourist' if one
goes to the Building

School or AWSchool which is provided for practice building. One's
efforts are saved temporarily for 3 days. Access to depositories of different types of object is
provided through billboards.


We found, the following drawbacks to working within the Acti
ve Worlds system.



slow response time, leading to delay in reacting to the movement of one's avatar and
objects;



a tendency to lose the other people that you are trying to work with, (one has no real
warning of what another’s intended actions are due
to a lack of subtlety in the depiction
of avatars);



a lack of peripheral vision (which can cause disorientation);



a seeming lack of coherence between the chat log and the three dimensional space
(perhaps due to delays in transmission);



the inability,

even for citizens, to introduce one's own object into an Active World;



the restriction of available objects to realistic copies of real world objects.


Whilst Active Worlds was never meant to provide a workshop for artists and designers,
nevertheless
it was useful to investigate CVEs, to discover in whether they could be so used,
and to suggest how they might be developed for more creative uses and users.


2.1.2

Building Babel II: Construction in Virtual Environments

In September 1998, we organised
an experimental workshop entitled 'Building Babel II'
which took place over three days at Coventry School of Art & Design. The workshop
explored the issues surrounding the construction and use of CVEs by the Art & Design
community. We selected eight parti
cipants who were asked to write a short piece explaining
what they thought they might do with a CVE and to consider why people might want to visit
such a space. They were from a mixed background including students, researchers and
lecturers. A report on t
he workshop can be found in Appendix A.6. It was successful in that
the participants were able to create their own avatars with behaviours and to observe their
participation in a shared virtual environment.


An important aspect of this Workshop was the at
tempt to integrate the real and the virtual
environments in order to facilitate collaboration. This aspect worked very well and bringing
participants together physically was of major benefit to the process. They developed an
13

understanding and awareness of

the others present and could associate each real person with a

virtual name, piece of geometry and a communication style. An interesting side
-
effect of
physical togetherness was that person
-
to
-
person communication was not restricted to digital
media and c
ould take place in the real space with participants shouting across the room to
each other. As this latter method was used a lot, it would appear that VRML CVEs have
some way to go before they can equal face
-
to
-
face communication.


The workshop went some
way to developing a VRML CVE environment to the point where
it was of use to artists and designers. Feedback from the participants was positive with most
of them saying that even if they had not managed to build their world completely, they had
nevertheles
s managed to achieve lesser goals: to learn about VRML in more depth, to gain
access to a server or simply to immerse themselves and observe this novel technology. Some
of the participants have continued to amend their worlds and the Building Babel mailin
g list
has continued discussing the highs and lows of both the technology and the workshop.


Should the workshop take place again we would like to take more time, provide more
technical support and explain in more detail the unique possibilities of CVEs
as opposed to
stand
-
alone VRML worlds. Building Babel can be found at http://nvrcad.coventry.ac.uk/. A
fuller report has also been published (Beach and Birtles 1999a).



2.1.3

The Virtual Carnival

The belief behind this project has been that artists and
designers will learn best if they are
motivated and have a creative aim. To emphasise this, and provide a fitting culmination to
the project, we organised a 'Virtual Carnival' entitled 'Farewell to Meat'
7
. It was held in
CyberTown on 6th December 1999.
Everyone who had contacted the NVRCADs was
invited to built a 'costume' or a 'float' in VRML (i.e. an avatar) and to connect to the
CyberTown server (http://www.cybertown.com/) at the prescribed time. For the participants,
it was aimed to provide the moti
vation for people to build their own VRML models, and
experience a significant multi
-
user VRML environment. From the point of view of
NVRCAD, it was hoped it would also tell us something about the reliability of these
emerging technologies for larger
-
scal
e operations and the possibilities for more performance
-
oriented events.


We used the Blaxxun VRML plug in for the 3D interaction (www.blaxxun.com). We
decided in advance that the Carnival would last for two hours and progress through four
separate locati
ons in the CyberTown 3D collaborative virtual world. It is difficult to know
who came only for the Carnival and who was a regular visitor (i.e. an 'on
-
looker'), but we
estimate that around 40 people came especially for the event. In most locations there
were
between 50 and 90 avatars present (including some technical 'guides', kindly provided by the
server organisation). Technologically, the event was quite successful, with many users
reporting that their equipment (in some cases just an 'entry
-
level' PC
) stood up well, though
some users reported serious delays when entering a new location.


The event was generally enjoyable, though it perhaps did not live up to its billing as a
'carnival'. However, we learned a lot about how to organise such events fr
om a creative Art
& Design point of view. A successful Carnival probably requires it to be an annual event,



7

It has been suggested that the origin of the word 'Carnival' is 'carne vale' (i.e. 'farewell to
meat'). Whether it
is the real origin we do not know, but it did seem an appropriate title for a virtual carnival.

14

and to be supported by prior tutorial work in colleges. We now have a clearer idea as to what
this should contain. An annotated list of user comm
ents can be found as Appendix A.7.


2.1.4

Avatars

For the Virtual Carnival we obtained an extended free licence for participants to use the
Spazz3D software <www.spazz3d.com> for building their avatar. A short tutorial on using
the software was written so

that a participant could create an avatar with some behaviour
within a few hours <www.coventry.ac.uk/nvrcad/carnival.html>. Some of the participants
involved in the Carnival had developed quite complex avatars. Examples of avatars designed
by some part
icipants can be found at
<www.kk.mah.se/students/wshop99/avatar/gallery/index.htm>. This has led us to consider
how artists and designers might approach the question of avatar design.


The participants of a CVE need to be directly visible to themselves an
d to other participants
(Benford et al. 1997). This not only creates a more realistic experience but also the
possibility of using conversational conventions (e.g. looking at the speaker). Although some
work has been published on the visual aspects of av
atars and their communicative functions
(Argyll 1988, Capin 1997) they refer mainly to aspects such as size, shape and visual
appearance. More important are the modelling of functions and behaviours which the user
can control and the other users can inter
pret.


A further complication is the presence, in certain virtual environments, of 'agents' as well as
user
-
driven avatars. Put simply, an agent is launched by a user but then has relatively
autonomous behaviour driven by its own internal program. While

there has been a steady
increase in the use of software agents for the assistance and guidance of new participants in
CVEs, this has been done without a common set of behavioural rules and vocabulary of
expressions. This has led to some difficulties in
understanding agents, and to confusion as to
whether one is encountering an autonomous agent or another user's avatar.


Keys issues in the design of avatars and agents are:



Identity of the Owner

If the owner wishes it, it should be possible to gather in
formation
about an avatar's owner, such as their identity, gender, location, interests.



Accessibility of the Owner

It should be possible to observe whether the owner of an
avatar is available for chat, would accept urgent messages only, or wants just to

talk to
selected friends. Also, information such as whether the owner has a webcam, an
audiolink, or whether text chat the only possible link.



Status of the Owner

One should be able to discover whether the owner of the avatar is a
guest user, a privile
ged user, a manager or monitor of the CVE. If the user is on 'pause'
mode, how long has that been in effect.



Avatars Functions and Characteristics

One should be able to discover what are the
actions within the CVE that can be performed by the avatar (e.
g. kinds of dialogue,
displacements, manipulations, expressions, gestures).


Avatars with behaviours can engage in non
-
verbal communication which plays an important
part in our social interaction and can be described as one of the following means (from
Arg
yle, 1975) :
-



Facial expressions (e.g. smiling);



Eye expressions (e.g. gazing, opening/closing eyes);

15



Body postures and gestures (e.g. waving, turning to face someone);



Hand postures and gestures as part of the discussion (e.g. for emphasis, negati
on, etc.)



Body contact (e.g. shaking hands)



Body colour and complexion (e.g. blushing)



Clothes, makeup and other aspects of appearance (e.g. casual, uniform).


Avatars with such behavioural characteristics should be able to communicate within multi
-
u
ser communities by employing 'body
-
language' and gestures that complement any textual
message (Earle and Beardon 1998). From recent communications, it would seem that there
is now quite a bit of interest in the Art & Design community (particularly among
the theatre
& performing arts disciplines) in the design of avatars. We would expect this topic to grow
in significance for artist and designers in the period ahead.



