Professional Documents
Culture Documents
Development Guide
A guide to understanding and producing for
Group Interactive Immersive Environments
Rodney Hoinkes,
Rob Krieger, Editor
Immersion Studios Inc.
Interactive Content Development Guide
Assistance for this project provided by the Minister of Canadian Heritage and the
CANARIE Applied Research and Interactive Media program
For the latest updates and revisions to this document, please visit http://icdg.imm-
studios.com/
All brand names and product names mentioned in this book are trademarks or
service marks of their respective companies. Any omission or misuse (of any kind)
of service mark or trademarks should not be regarded as intent to infringe on the
property of others. The publisher recognizes and respects all marks used by
companies, manufacturers, and developers as a means to distinguish their products.
Immersion Interactive Cinema and Poetic Dimensions are trademarks of Immersion Studios.
Director, Flash, Lingo, ActionScript and ShockWave 3D are trademarks of
Macromedia. PhotoShop is a trademark of Adobe. Gameboy and Gamecube are
trademarks of Nintendo. Playstation2 is a trademark of Sony. Internet Explorer,
PowerPoint, Xbox and DirectX (DirectSound) are trademarks of Microsoft. IMAX is
a trademark of IMAX Corporation. Maya is a trademark of Alias|Wavefront. 3D
Studio MAX is a trademark of Discreet. Digital Fusion is a trademark of eyeon
Software Inc. Creator is a trademark of Paradigm-Multigen. Vizx3D is a trademark
of Virtock. Designer is a trademark of Cult3D. Web 3D Repository is a trademark of
the Web3D Consortium. Modelbank is a trademark of Viewpoint. AutoCAD is a
trademark of Autodesk. GeForce2 is a trademark of Nvidia. RADEON 9800 Pro is a
trademark of ATI. Onyx4 UltimateVision is a trademark of SGI. PhotoModeler is a
trademark of Eos Systems. Polytrans is a trademark of Okino. PolyTRIM is owned
by Centre for Landscape Research, University of Toronto. ArcInfo is a trademark of
ESRI.
Contents
Acknowledgements ...........................................................ix
Chapter 1 Introduction
Introduction ............................................ 1-1
Content Domains....................................... 1-4
Background ............................................. 1-4
Who should read this guide? ......................... 1-6
How should this guide be read?...................... 1-7
Chapter 4 Technology
Introduction ...................................... 4-1
Software Systems ...................................... 4-4
Coordination System .................................. 4-4
Presentation Systems ................................. 4-4
Interactivity Systems .................................. 4-6
Data Management System ............................ 4-7
Facility Management System ......................... 4-7
Media Control Components........................... 4-8
Core Components ...................................... 4-9
Show flow.........................................................6-2
Audio/Video segments..........................................6-3
Conditionals ......................................................6-4
Group objects....................................................6-5
Device control ...................................................6-6
Unique production components ..................... 6-7
Show Flow ........................................................6-7
Creating and managing the show flow .................................. 6-9
Chapter 8 Conclusions
Conclusions ............................................. 8-1
Findings ................................................. 8-2
Future Research........................................ 8-5
Benefits of this guide ................................. 8-7
Acknowledgements
It is critical to extend my warmest thanks and appreciation to all those
who contributed to the contents and editing that have made this guide
possible. Of particular note is my editing partner on this project Rob
Krieger, who kept the ball rolling, stick poking, and editing moving
during the trying moments.
The larger group is recognized more fully in Appendix E, but I would
like to single out Shiraz Baboo, Arkyin Eng, Chris Quine, and
Matthew Goldblatt, of Immersion Studios and Tim Archer of Masters
Workshop, who took it upon themselves to travel the arduous path of
editing and contributing critical technical components of this guide.
I would like to thank the full Immersion Studios team who have over
the years achieved wondrous works of creativity and technical
brilliance that have ‘made it all come together’ and inspire others with
interest in this New Media approach.
Finally, I would like to thank my partner in founding Immersion
Studios and its CEO, Stacey Spiegel, who has been a constant source
of inspiration, challenge, and creativity in the evolution of these ideas
from day one. None of this would be possible without you!
Rodney Hoinkes, CTO
Immersion Studios Inc.
Introduction
We see the evidence of advances in content development using digital
tools every day. Fantastic worlds, battles, and special effects are now
commonplace in movies. Game developers routinely create richly
detailed and broad networked virtual worlds. Rich information spaces
multiply on the web daily. This content has led to and now serves
well-established distribution channels and formats like movies,
television, the World-Wide Web, traditional PCs and game systems.
Many books, courses, and online resources are also available to help
people express themselves in these digital media environments.
This guide focuses on an expanded set of ‘New Media’ or ‘eContent’
technologies that are starting to have a significant impact on the ways
we work, learn and play together. The expanded set of technologies
includes immersive environments, group interactivity, and rich media.
To understand why we would focus upon these capabilities, let us look
at each independently for its potential value:
Immersive environments. Being immersed in a setting, topic,
or experience has been clearly noted in research as improving
the sense of ‘presence’ which in turn has shown value in
focusing attention, amplifying perception, and expanding the
context of inquiry.
Group interactivity. Most complex situations in the world
today are structured, designed or decided upon by groups or
teams. Humans are largely social beings who gain comfort,
support and challenge from others around them.
Rich media. The growing capabilities of digital
representations have helped broaden our awareness of concepts
such as learning styles. This has demonstrated the value of
using different representational forms that allow for inclusive
situations versus language-dependant exclusion (e.g. legalese,
computer jargon).
For the past eight years, Immersion Studios has been a pioneer in the
field of immersive and group-interactive eContent. It has produced
over 14 large-scale, immersive group-interactive content productions
and developed a high-end intranet “theater” product to run and create
advanced eContent. It has also installed over a dozen group interactive
immersive environments (theatres, classrooms, etc.) worldwide in the
past four years. The popularity of its content and approach to advanced
eContent has led to a number of post-secondary curriculum courses,
the technology being used as a foundation for university research
Content Domains
This guide focuses on two content domains of emerging value to new
media content production. The first, is the educational context, in
particular the domains of “experiential learning”, “informal learning”,
and “simulation learning.” In this context, the underlying systems are
typically formalized and scripted with clearly targeted learning
objectives. The technology is often based upon pre-rendered computer
graphics for accuracy, perceptual quality (important for user
engagement), and commercial viability (in comparison to other
products/experiences).
The second content domain is focused upon professional applications.
Professionals are using immersive, group, and rich media
environments primarily to support decision-making. The
representation and engagement of political, financial, cultural issues
and multiple points-of-view are critical to this group. The underlying
systems must be more dynamic, highly iterative and extremely
interactive. The technology is typically based upon real-time computer
graphics for malleability.
Each of these domains focuses upon particular technology types,
interaction domains and user expectations. We focus upon these, as
they have been the early adopters of GIIE-type experiences. These
types of user contexts contain many of the same qualities that are
present and needed in more generalized content domains such as
entertainment, training, and simulation, clear future market domains
for these concepts.
Background
There are no development guides of this type to help developers
seeking the limits of high-end richly integrated and interactive
production. This guide focuses on the ideals of production that will
reach a significant, compelling, and truly useful level, rather than
simply what can readily be achieved using conventional digital media
approaches today. It establishes a bar for future development in
specialized venues and the challenges posed to wider distribution
through broadband networks.
Currently, GIIEs are limited to specialized venues. These venues are
usually found in museums, science centres and schools whose
primarily objective is educationally-oriented. Within the professional
and research realms, GIIEs have primarily been located within design,
planning and environmental-studies related universities and
organizations.
Audience Benefit
Educators Provides a model for future new media
applications to apply to education that
have proven learning value.
Supports ongoing research into rich
enabling technology.
Students Provides directions in new media that
may transform future professional’s
approaches, tools and capabilities.
New Media Demonstrates new approaches that have
Producers current market value and future
networked distribution opportunities.
Provides useful notes on new media
adaptation to broadband internet
distribution.
Technology Provides approaches to technology
Developers development, challenges, future
opportunities and areas of needed
research.
Suggests changes to development
methods.
Researchers Provides focus for future content and
broadband technology research.
Identifies key areas requiring future
research in perception and cognition.
Business Suggests opportunities and challenges
Development for future markets and distribution
Managers channels.
Suggests levels of technology and
production to support markets.
Provides information on end-user
expectations.
Understanding immersion
What does Immersion mean?
“If you have ever settled into a book for the evening,
leaving your body to occupy that of the protagonist, you
know that immersion is powerful.
If you have ever played make-believe with a child,
delving into the drama of the quest, you know that
immersion is fun.
If you have ever entered “The Zone” while crafting a
piece of prose, poetry or programming, you know that
immersion is productive.”
Linda Jacobson, Immersion Quest1
1 http://www.abbedon.com/electricminds/html/edg_vr_1668.html
Sensory Data
Immersion Immersion
Hollywood
Documentary
Blockbuster
2 Oetterman S (1997). The Panorama: History of a Mass Medium. Zone Books MIT Press,
Boston
6 Robbins, Edward. April 1994. Why Architects Draw. MIT Press (Cambridge, MA) ISBN
0-262-18157-6
7 Grasha, A. F. (1996). Teaching with style. Pittsburgh, PA: Alliance.
12 Diaz, D.P. & Cartnal, R.B. (1999). Students’ learning styles in two classes: Online
distance learning and equivalent on-campus, College Teaching 47(4), 130-135.
Model Description
Library A few people place material in a repository,
many draw on it.
Solicitation A few people place requests, many respond, for
example, a Request for Proposal system.
Team A small group working together on a project.
Community A Community of Practice or Interest, where
users seek out opportunities to share and help
each other in pursuits.
Process Complex systems that require combined skills
Support to handle exceptions.
13 Butler ,Timothy and Coleman ,David (2003). Suggest five fundamental models of
collaboration (figure reproduced from Collaborative Strategies newsletter).
14 Ouchi, William G. (1984). Teamwork and competition: how the Japanese computer
industry has developed. . Creative Computing Vol. 10, No. 8 / August 1984 / Page 145.
http://www.atarimagazines.com/creative/v10n8/145_Teamwork_and_competition.php
Example of a user being the first to discover something in this medical simulation
environment, gaining extra points and status. Discoveries in this environment are
shared with the team to encourage further inquiry and to reward this user (Immersion
Studios “Vital Space” IIC Show).
Within learning experiences, there is a greater divide as to the value
and role of competition, in particular with youth. Where winning is the
key goal of competition, only a few people can succeed while the
majority feels beaten and defeated. Where the focus can be upon the
means rather than the end (winning), many benefits can be observed:
Increased learning, performance or retention (simple drills or
speed-related tasks for personal improvement).
Measure oneself providing goals for improvement.
Sense of accomplishment through personal progress
measurement.
Monitoring method assisting in the identification of problems.
Cinematic
simple, 3x2 interactivity flow, where there may be three options that
each have two further options, the production effort can multiply
significantly.
Intro Intro
Conclusions Conclusion
15 Bitmaps and graphics that can be dynamically manipulated on top of running video.
Real-time
Face-to-face communications
Face-to-face meetings and interactivity have a great value as noted in
research cited in this guide. There is also evidence that the translation
to online environments is not without difficulty. This guide does not
directly answer the challenges to this aspect of the GIIE but does
identify a number of concerns for future evaluation.
Interactive transmission
Transmission changes within the interactive environment are largely a
pure technical challenge. Four key challenges exist in the broadband
transmission:
Firewall/NAT handling for security.
Bandwidth for transmission of interactivity data.
Latency in transmission of interactivity data.
Guaranteed concurrency of interactivity data/state.
Firewall/NAT (Network Address Translation), which hides a users real
address from the Internet, also means that establishing connections TO
a user is very tricky, and designs should be focused on a user
connecting TO the server (which is more standard in any case but this
is an issue). Also the use of not standardly used ports (80, 23 are
standard for web and mail) can lead to firewall issues and users
opening dedicated ports. Which ports are used for interactivity should
therefore be designed in a flexible manner and well documented for
users. This is also a problem for users who do not really know this
technical information, such that when the interaction does not work,
and they have to configure their own device (not easy or consistent
from one firewall/router to another), this may hinder adoption.
Interactivity itself does not need to be high bandwidth. Examples from
the ISI Sharks GIIE show clients requiring an average of 10Kbps for
both send and receive data in a typical 50-player 3d world. This does
not send all data about what someone is seeing, it only sends messages
that contain ‘changes’ (move item X 5m to the left) instead of sending
the whole item again in the new location (maybe thousands of
polygons).
Latency is the biggest issue facing broadband transmission of
interactivity data. If user selections and choices are not received in a
timely fashion, they may not be included where critically needed (may
come after a decision time) or users may be acting on an old state of
information (like the location of a target in a game). The sensitivity to
latency is usually a function of the type of content you are interacting
with. Reaction-based games are the highest demand on accurate and
rapid transmission of interactive data. For this type of content, many
systems include a path-prediction method allowing them to estimate
where things will be in the future even if they have not received
timeline updates. Latency measurements in the range of 100ms are
commonly desired for online gameplay. Decision-making
environments are usually less time-sensitive, where user reaction,
consideration, and feedback require more cognitive processing. In
these contexts, we have found that response (full round-trip) rates of <
500ms are typically sufficient.
Guaranteed concurrency is a final key concern with interactive data.
When the full final representation (e.g. a video) is transmitted you can
be more certain (depending on compression method) that the user will
be seeing something current. When interactivity data is transmitted
there are usually fewer feedback mechanisms to make you aware of
the current state (in the video example above you might see blocks
updating them becoming solid to know it is finished). This issue is
critical to assessment strategies of user learning success, and in
particular professional applications to know all users are ‘on the same
page’ in a discussion/revision.
Immersive transmission
Transmission changes within the immersive environment would
appear to pose one of the biggest potential problems for GIIE adoption
over broadband. Current three-screen immersive environments
Research to date
Previous research by Ritterfeld and Vorderer3 has directly focused
upon the Immersion Studios’ Cinema-based GIIE. That research has
indicated that enjoyment and presence, media competence and prior
interest, and perceived challenge and usability account for over one-
third of the educational impact of the immersion experience. Research
constraints did not allow for investigation of factors that might account
for the other two-thirds and that might include such things as prior
learning experiences, attention span, memory capacities, prior
knowledge, prior interest in science, among others.
3 Ritterfeld, U., Weber, R., Fernandes, S., Vorderer, P. (2003). Think Science!
Entertainment Education in interactive theatres. Computers in Entertainment (CIE),
Volume 2, Issue 1 (March 2004). Association of Computing Machinery. ISSN:1544-
3574.
Media determines
Usability
Competence
inhibits
Academic
Achievement
Educational
Challenge Presence Entertainment
Impact
Age
Prior interest
in Topic
4 Ritterfeld, U., Weber, R., Fernandes, S., Vorderer, P. (2003). Think Science!
Entertainment Education in interactive theatres. Computers in Entertainment (CIE),
Volume 2, Issue 1 (March 2004). Association of Computing Machinery. ISSN:1544-
3574.
5 Students’ prior interest in the topic does directly influences the educational impact, but
this has proven to be minimal in comparison to the role of entertainment.
Future research
To date research has found that the combination of effects in studied
GIIE educational content is a successful educational construct.
Furthermore, research performed in the development of this guide has
shown that broadband adaptation of this type of high-complexity
content need not critically diminish its successful perceptual goals for
most common network reasons. Future research into environmental
impacts is needed however to ascertain whether the social and
environment presence changes of a distributed group setting are
material to the broadest applicability of these concepts. This type of
research is both the most difficult but also the most insightful into the
fullest accessibility to this type of eContent.
Research to date
Within the realm of perception and cognition, most of the above
impacts have a role to play. As with the prior section on educational
GIIEs, we are focusing upon those aspects of user perception and
cognition in relation to content that may have the strongest roles to
play in the transition between Intranet and Broadband development.
Prior research has identified tools and representational issues, at least
for design and planning8. They have identified an essential base set of
tools for group interactive design inquiry:
Terrain tools (navigation over terrain, objects interpolating
over terrain).
Database access for non-visual attribute representation and
tracking.
Rapid 3D generation and ongoing manipulation of form from
2D data and attributes (parametric modeling using attributes
and dynamic number systems like a spreadsheet).
3D object geometric parameter manipulation (position, scale,
rotation, texture, colour).
Rapid kit of parts library of objects/forms to draw upon.
A similar set of tools can be defined for most disciplines. The
difference in a GIIE is the broad range of specific representations and
tools needed for a diverse team of participants. This interconnected set
of tools is not simply visual and thus points out potential problems for
certain network representational approaches that assume only visual
representations. An example of this type of problem can be seen in the
use of straight video compression (and sometimes scaling) to transmit
an experience from one location to another. These methods can have
significant impacts upon textual representations (documents,
spreadsheets, etc.) making them unreadable, and menus difficult to
use. For these situations, many video conferencing solutions offer a
high-resolution but low frame-rate alternative. Frame-rate is a critical
Future research
Current research has demonstrated a mixed view of professional
application of GIIE techniques from a group Intranet to a distributed
Broadband setting. A simple assessment of research to date might say
that the single-screen environment, which would be the most common
for users over Broadband, may be perceptually suitable. However
direct group interactivity that gets watered down in Broadband is
critical to many aspects of negotiation in these professional domains
Introduction
I think that we're also almost always dealing with
people for whom the computer is a bit of a mysterious
player in the decision-making process ... or the working
process. So they either have a sense that it's far too
complex and far too technical for them to really get
anything out of it. Or they have a feeling that somehow
it's incredibly automatic and to design something using
a computer, or to manipulate some big piece of
territory with a computer, is actually this kind of
automatic thing where you just press a button.
Coordination System
PD/AVS PD/GS PD/RS
(Audio-Video (Game (Registry
Sequencer) Server) Server)
ISSimEngine ISMultiSim
Facility
Management Core Components
System
ISNet
IsNetControl
ISSync
ISDevControl
PD/Multi-Flash ISWebControl
High-level diagram illustrates the components of a GIIE and their connection to one
another.