2.2

G
estural interfaces

The basic computers we use in HEIs still have a monitor, keyboa
rd and mouse, just the same
as they did twenty years ago. They may be faster and more powerful, but they are still
constrained by the boxes they come in. For many in Art & Design there is a desire to get 'out
of the box', to begin to transcend the limita
tions of this particular configuration of
components. The use of video projectors, for example, can make the output more communal,
and the addition of sound in CVEs provides a richer environment.


One of the input devices which, it was felt, might make VR

more accessible to more practice
-

and workshop
-
based disciplines was the data glove, or to be more general, gestural input
devices. We were aware of work being done in non
-
intrusive motion detection, via video,
but believed that greater accuracy might be

obtained by devices that directly monitor hand
movement.


One of the difficulties that arise with such devices is the definition of a set of hand gestures
and direct manipulations that is meaningful and relevant to the user, as well as being machine
class
ifiable. The interaction tasks that are usually most required in a virtual environment are
object manipulation, navigation, and command selection. A comparison of WIMP (Window
Icon Menu Pointer) and hand
-
based input solutions to these is given in Table 1.


WIMP interface

Hand Input interface

Menu and Object Selection

Pointing

Scroll bar

Pointing Up and Down

Mouse Click

Pointing to a Hotpoint

Tool Selection

Hand Transformation

Shortcut

Specific Gesture

Grabbing and moving

Direct Manipulation

Table
1: WIMP / Hand Input equivalence


From a user functional point of view, hand inputs can be classified into seven categories as in
Table 2.


16

Function type

Description

Examples

Non
-
intentional

reflexes, and reactions

protect one’s face

Meaningless

body re
lated hand functions and comfort
providing functions

catch a fly, scratch one's
nose

Direct manipulation

the hand reaches the object and manipulate
it

touch, grab, hold, push, pull,
move....

Transformation

the hand is transformed into a particular tool
f
or tele
-
control and other applications

hammer, puma robot hand,
micro
-
manipulator...

Written gesture

the hand draws a sign or writes a text on a
surface

draw a circle, write a
message

Static gesture

the hand performs static gestures (i.e.
adopts a postur
e)

stop, okay, point..

Dynamic gesture

the hand performs dynamic gestures (i.e.
adopts a succession of postures)

bye bye, come here, maybe

Table 2: The different hand functions from a performance point of view


The functions that have generated the great
est interest are:
-

Direct manipulation

This is the most natural and intuitive hand function. The hand
manipulates objects (virtual objects, interface objects and widgets) in the virtual environment
as if they were real objects. Direct manipulations are d
ivided into two classes, prehensile and
non
-
prehensile (Napier 1980). Prehensile manipulations involve holding an object between
the fingers and the palm, or between two fingers. Non
-
prehensile manipulations occur when
the object is not actually held by th
e hand, e.g. pushing or lifting.


Transformation

This is where the representation of the user's hand is transformed into a
particular tool to fit the requirements of a task. For example, editing the colour of a virtual
object. The hand is transformed int
o a paintbrush and the user paints the object in whichever
colour is selected.


Hand gestures

(static and dynamic) This is where the user’s hand expresses commands and
identifies or quantifies them by performing a particular gesture or by adopting a part
icular
posture. Static gestures (i.e. postures) are performed with no motion of the hand. All the
information is carried by the wrist and palm position and orientation and by the digits
flexion/extension. Dynamic gestures are performed with a motion of t
he hand.


A survey of user interfaces that have used gesture input can be found as Appendix A.8.


Hand functions do have some drawbacks. They lack precision and accuracy and it is difficult
to quantify a hand movement. As an alternative, if accuracy is r
equired, an interface object
such as a menu could be called up. Also, hand manipulations can only be performed on
objects displayed to the user.


In general, with gestural interfaces, logical and spatial navigation is replaced by pointing,
dialogues are r
eplaced by postures and short gestures, specific commands and short cuts are
replaced by gestures and manipulations. Table 3 gives the relevance of hand inputs according
to the interaction style that a user may encounter while interacting with a system.


I
nteraction Style

Duration

Hand Function

Suitability and example

Command Entry

Short

Posture, gesture

Average,
OK, quit

17

Menu and logical
navigation

Short iterations

Posture, gesture

Good,
pointing

Geographical
navigation

Short iterations

Posture

Good,
po
inting
.

Form
-
fill

Short to medium

Not applicable

Very poor,
sign
language

Natural language

Medium to long

Posture, gesture

Good,
bye, maybe

Direct Manipulation

Medium to long

Gesture

Very good,
pushing,
moving

Table 3: Suitability of hand functions for

different interaction styles


Compared to using a video camera for motion capture
8
, the glove method may be preferred
for the following reasons.



It is a proven technology, which can track the hand with less restriction on the kind of
gestures permitted.



There is no need for complex tracking algorithms, which generally rely on a hand data
model for the extraction of relevant data from the video image.



There is no problem of visual occlusion of the hand and hence no restriction of movement
on the user.

However, using a glove
-
input device implies the need for the user to wear an instrumented
glove. If the user is shifting from one task to the other, then this may be a major problem, at
which point video tracking might be more suitable.




8

See STEIM

18

3

The Visual Assi
stant (VA): a VRML
-
based application for the visualisation of
performance


The NVRCAD project discovered fairly early on that one of the major weaknesses holding
back VRML use is the lack of software that engages with the working practices of the various
d
isciplines of Art & Design. In these disciplines, as we have stressed several times, copying
physical properties (i.e. modelling) is frequently not the primary mode of representation
required. In order to respond positively to this situation, the NVRCAD
project has supported
the development of specific new software for the field of theatre visualisation.


The Visual Assistant has been designed to allow theatre students and staff to create simple
prototype 3D visualisations of a performance that are subseq
uently saved in VRML 2.0
format. It has also been designed with a range of teaching and learning situations in mind.
The software has been used in real workshop teaching situations with over 100 students,
theatre practitioners and tutors at Exeter School

of Arts and Design in Exmouth, at Central St
Martins School of Art in London, and at Malmö University College in Sweden. For a review
of the Visual Assistant see the CTIAD journal,
Outline

No 8.


The Visual Assistant software was designed for users who a
re primarily interested in the
theatre and are probably not particularly interested in computers. Experience of working with
such students led to some initial design objectives for the software.



It must be very simple to learn: you should be able to see

someone else using it for ten
minutes and then confidently use it yourself.



It must be simple to use: there should be no more than about 40 different tools available
in the application.



It should give meaningful results quickly.



What happens should
be like sketching: it should not matter if work is irretrievably
transformed or destroyed: you can always try again.



It should support 'process' rather than 'product': particularly processes that lead to clearer
understanding and better actions in the r
eal world.



It should present a believable 'language
-
game': when acting on the computer screen you
should be thinking as a theatre person.



It should support person
-
to
-
person dialogue; it should function as a common sketchpad to
support critical discuss
ions.



It should be able to lead to more detailed implementation.


The Visual Assistant acts as a bridge between 2D and 3D representations, allowing 2D
images (enhanced in Photoshop or other image processing applications) to be imported as flat
objects
within a 3D space. They thus behave rather like flat scenery on a theatrical stage.
Visualisations are built up in this way, but it would be wrong to think of them only as
prototype set designs. Experience has shown that people will use the software for

many
different visualisation tasks, and they are all accepted as legitimate. Some of the issues
involved in designing software for creative practice are discussed in (Beardon 1999).


Students will be typically set a visualisation task by their tutor and
they begin by collecting
about six two
-
dimensional images which they feel are relevant to the task. If resources allow,
they will show these to the group and there will be a discussion of them, simply as images.
19

They will then be introduced to the Visual

Assistant, usually for about 10 minutes in a lecture
theatre situation with a live demonstration projected onto a screen. They will then be asked
to take their images, and any others they care to use, into the VA in order to create a simple
visualisation
.


Over the remainder of the workshop they will be asked to show their developing VA model
to others, via projection, where both front and top views of the world can be seen and
discussed. If necessary, full 3D manipulation can be seen through a VRML brow
ser. The
ability to non
-
destructively explore various visualisation issues (for example, using greyscale
rather than colour on one image, or rotating the stage) is a particular advantage. Whilst the
VA will not replace the model box, and neither will it
do away with the need for more
detailed modelling software, it has the distinct advantage of encouraging more freedom in the
use of what is often seen as a very 'serious' technology.


Some of the issues that the VA has helped to address are those of:



o
vercoming visual 'block' (which is rather like 'writer's block');



representing the 'atmosphere of the stage' rather than the detail (or 'the space within which
the actors can do their job');



addressing the dominance of televisual close
-
ups (by forcing

a full
-
stage viewpoint);



providing the ability to improvise visually (see Beardon & Enright 1999).


The Visual Assistant project has been a large undertaking, involving funding and support
from a number of sources, but it has received significant sup
port from the NVRCAD project.
It is supported by a web site (www.esad.plym.ac.uk/va) from which it is possible to download
the software, access or download the User Guide, and access about 30 VRML 2.0 models
developed by students at the various workshops.


Insofar as it has been successful, the VA demonstrates that VRML can be made accessible to
relatively new users of computers, that it does not require any special hardware or software,
and that it can be used in such a way as to make teaching and learnin
g more visual and more
communal.


Notes on downloading and using the VA can be found in Appendix A.9.


20

4

Conclusions of the NVRCAD Project


At the outset of the NVRCAD project we anticipated a higher level of knowledge and
experience about virtual worlds
within the Art & Design community in UK HEIs than we
found to be the case. There seemed to be three main reasons for the relatively slow take up of
these technologies within these disciplines. The basic technology of a computer and
networking was often si
mply not available. There was no prior experience of 3D computing
upon which networked, interactive 3D computing could grow. Thirdly, there seemed to be a
misunderstanding of what use artists and designers might make of such technologies. VRML
was frequ
ently presented as a variation of CAD modelling, with the implied objective of
modelling real
-
world objects. This was not how those from fine art and the performing arts
saw the technology, and neither was it true of all designers. There was a much great
er
interest in exploring the potential of the virtual world paradigm, and in using the technologies
of communication through shared virtual worlds.


The NVRCAD project refocussed and concentrated on positive strategies to deal with the
situation it discove
red. Through direct contact with members of the Art & Design
community, via workshops, conference presentations and teaching at individual institutions,
we have realised that the problems facing those wishing to teach 3D computing in Art &
Design are consi
derable, but that creative motivation must be the starting point. Once
motivated, formal learning (e.g. tutorial
-
based) is not the natural approach taken (though it
may be necessary to some extent). A great reliance is placed upon finding existing models
,
learning how they work and adapting them to the purpose at hand. Many artists claimed to
work in this way.


Building upon the realisation that new and emerging technologies are more exciting, and
hence more motivating, the NVRCAD project began the explo
ration and promotion of certain
new technologies. The availability of the software for shared virtual worlds created a specific
set of opportunities. Through activities such as Babel II and the Virtual Carnival we explored
some of the technological and d
iscipline
-
dependent issues that will arise as these
environments become more accessible. Once again, the need for motivation and the dangers
of assuming too literal an interpretation of the content matter came to the fore. In a similar
vein, we conducted

some survey work into gesture
-
based devices such as data gloves.


As a further contribution, we have assisted in the production of significant new software, the
Visual Assistant. This software makes technologies such as VRML 2.0 accessible to
relatively
novice computer users, but within a context that continually refers back to the
principles and processes of their own discipline. The VA is freely downloadable, has been
tested in real learning situations and has been well reviewed.


There is clearly much
to be done in the future as new technologies emerge to a
knowledgeable Art & Design community that is larger than it was three years ago, certainly
better informed and much more eager for new media and techniques.


Nevertheless, it is the motivation to exp
lore virtuality in its most abstract forms that has
driven most of the work we have seen. To support this we need software that combines
simplicity of design while recognising the primacy of creative practice. These emerge as the
major themes of this NVR
CAD project.

21

Appendix A.1

NVRCAD Workshop '3D on the Internet'


The first NVRCAD Workshop, entitled '3D on the Internet', was held at Coventry University
on 23 September 1997 and attracted about 100 people. This workshop aimed to provide
positive examples

of VRML use in Art & Design and featured the work of six artists and
designers who had used VRML and related technologies (Birtles 1997).


There were presentations from:


Aharon Amir


Cut and Paste VRML.

This addressed the question, "How, as an artist
and designer, do you begin to learn a medium
as foreign as the mathematical descriptions of 3D provided by VRML?" The suggested
answer was to adopt the traditional approach of artists and designers

to observe and adapt.
That means finding source code, d
ownloading it, learning what it does by adapting it, and
then adapting it for you own purposes.


Steve Renford


Performing in a Shard World.

This described a project that involved a virtual poetry reading, which involved a mixture of
virtual and three liv
e poets. In addition, each member of the audience was represented by an
avatar which they could navigate through the space, visit linked spaces and carry on
conversations. In practice, the poetry reading tended to get ignored as users came to terms
with
the possibilities of navigation and encounter in virtual space.


Ivan Vaghi


Virtual Mannequins.

This paper described a system for building and animating 'virtual humans' (defined as 'fully
articulated 3D models of humanoid bodies'). By using VRML and Ja
va it was aimed to
realise an animation system for virtual humans that achieves real
-
time human
-
like fluid
movements.


John Paul Jones


J.P. Jones and his Singing Stones.

This described a BA project at UWC, Newport that involved an attempt to evoke the sp
irit of
a physical place (i.e. a particular place just outside Newport). Though the VRML world was
technically quite sophisticated, the creators (J.P. Jones and Dave Andrews) said that they had
adopted Aharon Amir's approach, had not learned VRML formally

and did not consider
themselves to be programmers.


Andy McGrath


British Telecom.

This was a talk from the commercial sector which stressed the need for meaningful content
within VR worlds, rather than just technical competence. It also demonstrated so
me of BT's
own work, include the TV programme 'The Mirror'.


Kate Allen


Virtual Sculpture.

This described the experience of a sculptor who came to VR through a series of problems
within her own physical work. 'Have your cake and eat it' uses the common
denominator of
food for a ritual to explore the combining of the virtual and the real. Kate both created a
VRML model of a cake and baked a real cake that looked very similar. She then arranged for
a tea at which both the real cake was eaten and the virt
ual cake could be explored (with
unexpected results).
22

Appendix A.2

Examples of VR in Art, Design and the Performing Arts


Galleries



The Abulafia Gallery <http://www.cgrg.ohio
-
state.edu/~mlewis/Gallery/gallery.html>
This is the most traditional of the
links, as it is simply a mock up of an imaginary gallery
space with some pictures placed in it.



Cio's Digital Gallery <http://www.2013.com/cio/gallery.html> Another example of using
VRML to build a gallery space to hang examples of your work.



Fabric
a Nacional de Moneda y Timbre <http://www.frimt.es/esp/museo/evisita.htm>
This is a fast and well though out museum of stamps and coins.



Galeria Virtual <http://www.lua.upf.es/~gvirtual/grdemo.ang.html"> This is an
interesting site which instead of repl
icating a real museum, creates a museum which takes
advantage of all the things you can do with VRML interaction to challenge
preconceptions about galleries. Use the viewpoints list if you find it difficult to navigate.


VR Art



VRML
-
Art <http://www.vrml
-
art.org> Here you will find art creations in VRML by
about sixty different artists.



Virtual Harmonium <http://www.parnasse.com/vrml.shtml> This has been around almost
since VRML began but is still a masterpiece of simplicity and design, utilising sound.




Pagan sound world <http://www.parnasse.com/paganSound.wrl.gz> Whilst the graphics
in this world might seem coarse grained, it is an alternative to reality and also makes use
of the spacialised sound capabilities of VRML.



Vibration Sound Gadget <http
://www.deanmoor.nl/users/roland/vibratron%5Cvibra.wrl>
When this was tested the sound didn't work. However, it is interesting even without the
sound!



Stelarc <http://www.stalarc.va.com.au/urlbody/intro.html"> This is the site which
belongs to the famou
s artist of the same name. This particular link is a VRML 'body'
which changes shape and colour depending on the pages that people have visited on
Stelarc's site.