An eighth component, production support tools are described
separately in Chapter 6. These components are specific to the
Immersion Studios’ approach to GIIEs. Others may use a different
approach but we have found this organization useful for development
and ongoing research and production flexibility.
All of these systems form the complete Immersion Studio Inc. (ISI)
Poetic Dimensions (PD) software suite. PD's core technology supports
a broad range of capabilities, which in other commercial and research
systems are only available individually and are not as closely
integrated.
Many commercial systems often lack one or more of the following
important features:
Multi-user triggering of events.
Triggering of events on multiple computers simultaneously.
Triggering of device events (e.g. lighting, doors).
Look-ahead buffering to ensure seamless transitions in audio
and video playback, or real-time animations.
Linking of multiple installation sites and clustering of servers
to provide scalable group interactivity.
Software Systems
Coordination System
Some GIIE environments are completely free-flowing and some are
require time-based control over actions and interactions. Where time-
based coordination of interactivity with immersion is required, Poetic
Dimensions/Audio-Video Sequencer (PD/AVS) is typically employed.
PD/AVS provides the core control over scripted group interactive
environments. A coordination system controls the flow of interactivity,
timing of events including device changes (lighting, video source
switching) and synchronization with immersive presentation systems.
PD/AVS is an entirely network-distributed system built on the core
technology of ISSimEngine. ISSimEngine encompass a high-level
object collaboration system designed for handling complex group
interactivity. ISSimEngine supports group objects that do not require
extremely low latency interaction. To support group objects that
require low latency interaction, the PD/GameServer support
application is available. PD/GameServer provides support for high-
performance multi-player games.
ISSimEngine and PD/GameServer allow applications to operate on
shared objects without any special knowledge of, or technical concern
with network transmission and synchronization over hundreds of
individual users or group decisions. PD/AVS can also handle group
interactivity across multiple installations or large groups (greater than
100 users) using ISMultiSim. ISMultiSim acts as a synthesizer of
inputs and relay server amongst multiple remote servers.
PD/RS is a small support server which acts to connect different
interaction and presentation systems together as may be needed for
broad multi-site network environments where collaborations may not
be pre-defined. An application announces itself to the registry server
and requests a list of possible appropriate remote servers or clients.
PD/RS then supports the brokering of authority and permissions to
allow for a direct connection to be made between these systems.
Presentation Systems
In this guide, presentation systems are categorized as any systems that
render content for the participants, that is, primarily concerned with
representations used for the immersive aspects of the GIIE. While
these applications are specific to Immersion Studios Inc., the
Interactivity Systems
Generally, the PD suite is not the final software tool for controlling
interactivity. It is more a framework over which interactive interfaces
can be built. This framework controls core timing mechanisms, group
interactivity mechanisms, scoring systems, and shared object control.
This is accomplished with a set of group object technologies discussed
previously under the coordination system and transmitted through the
core networking components discussed later. To support a range of
richer media capabilities without extensive technology development, a
series of media control components are also available.
The general philosophy has been to permit customized interactive
development purpose-driven to pre-scripted content and user groups.
This used to be accomplished with an in-house tool called Personal
Media Navigator (PMN). PMN required custom training and extensive
development to support the range of interactions desired in GIIE
content. This production process was significantly improved by adding
a simplified framework that sits underneath and supports more
conventional software tools such as Director , Flash or Internet
Explorer programming. By using standard, component-based
approaches to the underlying technology (COM and Xtras) and a well-
defined event process to integrate the remote communications needed
for group interactive experiences the production process improved
significantly.
While Director and Internet Explorer provide mechanisms (Xtras and
ActiveX respectively) to act as containers, they can also communicate
directly with the entire PD suite. If you use Flash however, it must
communicate through Director or Internet Explorer at this time.
A number of projects are also underway that require higher-levels of
interactivity than can be supported through these interpreted
programming tools. In these cases, direct C++ coding is used on top of
the same core framework. This is particularly true where high-
performance, real-time 3D simulations and games are required which
are not well supported in Director or the other interpreted tools. The
availability of the core framework allows for simple use of the
appropriate 3D or other representational engines that might be desired.
The Immersion Simulation Lab (ISL) GIIE presented a different model
of an interactivity communications framework. In this model, users do
not have a pre-scripted set of interactive moves they can make and
thus a controlled interactivity environment would be too restrictive. To
accommodate this, the PD/Controller (PDC) software was created. It
sits on top of the same code framework but does not structure
Core Components
The PD software suite uses several core components shared across
most applications. ISSync coordinates interactive show playback and
changes to audio, video, devices, etc. with time-based effects. ISNet is
a foundation network communication system operating upon TCP/IP,
therefore suitable for Intra-, and Internet applications. It handles all
low-level messaging between applications within the PD suite.
Computer Systems
Most GIIEs are networked, multi-computer systems. The setup and
organization of computers for a GIIE require careful consideration due
to practicalities and technical limitations:
Immersive servers use CPU and graphics, sound, and disk
resources at maximum levels. Visual servers in particular
usually run in full-screen mode.
Coordination servers often use network bandwidth extensively
and advanced game servers in particular can generate intense
CPU activity.
Interactive consoles are usually interfaces used directly by
participants and require an individual computer for each
participant. Some designs allow multiple (2 to 3) participants to
share a console or pass one around a group (in a wireless
installation).
Immersive servers
Due to the high-performance demands of immersive servers, most
require a fully dedicated computer for each screen of display. This is
why multi-screen synchronization is so critical when working with
immersive servers. Using high-end, SGI super graphics workstations
that can effectively display across multiple screens on one computer is
one notable exception to this guideline.
Running multiple screens on one computer divides the overall
performance of the system so that a machine, which might be 10 times
faster than another machine, is only a little more than 3 times faster
when it must render 3 screens of graphics. This performance issue
accounts for the emerging popularity of PC computer graphics
clusters.
The common servers run on individual computers include
PD/HDRender, PD/DMS, and PD/RT and PD/SAS.
Coordination servers
Coordination servers are a variable demand resource depending upon
the number of simultaneous users. Resources limitations must be
monitored primarily for networking bandwidth and CPU
maximization.
Depending upon the number of users, the following servers can be run
individually or multiple on a single computer PD/AVS, PD/GS
(PD/GameServer and variants), and PD/RS (a simple registry server).
Interactivity consoles
The greatest variable in number of computers is those that serve as
interactivity consoles. To allow multiple users to simultaneously
interact with each other and their immersive environment, every so
many users need their own computer. In fixed environments, these are
often racked computers linked to ruggedized touchscreen LCD
consoles. In more flexible environments, laptops or tablet PCs are
often used, usually in a wireless configuration.
The common interactive console applications include PMN and
PD/Controller.
PD/HDRender
(Master)
PD/HDRender PD/HDRender
(Slave) (Slave)
ISMultiSim
PD/AVS PD/GS
PD/HDRender
(Master)
PD/AVS PD/GS
PD/DMS
PD/AVS PD/GS
PD/RT PD/RT
(Master) (Master)
PD/RT PD/RT
(Slave) (Slave)
PD/Controller
PD/RT
PD/Registry Server (Master)
PD/Controller
Team composition
In producing a GIIE, the complete production team is organized
around the goal of delivering the goods to the users. What makes
producing a GIIE unique are the different disciplines involved, as
follows:
Software Development Production.
Game Development Production.
Video/Film Production.
Animation Production.
As well, common across all of these disciplines is the role to manage
the production.
The guide focuses on the integration required in producing a GIIE
rather than the intricacies of the production process itself. In general,
the production process is described in other text and documents —film
and video production, animation production, software development,
game development and project management. What is really of interest
here is how the different disciplines interact and integrate with each
other —this being where the guide does focus.
Team organization
A complete production team consists of the following organization:
Management.
Design.
Artwork.
Software.
Audio.
Quality Assurance.
Support.
Management
The management team consists of a diverse group of individuals with
many different responsibilities. During production of a GIIE, the
management team:
Ensures the GIIE meets target-marketing requirements.
Develops a plan to meet these requirements.
Executes the plan on time and on budget.
Design
The design team forms the creative heart of a GIIE. This team
develops, nurtures and cultivates a vision for the user experience. In a
GIIE, the design team also contributes to content research,
development and management.
Artwork
The artwork team wears a host of hats, creating the 2D and 3D
graphics and special effects for a GIIE.
Software
The software team creates the code and realizes the vision of a GIIE.
Audio
In general, the audio team designs, develops and manages three basic
components:
Sound effects.
Music.
Voice-over.
Audio can be developed either by a single audio house or by
specialists in each area.
Quality assurance
Although the industry has no unified, best practice methodology to
assure a GIIE meets an acceptable level of quality, there are de facto
approaches that can be followed (for example, module, system, alpha
and beta testing). The quality assurance team ensures that these
approaches are applied throughout the production lifecycle.
Support
The support team provides and manages a well-defined, stable digital
environment within which the production team can communicate and
share the work.
Management-related roles
Management-related roles within a management team consist of the
following:
Producer.
Project Manager.
Project Coordinator / Production Assistant.
Technical Lead.
Visualization Professional (ISL initiative).
Producer
The Producer manages the production at the highest level from a
business, strategic, creative and technology perspective. From a
business and strategic perspective, the Producer negotiates contracts
and financing and also communicates with external investors and
funding bodies. From a creative perspective, the Producer ensures that
the team maintains a consistent project vision. To that end, they work
closely with the Project Manager and the technical and design teams.
From a technology perspective, the Producer ensures that the
technology decisions are sound.
Project Manager
The Project Manager acts as the “communication hub of the
production team”. The Project Manager is the main point of contact
for feedback on project progress and any other project-related
information. The Project Manager manages the production team
ensuring that the project is completed on scope, on time and on
budget. The Project Manager maintains a clear vision of the project
and keeps the production focused in that direction.
Technical Lead
The Technical Lead manages all technology aspects of production.
They define, design, construct and evaluate technical production
Visualization Professional
The Visualization Professional works with real-time, real-world
simulation exercises in areas such as Urban/Regional Design and
Planning, Oil Exploration and Training initiatives. The Visualization
Professional is the content expert to both the client and the production
team, and requires an understanding of the real purpose behind the
modeling/simulation, as well as the specific techniques to properly
construct, render and display models to assist clients in this decision-
making endeavour.
Design-related roles
Design-related roles within a design team consist of the following:
Director or Creative Director.
Writer.
Content Manager.
Writer
The Writer writes a compelling narrative with a high degree of
interactivity required of a GIIE. Working with the Director, the writer
defines the story arc, adds dialogue and action ensuring the writing is
consistent with the characterizations, and the interactivity fits within
the story.
Content Manager
The Content Manager defines, acquires and tracks all content media.
Sometimes working with a Content or Media Asset Researcher, the
Content Manager ensures that all content is scientifically and factually
accurate.
Artwork-related roles
Artwork-related roles within an artwork team consist of the following:
Art Director.
Storyboarder.
Concept Artist.
Interface Designer.
2D Designer.
3D Modeler.
Texture Artist.
Animator.
Compositor.
Video Editor.
Art Director
As the conduit between concept and design, the Art Director maintains
the design integrity and visual style within a GIIE. Working with the
rest of the design team, the Art Director is a design advocate and
driving force of innovation and quality craftsmanship. The Art
Director mentors and sets creative agendas for the design team,
pushing boundaries, exploring new technologies and ensuring the
creative concept is realized from a visual perspective.
As well, the Art Director develops appropriate design solutions. The
Art Director is generally skilled in graphic design, illustration,
multimedia design and both interface and interactivity design.
As the manager of the artwork team, the Art Director must ultimately
translate the Director’s vision into visuals.
Storyboarder
Working closely with the Director and Art Director, the Storyboarder
designs and articulates the scenes in a sequence for either or both the
Big Screen and Consoles. Once the storyboard is complete, it is
submitted for internal and external approval. At this stage, the Concept
Artist begins sketching the designs.
Concept Artist
The Concept Artist must accurately convey the design of a media asset
before it is created. Working from the storyboard, the concept artist
creates black-and-white sketches, which are submitted for both
internal and external approval. Once approved, the concept artist
applies colour to these sketches and submits them again for internal
and external approval. Only at this stage does production actually
begin.
Interface Designer
The Interface Designer, generally an expert in graphic design and
interactivity, designs and implements the graphical elements of the
user experience. Of all the members on the production team, the
Interface Designer needs to play the role of user advocate, ensuring
that each team is focusing on the user experience.
The Interface Designer must also maintain consistent user interface
quality across the entire GIIE.
2D Designer
The 2D Designer, generally an expert in elements of typography,
illustration, photography and printing, implements the look of a GIIE
under the direction of the Art Director. The 2D Designer develops the
overall graphic layout and creates graphic elements such as templates,
menus, buttons, typography and color schemes. The 2D Designer also
performs photographic and image manipulation.
As well, the 2D Designer produces high-resolution artwork for
marketing materials such as brochures, advertising, signage, posters,
slide shows and other forms of printed or graphic communications.
The 2D Designer is a “visual problem solver” within the artwork team.
3D Modeler
The 3D Modeler creates models of 3D elements, for either pre-
rendered (high definition with potential for photo-realistic rendering)
or real-time rendered (game play). These models range from
characters to props to environments. The 3D Modeler also creates
rendering effects to depict 3D scenes either photo-realistically or
within a prescribed artistic design.
Sometimes a Character Modeler is required as part of the CGI
component of the artwork team especially when either photo-realistic
Texture Artist
The Texture Artist textures 3D environments, props and characters
within a GIIE. The Texture Artist creates and applies believable
photo-realistic textures to 3D models and environments.
Texturing involves applying images onto 3D models constructed from
polygons. The Texture Artist is virtually a 2D Designer who “skins” a
3D object creating compelling images that are painted onto a 3D
model.
Animator
The Animator creates motion for 3D elements including animating
characters, props, cameras, lights and environments.
Compositor
The Compositor combines all the elements of a scene, layering them
together for final output.
Video Editor
The Video Editor assembles footage and audio to convey the vision of
the Director – piecing shots together into a complete production. The
Video Editor creates Lyca reels. A Lyca reel is a preliminary, timed
movie using scanned storyboard images. As they become available,
storyboard images are replaced by various slug-shots and then by low-
resolution complete shots. Eventually, the Lyca reel becomes a
complete, low-resolution version of the GIIE.
Software-related roles
Software-related roles within a software team consist of the following:
Interactivity developer.
3D Programmer.
Tools Programmer.
Interactivity developer
The Interactivity Developer plans, designs, programs and integrates
aspects of the interactive multimedia components of the GIIE.
3D Programmer
The 3D Programmer plans, designs, programs and integrates aspects
of the interactive 3D real-time rendering components for the GIIE.
Tools Programmer
The Tools Programmer plans, designs, programs and integrates aspects
of any software tools required for the GIIE.
Audio-related roles
Audio-related roles within an audio team consist of the following:
Audio engineer.
Composer.
Mixing engineer.
Audio engineer
The Audio Engineer plans, designs and records all audio components
(except music) of the GIIE including dialogue and sound effects. As
well, the Audio Engineer edits all audio components including music.
Composer
The Composer plans, designs, creates and records all the music of the
GIIE.
Mixing engineer
The Mixing Engineer combines all audio components of the GIIE.
QA Lead
The QA Lead defines and manages the testing and acceptance process
of the GIIE. As well, the QA Lead creates the specification for, and
manages the testing laboratory. The QA Lead coordinates the testing
of the GIIE with the other members of the quality assurance team.
The QA Lead proofreads all reported defects and discards duplicate
and erroneous reports, often rejecting reports back to the reporting
tester, requesting clarification and/or testing.
Integration Specialist
The Integration Specialist ensures that the show flow between the Big
Screen and the Consoles is correct during the GIIE.
Usability Specialist
The Usability Specialist ensures that the user experience is sound,
measuring the ease of use of interactive systems based on the five
well-known criteria: user-satisfaction, effectiveness, efficiency,
learnability, memorability. Throughout the lifecycle of the GIIE, the
Usability Specialist performs usability tests including focus groups,
interviews, surveys and paper prototypes. Note that for educational
contexts it is imperative that usability testing with the target market is
performed and completed before a product is released.
Support-related roles
Support-related roles within a support team consist of the following:
Technical Support Specialist.
Network Support Specialist.
Introduction
The following section focuses on the types of technology used to
produce the Immersion Studios’ Immersion Interactive Cinema (IIC)
product. Where this technology is similar to other industry approaches
and products, we try to present the key differences. The material
focuses specifically on the Immersion platform, however much of the
information can benefit those pursuing other technology solutions.
Key concepts
Before discussing the IIC production process in detail, it is important
to explain some key concepts. These are core components of a
Cinematic GIIE show that relate specifically to organizing the
interactivity of the show. These components are very similar to those
of most programming languages, and those familiar with any type of
programming should be able to follow these interactivity concepts.
Show flow
Cinema shows with a fixed time and story design usually require a
structured representation of the interactive show flow. This is also
required to allow shows to run unassisted (no live host required) but
with some control and coordination.
A structured representation is typically one of the following:
Software program designed to run and control a particular
experience.
This is a custom program with some of the highest flexibility
but typically requiring the greatest technical skill to construct.
A description language designed to layout a show.
Typically, the languages are tuned to minimize complexity to
the types of activities needed for a timed media experience.
The language uses the professional terminology of a content
producer as opposed to generic terminology in a software
programming language.
The new media industry is largely divided in its approach. Many
interactive experiences are developed in programming languages, even
though they may be fairly high level such as Director Lingo and Flash
ActionScript. One of the most common description languages for rich,
synchronized media is SMIL (Structured Multimedia Integration
Language). Even basic HTML now supports enough features for
describing a content flow by linking timed “pages.”
Each of these languages supports the following show flow concepts:
Time-based triggering of events.
User-based triggering of events.
Playback of rich media content.