Char Davies 'Osmose' involves higher
-
order technologies to create virtual immersion and
a u
ser interface based on breathing. The graphics are also interesting, involving
transparency and visual ambiguity (Davies 1998).



Body without organs. <stark.kent.edu/~glittle/bwo_1.html>. Work by Greg Little.


Sculpture



The first sculpture in cyberspa
ce <http://tcc.iz.net/1owen.html> A piece of history. This
sculpture is derived from the very first objects made by the developers of VRML.



Philadelphia Museum of Art <http://www.narrativerooms.com/pogany/vr/index_a.html>
This is a lovely VRML world whic
h integrates into your browser window seamlessly.
Whilst not all of the sculptures are full 3D models it is just a really good use of VRML.
The gallery's theme is the series of sculptures, Mlle Pogany



Minneapolis Sculpture Garden <http://www.walkerart.o
rg/gallery9/walczak> Although
the sculptures in this garden are not actually 3D but 2D images imported when you click
23

on the spheres that represent them, this is executed in a sophisticated manner and the
interface is well thought out and looks good.



Pra
tt Institute <http://www.artnetweb.com/sculpctr/isctest/pratt/index.html>. A bit like
the link above but not as well thought out. The sculpture is not so well integrated into the
VRML


Jewellery



Project 925 <http://www.project925.com> claims to be a new

approach to selling
jewellery using 3D. The site allows clients to design a piece of jewellery from a set of
modular components.



Jigowatt <http://www.fabric8.com/jigowat/> has been running for two years and shows
how 3D models of jewellery can be pres
ented with technical information. This is a simple
yet practical application of 3D technologies.

Furniture



virtual.design
-
exhibition.com <virtual.design
-
exhibition.com> presents 3D models of
furniture from various design companies.



Charles monaco <ht
tp://www.3de.com/vrml2.htm> shows how a piece of furniture can be
examined and explained in VRML.

Graphic design



Aqualand <http://vrml.fornax.hu/aqua/> shows how VRML can be linked directly to
HTML to create a sophisticated multimedia presentation.


Th
eatre & Performing arts



ieVR Institute for the Exploration of Virtual Realities, The University of Kansas
<http://kuhttp.cc.ukans.edu/~mreaney/alpha.html>. The Institute's goal is to explore the
uses of virtual reality and related technologies, princip
ally within theatre production and
performance. The site contains various VRML models by students.



Back Hill Studios. Central St Martin's College of Art & Design. Theatre Design for
Performance. Contains various pieces of computer
-
aided modelling and
some VRML
models. Contains a link to the virtual_stages project
<http://www.stages.demon.co.uk/project/vrstage.htm> which involves developing new
software for set design. Cochrane Theatre
<http://www.stages.demon.co.uk/project/cochrane/cochrane.htm>



see
b theater: < http://seeb/net/> Links direct to VRML 2.0 files "for travel on stage
cyberspace" (An html page of explanation would have been helpful!)



Technical Theatre, Chaffey College <http://www.tstonramp.com/~sydmore/sydv.htm">



Virtual Drama Socie
ty <http:// www.virtualdrama.com/vds/research.htm">



Virtual actors <http://miralabwww.unige.ch/"> The Thalmann's site.



VRML Dream Perfomance <http://www.psd.k12.co.us/archive/libnet/1248.html">



CATS Computer aided theatrical score
<http://www.pisa.in
tecs.it/projects/CATS/description.html> . An ESPRIT project to
develop a system for creating and evaluating staging models of all sorts of performances
in the theatrical, television, film and advertising industries.

24



Computer Scenographics Studio <http:
//www.wfu.edu/~drp/frontdr.htm> . Darwin Reid
Payne uses various 3D modelling tools to produce scenography models.


Some useful collections of VRML objects



VRML object warehouse < http://ariadne.iz.net/~jeffs/vrmLab/Warehouse/> For anyone
trying to bui
ld a VRML model


a library of predesigned objects.



Construct.net <http://www.construct.net/worlds/> This is a collection of objects created
by Construct.com and features experimental architecture, behaviours, avatars etc.


25

Appendix A.3

NVRCAD Case St
udy: VRML modelling of real world objects


In order to explore the issue of realistic modelling, we decided to conduct an experiment to
see whether it was possible for a student to produce an acceptable VRML model of a typical
artefact within a sensible ti
me constraint.


We had made a link to the independent Ceramics Gallery, a significant gallery which was
also the base for the journal 'Studio Pottery'. The journal ran an interesting web site of
contemporary ceramics which included 2D images (see <http://
www.ceramic
-
society.co.uk/.>). Both the Gallery and NVRCAD were interested in discovering the
potential of VRML for modelling existing ceramic objects and making these models
available over Internet.


The approach was to select one or two ceramic items fr
om the Gallery and to appoint a
recently
-
graduated student of ceramics who had general familiarity and interest in 3D
computing but no prior knowledge of VRML. We employed the student for three weeks' full
-
time to develop VRML skills and create as realisti
c models as possible of the chosen pieces.


The Gallery suggested a number of historically interesting objects to model, but we felt that
they were geometrically quite complicated and it would be better to start with simpler
objects. We eventually decided

upon the simplest object we could find


a 1930's
Wedgwood ceramic mug shaped rather like a barrel, open at the top and with a curved
handle attached to the side. It was a plain cream colour with a simple inscription on the base.
As the original design

was inspired by geometrical simplicity, the object should provide the
least challenge to a VRML modeller.


The student, Andy Boucher, had considerable experience of three
-
dimensional CAD
packages, especially 3
-
D Studio, but had never looked at VRML code b
efore the
commencement of the project. We provided a dedicated PC running Windows 95, a version
of Netscape with the WorldView 2.0 browser plug
-
in and, as a modelling package, a current
copy of V
-
Realm Builder (v. 2.0). The comments that follow are large
ly from his report (for
the full report see Boucher 1997).


After completing the V
-
Realm Builder tutorials, the student produced some examples of
imaginary ceramic vessels. It was discovered that the builder could only create crude shapes,
which behaved in
consistently in the browser


sometimes the imported shape would be
completely inverted. It was found to be many times quicker to create objects in 3D Studio
and import them into V
-
Realm Builder. Such shapes were imported as indexed face maps, as
opposed t
o extrusions and although the files sizes for these were slightly bigger, the shapes
worked perfectly. This gave the opportunity to add more detail and thereby to create more
realistic objects.


The first example was a fairly simple cup with an animation
that spins the cup around its
centre point. The second example was a set of moveable vessels on shelving as they might
appear in a gallery. The third example was a real piece of ceramic, a Robert Dawson plate,
used to demonstrate how texture maps can be a
pplied.


The final, and main, example was the Keith Murray/Wedgwood mug we had originally
selected. Although it is a fairly simple vessel, in modelling terms even it has a fairly complex
geometry. The geometry for the mesh was constructed in 3D
-
Studio. The

form was then
imported into V
-
Realm Builder, where various details were added, such as the material and
26

the maker's stamp (which was a scan of the original). The main challenge however was
creating sufficient level of detail to be realistic, without becom
ing overly complex and
increasing the file size.


The student's main impression was that VRML and V
-
Realm Builder required the user to
know exactly what the finished result should be before it was possible to start. This is a
typical response to many soft
ware package by artists and designers who, in their normal
practice, are not used to starting with detail (Tovey and Harris 1999). Boucher also found
that the interface to V
-
Realm Builder had some clumsy and time
-
consuming aspects; for
example, navigating

around the constructed object can be very slow, there was no 'undo'
function and, though duplication is possible, it can only be applied to whole objects and not
parts of an object.


VRML browser plug
-
ins were found to be inconsistent in their rendering,

the main difference
being in the colours of materials and textures, but there was sometimes also a variation in
how shapes are interpreted and in the scaling of x
-

and y
-

co
-
ordinates. This inconsistency
emerged as a main drawback.


One of the main probl
ems encountered in creating realistic models in VRML was the lack of
shadows. VRML browsers create a series of high
-
lights and low
-
lights over the object, which
is adjusted according to the position of the viewer and the headlight. This means that, in
term
s of the Keith Murray mug, the region around where the handle joins the body of the
mug does not contain the tell
-
tale shadows that give those vital visual clues as to its form.


As a case study, the exercise taught us a lot. It was clear that the softwa
re support for VRML
was not well developed: the editors were crude, the browsers were inconsistent). The
technology, though promising, was failing in some basic respects if it was intended to be used
for realistic modelling for artists and designers. Vis
ual quality was poor and inconsistent and
the effects of lighting, often crucial, were largely ignored. There seemed little doubt that, at
this point in time, artists and designers would be better advised to use non
-
VRML approaches
if they wished to creat
e realistic models of existing artefacts.



27

Appendix A.4

NVRCAD Workshop 'Teaching VRML in Art & Design'


In April 1998 a workshop was held at the University of Teesside concerning the issues
involved in introducing VRML into taught courses within Art & D
esign (Edgar 1998). There
were six speakers, all of whom had been involved in some way in art, design or multimedia
courses that involved the student using VRML. Their various contributions are summarised
below.


Avon Huxor (Middlesex University)

Avon h
ad used VRML with postgraduate students at Middlesex. After being involved in
several projects he thought VRML mainly useful for collaboration. However, while multi
-
user worlds make this possible, in practice it is limited by such factors as lack of conte
nt, lack
of spatial sophistication, and inappropriate behaviour of those using the worlds. This is
disappointing because students need to be made familiar with these technologies as
teleworking becomes more accepted. At present (1998) he found it preferab
le to use two
dimensional collaborative systems such as BSCW (http://bscw.gmd.de) and ICQ
(http://www.mirabilis.com) and from there incorporate content into a multi
-
user space using
the Active Worlds 3D browser (http://www.activeworlds.com). Using an analo
gy with
physical work spaces, he demonstrated that careful thought was needed about how offices
and their spaces work, for example in exploiting chance encounters and delineating personal
space.


Taylor Nuttall (Innovation in Digital and Electronic Arts)

IDEA is a project run jointly by the Manchester Multimedia Centre, Manchester Metropolitan
University and the European Social Fund. It is aimed at any member of the public who feels
themselves creative and aims to bring together technology and creativity.

As a component of
this project Taylor teaches a VRML course consisting of three 6
-
hour sessions over a period
of three weeks. In addition the students are expected to put in 12 hours per week of their own
time. Students are required to be familiar with wr
iting HTML and general file and image
manipulation. Tuition is carried out on Apple Macintosh computers despite well known
difficulties with (lack of) browsers for this platform. This was a conscious decision made
because these students are more likely to
have access to this platform during and after
training. Students are encouraged to develop practical confidence with the system


for
example, being first required to install a browser. They then learn VRML via its code, using
only a text editor. Problems
encountered are incompatibility of different browsers and
platforms, and difficulty in finding good examples of code to give to students. The students
themselves report that they are only just beginning to grasp the possibilities of VRML when
the course en
ds.


Mike Phillips (University of Plymouth).

Mike teaches on the BSc MediaLab Arts course in the School of Computing where the
students are taught VRML as part of a wider course which also looks at other VR kit and
packages. While Mike and most of the
other tutors are trained in Art & Design, the course is
part of the Faculty of Technology and enjoys the support that this offers, particularly in such
terms as equipment, technical support, and access to relevant computer science courses.
Within the BSc
there are distinct strands of technical modules, creative modules and practice
-
based modules. Mike made many interesting points about the relationship of the 'creative' to
the 'technical' and stressed the need for a dual, and balanced, approach.


David
Smith (University of Wales College, Newport)

28

David was course leader for the BSc in Multimedia at UWC, which is a named award within
the School of Art & Design. He has supervised final year projects in which the students used
VRML extensively. His talk w
as entitled 'The Blind Leading the Partially Sighted: VRML as
a Case Study in Teaching Advanced Digital Techniques in Art & Design' in which he invited
educators to consider how they will cope in the future as it becomes more and more likely
that students
will wish to concentrate on technologies that the educators themselves know
little or nothing about, and which they will frequently have neither the time nor the resources
to learn.


Alan Peacock (The University of Hertfordshire)

Alan is a team member o
f a course named 'Virtualities and Visualisation' run jointly by staff
from the Faculty of Art & Design and the Faculty of Engineering and Information Sciences at
Hertfordshire. The course involves students from both the technology and the creative arts
st
reams. Though VRML has a tendency to be seen as technical, the creative aspect of the
course means that students have to present and explain their work to tutors, despite the time
that this takes. This facilitates the student/tutor discourse that is trad
itional in Art & Design.
It also makes the responsibility for the demonstration and execution of the work lie with the
student. Again from experience, this was only found to be practical if the tutors specified at
the outset the hardware/software configu
ration that all students must use to demonstrate their
work.


Clive Fencott (University of Teesside)

Clive also teaches VRML to a range of computing and multimedia students. His contention
was that the most fundamental concept that students must understa
nd when designing virtual
spaces is the need to provide a feeling of presence for the user. This involves recognising
that, in many instances, they must represent carefully chosen aspects of a space rather than a
complete replication of all physical attrib
utes. He presented, as an example, a functional
model of a cable car system which was driven by gravity and water tanks that could be filled
and emptied.
29

Appendix A.5

Multi
-
user environment software


The following four examples of client
-
end software were

investigated by the NVRCADs.
Practical information about downloading and installing these systems is available at
http://vr.tees.ac.uk/rachael/home.html



Sony’s ‘Community Place’ http://vs.spiw.com/vs/

There are three pieces of software that make up
this suite (Community Place Bureau,
Community Place Browser and Community Place Plug
-
in) and it is not altogether clear at
first which one needs. To browse in Community Place multi
-
user worlds, one can choose
between the Community Place Plug
-
in or the Brow
ser. The Plug
-
in is installed as a helper
application for Netscape, and Community Place must be set as the default VRML browser in
Netscape’s preferences. Community Place Browser is a stand
-
alone application and needs no
other browser software. Choosing th
e browser rather than the plug
-
in would be the
recommended option, especially if one has other VRML plug
-
ins that one prefers to keep as
the default VRML browser in Netscape’s preferences. Community Place Bureau is for those
people who want to run their o
wn server, i.e. building and providing worlds for others to
browse.


Downloading was very slow from the Sony site. In addition, one will probably need to
download and install the JDK (Java Developers Kit) or the JRE (Java Runtime Environment)
if one choos
es the Community Place Browser option (not altogether clear on the Sony site).
Rather than take the hyperlink to the Java site, which might be slow, a better option might be
to find a mirror site in the UK (e.g. Imperial College).


Installation is simple
. Both the JDK/JRE and the browser/plug
-
in are .exe. files, so double
clicking will start the installation process. Once installed, Community Place is different in the
way that information is passed to your machine when it makes a connection to a multi
-
use
r
world. Some of Sony’s multi
-
user worlds need to be downloaded from web pages onto your
machine before you can make a connection to them. This means that the world resides on
your machine permanently. If one opens a downloaded file (usually named main.wrl
) in
Community Place, this will trigger an IP connection to the server at which information for the
world exists. This is ostensibly faster as only information about people in the world has to be
passed to your machine after your first visit. Everything el
se resides on your machine.


You may decide that Community Place has the most stylish interface and is the easiest to use.
However it lacks some of the functionality of Blaxxun and Active worlds, e.g. teleporting,
telegrams, etc.




Circle of Fire’s ‘A
ctive Worlds’ http://www.activeworlds.com/

Active Worlds was easy to download and install. The web page makes it clear that there is
only one piece of software and this can either be downloaded in one go, or can be
downloaded onto two floppy discs. The p
age informs you that the average download time is
12 minutes but it is often faster. It is installed in the usual manner and is a browser in its own
right, i.e. it does not rely on Netscape.


Connection with Active Worlds is initially quick although it s
eemed slower to load texture
files than Blaxxun. Whilst the interface is not as stylish as Community Place, Active Worlds
offers more features. In terms of fostering a community it is fairly sophisticated, offering two
levels of participation. One can visi
t anytime and without limit as a ‘tourist’ but in order to be
able to send and receive ‘telegrams’, choose from a wider range of avatars and to ‘build’ on a
30

plot of land, one must become an active worlds citizen (which costs $US 19.95). Another
interesting

feature is the web page window, which gives information about the world you are
in and its community.