The IIC uses the Poetic Dimensions/Audio-Visual Sequencer
(PD/AVS) description language for its show flow, as this is typically
easier to teach and learn than a full programming language. All of the
other common approaches listed above lack support for one or more of
the following:
Multi-user triggering of events.
Triggering of events on multiple computers.
Triggering of device events (for example lighting, doors).
Look-ahead buffering to ensure seamless transitions in audio
and video playback.
A more detailed discussion of these critical functions is provided in
Chapter 4 dealing with technical factors.
The PD/AVS description language supports the following key
concepts that are the most critical to understand:
Audio/video segments.
Conditionals.
Group objects.
Device control.
Audio/Video segments
In an interactive show, there is no single linear playback concept. The
show may branch and diverge or shift direction and timing based on
interactive events. The PD/AVS language lets you control by breaking
up aspects of each show into segments and only providing linear
timing within each.
Segments are typically defined by a sequence of time until the next
decision-point that will change the course of the overall show.
To those with programming backgrounds, a show flow can be seen as
a flowchart with each segment being a subroutine for the purpose of
logically organizing events in a show. Each segment ends with either a
link to a single other segment, or with a set of conditionals that allow
show flow to branch based on interactive inputs.
Within each segment, timed events can be inserted to control:
Devices (for example lighting, audio levels, electronic doors,
fog machines).
Special interactive events, if appropriate (for example special
sounds, unlocking features based on users achieving particular
goals).
Conditionals
As discussed under segments, events and changes in show flow can
occur through the use of conditionals.
A conditional is defined in PD/AVS as an event that occurs only if a
particular condition is met. In its current form, most conditionals are
simple numerical comparisons.
A simple example for an interactive decision on a right or left
direction might be:
if chooseLeft > chooseRight then goto segment “DriveLeft”
if chooseRight > chooseLeft then goto segment “DriveRight”
if chooseLeft = chooseRight then goto segment “DriveStraight”
Production Note: In a single-user application environment, the user
could simply be presented with two arrows, a left and
right, and therefore only those two conditions could
occur. In a multi-user environment, there is no control
that says the totals must be one way or another and
thus the equal conditions must also be considered and
handled.
Group objects
Interaction requires decisions to change actions. Within the IIC
system, decisions are made through conditionals. Conditionals require
an evaluation of data in the form of numeric variables. To support
multi-user group interactivity, these variables need to be shared or
synthesized across the network. This is accomplished through the use
of group objects in the IIC terminology.
These objects are globally defined, updated and synchronized. It is the
values in these objects that are tested within PD/AVS conditionals.
Each value that is needed globally or to be compared and updated
remotely is established as a group object.
Production Note: Group objects are only controlled in an absolute
manner through a group object server, not through a
client. While a particular client may set a group
object’s value, it could just as easily be set or
modified by other clients. Typically group object
values are managed in one of two ways – value
modifiers or shared value averaging. Value
modification is typically done through an increment
(which can be negative). For a straight vote, each
client might simply increment the selected object by
one (+1). If a dynamic vote were required, allowing a
user to “change” their vote until a time limit is met,
all subsequent votes would decrement (-1) the
previous vote and increment the new voted on object.
Another common group object management
technique is value averaging. This allows individual
clients to set an object value they might desire (for
example selecting how fast you would like the ship to
go.) The group object can automatically average these
Device control
Within the IIC, additional dimensions of an experience can be
manipulated through device control. This can control electronic
devices such as audio mixers, lighting systems, dry ice generators, or
water sprayers. This can also control networked computer systems,
stop or launch applications or set machine parameters (audio levels,
display brightness).
Device control can follow as a time-based event (turn off lights 30sec
into the show), or as a conditional event (flash the lights if the
submarine is hit and damaged more than 50%).
Immersiv e Immersive
Immers ive Immersive
Screen Screen
Video Audio
Interactiv e Playback
Content Content
Sprites QA
Show Flow
Show and G roup
Flow Interactivity
Creatives QA
Console
Console Console Console
Interactiv ity
Networking Content Interactivity
QA
Show Flow
Establishing the core show flow is one of the first key production
activities needed for an IIC show. A show flow establishes the key
segments of the production based on the interactive options desired by
the show’s creative staff.1
Segments in an IIC should first be broken up based on show
interactivity. Wherever you wish to offer choices to the audience is
considered an interactive branching point. For example, allow the
audience to turn left or right, pick an ending, or choose a hypothesis to
explore. At these points, the previous segment should be ended and
two (or more as needed) segments should be added to handle each
optional media segment.
Intro Intro
Conclusions Conclusion
STEP 1: Establish Rough STEP 2: Establish Final STEP 3: Establish Final Show STEP 4: Establish Final Show
Show Flow: Show Video: Flow (picture lock): Audio:
Rough Segments Shots Segments Match Show Flow Segments
The following illustration presents the steps to create and manage the
show flow. For each step there is one or more deliverables required.
Step 1 Establish Rough Show Flow: Rough Segments
Deliverables
1 Show Flow Diagram: establishes the number of segments in
the show and where branching occurs, as previously shown.
2 Show Segment Name / Length List: establishes the number of
segments in the show, the length of each segment and what is
shown on both the Consoles and the Big Screen during each
segment. Note that at this point in the project, the timings are
only approximate and as such are shown in seconds and
minutes. The following table provides an example of this list.
Console Networking
Console networking involves defining five shared show
communications types:
1 Branching conditional group objects.
2 Shared multi-user group objects.
3 High-performance game interactivity.
4 Scoring methodology.
5 Time synchronization events.
Scoring methodology
If a Game Server is used, it routinely handles and tracks scoring. In
cases where a special server is not used and a shared scoring system is
desired (where high scores need to be displayed), there are two
methods of tracking individual scoring:
1 Built-in server-side group object scoring.
2 Custom client-side group object recording of scores.
The group object server provides some common methods of assigning
and tracking scoring based on the following criteria:
Interacting with a group object.
Spending time interacting with a group object.
A score can be assigned to a group object, and any time a client
interacts with that object (changing its value – not simply reading
values), that client will receive a set score. This can be set to only give
a score the first time it is used or each time. The “first time” counter
can be reset within the show script. Secondly, a score can be obtained
by spending a sufficient time interacting with a group object. This is
used to represent cases where a user should get a score for sufficiently
researching a topic (spending time reading something).
If these automated methods are not sufficient, client-side (but shared)
scoring can be used by creating a discrete score group object for each
client. From that point the client can increment its score as is
appropriate to its local scoring algorithms.
Only through early sample testing on the target systems (or at least the
lowest performance ones) can sprite complexity be safely validated at
this time.
Updating sprites must be done through a game server, as these are
typically high-performance update items. Some game servers translate
between group objects and sprites in some cases without additional
coding, however, it is just as common to have to create custom code.
Console Content
When developing content-oriented interactive experiences (often the
most educational or accurate simulation-based), or any broad audience
experiences, interactivity and content will change often. It is best to
accept this from the outset and plan intelligently for it.
Content used in interactivity, for example, videos that can be looked
up, HTML information in a virtual guide, a picture of a creature, help
instructions or captions can take many media forms. A common
approach to organizing data and managing changes to it becomes
critical if, for example, user testing shows base assumptions to be
incorrect or science advisors inform you that your data is outdated.
To help manage this process, the following three standards are
routinely used in IIC production:
1 Variable data naming.
2 Linking variables to their data.
3 Locating external variable data.
In IIC production, variable data is applied to as much media as can
reasonably be accommodated in a project, not only reference material
(pictures or text) but also buttons, text labels, help files, captions, etc.
Also, all key timing elements in a show that are managed by the local
client (instead of the PD/AVS show flow) should be made as
variables. You may be surprised at what you may need to tweak when
you see hundreds of real users experiencing your show.
Each piece of variable data is given a name or “variable”, and it has a
value that may be a simple number or string, or be an external file.
Appendix D –IIC Object, Variable and Data Referencing Standards for
Console Content describes the method employed in most IIC
productions. If you already employ coding standards, you may wish to
use your own, as these are simply provided as a guide.
Console Interactivity
IIC console interactivity is typically produced using a variety of
production methods including:
Director.
Web Browser.
Flash (within Director or a Web browser).
Direct C++ coding.
Most development work to date has occurred within Director for
several reasons:
Extensive support for plug-ins through its “Xtra” concept.
Extensive media type support (digital video of many formats,
web pages, etc.).
Reasonably powerful coding language (Lingo).
Direct 3D support (Shockwave 3D).
In some cases, Flash has proven to be a better tool for content control
or animation. However, Flash does not have the same Xtra support as
Director. It must be controlled through a container application that
hosts the networking and utility tools of the IIC. Director is the richest
container application that can serve this purpose. Internet Explorer
using ActiveX and JavaScript is also a reasonable capable container
application.
If you want high-performance 3D, Director’s Shockwave 3D can be
very limiting or under-performing. In these cases, development has
gone directly into C++ code and has used IIC support tools through
COM or DLLs. For 3D, many commercial and OpenSource engines
are available.
When designing interactive consoles, a key consideration is how they
handle group objects. Group objects can be copied into a local variable
when changed, thus allowing the rest of the code to behave normally
or, they can be read and modified directly ensuring an accurate current
state at all times.
A new capability of the IIC worth considering is using multi-site
group objects. Multi-site group objects are objects that have a local
group context (for example, current to all users in the cinema, or on
this server) and a universal group context (for example, across all
servers or all connected cinemas). In this case, you must manage and
decide between the two states of each group object, modifying and
displaying the correct one at the correct time.
8-10Mbps and stereo audio are the norm. Frame rates of 29.97fps for
content shot with video cameras and CG content at 29.97fps or 30fps
are acceptable.
There are several differences and challenges to production that must
be noted for the DMS as opposed to HDVS including:
DMS does not support seamless audio/video junctions.
This means that each segment jump will be noticeable to the
users. Minimize the number of show segments or use fades in
and out wherever possible so the “jumps” between clips are not
as noticeable.
DMS does not support direct sprite overlays.
Sprites must either not be shown during digital video
sequences or be shown adjacent to the video. Video resolution
must be further reduced and the video must playback at less
than full-screen.
As well, note that there are issues when converting from a 4:3 aspect
ratio to a 16:9 aspect ratio. Please see the section entitled Converting
4:3 to 16:9 aspect ratio later in this chapter.
Production Note: To take advantage of the DMS ability to show more
dynamic graphics (including 3D) through Director, a
customize-coded DMS server must currently be built
using the core DMS as its foundation. A future
version allowing simple cast plug-ins is a likely
improvement to this development process.
Customization of the DMS code is also required for
sprite construction at this time although a future
version will allow external configuration identical to
that of the HDVS system.
35mm film
Generally, the time, cost, and quality issues using 35mm (analog) film
for this type of production does not recommend this content capture
method. HD Video has proven to be a much less costly and cleaner
choice where the option exists.
Triple Screen
3x4:3 aspect ratio
(3072x768 total)
aligned carefully between the initial path and the alternative branching
paths.
Two approaches used to date include:
1 Method one involves creating a series of splines and aligning
the connection points to create a smooth tangent (in and out
tangents are the same) between the connecting paths.
2 Method two is more suitable for quick multi-path segments
where the action takes place in a specific area (for example
multiple attacks on a monster). Free (not connected to
geometry) cameras are used and animated through time in
Free-Form without paths. Align the Cameras to each other (the
point where they diverge.). With Animate on, the Cameras are
moved with respect to their time and position along the
sequence. To connect a camera to the next connection point, go
to the frame of convergence and Align the camera to the other
camera.
Whichever method is chosen, the following steps are then necessary
for testing the end result and tweaking for proper alignment.
1 In both instance, check for the proper connections by sliding
the time-slider (with everything hidden except for the cameras
and paths). This should give a quick check to see if cameras
are linked to each other and the individual cameras are not
animated by accident.
2 Render each camera at divergent/convergent points to compare
each camera in an image-processing program (such as
PhotoShop).
3 Render the center cameras in shaded mode to see if they are
connecting properly (60 frames before and 60 frames after
merge) to see if the connection and motions are smooth.
4x3 16 x 9
This results in a loss of picture, of course, but does not modify any
data in the picture (no distortion to scaling/stretching).
2 Compression.
The second approach is to compress the image.
4x3 16 x 9
With this technique the entire picture is preserved. That is, every pixel
is still displayed, but is compressed into the smaller area. The result is
that an image appears shorter and fatter, although the resolution is still
1024 x 768.
Immersion uses the compression technique when changing an aspect
ratio in its productions, valuing the ability to see the full image over
the resulting distortion of the image.
Production Note: AC3 files are exactly the same file size as an
uncompressed stereo audio file. Once you produce an
AC3 file, you can confirm that it is the correct length
to match the 30fps segment using the following
calculation:
#of video frames (30fps) = (AC3 File size in Bytes - 44)/5880
To confirm the audio files prior to AC3 encoding simply divide by
2940 (instead of 5880) as the individual channel files are mono.
(These sizes are based on 44.1kHz sample rates).
For consistency in production, use standard file naming for the 6
channels of audio. One standard file naming convention you might
want to consider is to use filenames composed of the segment name
and the channel abbreviation, for example intro-FL.wav (for intro
segment, front left channel). The channel abbreviations typically used
are FL, C, FR, RL, RR, and SUB (or LFE – Low Frequency Effects).
Production Note: Be very careful with checking these numbers and
following a rigorous procedure. Many listeners can
readily detect faults of only three frames as a lip sync
error.
etc. for each of the key 'chunk'. The wav files MUST be the exact
same length as the video to be captured (or the closes number of
samples just over the video time).
We run video ultimately at a pure 30fps progressive (not 29.97 NTSC)
so no drop frame (NON-DROP FRAME) as well. To
mathematically double-check things, 2940 bytes of MONO 44.1kHz
audio = 1 video frame, so a 1sec segment (30fps video) should be
88,200 bytes in size (+ a wav header which is usually 44 bytes,
meaning a pure filesize per channel of 88,244 IF the audio was for a
1sec video).
filesizeOfAudio = time(in sec)*30*2940 + 44
Always provide a test set of files to ensure proper timing/sizing prior
to full production.
Introduction
Real-time Group Interactive Immersive Experiences (GIIEs), such as
the Immersion Simulation Lab (ISL) are primarily used for high-value
professional design, simulation, or training purposes in such
disciplines as (but not limited to):
Design and planning.
Oil exploration.
Automotive design.
Pharmaceutical and chemical processes.
Training and simulation.
Military.
Aircraft.
Ship.
Nuclear plant operations.
These environments are increasing used for less professional and more
educational and entertainment purposes, such as:
Highly dynamic exhibits.
Immersive video games.
This chapter deals with the above real-time GIIEs in a general way,
indicating many of the characteristics shared amongst these
environments. Bias may be shown towards design and planning
applications as these are the most common currently under study.
After the generic detail is provided in this chapter, some specific detail
for design and planning, and exhibits and games is also given.
Key concepts
The Immersion Simulation Lab (ISL) shares many characteristics with
other real-time GIIEs. There are also, a number of differences
particularly where the ISL focuses upon the needs of design and
planning industries. These differences will be discussed in detail after
the shared characteristics are described.
The following concepts are critical to real-time model data linking and
sharing:
Object naming.
Object attributes.
Flat scenes versus scene graphs.
Object naming
Naming components is a key activity in all CG modeling
environments. It can be treated either in a very casual or very strict
manner, depending on the organization or on the number of CG
specialists working on a project. Names of data objects are the primary
method to access the data and control it interactively.
Naming occurs at three levels depending upon how the applications’
data is organized:
Files.
Layers.
Objects.
Object naming using common language rather than technical is
critically important. End users of a GIIE will use these names to select
objects or data to interact with or create links to. Non-intuitive (i.e.
FM-01-LXD-0443) or generic (i.e. BOX01) naming can hamper or
even eliminate interactions altogether. Ideally, naming has been found
to focus on the common name, domain-specific type of data, its
position or scale, and perhaps a version or option identifier (i.e.
HOUSES-SE-2story-optionA, or POSTOFFICE-45m-v1).
Another important use of naming is in controlling complexity. To
simplify models, change options or show detail when focused on one
area, a user can manipulate data by its name organization or through
direct geometric manipulation. Naming provides a mechanism to
group data together to quickly be manipulated.
Object attributes
Some data creation applications can associate specialized attributes to
geometry. This can be used to specify information not typically
inherent in the graphical representation (i.e. material composition,
property value). Many immersive environments can only manage
geometric data and strips out attribute data in conversion.
Attributes can be used for many valuable purposes including:
Identification/searching for particular data that basic naming or
geometry may not reveal.
Automated geometric generation or transformation based on
attributes (i.e. extrude building footprints by the value of their
height attribute).
Direct database linking or relational links to broad databases
not contained within the geometric descriptions themselves.
Full
Experience
Project Goal QA
Definition
Interactivity Interactivity
Simulation Interactivity
Device Content
Coding Options
Coding Linking
Conventional components
Immersive visual content
Working on real-time projects involves a number of tasks specific to
its visual immersive content:
Immersive model considerations.
Model sources.
Optimizing for real-time.
Model assembly.
Automated routines
General-purpose, immersive environments are common in professions
where work is undertaken between different teams and there are few
standards for immersive collaboration.
Most professional disciplines involving work of any contextual
complexity rely upon a range of base data sources. For large-scale
contexts, GIS data is a common geometric input, as is remote sensing
data (aerial photography, satellite imagery) common for imagery, and
forms the base for many automated routines.
Where sufficient base data may not exist, it is often necessary to use
digitizing and manual crafting. As many uses of a GIIE involve the
creation and investigation of future environments (buildings, cars, oil
wells, etc.), crafting is also a necessity. Many tools can create new
geometric models, including 3D Studio MAX, Maya, and AutoCAD.
The most critical aspect with many of these tools is that they are
designed for computational rendering or construction drawing. This
means that they often focus on complexity and quality of rendering or
1 http://www.multigen.com/products/index.shtml
2 http://www.vizx3d.com/
3 http://www.cult3d.com/
4 http://www.web3d.org/vrml/oblib.htm
5 http://www.replica3d.com/
6 http://modelbank.viewpoint.com/
7 http://www.3dlinks.com/links.cfm?categoryid=9&subcategoryid=90
8 http://www.clr.utoronto.ca/LINKS/GISW/origarticle.html, GIS World, July 1995
entitled: 3D for Free - Toolkit Expands Visual Dimensions in GIS
9 Such as PhotoModeler, http://www.photomodeler.com/
10 These numbers are derived from a number of sources and are not all directly comparable
(as the methods of determining these numbers varies greatly). They do give a good
indication of the magnitude of change in capabilities over time however.
PC Graphics
Hardware
Nvidia GeForce2 25M 800K
(2000)
ATI RADEON 9800 380M 12.6M
Pro (2003)
Super Graphics
Workstation
SGI Onyx4 300M to 4800M 10M to 160M
UltimateVision
(2003)
several GBs available. This total memory is not usually available for
textures alone. The same video RAM also holds the screen data and
depth data for 3D (and sometimes other data). To provide seamless
animation, most software also uses double (or more) buffers (memory
that is screen-sized but non-visible) to draw off-screen and then just
display the final image. This multiplies the basic screen RAM used (by
close to two, three or more times).
While the number of graphics options used in rendering significantly
impacts these numbers (and hardware optimizations can minimize this
too), a simple set of math might indicate how much of the video RAM
is used just for basic rendering and display before textures are loaded:
Video RAM for Rendering = screen pixels * (pixel colour
depth*2+16)/8
Pixel colour depth is usually 24- or 32-bit (8 bits of Red, Green, and
Blue). This number is multiplied by 2 for double buffering. The extra
16 is a basic assumption of the number of bits per pixel used for depth
(Z) information. To convert from bits to bytes divided by 8.
Therefore, a standard 1024x768 screen in 32-bit colour might
consume:
1024*768*(32*2+16)/8 = 7,864,320 bytes (approx. 7MB) of
video RAM
And this memory consumption increases to 13MB for a 1280x1024-
resolution screen and 19MB for 1600x1200. If you only have a 64MB
graphics card, screen resolution consumes a significant portion of the
available memory.
The amount of memory used for a texture can simply be calculated as:
width*height*(pixel colour depth)/8
Typically, graphics systems are optimized for textures with a width
and height that are powers of two (2,4,8,16,32,64,128…1024 pixels)
per side (width and height can be different, just that each should be a
power of two, e.g. 32x128). If you always use large 1024x1024x32-bit
textures, then each texture would consume 4MB of video RAM.
As with polygons, if you reduce the pixel count or their colour depth,
you can fit your models entirely within the graphics card memory and
render at maximum speed. There are also methods of using
compression textures in video memory (much like JPEG images) on
some systems, but these must be pre-generated and are often used only
for video games.
Model assembly
The above section describes a number of ways of obtaining or creating
models for a real-time immersive environment. The model assembly
often uses data from a number of sources that must be assembled into
a single environment.
Model assembly can be the weak link in the visual data production
phase for the following reasons:
Model conversion problems.
Model coordinate problems.
Image rectification.
Model meshing problems.
Model complexity problems.
While model data sources can often be planned out and well
structured, the model assembly process can make or break an entire
production process. To ensure success, it is important to establish file
format standards that all data contributors must meet before beginning
a project. Next, it is wise to test all formats and software systems early
to catch problems before they disrupt production deadlines.
Converting a model from one file format to another can present
another large black box of problems. Proprietary formats are the most
difficult to manage while more generic ones often result in the loss of
some specific data. Ideally, you want to work with a format that at
least supports:
Hierarchical scene descriptions.
A variety of polygonal forms (strips, quads, etc. not just
triangles).
Texture coordinates.
Normal coordinates.
Material properties.
Animation key-frames.
Camera export for static views and animations paths.
There are few general data translator tools on the market. Most
software packages come with a number of built-in data translators or
sometimes have the option to buy or create plugins to natively
11 http://www.okino.com/
a city) and at the same time other data with high precision on a smaller
scale (a circuit board, for example), you find data gets rounded off and
you lose precision. A common way to overcome this problem is to
move a model that is in large UTM (or similar) coordinates closer to
the (0,0,0) axis center. This technique is called making a local origin
point. Such conversions must be done with software that provides
double-precision (64bit) support otherwise; the conversion can be
corrupted in the translation process itself.
Once most geometric data is organized and placed, often you must
perform a number of image rectification steps. Surface textures
gathered from imprecise or non-orthogonal sources (camera photos,
satellite images or aerial photographs) often need correction to match
the Cartesian coordinate space (or UTM) that the whole model is now
in. This correction is sometimes done in the model creation phase but
may need re-adjustment or initial adjustment after model assembly.
At this point, all key model components should be in the same
coordinate space with matching scales. You should see roads in the
right place relative to buildings or trees in the correct locations. Often
however, the separate pieces of a model do not quite align correctly. A
terrain model sampled at a 30-meter grid often will not match a local
riverbed model crafted at a 1-meter accuracy. Furthermore, many
models are originally built or designed for a flat model and do not
properly sit on newly introduced surfaces under them. A series of
meshing activities is often needed including:
Interpolation – making objects sit in the Z-axis on top of one-
another properly.
Meshing – connecting overlapping surfaces or gaps in meshes.
Alignment – adjusting the position of objects that may have
been estimated from less precise base data now that detailed
model components can be referenced.
2 Bi-directional linking.
Links and searches from the sensory to data immersion
environments work both ways, allowing data inquiries to
feedback and change the sensory models. This model typically
requires a built-in or external database capability. Other types
of data links are possible and can be of great value for
example, Geographic Information System (GIS) or links to
spreadsheets.
During the production phase of a GIIE, geometric data gets linked to
external applications. Uni-directional linking can be done on a one-to-
one basis, matching any object in the GIIE by name (or attribute ID) to
an application and its startup parameters:
HOUSES-SE-2story-optionA=iexplorer.exe
http:///www.asite.com/search_buildingsDB?keys=housing&height=2
For bi-directional linking, a more formal link must be established
between attributes in the real-time model and attributes in the remote
application. In addition, a formal communications protocol must be
understood and supported, protocols like ODBC, DDE or SOAP.
Within each of these data types, there are a myriad of file formats and
standards, some well supported (DXF, DWG, E00, etc.) but many
more that have only partial (only old versions or partial data type
support) capabilities in any particular package. This makes data type
linking one of the most time consuming and constantly-changing
aspects of working digitally on multi-source design and planning
projects.
Focus on assembly
In many instances, software tools for real-time simulations are still just
in their infancy, though there are tools that can be learned in very short
order. However, the greatest challenge at the moment is the data that is
available to work with. With this, though there are a multitude of tools
available there is not the capability to integrate all the required data at
this point. As such we need a broad spectrum of capabilities to support
all that is required, and these people are not readily available since
there are too many capabilities for one person to master.
The data sets can be seen as the raw materials to put into the modeling
structures. However, unless the raw materials are there and in the right
form, an exorbitant amount of time is required hunting and gathering
the data – from cities and GIS and then trying to get the architect to
provide a file that actually works. There are no particular standards on
which to construct data sets and therefore models come in any and all
types of forms.
With so many architectural model or geometric modeling packages
without any standards to follow, there is much work to assemble data
sets. The architectural model and geometric modeling packages range
from Maya at one extreme to AutoCAD and the other base
design/modeling packages at another. Generally DXF is used as the
standard file format to share the data sets. However, exporting from
one package to another creates spurious results, depending upon the
coordinate systems being employed as well as the scale of the model
being constructed. Here as well, surface normal issues can easily arise
depending on how a model is constructed, and some systems or even
the way operators work insert multiple polygons that are not visible or
known at the time (for example, the system makes six copies of the
same surface, three of which point one way, and the other three are
pointing another way).
For a moderate sized urban context (several square blocks), it can take
anywhere from five days to two weeks to get the raw materials in a
form that can be properly integrated. In many instances when data is
received it is incorrect in any or all of the following ways:
Normals are all wrong (backwards).
Model is the wrong scale.
Incomplete data sets.
Another factor impacting the assembly of data sets for modeling is
there is no single package available that is designed to provide this
capability – a hybrid meta-package that has elements of all of the
technologies for the purpose of assembling data sets. One of the most
useful package that we have found today is PolyTRIM12. Without this
12 PolyTRIM is a research tool developed by the Centre for Landscape Research at the
University of Toronto. See http://www.clr.utoronto.ca/POLYTRIM/polytrim.html
Developing a 2D model.
Developing a 3D model.
transparency is used to ‘mask’ out (eliminate the drawing of) the non-
rectangular portions of the shape. Modelers must be aware however, of
the different (and sometimes incompatible) methods of specifying
transparencies in a texture. Some systems (many 3d modelers like 3d
Studio MAX) require one texture for the colour data and another
texture to specify the transparency (usually a grey-scale image) where
black is entirely see-through, and white is solid (no transparency).
Other systems use a single, usually 32bit (8 bits of red, green, blue,
and alpha) texture to achieve the same effect. This can require an extra
data conversion step for textures.
Shadows are another critical representational capability for most
architectural 3d systems. Shadows assist in defining edges, light access
and patterns. Most systems now support volume shadows which
accurately project geometry volumes through space. This
unfortunately does not work for real-time systems where geometry is
only defined in a texture (a tree or roof gable), where the attached
rectangular polygon would be used for shadow calculation. In these
cases alternative methods such as projective shadows must be
employed. Most current systems do not support such variable hybrid
methods within a single real-time environment and therefore shadows
remain a challenging and limited capability for highly integrated
design and planning work.
Most systems supporting architectural work also have some levels of
lighting support. Most of these systems operate in a non-real-time
rendering mode as lighting calculations, bounces, and reflections of
the complexity in most buildings is beyond real-time graphics
capabilities (which are often limited to 8 or fewer dynamic lights).
Non real-time systems (such as many radiosity systems) can often be
used to render light maps into or overtop of surface textures. These
resultant textures can be viewed but typically not changed in real-time
environments. Changes in pixel shader technologies of leading-edge
graphics cards are starting to make some levels of dynamic complex
lighting possible.
Interactive moves
Prior research has identified tools and representational issues, at least
for design and planning13. They have identified an essential base set of
tools for group interactive design inquiry:
14 For more information on height maps and terrain modeling see the following:
• http://www.qeradiant.com/manual/Terrain_Manual/pages/height_maps.html
• http://home.planet.nl/~monstrous/terrain.html
15 For more information on BSP Trees and Portals see the following:
http://www.webcom.com/~petrich/games/duality/dualism_articles/portbsp.html
ftp://ftp.sgi.com/other/bspfaq/faq/bspfaq.html
Conclusions
As outlined in the introduction, this guide has three main purposes:
1 Expand awareness, interest and capabilities for production of
group interactive and immersive environments.
2 Elaborate the issues and challenges surrounding these
environments.
3 Describe key areas for future research and development.
Each chapter of this guide has played a role in presenting and
elaborating the concepts that have shown themselves to be the most
important towards these three goals.
We started this guide with a discussion of what makes a group
interactive immersive experience valuable and a key indicator of the
future evolution of new media in professional, educational, and
entertainment experiences. We spent time examining the key findings
from research into the key user perceptions and cognition when
engaged in these environments. This examination revealed the most
important challenges and roles for a Group Interactive Immersive
Environment (GIIE) as a new media form.
The second part of this guide delved deeply into the technical structure
of a GIIE and the production steps necessary to fully exploit them. The
guide has reviewed the technology and technological concepts that
have been created to allow for this holistic integration of components.
Finally, the guide has reviewed in some detail, the process, knowledge,
and skills necessary to produce content of this form. Within each of
these steps the guide has examined in more detail the changes
necessary for adaptation to broadband distribution that can open the
markets and user accessibility to this rich new media approach.
With the explorations and knowledge presented through this guide, we
have arrived at a number of conclusions about GIIEs. In particular, we
can look at the opportunities surrounding their release from dedicated
venues to broader use over the Internet.
Findings
While there has been a strong focus upon the full sensory quality of a
GIIE experience, it has become evident that the overall experience is
far greater than the sum of its parts:
Immersive visuals.
Immersive audio.
Rich media content.
Personal interactivity.
Group interactivity.
Social interplay of participants (non computer-mediated).
In most media environments only one or two of these are engaged at
any one time, while in a GIIE, a wide range are available and exposed
over the course of an experience. Many viewers of this new media
have expressed that it is “too much” for a person to take in. In fact,
early user testing had revealed similar concerns. What has been found
over the years however is that the concept itself is fine but the specific
instances of experiences can try to do too much, too quickly and in too
complex a fashion. While there are huge challenges in designing
“good” usable systems for a single user using a small number of types
of media, the challenge is only amplified in a GIIE. Careful attention
to design and most importantly, usability testing has proven to be key,
even for experienced designers.
Of a second note on the concept of ‘too much’, is the changing nature
of media engagement and access in the broader world around us. This
may not be perceived by all, and we have certainly found that age
plays a role as the nature of media and interaction has changed
radically over the past 15 years. It is not uncommon today to see:
Rich media engagement reaching new levels through the
expansion of the Internet. Recent studies have now placed up
to 75% of Americans online, while passive TV consumption is
dropping yearly.
People watching TV while exploring related content on the
Web. There are many examples of enhanced TV that pushes
this approach.
Most video games now having a multi-player online mode with
game-play interaction and chat and even voice interaction.
Future Research
Unfortunately, one guide such as this cannot answer all questions,
even when restricted to the GIIE context. It has however been the goal
of this guide to identify key questions, even where answers are not yet
available. These questions can then be used to direct future research
efforts more effectively. We have classified the research questions into
three categories, which we have organized this guide around:
perception and cognition, technology, and production.
Some of the biggest challenges and opportunities lie within the realm
of perception and cognition. As we identified in the findings, a clear
understanding of how “same-place” versus “different-place” groups
impact the utility and acceptance of GIIE content can have a large
effect upon the scope of utility, accessibility of the concepts and
market potential for this technology and content production. We
believe there are two key aspects that should be carefully considered in
establishing any such research programs:
1 An assessment of the role played (the magnitude of impact) in
the overall value of the experience by “same-place” versus
“different-place” user distribution. This parallels the
assessment of the ‘resolution/bandwidth’ study reported in this
guide.
2 A more detailed assessment of what professional conditions
and activities (in particular) are appropriate for broadband use.
This will require a running application environment and real
tests under real-world conditions to ensure the goals and
needs/desires of the users are honestly reflected. This will
necessarily be domain-specific.
Summary
The present study investigated factors influencing the value of an
interactive, immersive Entertainment Education experience about sea
lions or dinosaurs provided to over 450 male and female students in
grades 5 and 9 in Charlottetown, Prince Edward Island. Students
completed questionnaires immediately after the show and after a delay
of one week. Their high ratings of enjoyment, presence, and acquired
knowledge generally confirmed results of a prior study of the sea lion
show (Ritterfeld, Weber, Fernandes, & Vorderer, 2003). Teachers,
who had accompanied their students, also rated the experience
positively both immediately and after the one week delay. They
believed that students benefited most from the combined large screen
and interactive component, that all students enjoyed the experience,
and that even students with special needs would benefit from it.
Enjoyment, interest in learning more, and presence were highly
correlated for both shows, and in general the two different shows led
to the same pattern of appraisal. Girls responded significantly more
positively to the presentation than did boys, and grade 5 students
responded significantly more positively than did grade 9 students, but
these differences were small from a practical standpoint. Although
student-self-assessed and teacher-assessed learning was high, learning
of specific facts was not well-reflected in the results of a 32-item yes-
no questionnaire when compared with a control group of over 200
students who had not experienced the presentation. Similar findings
were also observed by Ritterfeld et al. (2003) suggesting that the
educational value of the presentation lies not in acquisition of new
facts but rather in the establishment of new perspectives. One primary
variable of interest was degree of resolution or visual clarity of the
immersive presentation. For the dinosaurs show, one third of the
students viewed the show under low, medium, or high visual
resolution. Mean student ratings on several scales (e.g., “the show was
fun”) varied with increasing degree of resolution; however, these
differences were negligible practically speaking. Small and moderate
degradations in quality of a large-screen display of an interactive
Entertainment Education presentation had little impact on enjoyment,
learning, and sense of presence. High appraisal of the immersive
interactive educational experience by students and teachers regardless
of the level of resolution could warrant more widespread availability
through broadband distribution at lower resolution than is technically
possible under more costly presentation conditions.
Background
The expansion of multimedia technology has resulted in the
development of innovative Entertainment Education which blends
entertaining experiences with educational goals. In recent research
(Ritterfeld, Weber, Fernandes, & Vorderer, 2003), 656 children
participated in a science based immersion show in an Immersion
Theatre environment which combined a movie-like cinematic
presentation on a large screen with a computer-game-playing
environment on a personal screen. Through interplay between
information presented on the global screen and individual game play,
students participated in solving a real scientific problem about sea
lions. Results indicated that the Entertainment Education immersion
experience had positive educational effects, for example on enjoyment
and sense of presence. However constraints of the research did not
Method
Procedure
Acquisition of the participation of 38 classes in the study entailed
contacts with the school boards, principals, and teachers; creation of
schedules and clearance of transportation services (school buses).
More specifically, the Director of Program Services of the Eastern
School District was contacted by telephone and by letter to request
permission to contact the elementary and junior high school principals
in the PEI Eastern School District to participate in the study. Upon
receiving permission to contact the principals, an information packet
was sent to the school board and principals at all applicable schools.
The packet contained: (1) a letter describing the research and
requesting permission to recruit students within the school, (2) the
student consent form, (3) parent consent form, (4) the teacher short
term impact questionnaire, (5) the teacher long term impact
questionnaire, (6) the student short term impact questionnaire and (7)
the student long term impact questionnaire. Once permissions had
been obtained, classes were assigned to control and experimental
conditions. When actual testing took place, two experimenters and
typically one teacher were on hand to organize the testing of each
single class of students in the immersion theatre. Teachers oversaw
any questionnaire completion conducted in the classroom.