Blaxxun http://www.blaxxun.com/

The Blaxxun web pages may seem slightly confusing but a little perseverance shows that you
need to download CCPro.
Blaxxun is fairly quick to connect to the main server and to
download worlds. The window for the world is bigger than that of Active Worlds and hence
it is easier to navigate. Although you can resize the Active Worlds world you will lose other
information
shown at the side. The availability of many different ready
-
made avatars is an
advantage of Blaxxun, as well as the high quality of the Blaxxun home world. A nice feature
is being able to beam to people (i.e. click on an avatar to relocate yourself to a po
sition facing
them). Otherwise, we found it difficult in Active Worlds to find each other (the compass is
difficult.) There is a menu for gestures, which is another nice feature.




VNET http://ariadne.iz.net/~jeffs/vnet/

You don't need an extra piece o
f software to run VNET worlds, only a proprietary VRML
browser such as CosmoPlayer (available at :http://cosmosoftware.com/). VNET is not really
an example of multi
-
user software. That is, it will only look at simple VRML files that are
not designed to be

visited en masse. What VNET does is create a new interface at the bottom
of the screen by creating an EAI (External Authoring Interface) in Java which extends the
VRML and turns it into a multi
-
user world. Academic users may be frustrated, however, as
an
y kind of firewall seems to result in it failing to work.


31

Appendix A.6

NVRCAD Workshop 'Building Babel II'


This experimental workshop aimed at providing participants with an opportunity to build
collaborative multi
-
user spaces to their own specif
ications in a mutually supportive
atmosphere. The support mechanisms were important as those involved were expected to
have little working knowledge of CVEs and we remained unconvinced about the ease
-
of
-
use
of the CVE interface. Over a three day period we
provided participants with a server and a
number of machines running a CVE client and other relevant software, and these were
provided within an Art & Design studio scenario with which participants would be well
acquainted.


We decided to use a VRML based
CVE as from a technical perspective this offered
adaptability, stability, a fully defined specification and a wide variety of resource material.
We considered VNET, Community Place and Blaxxun servers and settled on the Blaxxun
server and client. Despite
its complex interface, involving a series of HTML frames and
embedded files, Blaxxun offered a stable server with support material and examples. It
would also allow participants to upload a VRML world which could then be immediately
shared.


By deciding

to provide a high level of control for the participants (i.e. access to the server),
we incurred a security risk by providing access to configuration files and scripts from where
it is relatively simple to halt the server. We chose to trust that particip
ants would not wish to
jeopardise their own work by making untried, frivolous or malicious changes. Anyone
accessing the experiment live from outside the working group (i.e. via the World Wide Web)
was not given write permission.


Trust was also a fundam
ental consideration when planning the physical setting of the
workshop. Since the workshop was intended to investigate collaboration, we conceived of
this as both 'real' and 'virtual'. CVEs are generally investigated from the point of view of
remote workin
g , but we believed that our participants would benefit from meeting each other
face to face and also from meeting us. We believed this would encourage meaningful
dialogue between participants and any trust and rapport that formed would be transferred into

working in the 'virtual' world. This scenario would also help provide a support mechanism for
those who lacked confidence technologically. We cleared a clay workshop and used the
dividing boards and desks common to Art & Design studios to break up the sp
ace into a
variety of re
-
arrangeable niches for computers and people.


In order to help the participants navigate around the CVE, we provided an initial virtual
structure for the participants which reflected the real environment spatially. There was a
glo
bal public world which one entered and from which one could claim one's own private
space. Participants could enter their space through a 'gateway' by moving into the proximity
of some kind of 3D representation of their individual space (a 3D icon), rather

like moving
around the real studio and then entering the divisions of space that belonged to themselves or
others. From this point the real and the virtual began to diverge.


We found that in terms of navigation the simplest arrangement of the 'gateways'

was to place
the 3D icons at ground level, in a line immediately in front of the viewer as they entered each
space. It became apparent during the workshop that people were creating far larger and
complex worlds than we had anticipated, resulting in the s
ize of the public home world
32

becoming so large that it caused system crashes. A separate file was therefore created in
which participants could place a small representation of their private space.


On their desktop participants were provided with a text
editor and various VRML builders
such as Spazz3D and CosmoWorlds in order that they have as wide a choice of software as
possible to create their world. Once the participants had constructed a VRML world they
uploaded it using a CGI script provided. Once

uploaded the participant could view their
creation in the Blaxxun client. In an adjacent frame we provided a set of files with examples,
information about Babel and construction information.


An explanation of the technology took up most of the first day

of the workshop. We
considered that it was very important for the participants to begin building their VRML as
quickly as possible and so they were encouraged to make simple pieces of geometry by
whatever means and upload them to the space they had chosen
. In both the 'real' and the
'virtual' environment more spaces were provided than were required so that the individual did
not feel forced into collaboration unless they chose it.


By day two participants began to understand the limits and possibilities
of the virtual spaces.
Those with less experience tended to construct less complex ideas and to stay away from
complex scripting. They tended to work in an organic manner responding to the tools that
they were using, rather than implementing a fixed idea.
By this time it had become clear that
some of their original ideas had been modified in light of what they felt that technology could
support. At the end of day two we gathered to review the progress made. It became clear that
some of the participants did

not have an understanding of the difference between a
collaborative and a stand
-
alone VRML world or, if they did, they did not create a work which
took advantage of this aspect of a CVE. We suspect that this was due to two reasons. Firstly,
as suggested,
the participants did not understand that in a VRML CVE they have the ability
to create behaviours and events which all participants present in a world can see at the same
time. Secondly the participants, if they did understand the principle of shared event
s, they did
not always realise that they had to script these events so that they could be shared.


By day three participants began to try to add events and to implement the more complex
elements of their work, and they hit problems. We had underestimated
the level of technical
support which would be needed to deal with this, even for such a small group. Although most
individuals were by now comfortable with creating the geometry, the complexities of
scripting and passing events meant that much of the suppo
rt group's time was spent
answering what we had previously considered to be relatively simple questions. To
compound this, technical problems began to occur as the server struggled to update the work.
Participants became confused as to which version they w
ere viewing, due to caching
problems on their browsers.


The event ended on a high, however, with all participants gathering to review the work that
had been created. This virtual meeting of all the participants in the space was so successful
that perhap
s we should have introduced this from the beginning of the workshop as a starting
point to collaboration. If we had explained shared events and the scripting involved
immediately after participants had had a chance to familiarise themselves with the softwa
re
and environment, we might not have had so many problems. Only at this point did the extent
of the problems of calling in files as links become clear as the global entry world became an
unintended conglomerate of all the worlds, a kind of unplanned virtu
al collaboration.

33

Appendix A.7

User comments from the NVRCAD Project 'Farewell to Meat' (Virtual
Carnival)


The following is edited quotes from some of the participants.


The Carnival did succeed in generating a sense of 'community':

"I was surprised at

the level of bonding that happened between the group over the
session. I feel that this was strengthened by the activity of moving onto different worlds as

a group. This seemed to help establish a group identity."

"…the sense of people gathering from all
different (real) locations was strong

it really
felt like gathering together and meeting new people"

"The idea of going from location to location was a very good one

and gave it a sense of
procession."


While the intention was not to be prescriptive and
to allow the event to form itself, this was
not very successful and nothing happened to give a sense of 'event' (perhaps because no
-
one
designed any floats, so there were only avatars):

"I also feel that there was much confusion at the beginning, there see
med to be no sign of
a carnival what so ever. Floats may be an idea if the experiment is conducted again."

"Initially, no
-
one seemed to know what was going on and people were asking 'Where is
the carnival?'. I think, if we were to do it again, then we wo
uld need to build several
large floats that would provide clear focal points in the world."

"Even though the world were geographically small people still could not meet. Everyone
was there for the Carnival and yet no coherent structure appeared. In the rea
l world
there is an established power structure from organiser to participant because of the lack
of physical constraint its application in the CVE is not possible."


In addition to building floats, it seems that the idea of a fixed route could be benefici
al.

"… like having moving viewpoint (see in the Funfair world you will find that you can get
on the Ferris Wheel). Instead of making a choice we could extend this to have an invisible
train that people can get on and which carries them along the procession

route"


Some users seemed to lack basic knowledge of navigation within VRML worlds.

"People did not seem to be aware of the functions of the browser such as the 'beam to'
function."

"A lot of users were clearly novices in the VR world. Maybe this is inev
itable, but it
would be nice to think there could be greater knowledge about using the interface."


The technology itself was found to have some limitations.

"…difficulties with the interface and balancing chat and navigation meant that people
found it dif
ficult to orient and organise themselves."

"I found it frustrating that I could not visually put peoples handle / nickname to their
avatar. This meant that often I didn't know if I was actually talking to the person in front
of me or not"


Reports on netw
ork/processing speeds varied.

34

"…at times it was overloaded just doing the rendering. I couldn't get in to move my
avatar or even type any chat. This was particularly bad at 'Stonehenge', where I never
got away from my original position (for 15 minutes!).
"

But, also,

"I found that the technology and software worked 100%"

"My laptop has GL support and therefore the 3D rendered very fast and that it played the
audio in the Stonehenge world. The software was very robust."


A possible solution was suggested,

"
In Blaxxun there is a cache setting that allows you to specify that any downloaded
version of VRML geometry is only checked for once every 7 days. By choosing this setting
any visited site or loaded avatar would be retrieved from the cache and not from the

server (If people visit the event prior or we have a download time at the start of the
event)"

35

Appendix A.8

Survey of User Interfaces with Gesture Input


Put That There

This was the first implementation of a gesture
-
based system and was
produced at the
MIT Media Laboratory. The user sat in the centre of a specially equipped
room in which the objects appeared on a wall
-
size display. Through a combination of
pointing and speaking, the user could name, move, copy and erase objects (Bolt 1980). It was
the f
irst demonstration of the feasibility of a system based on spoken instruction and pointing
gesture.


GlowFlow

and
VideoDesk

GlowFlow

was an aesthetic light
-
sound environment controlled
by computer and driven by participant actions (Kruegger 1983). The
G
lowFlow

was actually
an empty rectangular room with a display (consisting of a suspension of phosphorescent
particles in water) which was turned on by floor pads on which the participants could walk.
Later, Kruegger developed
VideoDesk

(Kruegger 1991) whic
h consisted of a table with a
camera mounted above it. The camera was aimed down towards the user’s hands, which were
always resting on the desk’s surface. The silhouette images of the hand appeared on a
monitor on the far side of the desk and could be sup
erimposed on any application. The user
could perform any operation that could be accomplished using a mouse.


ATARI

A flexible user interface which uses gesture inputs was developed at the ATARI
Cambridge Research Centre in the 1980s. The device was based

on a single
-
finger pointing to
manipulate displayed virtual objects and was developed for the ATARI range of family
computers.


University of Illinois

At the same time there was work at the University of Illinois on a
graphical interaction technique ba
sed on gestures (Konneker 1984). This ‘simple’ system
recognised gestures performed with a pointing device. Text editing was suggested as a
potential application.


VPL

The first hand
-
input interface was suggested as a possible application of the
Z
-
Glove
,
the forerunner of the VPL
Dataglove

(Zimmerman et al 1987). It is the availability of the
DataGlove

that made most of the later research into hand gestures possible.


Carnegie
-
Mellon University

Research into the human aspects of gesturing was carried o
ut at
Carnegie
-
Mellon University (Hauptmann 1989). The experiments were based on a set of
video cameras filming users and transmitting pictures of their hands to a hidden human
operator. The research suggested that there are “intuitive, common principles i
n gesture
communication”, that there are no experts in gesture communication, and that gestures are
equally accessible and usable by virtually all computer users.


AT&T Bell Laboratories

At AT&T Bell Labs
9
, a synthetic visual environment was
developed co
mbining hand and voice input (Weimer et al 1989). By using the recently
developed
DataGlove

they were able to bring a 3D object into the visual space of the user
and to include it in the user’s hand space. This is the first project where a computer model o
f
the user’s hand is continuously displayed. The user can operate a control panel, select objects,
and perform basic operations.





9


Now Lucent Technology.

36

Rutherford Appleton Laboratory

At RAL, a window manager has been set up with hand
gesture control. Eight gestures were defin
ed and there are future plans to integrate speech
input. For example, a flat hand on a window would move that window around the display
area and a flat hand moving backward would bring a window from the background to the
foreground.


Whole Hand Input

Thi
s MIT
-
Media Lab project investigated the use of hand gestures as the
sole input channel (Sturman 1991) and describes a taxonomy for hand gestures dealing only
with how they are performed. In this project, gesticulations and direct manipulation of
widgets
are the main applications proposed and implemented (e.g. the control of a six
-
legged
robot, a construction crane and an animated muppet). There is also an attempt to develop a
music instrument, based on hand
-
input.


Glove Talk

The
Glove Talk Pilot Study

uses its own set of gestures which were chosen for
their easy classification (Fels 1990). It uses an interface composed of a VPL
DataGlove

and a
DECtalk

speech synthesiser to implement a hand gesture to speech synthesis system.


Charade

As with many ot
her projects,
Charade

is based on a VPL
DataGlove

and a
polhemus 3D tracker, connected to a Macintosh running a real
-
time algorithm. The system
was designed with an active zone in the form of a projection of computer display on a wall.
Gestures are only in
terpreted if the user’s hand is pointing towards the projection. All
gestures are defined as being a series of 3 postures i.e. a start
-
posture, a hand displacement
and an end
-
posture (Baudel et al 1993).


Visual Gesture Recognition

In this project the au
thors use a non
-
standard set of gestures,
selected from the American Sign Language but performed single handed and slightly
modified (Davis et al 1994). It is essentially a computer vision gesture recognition method,
which allows users wearing a specially
marked glove to interact with a computer system.


Finally, different systems have been developed for a wide range of applications and products.
N. Hataoka works on a multimodal interface based on speech and pointing gestures that has
been applied to a hous
ehold interior design system. P. Pook has developed a gesture based,
robot control system in which the user performs pre
-
recorded gestures, which are then
translated into robot motions. Mitsubishi have developed a simple interface for domestic
electronic
appliances based on the manipulation of virtual menus appearing on the TV screen.
This is a first attempt to introduce a commercial product that includes hand gestures and
includes a system to track the user’s hand using a video camera.


37

Appendix A.9

User

Notes for the Visual Assistant


The Visual Assistant is software to support visualisation in the field of theatre and
performance. It is designed especially for casual computer users. It currently runs on Apple
Macintosh computers only, and comprises ab
out 20,000 lines of 'C' code developed using
Metroworks CodeWarrior. A PC version is under development using Borland C++ and
should be ready during 2000. The latest version (February 2000) is v.1.5.


The VA presents the user with a darkened area repres
enting a 3D performance space. The
dimensions of this space are 1024 units wide x 1024 units deep x 768 units high, and the
viewing position is 768 units in front of the nearest edge. 2D image objects can be brought
into this space and moved around: as t
hey are moved backwards they will appear smaller and
become obscured by objects in front of them. The space may also be viewed from directly on
top (i.e. a floor plan). Other viewpoints can only be seen by saving the model and viewing it
in a VRML browse
r.


The VRML version can be viewed, on
-
line or off
-
line, by using a VRML plug
-
in to an
Internet browser. A free beta
-
version of a VRML plug
-
in for the Apple Macintosh was, at
the time of development, obtainable from <cosmosoftware.com>. It is advisable t
o have
64Mb of RAM and to set the memory requirements of Netscape to at least 25Mb to get
acceptable performance from this VRML browser.


The application always fills the entire monitor (giving the impression of a darkened theatre
within which objects and
lights are introduced) and will resize the world to the available
space. Saved folders are typically around 1Mb each and this includes both a model for
loading into the Visual Assistant (stage.va) and a VRML version of the model (stage.wrl plus
a number o
f GIF and JPEG files). There are plans to enable the VRML output from the
Visual Assistant to be transferred to the virtual
-
Stages
10

software being developed by Chris
Dyer (Dyer 1999) where it can be gradually converted into a working 3D model of a set
loc
ated within a particular theatre.