Participating schools received a small monetary contribution to their
library as a token of appreciation for their cooperation.
Approximately one to five days prior to the immersion experience, the
students filled out a preliminary questionnaire in their class to develop
a personal code number, and to collect demographic information of
age, and hobbies and interests (Appendix A). On the day of the
immersion experience, students and their teacher arrived by bus at the
Atlantic Technology Centre (ATC) and were escorted to the
Immersion Theatre.
The Theatre is a purpose-built facility, sloping to a two-story height of
the wall supporting the screen. It is sound-treated and carpeted. There
Design of Experiment
A primary variable of interest was the level of resolution of the show.
The high resolution version was HD Intranet quality and the same
quality that had been experienced by participants in the study of
Ritterfeld et al. (2003). Two additional levels were compared. The first
designated “medium resolution” was perceptibly reduced in resolution
when compared directly with the high resolution version, however,
without the comparison, it might be hard to imagine a superior quality.
The third level, designated “low resolution”, was perceptibly of
reduced resolution without the need of comparison to a HD standard,
although discernability of information was not impaired. The
technology underlying the creation of the levels of resolution of the
shows was developed by Immersion Studios where additional
information can be obtained.
Testing of pilot classes of grade 6 students, one for each of two shows
for each level of resolution (approximately 150 students) enabled the
checking of the procedures for running the experiment.
The testing session in the Immersion Theatre lasted about 45 minutes,
whereas the presentations were approximately 22 and 25 minutes in
duration. Upon arrival and settling down of the class in the theatre, the
first 22 students were assigned to a separate terminal, with the
remainder (if any) required to double up. A technologist familiar with
the Immersion Studios technology initiated the first run for the day,
and, subsequently, the experimenters ran a second, third or fourth
show depending on the day’s schedule. Immediately upon completion
of the presentation, the students were asked to complete a
Table 1a:
Experiment Design
Sea Lions Dinosaurs
low res medium res high res low res medium res high res
Grade Glen Stewart Glen Stewart Glen Stewart Spring Park Spring Park Spring Park
5 Sherwood Sherwood Sherwood St. Jeans West Royalty West Kent
Grade Birchwood Birchwood Birchwood East Wiltshire East Wiltshire East Wiltshire
9 East Wiltshire East Wiltshire East Wiltshire Stone Park Queen Charlotte Stone Park
Table 1b:
Experiment Design: number of participants in each of three successive questionnaires
Sea Lions Dinosaurs
low res medium res high res low res med res high res
1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
Grade 26 26 25 27 27 26 24 24 24 24 23 23 23 22 21 19 16 23
5 24 24 23 20 20 20 20 19 19 23 23 20 26 26 21 17 15 14
50 50 48 47 47 46 44 43 43 47 46 43 49 48 42 36 31 37
Grade 24 24 24 26 26 26 23 25 21 20 29 20 26 21 25 32 31 18
9 23 22 17 29 24 24 22 22 22 23 19 18 31 31 29 28 28 28
47 46 41 55 50 50 45 47 43 43 48 38 57 52 54 60 59 46
Control conditions
There were also 201 students who completed the first and final
questionnaires in order to determine a baseline for the amount known
about the topics of Sea Lions and Dinosaurs and to obtain
demographic data. These students also viewed one of the shows, but
only after the questionnaires were completed. There were
approximately 50 students in each Grade (5 and 9) x Show (Sea Lions
/ Dinosaur) combination.
Pilot testing
Prior to testing of the Grade 5 and 9 classes (approximately 150
students) in the main experimental design, six Grade 6 classes were
presented with one of the two shows in one of the three resolution
conditions in order to test the basic methodology and instrumentation.
Grade 6 was chosen for the following practical reasons. Falling
between the two grade levels in the main experiment, it would
sufficiently represent each. Second, same schools could provide the
grade 6 classes that were providing some of the other classes in the
main experiment thus avoiding the approach to additional schools and
principals for Grade 5 control classes. Finally, the procedure of
transporting students from several of the schools involved in the main
experiment would receive a “dry run.”
Results
Each student who participated in the main experiment produced three
sets of data: (1) in their school classroom, preliminary demographic
data, 1 to 5 days prior to the Immersion presentation (2) in the
Immersion Theatre, immediately after the Immersion presentation, and
(3) in their school classroom, exactly 1 week after the show. As well,
the teachers of students in the main experiment completed two
questionnaires to give their impressions of the experience of the show
for their students, (1) immediately after the Immersion presentation
and (2) approximately a week after the Immersion presentation.
The study by Ritterfeld et al. (2003) carried out the student and teacher
assessments immediately after the Immersive experience, and a post-
test for a small proportion (< 20%) of both students and teachers. We
will begin with the student and teacher questionnaires following the
immersive experience (referred to as Short term impact and Study 1
for students and Study 2 for teachers by Ritterfeld et al. (2003).
Table 2(a)
Mean rating for each question immediately following the presentation of Dinosaurs
Each cell shows the mean rating, the standard deviation, and the number of respondents.
consistent pattern inconsistent pattern personal trait potential confound
The show was fun. 4.81 4.69 4.97 4.82 4 .26 4.39 4.52 4.39
.58
.70 .62 .18 .74 .79 .83 .79
110
43 36 31 35 38 33 106
I learned a lot during the show. 4.36 4.11 4.35 4.28 3.75 3.55 3.63 3.64
.69 .89 .91 .83 .97 1.10 .86 .98
42 36 31 109 36 38 33 107
There were too many pictures, information and 1.78 1.78 1.32 1.64 1.64 1.63 1.52 1.60
games in the show.
1.08 1.27 .83 1.1 .99 .97 1.06 1.00
41 36 31 108 36 38 33 107
I figured out how to play the games. 4.33 4.33 4.70 4.44 4. 3.97 3.91 3.96
.93
1.12 .93 .53 1.29 .85 1.33 1.16
109
42 36 31 36 38 33 107
At home I love to listen to loud music. 4.29 3.60 3.87 3.94 4.5 4.34 4.30 4.38
1.31 1.65 1.25 1.43 .65 .97 1.02 .89
42 35 30 107 36 38 33 107
I often surf the internet. 3.30 3.56 3.43 3.43 4.31 4.32 4.24 4.29
1.62 1.44 1.53 1.53 .86 .96 1.06 .95
42 34 28 104 36 38 33 107
I got upset during the show. 1.29 1.06 1.03 1.14 1.39 1.35 1.30 1.35
.81 .34 .18 .56 .90 .75 .88 .84
42 34 30 106 36 37 33 106
I wanted to learn why the dinosaurs became extinct. 4.21 3.79 3.37 3.84 2.89 3.22 3.79 3.28
1.18 1.59 1.54 1.46 1.16 1.32 1.08 1.24
42 34 30 106 36 37 33 106
My grades in school could be much better. 3.50 2.80 3.25 3.1 3.75 3.11 3.24 3.36
1.48 1.49 1.38 1.47 1.16 1.41 1.44 1.36
38 35 28 101 36 38 33 107
Today I would have loved to play longer. 4.36 4.66 4.87 4.60 4.47 4.58 4.55 4.53
1.26 .73 .51 .95 .70 .79 .87 .78
42 35 30 107 36 38 33 107
After playing, I want to learn more about dinosaurs. 4.17 3.34 3.97 3.84 3.08 3.21 3.13 3.14
1.18 1.43 1.05 1.28 1.18 1.40 1.21 1.26
41 35 29 105 36 38 32 106
I often did not know if I would have to look at the big 2.9 2.15 2.34 2.50 2.33 2.21 2.53 2.35
screen or the small screen.
1.55 1.48 1.37 1.50 1.31 1.30 1.50 1.36
41 33 29 103 36 38 32 106
The show was boring. 1.17 1.52 1.31 1.32 2.08 1.54 1.76 1.79
.58 1.06 .93 .86 1.16 .93 1.15 1.09
42 33 26 101 36 37 33 106
When I came in, the topic of the show did not interest 1.65 2.09 1.63 1.79 2.81 2.59 2.25 2.56
me.
1.21 1.42 1.19 1.28 1.37 1.24 1.41 1.34
40 35 30 105 36 37 32 105
The teacher should take us to more shows like this. 4.66 4.62 4.77 4.68 4.58 4.74 4.76 4.69
1.02 .70 .90 .88 .80 .72 .75 .76
42 34 30 105 36 38 33 107
After playing, I want to learn more about the oceans, 4.33 3.61 4.10 4.04 2.94 3.32 3.33 3.20
animals and weather. .
1.20 1.30 1.03 1.22 1.07 1.25 1.14 1.16
42 34 30 106 36 38 33 107
I am an excellent student. 4.14 3.91 4.07 4.05 3.53 3.89 3.91 3.78
.93 1.09 1.05 1.01 1.06 1.06 1.23 1.12
42 35 30 107 36 38 33 107
I want to play the show again. 4.52 4.43 4.93 4.61 3.92 4.23 4.36 4.19
.99 .98 .26 .87 1.27 1.04 1.14 1.16
42 35 30 107 36 38 33 107
At first I did not really care why dinosaurs 2.76 2.77 2.65 2.73 3.37 2.92 1.97 3.08
disappeared from the earth.
1.59 1.54 1.50 1.53 1.24 1.24 1.47 4.32
42 35 31 108 35 38 33 106
During the show I sometimes got angry. 1.62 1.14 1.13 1.33 1.44 1.45 1.25 1.39
1.31 .60 .73 .99 .88 1.06 .80 .92
42 35 30 107 36 38 32 106
I often play computer games. 4.05 3.53 3.55 3.74 3.66 3.95 3.97 3.86
1.43 1.65 1.55 1.54 1.26 1.06 1.33 1.21
42 36 29 107 35 38 33 106
I am often bothered by loud noise. 2.29 1.97 1.84 2.06 1.47 2.08 1.64 1.74
1.60 1.46 1.16 1.44 .70 1.30 1.25 1.13
42 36 31 109 36 37 33 106
In the show it was hard to know what to do. 2.62 2.14 1.97 2.28 2.39 2.76 2.36 2.51
1.56 1.38 1.27 1.44 1.29 1.24 1.39 1.31
42 36 30 108 36 38 33 107
I got distracted by stuff on the big and small screen 2.56 1.59 1.58 1.96 2.09 2.34 2.15 2.20
to look at.
1.57 1.02 .96 1.32 1.09 1.21 1.48 1.26
41 34 31 106 35 38 33 106
I wanted to investigate why the dinosaurs became 4.26 3.46 3.80 3.87 2.72 2.74 3.30 2.91
extinct.
1.17 1.42 1.54 1.39 1.16 1.46 1.31 1.34
42 35 30 107 36 38 33 107
During the show, I forgot about everything else. 3.93 3.75 3.97 3.88 3.08 3.24 3.59 3.29
1.40 1.44 1.33 1.39 1.30 1.32 1.41 1.35
42 36 31 109 36 38 32 106
Today I learned important things about dinosaurs. 4.57 4.31 4.65 4.50 3.36 3.58 3.91 3.61
.77 .95 .87 .87 1.22 4.27 1.18 1.23
42 36 31 109 36 38 33 107
I felt that I was part of the show. 4.20 3.72 3.81 3.93 3.28 3.66 4.06 3.65
1.15 1.56 1.35 1.36 1.32 1.15 1.09 1.22
42 36 31 109 36 38 33 107
I would tell my friends to play. 4.10 4.00. 4.55 4.19 3.56 3.79 4.18 3.83
1.45 1.31 .99 1.29 1.36 1.26 1.21 1.29
42 36 31 109 36 38 33 107
Table 2(b)
Mean rating for each question immediately following the presentation of Sea Lions
Each cell shows the mean rating, the standard deviation, and the number of respondents.
(Resolution variable invalid).
The show was fun. 4.87 4.74 4.67 4.76 4.54 4.54 4.4 4.50
.41 .61 .86 .65 .72 .75 .74 .74
38 35 36 109 46 46 47 139
I learned a lot during the show. 4.11 4.34 4.47 4.30 4.38 4.28 4.38 4.35
1.13 .97 .84 1.00 .65 .72 .77 .71
38 35 36 109 45 46 47 138
There were too many pictures, information and 1.79 2.11 1.72 1.87 1.41 1.80 1.85 1.69
games in the show.
1.26 1.23 1.34 1.28 .78 1.07 1.12 1.01
38 35 36 109 46 46 47 139
I figured out how to play the games. 4.42 4.03 4.28 4.25 4.46 4.29 3.91 4.32
.95 1.32 1.21 1.16 .72 .83 1.33 .83
38 35 36 109 46 46 33 139
At home I love to listen to loud music. 3.29 3.57 3.89 3.57 4.24 3.89 4.23 4.12
1.52 1.63 1.39 1.52 1.14 1.39 1.15 1.23
38 35 35 108 46 45 47 138
I often surf the internet. 3.82 3.36 3.89 3.70 4.04 4.11 4.17 4.11
1.39 1.71 1.39 1.50 1.25 1.02 1.22 1.16
38 33 35 106 46 46 47 139
I got upset during the show. 1.32 2.26 1.41 1.66 1.80 1.84 1.72 1.79
.82 1.64 1.04 1.27 1.17 1.15 1.17 1.16
37 34 32 103 46 45 47 138
I wanted to learn why the sea lion population went 4.11 3.91 4.0 4.01 4.02 3.78 4.17 3.28
down.
1.07 1.40 1.26 1.24 .97 1.05 .96 1.24
37 35 35 107 45 46 47 106
My grades in school could be much better. 3.18 3.0 2.45 2.89 3.93 3.80 3.64 3.79
1.47 1.41 1.48 1.48 1.12 1.31 1.36 1.27
38 33 33 104 46 46 47 139
Today I would have loved to play longer. 4.55 4.66 4.24 4.49 4.51 4.74 4.38 4.54
1.01 .54 1.44 1.06 .79 .53 .95 .78
38 35 34 107 45 46 47 138
After playing, I want to learn more about science. 3.37 3.35 3.53 3.42 3.50 3.24 3.36 3.37
1.32 1.45 1.31 1.35 1.15 1.23 1.26 1.21
38 34 34 106 46 46 47 139
I often did not know if I would have to look at the 2.38 2.45 2.77 2.53 2.22 2.35 2.35 2.31
big screen or the small screen.
1.50 1.37 1.50 1.45 1.20 1.27 1.40 1.29
37 33 34 104 45 46 46 137
The show was boring. 1.39 1.27 1.26 1.31 1.61 1.65 1.73 1.66
.84 .67 .90 .80 .95 1.04 1.05 1.01
36 33 34 103 46 46 45 137
When I came in, the topic of the show did not interest 2.20 1.80 2.43 2.14 2.69 2.85 2.53 2.69
me.
1.47 1.30 1.63 1.48 1.36 1.37 1.27 1.33
35 35 35 105 45 46 47 138
The teacher should take us to more shows like this. 4.63 4.94 4.91 4.82 4.62 4.82 4.40 4.61
.79 .34 .38 .56 .78 .65 1.08 .87
38 35 34 107 45 45 47 137
After playing, I want to learn more about the 3.92 4.23 4.46 4.19 3.64 3.67 3.74 3.69
oceans, animals, and weather.
1.02 1.00 .85 .98 1.14 1.08 1.06 1.09
38 35 35 108 44 46 46 136
I am an excellent student. 3.92 3.74 4.18 3.94 3.48 3.65 3.98 3.71
1.12 1.16 1.0 1.10 1.03 1.30 .99 1.13
37 34 34 105 46 46 47 139
I want to play the show again. 4.57 4.53 4.80 4.63 4.37 4.63 4.13 4.37
.99 .93 .76 .90 1.00 .88 1.24 1.07
37 34 35 106 46 46 47 139
At first I did not really care why the sea lion pups 2.53 1.59 1.80 2.00 2.28 2.35 1.94 2.19
died.
1.62 1.10 1.35 1.43 1.15 1.30 1.19 1.22
38 32 35 105 46 46 47 139
During the show I sometimes got angry. 1.81 1.88 1.45 1.72 1.67 1.76 1.83 1.75
1.37 1.39 1.09 1.29 1.22 1.13 1.17 1.17
37 32 33 102 45 45 47 137
I often play computer games. 3.84 3.76 4.03 3.88 3.59 3.59 4.02 3.73
1.36 1.60 1.36 1.43 1.26 1.31 1.13 1.24
37 33 35 105 46 46 47 139
I am often bothered by loud noise. 2.49 2.39 2.09 2.33 2.11 2.30 2.02 2.14
1.39 1.60 1.46 1.48 1.02 1.39 1.31 1.24
37 33 34 104 46 44 46 136
In the show it was hard to know what to do. 2.24 2.31 2.29 2.28 2.02 1.80 2.23 2.02
1.30 1.47 1.30 1.34 1.11 .94 1.24 1.11
38 32 35 105 46 45 47 138
I got distracted by stuff on the big and small 2.22 2.0 1.94 2.06 1.85 1.84 1.80 1.83
screen to look at.
1.44 1.37 1.25 1.35 .94 1.09 1.11 1.04
36 33 34 103 46 45 46 137
I wanted to investigate why the sea lion pups 3.76 4.33 4.37 4.14 3.79 3.80 3.93 3.84
died.
1.48 1.02 1.09 1.25 1.25 1.16 1.31 1.23
38 33 35 106 46 45 46 137
During the show, I forgot about everything else. 3.50 3.45 3.49 3.48 3.50 3.78 3.50 3.59
1.59 1.42 1.58 1.52 1.41 1.37 1.22 1.33
38 33 35 106 46 46 46 138
Today I learned important things about the ocean. 4.45 4.61 4.73 4.59 4.28 4.29 4.24 4.27
1.01 .90 .80 .90 .89 .82 .90 .86
38 33 33 104 46 45 46 137
I felt that I was part of the show. 3.62 3.67 4.29 3.86 3.76 3.85 3.85 3.82
1.46 1.55 1.23 1.44 1.32 1.17 1.19 1.22
37 33 35 105 46 46 46 138
I would tell my friends to play. 3.97 4.36 4.34 4.22 4.15 4.02 3.78 3.99
1.35 .99 1.21 1.20 1.03 1.32 1.28 1.22
38 33 35 106 46 46 46 138
and three between groups factors of grade (5 vs. 9), show (Sea Lions
vs. Dinosaurs) and gender (male vs. female) each having two levels.