Free downloadable versions are available from the project's web
-
site
<www.esad.plym.ac.uk/va/> where there are also examples of models developed by users.
The Apple executable version currently occupies under 600K, makin
g it transferable on
floppy
-
disk if desired. To run it needs at least 8Mb RAM. The VA runs on any PowerPC or
G3 processor but it will not run on the older 68000 series models (Classic, LC, Performa).
Graphics performance will depend upon processor speed

and VRAM (233 MHz with 4Mb
VRAM seems to work acceptably).


The tools available in the Visual Assistant are listed below.





10

Soon to be re
-
launched as Open_stages.

Tools for creating new objects are:



import image from an existing file


create a rectangle


create a circle


create a per
son using a standard
image


draw a straight line


draw freehand with a pencil


paste a previously copied image
(from another application)

38


make a duplicate of an existing object

Tools for moving and resizing images



move an object horizontally
and
vertically (do not change depth)


move an object horizontally and
change depth (but not height)


resize an object

Tools for editing images:



colour image keeping contours (using
chosen colour)


make into a silhouette


make lighter


make d
arker


on/off transparency (make all white
pixels transparent)


use an eraser on an image


flip the image left
-
to
-
right


flip the image top
-
to
-
bottom


turn the image into a pattern that
covers the object


resize the pattern


delete an object


Other Tools include:



copy a colour from the screen


export an image in JPEG


project an object onto the floor


group two or more objects


play a sequence of settings


get information about an object or
scene


read/write a text note for a
n object or
scene


switch the light filter on/off and set
features


move the light filter



39

References


Argyle, M. (1975)
Bodily communication
, second edition. Methuen & Co. Ltd., London, UK.

Ashdown, N. (1996) The Virtual Reality Modelling Langua
ge in Art & Design Higher Education. SIMA
Report No. 17. JISC Advisory Group on Computer Graphics, Loughborough.

Baudel, T. and Baudoin
-
Lafon, M. (1993) Charade: remote control of objects using free
-
hand gestures.
Communication of the ACM

36
(7) pp. 28
-
35.

Beach, R. & Birtles, L. (1999a) Building Babel II: Construction in Virtual Environments. Proceedings
Computers in Art and Design Education '99 (CADE'99), University of Teesside, April 1999, pp. 211
-
14.
See also: http://vr.tees.ac.uk/rachael/articles/
babel/babel.htm

Beach, R., & Birtles, L. (1999b) Supporting Creative 3D Computing in the Art and Design Community. Proc
IDATER 99, International Conference on Design and Technology Educational Research and Curriculum
Development '99'. Loughborough Univers
ity, June 1999, pp. 206
-
212.

Beardon, C. (1997) Strategies for the Effective Use of Computers and Communications Technologies in Art &
Design. JISC Advisory Group on Computer Graphics, Technical Report No. 31. 28pp. (ISSN 1356
-
9066)

Beardon, C. (1999)
The design of software to support creative practice. Roberts, P. & Norman, E. (eds)
IDATER 99. Proc. IDATER 99 Conference, University of Loughborough, August, 1999, pp 29
-
34.

Beardon, C. & Enright, T. (1999) Computer
-
based improvisation: some experience
s of using the Visual
Assistant in teaching. Digital Creativity, 10(3), 153
-
166.

Benford et al. (1997) Embodiments, Avatars, Clones and Agents for Multi
-
User, Multisensory Virtual Worlds,
Multimedia Systems, Springer
-
Verlag, Berlin Germany.

Birtles, L. (1
997) 3D on the net. Digital Creativity, Vol. 8, No.3/4, pp.168
-
172.

Bolt, R.A. (1980) Put that there: voice and gesture at the graphics interface.
Computer Graphics

14

pp.262
-
270.

Boucher, A. (1997) An investigation into constructing virtual ceramics.
http://www.esad.plym.ac.uk/NVRCAD/projects/ceramics/report.htm

Boyd Davis, S. (1996 ) The design of virtual environments: with particular reference to VRML. JISC Advisory
Group on Computer Graphics, SIMA Report <http://www.agocg.ac.uk/reports/virtual/vr
mldes/title.htm>

Capin, T.K. (1997) The HUMANOID environment for interactive animation of multiple deformable human
characters, IEEE Computer Graphics and Applications, March 1997,

CHEAD (1995) Information Technology Survey, September 1995. CHEAD, Winch
ester, UK.

Costello, P.J. (1997) Health and safety issues associated with virtual reality: a review of current literature.
JISC Advisory Group on Computer Graphics, Technical Report No.37.
<http://www.agocg.ac.uk/reports/virtual/36/report37.htm>

Costel
lo, P.J. and Bee, S.T. (1997) 3D Model Databases: The Availability of 3D Models on the World Wide
Web. JISC Advisory Group on Computer Graphics, Technical Report No. 36.
<http://www.agocg.ac.uk/reports/virtual/36/report36.htm>

Cruz
-
Neira, C., Sandin, D.J
. and DeFanti, T.A. (1993), Surround
-
screen Projection
-
Based Virtual Reality: The
Design and Implementation of the CAVE, in Proceedings of ACM SIGGRAPH '93, Anaheim, CA.

Davies, C. (1998) Osmose: notes on being in immersive space. Digital Creativity 9(2)

65

74.

Davis, J. and Shah, M. (1994) Visual gesture recognition.
IEEE Proceedings Visual Image Signal Processing
141
(2) pp.101
-
106.

Dyer, C. (1999) Virtual_Stages: an interactive model of performance spaces for creative teams, technicians and
students.
D
igital Creativity

10
(3) 143

152.

Earle, N. & Beardon, C. (1998) The Role of Obligation within Virtual Encounters. Paper presented to the
'Collaborative Virtual Environments' (CVE'98) Conference, Manchester, June 17
-
19, 1998.

Edgar, R. (1998) Teaching VRML
and Java in Art and Design. Digital Creativity, Vol. 9, No.3, pp.182
-
4.

Edgar, R & Salem, B. (1998) VRML Multi
-
user Environments. Outline, Issue 6, Autumn, 1998, pp.32
-
4.

Fels, S.S. (1990) Building adaptive interfaces with neural networks: the Glove
-
Tal
k Project study. PhD Thesis,
Department of Computer Science, University of Toronto, Ontario, Canada, unpublished.

Hauptmann, A.G. (1989) Speech and gestures for graphic image manipulation.
Proceedings of CHI'89,

May
1989, pp.241
-
245.

Konneker, L.K. (1984
) Graphical interaction technique which uses gestures.
Proceedings of the IEEE First
International Conference on Office Automation
, New Orleans, LA, USA, pp. 51
-
55.

40

Kruegger, M.W. (1983)
Artificial reality
. Addison
-
Wesley, Reading, MA, USA.

Kruegger, M.W.

(1991)
Artificial reality II
. Addison
-
Wesley, Reading, MA, USA..

Levy, P. (1997) Welcome to virtuality. Digital Creativity 8(1) 3

10.

Reast, G. (1997) Provisional Results from the CTI Art and Design Questionnaire. Outline: the CTIAD
Newsletter Issue 3:

June 1997

Sturman, D.J. (1991) Whole hand input. PhD thesis, Media and Arts Section, School of Architecture and
Planning, Massachusetts Institute of Technology, MA, USA, Unpublished.

Tovey, M. and Harris, G. (1999) Concept design and sketch mapping.
The
Design Journal

2
(2) 32
-
42.

UCAS (1997) The Big Official UCAS Guide to University and College Entrance 1997. UCAS & 'The
Independent'.

Weimer, D. and Ganapathy, S.K. (1989) A synthetic visual environment with hand gesturing and voice input.
Proceedings of

CHI'89
, May 1989, pp.235
-
240.

Zimmerman, T.G., Lanier, J.Z., Blanchard, C, Bryson, S. and Harvill, Y. (1987) A hand gesture interface
device.
Proceedings CHI + GI '87
, pp. 189
-
192.