Regarding the Resolution variable, if low resolution led to lower levels
of enjoyment or learning then obviously there was an educational cost
to employing lower level equipment in an educational setting. If,
however, this degradation variable was statistically negligible, then
less costly, more widely distributed Immersive programming for
education could be warranted. The data are available from the UPEI
researchers but are not included here as a simpler approach to the data
analysis, following Ritterfeld et al. (2003) was taken which reduced
the number of variables for consideration from 29 to 15. Separate
analyses for Dinosaurs alone were also conducted.
Ritterfeld et al. (2003) did not report their data as a function of the
independent variables of gender and grade, nor did they provide the
data for each of the test items as in Table 2b (Table 2a refers to the
data from Dinosaurs, which was only investigated in the present study
and not by Ritterfeld et al.).
Ritterfeld et al. (2003), however, pointed out that for all but one of the
29 questions, each question was a member of a pair of questions
examining the same concept, for example “I am excellent student” and
“My grades in school could be better” test the self-assessment of
scholastic performance, or, as another example: “I wanted to
investigate why the sea lion pups died” and “I wanted to learn why the
sea lion population went up”. For purposes of comparison with
Ritterfeld et al. (2003), the 28 questions were recoded into pairs that
tested the 14 concepts specified by Ritterfeld et al. (2003) as shown in
Table 3.
The overall means for the two shows appear in Table 4.
Table 3.
Recategorized rating dimensions in accordance with Ritterfeld and Vorderer (2003). The term “recoded” refers
to a reversal of sign to give both scales the same high and low values on the dimension
Achievement 9. My grades in school could be much better. (Recoded)
17. I am an excellent student.
Interest in topic prior to attending the show. 14. When I came in the topic of the show did not interest me.
19. At first I did not really care why the dinosaurs disappeared
from the earth.
Interest increased during the show. 11. After playing I want to learn more about science.
16. After playing I want to learn more about evolution, survival
and dinosaurs.
Table 4
Students: Questionnaire Immediately Following Immersive Presentation
Descriptive Statistics
Sea Lions Dinosaurs
Variable Mean N Mean N
SENSS 3.19 240 3.32 208
ACHIEV 3.84 239 4.15 212
MEDCOMP 3.88 241 3.83 207
ENJOY 4.57 245 4.59 213
PRESEN 3.7 243 3.69 205
CHALL 4.36 240 4.40 217
TOPIC 2.29 239 2.53 207
NIGEMO 1.74 232 1.30 214
USABIL 4.07 243 3.91 216
INTERPL 2.13 233 2.24 205
MOTIVAT 3.98 240 3.48 210
INTERES 3.66 242 3.55 210
LEARN 4.36 240 4.01 215
PLAY AGAIN 4.39 240 4.36 211
RECOMM 4.49 245 4.4 214
Enjoyment
The Enjoyment variable aggregated the judgments that the show was
fun and that the participant would have loved to have played the game
longer. Grade 5 children (Mean = 4.67) significantly enjoyed the
experience more than the Grade 9 students (4.50), although Enjoyment
level was exceedingly high and almost at ceiling (4.58, SD. 0.66) on a
5-point scale, F (1, 23) = 7.3, p < .01. Of considerable interest is the
fact that for the Dinosaur show, Enjoyment increased with
increasing Resolution (4.47, 4.60, 4.73 respectively), but the same
pattern was not observed for the Sea Lion show (4.61, 4.67, 4.43).
The interaction of Resolution and Show was significant, F (2, 23) =
4.30, p < .05 and is attributable to the flatter function for the Sea
Lion show as compared to the steadily increasing function for
Dinosaurs.For Dinosaurs the effect of Resolution is consistent with
expectation, that the highest resolution would be associated with the
highest enjoyment level. However, it is clear that even the poorest
level of resolution here leads to high ratings of enjoyment, and the
effect of resolution, although significant for one of the shows has an
effect of just over one-quarter of a scale value on the 5-point rating
scale.
Presence
Presence is one of the key effects of the Immersion experience, and is
in a sense a measure of Immersion. It represented the aggregate of the
degree to which all else was forgotten during the show and the feeling
that the participant was part of the show. The overall rating was high
(3.69) and almost identical for both shows. While it might have been
expected that the Resolution variable would have influenced the sense
of Presence, this was not the case. There were no significant effects of
Resolution. The only significant effect was that of Gender, with
females (3.80) expressing higher levels of Presence than males (3.57),
F (1, 23) = 5.85, p < .05. The effect was stronger in the Dinosaur
show, and the interaction between Gender and Show approached
significance (p < .054). Females may be willing to report higher levels
of Presence or they may be more willing to “let themselves go”.
Regardless, the effect is again on the order of .25 of a scale value.
Challenge
The Challenge factor aggregated the degree to which the show was
boring, and the excess of information in the show. The mean value
overall was high (4.29) indicating that students were neither bored by
nor overloaded with information from the show. There were no
statistically significant effects of this variable. Clearly, low resolution
did not lead to greater levels of boredom or information overload.
Motivation
The Motivation variable measured the desire during the show to
investigate the problem of the sea lions or the dinosaurs, respectively
(e.g., why the dinosaurs became extinct). Grade 5 children (3.96) were
more motivated than were Grade 9 students (3.56), F = 22.02, p <
.001. Motivation was greater for the Sea Lion than the Dinosaur show
(3.98 vs. 3.48), F = 18.9, p < .001. The effect of show however was
strongest for Grade 9 for whom Sea Lions led to a much higher
Motivation (3.91) than did Dinosaurs (3.11). Comparable values for
Grade 5 were 4.07 and 3.86 respectively, F (1, 23) = 7.3, p < .01.
There was a complex interaction with Resolution, F (2, 23) = 3.10, p <
.05, with Grade 9 students increasing motivation with Resolution on
the Dinosaur show, due largely to a low value on the low resolution
condition (2.8). A more modest incline was shown for Grade 5
students on the Sea Lion show. Since the effect is not found for the
other two grade x show conditions, there is only very modest cause to
suggest that higher resolution increases interest in the scientific issues
of the immersion show. Finally, Gender also played a role in
Motivation in interaction with the show. Females were relatively more
motivated for the Sea Lion show than were males, though equally
motivated for the Dinosaurs as were the males, F (1, 23) = 4.48, p <
.05.
Negative Emotions
Negative Emotions measured anger and upset during the show. It
might have been expected that if the Resolution variable were to have
an impact, it would be here. One has only to recall interruptions and
poor transmission on television or broadcasts in an airplane to
appreciate this possibility. There were in fact several significant
effects observed. Sea Lions led to more negative emotions than
Dinosaurs (although whether this had to do with empathy for sea lions
as compared to problems with the transmission is not clear). The mean
negative emotion score for Sea Lions was 1.74 and for Dinosaurs was
1.30, F (1, 23) = 24.05, p < .001.For the Dinosaur show, the degree
of Negative Emotions decreased with degree of increasing
Resolution (1.45, 1.23, and 1.16 for increasing Resolution). This is
as would be predicted. The same pattern was not obtained for Sea
Lions (1.58, 1.92, and 1.61), consistent with the reduced validity of the
resolution variable here. The interaction between Show and
Resolution was significant, F (2, 23) = 3.79, p < .05. There was also
an interaction with Gender and Show. Females became relatively more
upset with the Sea Lion show (1.23, vs. 1.89), possibly indicative of
the emotional content of the narrative (compare males 1.37 vs. 1.56), F
(1,23) = 7.47, p < .01. Grade, Gender, and Show was also significant,
with the effect for females being most pronounced in Grade 9 (1.21 vs.
1.95) as compared to males (1.49 vs. 1.47) who show no difference at
this grade, F (1, 23) = 4.00, p < .05.
Usability
Usability referred to the ease of figuring out how to play the game and
in knowing what to do during the show. The overall Usability was
high (3.97). Surprisingly, the older and younger students fared equally
well, however, the Grade 5 students found the Dinosaur show
relatively more easy than the Sea Lion (3.92 vs. 4.06) show and the
Grade 9 students found the opposite (4.14 vs. 3.72 for Grade 9 Sea
Lions vs. Dinosaurs), F (1, 23) = 12.14, p < .001.
Interplay
Related to Usability was the Interplay factor which measured the
confusion regarding when to look at the big and small screen, and the
degree of distraction created by the two screens. Males reported less
confusion (1.98) than females (2.36), F (1, 23) = 15.58, p < .001. This
is consistent with the literature on gender and spatial ability and it
would also be consistent with the female willingness to admit
difficulties in spatial tasks. Whether there is a real gender difference
here bears further investigation.
Learn
The Learn factor aggregated the self-assessed amount of learning
during the show including the learning of specific important
information about sea lions or dinosaurs respectively. The overall
mean amount of Learning was high at 4.20. Not surprisingly, Grade 5
students reported higher Learning than Grade 9 (4.42 vs. 4.20), F (1,
23) = 32.31, p < .001. Females reported more learning than males
(4.27 vs. 4.11), F (3.85, p < .05. There was more Learning reported for
the Sea Lion show than the Dinosaur show (4.36 vs. 4.01), F = 20. 31,
p < .001, although this was mostly attributed to the Grade 9 students
who felt they learned much less in the Dinosaur show (3.62 vs. 4.31
for Grade 9 students, and 4.40 and 4.43 for Grade 5 students), F (1,
23) = 13.42, p < .001). Finally, females overall felt that they learned
more in the Sea Lion show (4.00 vs. 4.5 for females and 4.02 vs. 4.20
for males), F (1, 23) = 6.00, p < .05. Resolution however played no
role in the self assessed amount learned in the show. It might have
been the case that the highest resolution would have led to the highest
judged assessed amount of learning.
Recommend
The Recommend variable reflected the interest of the participant in
recommending the Immersion experience to a friend and the wish of
the participant to be provided similar learning experiences by the
teacher. The overall rating was high (4.38). The mean values
increased with the degree of Resolution (4.25, 4.42, and 4.49). In
other words, there was a greater tendency to recommend the
experience with increasing quality of the presentation. However
the effect was not large, but it was significant, F (2, 23) = 3.13, p <
.05. Moreover, this effect was confined to the Dinosaur show (4.22,
4.32, and 4.62 for Dinosaurs vs. 4.29, 4.52, and 4.36 for Sea Lions).
The interaction of Resolution and Show only approached significance,
F (2, 23) = 2.83, p < .06 and statistically does not rule out attention to
the effect of Resolution on the Recommend reaction.
The younger Grade had a significantly higher value (4.49) than the
older Grade (4.28), F (1, 23) = 8.19, p < .004. Finally, the females had
higher Recommend responses for the Sea Lion show (4.50 vs. 4.33) as
compared to males (4.24 vs. 4.40), F = 5.75, p < .05, though it is
difficult to interpret this partiality.
about the topic of the show (be it survival of sea lions or dinosaurs,
respectively).
The mean Increased Interest was relatively high (3.61). Not
surprisingly, the Increased Interest was higher for Grade 5 (3.88) than
for Grade 9 (3.37), since the Grade 5 students have more to learn than
Grade 9 students, F (1, 23) = 29.86, p < .001. The effect was stronger
for Dinosaurs (3.94 vs. 3.16) than for Sea Lions (3.81 vs. 3.54)
suggesting that the Dinosaur show may lead to lower self-assessed
learning for the older than younger students more so than the Sea Lion
show. The Resolution variable also interacted with the grade and
gender, F (2, 23) = 3.84; but no particular pattern is evident other than
for Grade 5, there is a small systematic increase with Resolution.
Sensation Seeking
This variable represented an interest in loud music and noise. Grade 5
children sought this sensation significantly less than Grade 9 students
(3.75 vs. 4.11) F (1,23) = 13.26, p < .001, and those in the Dinosaur
show had a significantly higher value (4.13) than those who saw Sea
Lion show (3.78), F (1, 23) = 13.62, p < .001. In view of the fact that
the difference due to show can result only from unfortunate sampling
and is of the same order of magnitude as that for the difference
between grades, it is unclear what to make of the difference between
grades. The direction of this difference is consistent with that
associated with the recent literature on adolescence which
characterizes adolescents as sensation seeking. The variable did not
interact with any others in the analysis.
Achievement
This variable was based on two self-assessments of excellence as a
student and grades. Ritterfeld et al. (2003) had suggested that poorer
students benefitted most from the immersive educational experience,
again based on self-assessed achievement. It was therefore possible
that better students would be less affected by the resolution variable
than the poorer students would. If the best students by chance ended
up in the high resolution condition, the chance of revealing effects of
high resolution would be reduced (i.e., good students could learn by
and handle any presentation condition). It did so happen that those
students with highest self-assessed achievement were significantly
more prevalent in the two highest categories of Resolution. Mean
Achievement for Low Resolution was 3.067 (.08 SE), for Medium
Resolution was 3.31 (.08), and for High Resolution was 3.40 (.09).
The Resolution factor was statistically significant, F (2, 23) = 4.41,
p < .05, and was attributable to the difference between the low and
high resolution groups (Tukey HSD = .33, p < .01).
The effect interacted with Show because for Dinosaurs, the group that
received the Medium resolution rated themselves higher than any of
the remaining 5 groups. Nevertheless, for both shows, both groups
receiving the Low Resolution rated themselves lower than both groups
receiving the High Resolution, F (2, 23) = 3.07, p < .05.
This distribution of achievement decreases the chances of observing
an effect of immersion because those most likely to benefit from high
resolution (the lower achievers) were not well represented in the high
resolution condition and were disproportionately distributed in the low
resolution condition. Thus, if effects of resolution do emerge, they are
likely to be strong effects that counteract this achievement variable.
The argument is based on the validity of Ritterfeld et al.’s conclusion
that the Immersion experience benefits the poorer students.
Mean Achievement was statistically higher for the Grade 5 students
than for the Grade 9 (3.41 vs. 3.11), most likely due to a greater
concern for grades with increasing grade level, F (1,23) = 10.32, p <
.001.
Media Competence
Media Competence referred to frequency of playing computer games
and surfing the Internet. The overall level was relatively high at 3.85.
Here it would be expected that older students and males would give
higher values than younger students and females and other effects
would be the result of unfortunate sampling. Indeed, Media
Competence of Grade 9 students exceeded that of Grade 5 (3.99 vs.
3.69), F (1, 23) = 7.07, p < .01, and males (4.0) exceeded females
(3.73), F = 6.17, p < .05. The gender variable also interacted with
show, with only a negligible difference between males and females in
the Sea Lion show as compared to the Dinosaurs show, F = 4.04, p <
.05.
Table 5
Teacher’s Assessment (N = 25) Charlottetown and Comparison Data From USA Study
Charlottetown Teachers USA Teachers
N Mean S. D. N Mean S.D.
Table 6
Follow up Appreciation of the Students for Sea Lions
Comparison data from Ritterfeld et al. (2003) Included
Sea Lions Charlottetown Sea Lions USA
Delay of 1 week Delay of 1 month
(Ritterfeld et al., 2003)
Valid N % Agree Valid N % Agree
I would love to play the game again 270 96.7 126 85
I learned quite a bit during the show 267 86.9 121 77
The show helped me to understand how 270 80.4 125 70
scientists work
I talked to friends or relatives about the 270 79.7 126 66
show
The show encouraged me to become 268 67.2 124 61
more interested in science
I still have vivid memories of the show. 263 81.7 126 59
I want to learn more about El Nino 266 60.5 126 52
The show helped me become more 267 55.8 124 42
interested in biology
The show encouraged me to become 267 48.7 125 35
more interested in geography
Mean performance on the factual items for Sea Lions for both
experimental and control groups was low at 58%. Chance performance
is 50%. Had many facts been conveyed during the show, the score of
the experimental group would have been much higher. Mean
performance for both of the groups tested on Dinosaur information
was higher at 68% and 65% for experimental and control group
respectively, but this increase over the Sea Lion study reflects general
knowledge or ease of the test. The difference between experimental
and control groups is negligible and the difference between means was
not significant in a t-test. The r2 on the 32 items for both Sea Lions
and Dinosaurs was moderately high again suggesting that participants’
were relying on general knowledge rather than facts acquired during
the test.
More detailed analysis of the questionnaire in terms of those items for
which knowledge had been presented during the test may provide
useful data, however, the overview here suggests that the educational
benefit of the show is not in the communication of facts that can be
tested through yes-no questions. The more detailed analysis reveals
that for several questions, exposure to the show led to much higher
scores, but this was not always the case, and indeed some scores were
lower with exposure to the show, though not to the same extent.
Ritterfeld et al. (2003) also reported low scores on these questionnaire
items. They found that control questions led to higher responses and
argued that the students had been so inspired by the Sea Lion show
that they had actually sought out additional information in the
intervening month between the show and the questionnaire. This
explanation seems less likely given the high correlation between the
control group and the test group in the present study. The breakdown
for facts contained in the show and those not contained in the show
further supports this claim. At first glance the low scores on the test
for those who have seen the shows challenges the view that much was
learned through the show. There are many kinds of learning and the
kind tapped by the yes-no questionnaire is not that which is acquired
through the immersion experience. Learning how to do something,
broadening a perspective and changing an attitude cannot be easily
tested this way.
Performance for the experimental and control participants for both
shows is shown in the figures.
Dinosaurs
Table 7a
Teacher’s Report One Week After Sea Lions (N = 8)
Discussion Topic
Class Discussion Topic After the Show Per Cent Responded Yes
Technology of the show 75
Aesthetics of the show 75
Causes for the sea lion population decline 25
How to solve scientific problems 12.5
Global warming 37.5
Pollution to the oceans 25
Commercial fishing 12.5
Food chain 0
Table 7b
Teacher’s Report One Week After Dinosaurs (N = 10)
Discussion Topic
Class Discussion Topic After the Show Per Cent Responded Yes
Technology of the show 70
Aesthetics of the show 40
Extinction 30
Animal Adaptation 10
Dinosaurs 40
Pangea 10
Earths’ Periods 0
k/t event 30
Meteor Impact 10
Nothing related to the show 9.1
Table 8
Teacher’s Judgment About Suitability of Immersion Shows (Sea Lions and Dinosaurs) for
Particular Student Groups One Week Later. Question stated: For which students do you
believe the show was especially valuable?
Conclusion
The effect of resolution, though modestly apparent from a statistical
standpoint particularly for the Dinosaur show, was for all practical
purposes negligible. To review Enjoyment increased modestly with
Resolution for Dinosaurs (from 4.47 to 4.73) but showed the opposite
pattern (but more weakly, 4.61 to 4.43) for Sea Lions leading to a
significant interaction of Show with Resolution. Similarly, Negative
Emotions decreased with increasing Resolution for Dinosaurs (from
1.45 to 1.16) leading to a significant interaction with Show and
Resolution. This effect was retained in the analysis of variance for
Dinosaurs alone. There was one significant main effect of Resolution
and that was for the willingness to Recommend the show (increase
from 4.25 to 4.49 with increasing Resolution) but again the effect
interacted with Show and was most prominent for Dinosaurs (range of
.40 as compared to range of .13 for Sea Lions). The Grade 5 students
wanted to play the show more for the highest resolution, and for both
Grades, the lowest resolution led to the least desire to play the show
again. As a potential confounding variable, it so happened that the
highest self-assessed achievement was disproportionately found in the
students who received the highest Resolution condition. If high
achievers are least likely to benefit from high Resolution, this could
account for weaker effects of Resolution than would otherwise have
appeared. Nevertheless, additional analysis revealed that achievement
was not significantly correlated with any other of the 14 dependent
measures, so its confounding effect, if any, is negligible.
That resolution of the presentation a negligible effect on the many
variables examined in the present study may well be consistent with
several theories of multimedia perception and learning. As described
by Ritterfeld et al. (2003), several theorists have also pointed to the
importance of visual imagery in the absorption of an audience into a
narrative (e.g., Green & Brock, 2003). Immersion into the narrative is
referred to as transportation, and absorption of the narrative is referred
to as transformation. Together transportation and transformation lead
to learning. Again, the extent to which visual resolution influences
absorption into a narrative or absorption of the narrative may have less
impact than many other variables that comprise the multimedia
material presented. From another perspective, the Congruence-
Associationist theory proposes that multimedia perception elicits the
creation of a working visual narrative (Cohen, 2001, in press) in an
audience member who aims to create a story to explain the sensory
data received. The establishment of such a narrative depends on
external cues from the media presentation that prompt the cognitive
Acknowledgment
The project was funded under a contract from Immersion Studios Inc
and CANARIE Assistance. Cooperation of Rob Krieger and Dr.
Rodney Hoinkes of Immersion Studios Inc is gratefully appreciated.
As well, at the Atlantic Technology Center, the assistance and
cooperation of John Hughes and Christina Fontaine are acknowledged.
In regard to the contributions of the individual contributors to the
present document: Sandy MacAulay was the primary contract
negotiator during the formative stages of the project. Annabel Cohen
supervised all aspects of the research and is the primary author of the
report and included full consultation with research assistants on the
project Betty Bailey, Kelti MacMillan and Jennifer Ghiz who each
read and commented on an earlier draft. All three worked in
remarkable coherent collaboration so as to complete the project within
a reduced time line that arose for reasons beyond the control of the
team. Kelti MacMillan and Jennifer Ghiz took the primary
responsibility in testing the participants, with Kelti MacMillan
assuming the role of project manager for this aspect. Betty Bailey took
the primary responsibility in writing the ethics reviews and handling
their revision. She also assisted with testing and played a major role in
data entry and analysis. Jennifer Ghiz also assisted with data entry and
analysis. Kelti Macmillan managed the complex organization of the
data collection from the schools. The cooperation of the
administrators, principals, teachers, and students of the Eastern School
District of Prince Edward Island, and the teachers from the Glen
Steward, Spring Park, Sherwood, St. Jeans, West Royalty, West Kent,
Birchwood, East Wiltshire, Stone Park, and Queen Charlotte Schools
is gratefully acknowledged.
References
Cohen, A.J. (2001). Music as a source of emotion in film. In P. Juslin
& J. Sloboda (Eds). Music and emotion (pp. 249 - 272). Oxford:
Oxford University Press.
Green, M. C. & Brock, T. C. (2003). In the mind’s eye:
Transportation-imagery model of narrative persuasion. In M. C. Green
et al., (Eds). Narrative impact: Social and cognitive foundations (pp.
315-342). Hillsdale, NJ: Erlbaum.
Ritterfeld, U., Weber, R., Fernandes, S., Vorderer, P. (2003). Think
Science! Entertainment Education in interactive theatres. Unpublished
paper. University of Southern California.
Executive Summary
Evaluations to assess the perceived advantages and disadvantages of a
single-screen version of Poetic Dimensions software compared with
the existing immersive, three-screen version were conducted at the
University of Toronto and the University of Waterloo respectively.
The intent of developing a single-screen version is to make it available
over a high-speed Internet connection. Since the single-screen version
of Poetic Dimensions is in its formative stages and the demo was a
simulation running on a stand-alone PC, the evaluations used a focus
group methodology. The groups were shown both versions of the
software and asked to comment on the perceived differences, the
advantages, and the relative merits of each in assisting users to make
informed decisions in a planning exercise.
Additionally, the groups were asked to assess the importance of group
dynamics in the immersive version and how a distributed format
would affect the decision-making process.
The initial intent of the evaluators was to have at least twenty
participants in each group, but it was extremely difficult to find expert
subjects who had sufficient experience with the software and could be
available for 2 ½ hours. Most of the participants had extremely
demanding schedules, and we would like to acknowledge their
generosity in making time available for this project. The evaluations
were conducted on October 30 and November 26 respectively with
seventeen participants. The first group had considerable experience
with the immersive software, having used it in the spring of 2003 to
develop a position on the impact of proposed development on the
Lakeshore on Fort York, in Toronto. The second group of participants
was primarily new to both versions.
The evaluations went well with excellent participation and valuable
input from both groups. It should be noted, however, that the groups
were quite different, with the Waterloo group being more focused on
how the software could assist them in their daily planning activities.
As such, they were more forthcoming on questions pertaining to their
perceptions of the existing software than on discussing the potential
advantages and/or drawbacks of a broadband application.
Notwithstanding, the following major conclusions were reached:
The immersive, interactive, three-screen version of the
software has the biggest “wow” factor, but the single-screen
version has the benefit of widespread dissemination and
accessibility (fewer time and space constraints).
Purpose
The purpose of the evaluation was to explore the perceptual issues
around the delivery of high-end immersive and group-interactive
experiences over next-generation broadband networks (as exemplified
by CA*Net 3 and 4). To undertake the evaluation, Immersion Studios
developed a single-screen version of their Poetic Dimensions software
to simulate broadband delivery. The evaluation was designed to
encourage participants to articulate their perceptions of how their
ability to identify issues and to make and share decisions was affected
by the transition from a three-screen 135-degree immersive
environment to a single-screen version of the same environment. It
was hoped that feedback from the evaluation would determine the
feasibility of and critical next steps in the development of a
commercially available broadband application.
Methodology
Two groups were invited to participate in a 2 ½ hour session that
provided the opportunity to compare a 3-screen immersive version of
the software in a group setting with a single-screen version. Every
attempt was made to ensure that the functionality of the software was
similar in each version, certainly to the extent that the group could
explore the environment from multiple perspectives, instruct the
operator to move about the site from both aerial and eye-level
perspectives and add and/or remove elements of the environment in
order to facilitate discussion and ideation. The evaluator and the
operator had pre-determined the “path” that would be taken through
the environment, so that it could be replicated in each version. Each
demonstration began with an overview of the project, a brief
explanation of the purpose of the evaluation, and the guidelines for
responding to questions and probes. The duration of the focus group
for each version was approximately 40 minutes, followed by a
brainstorming session of twenty to thirty minutes designed to explore
potential uses for the broadband delivered version of the software.
Questions were prepared in advance of the sessions. While set
questions and probes had been identified, the intent was to allow the
group as much freedom as possible within the time constraints to
express their opinions in a free-flowing, spontaneous fashion. The
evaluator asked and was granted permission to tape the sessions.
It should be noted that the participants were not able to run the
software themselves or to communicate remotely through some form
of communication tool.
Group 1
The first focus group session was conducted at the Centre for
Landscape Research at The Faculty of Architecture, Landscape and
Design at the University of Toronto. The invitees were part of a group
that had worked with Professor John Danahy at the Centre in the
spring of 2003. The group was composed of city employees from the
planning, urban design and mapping departments (4) and community
activists from the Friends of Fort York group (5), who had come
together to explore the impact of proposed high-rise waterfront
development on Fort York. As such, the group was quite familiar with
the three-screen immersive version of the software. The lab at the
Centre for Landscape Research is not very large, but all nine
participants were seated and able to see the screens clearly, although
they were not ideally seated to interact with one another.
Description of Session
The session began at 6:40 PM with an introduction from Rodney
Hoinkes, Chief Technology Officer at Immersion Studios, who
explained that initially the computing power required to create the
environments was so great that expensive, dedicated systems were
needed. With recent improvements in the power of desk top and lap
top systems and the potential for ubiquitous high speed access, the
ultimate goal for Immersion Studios is to have a version of their
software that can run on a home computer, with users sharing files
over the Internet.
As John began the walkthrough, reminding the participants of the
issues they had met to discuss in the spring, the evaluator initiated a
short discussion on how the issues would have been approached prior
to the availability of the immersive environment. Respondents
indicated that paper documents and foam core models would have
provided the main points of reference for discussion. When prompted
to describe how the immersive software differed from these traditional
references, the group offered the following:
A foam model is tougher. You can’t look inside it. There is no context.
Size is an issue. The model shows a view seen in no other view – it
isn’t real.
It’s difficult to lug a model around.
A drawing captures a moment in time and place, as interpreted by the
artist. The point of view is dictated by the elevation of the drawing.
There is no sense of how the model/drawing relates to anything else in
the environment.
I have no confidence in a paper model.
It is very difficult to make a case with a model.
Specific comments on the advantages of the immersive software over
traditional methods were:
You could see things from eye level, as though you were actually in the
environment.
You are aware of surrounding elements. You can ask to have a closer
look or walk around a building and see it from all angles.
John could add/remove storeys, move buildings closer or farther
away, so that you could see the impact immediately. You didn’t need to
wait for someone to re-render the drawings.
You could brainstorm solutions and see them right away.
decisions on more specific issues. They noted that they had seen other
environments at the lab which had considerably more detail. John and
Rodney pointed out that the software has the capacity to handle much
more detail.
Participants were asked to articulate what advantages/disadvantages
they perceived in having the 135-degree immersive environment.
Since they had not yet seen the single- screen version, the purpose was
to have them concentrate on and evaluate the experience of being in
the immersive environment, rather than focusing on a specific task.
I’m sitting dead centre of the screen, and I don’t see any advantage in
having the side screens. If anything my focus is drawn away from the
main information.
This comment elicited many counter arguments.
It’s not the same information. One screen would be very flat, like
looking at a TV screen.
You wouldn’t get the same sense of being in the environment and
having the context of the things you can see in your peripheral
vision…
The respondent asked John to illustrate this point with a particular
view of Fort York that showed the Gardiner Expressway on one of the
side screens.
…Now imagine what that would be like without the side screen. It’s
not the same at all. You wouldn’t have all the information. You
wouldn’t see what you would see or be aware of if you were actually
there.
It’s important to have the peripheral stuff, but it is more important to
be able to explore the environment and to be able to change the point
of view and buildings.
It’s so important to be able to change things quickly without having to
send people out with the equipment.
Participants were asked their opinions on the perceived
advantages/disadvantages of being in a group setting and having the
ability to see, hear and respond to each other in real time.
It’s very important. We were able to talk back and forth, share
reactions.
We could get input from different directions, try out scenarios really
quickly.
I am seeing the same thing, but now it’s been squashed into a single
screen. Right?
I don’t think so. It’s different.
This was most interesting. John and Rodney explained how trying to
stitch the three screens into one would cause huge distortion and that
what they were seeing now was equivalent to the centre screen of the
earlier demo. Two thirds of the information was no longer available to
them.
Let me look around. I want to turn my head. Show me what happens
when I turn my head.
I think there’s a risk that with that narrow aperture, you will lose your
sense of the whole.
It might be OK for home use, once you’ve seen the large screen.
It could be scary for those who are not used to thinking in abstract
ways.
I see this as a quite different use of the technology. If you accept the
limitations, it can be quite useful, because you could make it more
available, more democratic.
At this point there was considerable discussion about how the product
might work in a distributed version. Several scenarios were
considered.
1 Interested parties could go to a website and download the
model to their computer. They could explore the environment
and then either e-mail or phone in comments and suggestions.
2 A more sophisticated version would allow users to request
changes to the environment, such as changing the height or
density of structures. All users would see these updates and be
able to comment.
3 An even more advanced version would allow
developers/planners to prepare and “drop in” their plans to the
existing environment. Rodney explained that all of these
scenarios were possible. There were considerations, such as the
current lack of standards and file compatibility that would have
to be addressed:
Who would be responsible for the integrity of the model? It would
need to be one person.
I can imagine there will be all sorts of problems with the accuracy and
currency of the environment.
This could be made available from the city website. Everyone could
see the same thing and make comments.
Having the power to change the model is the most important feature.
It’s not enough to just walk around and look at things.
Who is in control with a distributed model?
Hijack the bus!
A developer should be able to drop a piece into the environment. Look
at how much time we would save.
The group was asked if they felt that a distributed version would be
useful in terms of enabling larger numbers to have input into planning
decisions. Had it been difficult for their group to meet, given the busy
schedules of all participants?
It would have been very difficult if John hadn’t agreed to be available
to us twenty-four hours a day.
I can see somebody not being able to make a meeting because of kids
or whatever, going home and looking at this and then being able to
make comments and contribute.
I think you could use it as preparation for a group meeting.
What about the level of participation? You can’t know whether
someone is serious or just on the periphery.
Is it a problem for the city to get community input because of
scheduling issues?
There are always people who are upset that they can’t make a
meeting. Anything that we can do to make information available to the
community is good. It’s important to connect the community to
ongoing information.
The discussion again turned to level of detail and the possibility that at
both ends of the spectrum – the developers and the general public –
credibility could suffer if an appropriate level of detail in textures,
shadows, etc. was not made available.
A suggestion was made:
Perhaps what we need is an analogy to the building model; first you
need a box which shows the zoning envelope, then you show cast
shadows; then you add detail.
If this was to be made available in a distributed mode, would the
general public have enough knowledge to know what they are looking
at?
We probed for how important the group dynamic was versus the
distributed model where communication could be synchronous or
asynchronous, but not face-to-face.
Ideally, bring people together, but failing that, it’s better than no
community involvement.
It’s sort of like a speaker phone. You can’t see the person’s real
reaction. They could be agreeing with you, but making faces at the
same time.
The benefit is in the numbers you could reach. I mean, how many
installations are there like this?
I think what you have to be aware of is that many decisions are made
without information. There are so many other things going on. I might
be there to discuss the design intent, but there are all kinds of hidden
agendas; it could be money or politics.
Do you think that you could make the same decisions using distributed
software as you did with the immersive 3-screen version?
Yes, I believe that we would have moved in the same direction. I can’t
assume that we would have come to the same conclusions.
It was useful to sit in front of the screen together.
Individual decisions and discussions would probably have taken
longer. You lose the ability to have that back and forth dialogue.
Do you think that there is a commercial application for this software?
Yes, I think it is viable. It could be used by city planners, architects,
etc.
I think it would be very useful when you had multiple projects on the
go.
It needs to be tied to an underlying database. It needs to interface with
related information.
Its portability is very attractive. You could take it to reluctant
observers.
It needs to be widely available to the community. We need something
like this to democratize the decision-making process.
I think we all agree that the best is the immersive, 3-screen version in
a group setting, but because it reaches so few, a distributed version
would be useful.
The session wrapped up at 8:40 PM with thanks to all participants.
Group 2
The group consisted of planners from the City of Kitchener and the
region of Waterloo. In all, there were eight participants. Professor
Thomas Seebohm led the discussion with Kevin Moule, the systems
manager and applications programmer running the software. Rodney
Hoinkes from Immersion Studios provided the introduction and
context for the evaluation. Two graduate students were present as
observers and to answer questions on the proposed buildings they had
prepared for insertion into the environment. This was a very different
group from the one at U of T, being quite homogeneous both in
occupation and age. All are directly involved in planning activities on
a day-to-day basis. From their questions and comments, it was clear
there was a high level of knowledge of and comfort with technology.
The planners at the City of Kitchener regularly work with software
that allows them to create a simple massing model of a development
proposal and to insert that into a model of the existing context where
the latter is also a simple massing model.
The planners use the aforementioned software to evaluate the visual
impact of the height and massing of a proposal. Throughout the
session, it was noted that some of the planners were comparing their
existing software with the Poetic Dimensions software. Their focus
was narrower than the U of T group and discussion about a broadband
version of the software was harder to elicit.
Description of Session
The session was conducted in a lab at the School of Architecture,
Faculty of Environmental Studies at the University of Waterloo. A
three-screen set-up, similar to the one at U of T, had been set up with
an actual viewing surface of approximately 135 degrees. The screens
were custom built in the engineering department at the university and
have been coated with an anti-reflective paint. The coating was
employed to reduce “bounce” in the stereoscopic projection. With the
front-projected three-screen version, it had the effect of making the
two side panels appear darker than the centre panel. The most
beneficial viewing point was directly in front of the middle panel.
Accordingly, throughout the second part of the session when we
investigated the three-screen projection, we encouraged participants to
rotate, until each had the opportunity to view the environment from the
most advantageous position. Without further testing, it is difficult to
know how much this phenomenon may have affected their perceptions
of the three-screen set-up.
We want to start out from the bird’s eye view, because we are
planners, we need to see the big picture, but to feel it on the ground
level, you need the detail.
You need more (detail) than in the existing ground level. It becomes
important when you have to consider things like parallel parking and
landscaping.
If you were trying to develop a vision, rather than a specific plan, do
you need the level of detail you have been discussing?
Can you add dynamic elements, for instance, could you superimpose
items like an LRT on the existing environment, so people could see the
train running through it? That would be amazing. (Rodney answered
in the affirmative.)
It depends on the audience. Politicians, bankers and Joe Public would
need the level of detail to relate. They would find it very compelling
and infinitely more valuable than anything we can currently do.
You could show them the existing environment and then say “Let’s
work with you” and begin the dialogue.
At this point, we took a short break and moved to the three-screen
immersive software.
The evaluator asked the group for their reactions to the three screen
environment.
I didn’t notice there were three screens. I am still focused on the
centre one.
Does it have to be three screens? Two would be good. The third
doesn’t add anything.
Could the single screen just be panoramic?
Do you believe that you are seeing the same information?
It’s more realistic. It fills your peripheral vision. I’m not sure if the
information is the same. I think so.
I think you are seeing more. Now I see it. I think there’s more
information. It makes you feel more as though you are actually there.
This would be great for showing store owners what their storefronts
look like. Most of them have no idea. It makes it so realistic.
It makes you move your head more, unless you are right in the middle.
The public needs this. The public needs to feel they are there. We
(planners) can fill in the gaps. This would make the public feel more
relaxed. It would give them a sense of comfort.
The one screen makes you feel that you are watching a movie; this is
more real.
It makes your eye denser somehow.
Are you moving faster through the environment? I felt the other was
too slow, but it could be just that there is more to look at here and it
keeps the eye involved.
Yes, it’s more comfortable than the single screen, because the eye has
so much to look forward to. There’s more in the space.
It feels very real. I became more conscious of the things I would do
normally (on a site visit), like looking to the left or right as I went
around a corner. I feel I want to do those intuitive things now.
At this point Rodney switched off the two side screens. The evaluator
asked how the group was affected.
I didn’t notice going back to the single screen.
His imagination filled in the two screens.
I don’t think it is important how many screens there are; it’s how you
discuss the issues.
It’s not critical to have three screens with developers, etc. Once the
“Wow” factor is over, the real power is in the tool having
manipulability.
That’s the most important thing, to be able to change elements of the
environment as we are looking at them and discussing them.
But it would be good to have the “wow” factor at final presentation.
The group responded to a question about how important it would be to
be remotely connected. Did they have difficulty bringing a group
together?
It’s not so much an issue of travel time; it’s more a function of being
too busy.
I don’t think it would be important for this group to have an Internet
connection to a version you could manipulate remotely.
It could be used as a method of doing your homework for a planning
session. There might be a better sense of coordinated control. On the
Internet, you’d be jumping back and forward. It might be too crazy.
The group saw the main advantages of the software to be:
The ability to do infinite numbers of changes and adjustments quickly
and easily. A paper plan takes too long. People are more likely to
Findings
It is important to acknowledge some contributory factors in the
evaluations, before attempting to outline the major findings.
The two groups were quite different, both in demographics and in their
familiarity with the software. For many of the Waterloo participants,
the demo session was their introduction to the software, whereas the U
of T group had worked together with the immersive software over a
period of weeks.
There was a difference in the way that the single-screen software was
presented to the two groups. At U of T, the participants saw the
software demonstrated on a laptop computer, to more nearly simulate
the experience of a single user; at Waterloo, the image was projected
onto the centre screen of the three-screen set-up, because there were
too many participants to gather around a single computer. The
differences may have affected responses. It was not possible at the
time of testing to fully simulate the experience of being a single user
connected through broadband technology to other networked users. As
such, we asked the respondents for comments on the usefulness and
feasibility of a concept. The information gathered through the sessions
has confirmed the perceived value of a single-screen version of the
software utilizing broadband networks, and more clearly defined the
next stages of its development, as well as identifying several
interesting areas for further research.
their seats, eyes on the screen and there was little unrelated discussion.
Although the sessions ran slightly longer than planned, none of the
participants left before the end.
Advantages
Both groups alluded to a “Wow!” factor with the three-screen version.
They felt that being together in the same physical space, with the
opportunity to manipulate the information and discuss changes to the
environment in real time was the preferred option. Participants
expressed the opinion that the immersive version provided the most
realistic experience, the closest to actually being there, while still
offering the benefits of manipulability. Several participants who are
not trained planners or architects explained that the software promoted
confidence to examine a proposed development because they could
walk through the environment, view it from multiple perspectives and
understand the relationships amongst the elements.
Many participants from both groups expressed the opinion that this
type of software would increase public confidence in the planning
process. It was felt that models, plans and artist’s renderings do not
provide enough context for the average viewer who cannot visualize a
finished building from a two-dimensional plan.
Several participants felt that this software would be very useful for
final presentations to small groups of proposed developments.
Disadvantages
Participants in both groups acknowledged the constraints of space,
time and money in setting up such a system and its limited
accessibility. While there was nobody who indicated the immersive
version was not useful, some participants were unable to distinguish
between the three-screen and the projected single-screen versions.
There was discussion amongst both groups about the level of realism
required to make the software credible, particularly with the public.
Both groups suggested a graduated approach, where wire frame and/or
massing models could be used in early planning with experienced
personnel. Graphic sophistication could be added for presentation to
those less able to think in abstract three-dimensional forms. These
discussions were prompted by both the three-screen and single-screen
versions.
Conclusions
Overall it was perceived that the three-screen immersive environment
made for the most realistic and interactive experience, because of the
engagement of the peripheral vision that made it most analogous to the
real experience of walking through an environment, but clearly the
groups recognized its limitations in terms of cost and accessibility.
Single-screen version
Advantages
The perceived advantage of a distributed version for both groups was
for the ability to provide a publicly accessible website, perhaps as an
extension to the municipal planning authority, where plans under
consideration could be downloaded, examined and commented upon.
The U of T group described differentiated levels of manipulability,
depending on the expertise of the contributor. At the highest level,
developers and planners could amend the plan or even insert their own
concepts into the existing environment. It was felt by the planners that
this would encourage more creative discussion in the time saved on
paperwork and administration. However, there was concern about
standards and the manageability of multiple versions of a plan.
The ability to change elements in the plan instantly to test hypotheses
rather than suffer the current delays in creating iterations was seen as a
very powerful advantage for both versions of the software. This
advantage was the one referred to most often by both groups. The
second group kept coming back to it, perhaps because their demo
environment only allowed the user to remove and substitute elements,
not change the nature of the elements themselves.
Several references were made to using the single-screen version of the
software as a tool that would help users to prepare for planning
meetings and discussions. This preparation was seen as particularly
valuable when confrontational negotiations were anticipated.
Disadvantages
Initially, members of the Waterloo group indicated that they did not
see value in a broadband application for the software; however on
further prompting they conceded that it might be a useful vehicle to
promote community involvement in planning decisions. There was
also consensus that being together physically for discussions,
especially when those discussions are confrontational is the most
With the University of Toronto group, there was a strong feeling that
this software in its distributed format would democratize planning and
make public consultation more accessible and feasible, although
concerns were expressed about how to manage the process to maintain
credibility.
The Waterloo participants were more reluctant to embrace a
broadband version of the software. While they could clearly see the
application of the software in their jobs, they initially rejected the
concept of a broadband application.
Of interest and a possible avenue for further research was the
confusion expressed by some of the participants as to the differences
between the three-screen and the single-screen version. Some
respondents were unable to discern that the information they were
seeing was significantly different. Although they were seeing one third
of the information available to them on the three-screen immersive
projection, they believed it was the same, just compressed to fit on the
smaller screen. Others immediately recognized the differences. With
the first group, the confusion seemed to lie in a clear division between
those who worked with abstract visualization as a matter of course and
those who did not. However, the Kitchener/Waterloo participants
expressed the same confusion and they are all planners. It would be
interesting to pursue research to ascertain whether they actually did
not perceive the difference in the amount of information, or whether,
as one respondent explained, “Imagination fills in the information
from a memory of what was/should be there.”
Finally, respondents indicated that they believe there is a future for a
distributed version of the software. They would definitely like to see
developments in the software that would make it a simple, user-
friendly tool to improve community access and input in planning
decisions.
Although the tests conducted by Egenuity and Immersion Studios
were simulations designed to replicate the performance of broadband
performance, The University of Waterloo has already installed
software at the municipal offices to facilitate collaboration between
their organizations. The installation and established relationship
between the university and the municipality could provide an ideal test
bed for further development and for ongoing research to assess how
well the software meets its perceived capabilities to improve
community access to and input on public planning.
Acknowledgements
Assistance for this project was provided by the Minister of Canadian
Heritage and the CANARIE Applied Research and Interactive Media
program. CANARIE implements and manages the national research
and education broadband network CA*Net 4. In addition, it
administers and supports research and development initiatives in e-
Content, e-learning, e-health and e-business.
Special thanks go to Professor John Danahy of the University of
Toronto’s Centre for Landscape Research and Professor Thomas
Seebohm of the University of Waterloo’s Faculty of Environmental
Studies who generously provided of their time and facilities to make
the evaluation a positive experience for all.
Appendices
Questionnaire 1
Comparison of Immersive and PC Environments: Focus
Group Questions for Immersion Studios October 30,
2003
Opening
Thank you so much for agreeing to attend this session and help us with
the evaluation. As you know, Immersion Studios has been a leader in
the field of developing immersive technologies for a variety of
purposes. Since Rodney is with us, I will ask him to give us a brief
history of the development of the software you will use tonight and tell
you a bit about the CANARIE contribution, their requirements and the
purpose of tonight’s evaluation.
Introduce Rodney – Chief Technology Officer of ISI and former
professor at U of T.
Thanks, Rodney. We have one last little piece of business. I would like
to record the session tonight for the sole purpose of being able to refer
to the tape to ensure accuracy in my report for ISI. No names will be
used in the report, only a general description of the number of
attendees and the session. Once the report is written, a copy will go to
ISI for their records. No other copy will be kept.
Comparison of Immersive and PC Environments: Focus Group
Questions
Note: As we do not want to lead the subjects, the probes in red will
only be used if the subjects lead the discussion into the specific topics
covered by the probes.
In contrast, the probes in blue will only be used if the subjects do not
mention a particular topic in the context of their responses.
Introduction
Goal of session (i.e., to compare single-screen version to
immersive environment) Not usability of the software.
Who we are
Names of participants, roles
Issues explored by Friends of Fort York
Can you tell me a bit about your group, Friends of Fort York?
What kinds of issues has your group looked into using the
immersive environment?
How did you explore these issues in the past prior to the
availability of the immersive software?
How might you use this at one of your meetings, if one of the
group members brought it in on a laptop?
• Would you find it useful to have this kind of access? Why?
• Would it have been useful at your meetings a few months
ago when preparing your case?
• Would it be useful to bring the laptop version to any other
meetings?
Comparison of single-screen Version and Immersive Environment
(before seeing immersive environment)
What benefits, if any, are there using the single-screen version
vs. the immersive environment for you or your group (based on
their memory of the immersive experience)?
Possible Probes, depending on their answers:
• Would you benefit in using the single-screen version in any
way from a singular/personal exploration of the space?
• Are there any benefits to being able to control the
exploration yourselves?
• Would your group benefit from being able to return to the
environment multiple times?
What drawbacks, if any, are there using the single-screen
version vs. the immersive environment for you or your group
(based on their memory of the immersive experience)?
Probes if necessary:
• Does it give you identical views?
• Could you use it to make decisions?
Exploration of Immersive Environment and Comparison
• Let’s go to the immersive environment now and explore it.
John will operate it for the group. (Direct John to what you
want to see so that you can make a comparison to the single-
screen version.)
• How would you compare it to the single-screen version now
that you’ve had a chance to see them side-by-side?
Probes if necessary:
• Do they both give the same information?
• Are there things in the immersive environment that you did
not see in the single-screen version?
• Do they both allow you to make decisions or draw
conclusions? Do you draw the same conclusions?
• Does this environment change the dynamics of the group at
all?
Questionnaire 2
Comparison of Immersive and PC Environments: Focus
Group Questions for Immersion Studios, November 24,
2003
Opening
Thank you so much for agreeing to attend this session and help us with
the evaluation. As you know, Immersion Studios has been a leader in
the field of developing immersive technologies for a variety of
purposes. Since Rodney is with us, I will ask him to give us a brief
history of the development of the software you will use today and tell
you a bit about the CANARIE contribution, their requirements and the
purpose of today’s evaluation.
Introduce Rodney – Chief Technology Officer of ISI
Thanks, Rodney. We have one last little piece of business. I would like
to record today’s session for the sole purpose of being able to refer to
the tape to ensure accuracy in my report for ISI. No names will be
used in the report, only a general description of the number of
attendees and the session. Once the report is written, a copy will go to
ISI for their records.
Comparison of Immersive and Single-screen Environments: Focus
Group Questions
Note: As we do not want to lead the subjects, the probes in red will
only be used if the subjects lead the discussion into the specific topics
covered by the probes.
In contrast, the probes in blue will only be used if the subjects do not
mention a particular topic in the context of their responses.
Introduction
• Goal of session (i.e. to compare single-screen version to
immersive environment). Not usability of the software.
• Who we are
• Names of participants, roles
Focus group participants’ introduction
• Introductions
• Can you tell me a bit about your group?
• What kinds of issues does your group consider?
Probes if necessary (i.e. if subjects don’t discuss):
• What is involved? (e.g., number of people, tasks)
• What information are you looking for?
• How do you use the information?
• How do you explore these issues currently
Probes if necessary
• What types of media/materials/documentation do you use?
• How many people are consulted/involved?
Introduction
One wide-angle lens only allows for one center of perspective, radiating from the
center screen.
Although the complete image is in three separate 4:3 images, there is
only one center of perspective in the center screen. This causes
unreasonable stretching in the side screen images (e.g. wide-angle
distortion) and simply does not look correct in the 3-screen theatre.
In the diagram below is the proper rendering, using three separate
cameras with therefore three centers of perspective.
Correct
Introduction
When working in any production development environment it is
paramount to follow standards and none is more important to be
following than well-defined naming conventions.
Object naming
Naming components is a key activity in all CG modeling
environments. It can be treated either in a very casual or very strict
manner, depending on the organization or on the number of CG
specialists working on a project. Names of data objects are the primary
method to access the data and control it interactively.
Naming occurs at three levels depending upon how the applications’
data is organized:
1 Files.
2 Layers.
3 Objects.
Object naming using common language rather than technical is
critically important. End users of a GIIE may end up using these
names to select objects or data to interact with or create links to. Non-
intuitive (i.e. FM-01-LXD-0443) or generic (i.e. BOX01) naming can
hamper or even eliminate interactions altogether. Ideally, naming has
been found to focus on the common name, domain-specific type of
data, its position or scale, and perhaps a version or option identifier
(i.e. HOUSES-SE-2story-optionA, or POSTOFFICE-45m-v1).
Another important use of naming is in controlling complexity. To
simplify models, change options or show detail when focused on one
area, a user can manipulate data by its name organization or through
direct geometric manipulation. Naming provides a mechanism to
group data together to quickly be manipulated.
1 INI files are a standard notion for most developers under Microsoft Windows. While
Macromedia Flash can have a script written to parse INI format files, it is much easier to
use Flash INF files which are slightly different in formatting, using a ‘&’ between
variable/data pairs instead of carriage returns.
University of Toronto
Mr. John Danahy
The University of Toronto, Faculty of Architecture, Landscape and
Design, Centre for Landscape Research is a world leader in real time
visualization as a design and planning tool. The CLR currently
operates an Immersion Simulation Lab, and continues to develop and
experiment with a variety of Urban Planning, Architecture and
Landscape Architecture projects. The City of Toronto, City of Ottawa,
and National Capital Commission are frequent clients utilizing the
immersive and interactive capabilities of the CLR. The CLR will be
providing content and content editing for urban models being used in
professional decision-making as well as consulting on evaluation goals
and methodologies for professional design and planning applications.
UofT has provided regional planning content, content editing and test
groups, as well as its Immersion Simulation Lab for testing to this
project.
University of Waterloo
Professor Thomas Seebohm
University of Waterloo, Integrated Centre for Visualization, Design
and Manufacturing, is conducted a variety of large scale planning and
design exercises for the Region of Kitchener/Waterloo, using its
Egenuity Inc
Ms. Marilyn Welsh
Egenuity Inc. is a full-service e-learning firm offering consulting
services to education and training organizations in the areas of
strategic planning, project management, instructional design, training,
and evaluation. Its principal officer, Marilyn Welsh, has over twelve
years experience in multimedia and e-learning development and is a
published author and frequent speaker on the use of technology in
education. She has advised educational institutions and private
organizations in Canada, the USA, Mexico, South America, the UK,
Europe, Australia and New Zealand. Egenuity conducted the
professional evaluation portion for this study.
Sheridan College
Mr. Avrim Katzman
Sheridan College, SCAIT, working with its Immersion Simulation
Lab, has been conducting research in a variety of new media areas;
including interface design and functionality, group interactive content
development, and animation techniques.
Centennial College
Mr. Nate Horowitz
Centennial College’s Centre for Creative Communications offers
comprehensive teaching in Interface Design, Video Production and
post, 3D animation, Online Writing and Editing; Instructional Design
and Information Architecture. The faculty and students, including
adjunct professors have all worked as developers and designers in
broad bandwidth applications within the private sector for at least 5
years.