An Introductory Guide To Scientific Visualization

User Manual: Pdf

Open the PDF directly: View PDF .
Page Count: 169

Download
Open PDF In Browser	View PDF

An Introductory Guide to
Scientific
Visualization

R. A. Earnshaw
N.Wiseman

An Introductory Guide to
Scientific Visualization
With 72 Figures

Springer-Verlag Berlin Heidelberg GmbH

De. Rae A. Eamshaw
Head of Computer Graphics
University of Leeds
Leeds I.S2 9JT, U. K .

Nonnan Wiseman
NERC Computer Services
Kingsley Dunham Cent~
Keyworth
Notts NG12 5GG, U. K.

Front covrr plate $hoW$ an image of Saa Frandsco with sc:vc:ntl multi-$pcctntl
analy5c:S bc:1ng c:mied out. Counesy of Silicon Graphlcslnc., AVIRlS data, and the
Jet Propulsion lJIbonnory.

ISBN

978-~2-63470-3

libnuy of Congress CatalogiJlg·in-Publication Data
Earnshaw,RaeA.An introductO!:")' gulde to sclcntlficvlsualization/RA.Eamshaw,
N. Wlseman. p . cm. Includa bibliographical refcrences and Index.
ISBN 978-3-642-63470-3
ISBN 978-3-642-58101-4 (eBook)
DOI 10.1007/978-3-642-58101-4

1. Science-Methodology. 2. YisuaUzation-Data proces!lng. 1. Wlsc:maa, N.
(Norman) II. Title. Q175.E2H 1992 502.8-dc20 92-10987 CIP

This work is subject 10 copyright.AlI rights are reserved, whether the whole
part of the material is concemed, spcclfically the rtghts of translatlon,
reprinting, reusc: of ilIustrations, recitation, broadcasting, rcproductlon on
microfilms or in aay other way, and storage in data banks. Duplication of this
publication or pans thereof ls pcrmitted only under the provlsiofl5 of the
~rman Copyright Law of Septembc:r 9, 1965, In II.!; currcnt version, and
pcrmission for usc: must always bc: obtained (rom Springer-Yerlag. Yiolations
are llable for prose<:ution under the ~rman Copyright Law.

CI Springer·Yerlag Berlin Heideibc:rg 1992
Originally publi.~hed by Spri nger- Yerlag Berlin Heidclberg New York in 1992
Softcovcr repri nI of lhe hardcovt.'T Ist edilion 1992
The usc: of registercd names. trademarks, etc. In thls publlcation does nOI
imply,even in the absence of a specific statement, that such names are c:xempt
(rom the relevant protective bws and regulations and there!ore free for general
"~ .

Cover Design: H. l.oplca, Iivesheim
Typnctti.og: K + Y ~'otosaIZ GmbH, Bee rfc1dcn
3313140-5432 - Printed on acid-frec paper

Foreword

Visualization has been the cornerstone of scientific progress
throughout history. Much of modern physics is the result of
the superior abstract visualization abilities of a few brilliant
men. Newton visualized the effect of gravitationa.l force
fields in three dimensional space acting on the center of
mass. And Einstein visualized the geometric effects of objects in relative uniform and accelerated motion, with the
speed of light a constant, time part of space, and acceleration indistinguishable from gravity. Virtually all comprehension in science, technology and even art calls on our
ability to visualize. In fact, the ability to visualize is almost
synonymous with understanding. We have all used the expression "I see" to mean "I understand".
Modern science part departs from the closed theories of
the last century and demands computer simulations to
understand real world situations. Scientific Visualization is
the eyes through which these simulations are viewed, from
electrochemical bonds to simulated interstellar jets associated with black holes.
Scientific Visualization is of value beyond strictly scientific applications, however. The same technology is now
used in such diverse applications as clothing design, industrial design, automobile and airplane design, genetic
engineering, chemical and drug design, oil and mineral exploration, chemical and nuclear power plant design, and
motion picture special effects and animation. It is rapidly
becoming a requirement for virtually all disciplines that
deal with geometric things.
What is Scientific Visualization? It is a set of software
tools coupled with a powerful 3D graphical computing envi-

Foreword

ronment that allows any geometric object or concept to be
visualized by anyone. The software provides an easy to use
interface for the user. The hardware must be able to manipulate complex, geometrically described, 3D environments
in motion, color and with any level of "realism" called for
to better communicate the essence of the computation.
Scientific Visualization is in its infancy, but the technology is sure to revolutionize scientific education. I believe
that the requisite 3D graphical processing capability will be
built into all personal computes within the next five years.
And by the year 2000, I am confident that even the home
digital television will combine such 3D graphical processing
capability with digital video and audio. Then, even complex
scientific textbooks will be viewed interactively on the
home screen, with video clips depicting a lecturer, mathematical experiments run in and visualized on the "TV" and
the student able to guide the learning process. But until
then, such books as this will guide the way.
May 1992

James H. Clark
Chairman, Silicon Graphics Inc
Mountain View
California, USA

Preface

Scientific visualization is concerned with exploring data
and information in such a way as to gain understanding and
insight into the data. This is a fundamental objective of
much scientific investigation. To achieve this goal, scientific
visualization involves aspects in the areas of computer
graphics, user-interface methodology, image processing, system design, and signal processing.
This guide is intended for readers new to the field who
require a quick and easy-to-read summary of what scientific
visualization is and what it can do. Written in a popular and
journalistic style with many illustrations, it will enable readers to appreciate the benefits of scientific visualization and
how current tools can be exploited in many application areas. It will be invaluable for scientists and research workers
who have never used computer graphics or other visual
tools before, and who wish to find out the benefits and advantages of the new approaches.
This guide is concerned to answer the questions which
the newcomer to visualization may wish to ask concerning
what it is, what it can do, what facilities are available, and
how much it costs. Points of contact for further information are also provided.

VIII

About the Authors

About the Authors
Dr. R. A. Earnshaw
University of Leeds
UK

Mr. Norman
Wiseman
Natural Environment Research
Council, UK

Rae Earnshaw is Head of Computer Graphics at the University of Leeds, with interests in graphics algorithms, scientific visualization, display technology, CAD/CAM, and human-computer interface issues. He has been a Visiting Professor at Illinois Institute of Technology, Chicago, USA,
Northwestern Polytechnical University, China, and George
Washington University, Washington DC, USA. He was a
Director of the NATO Advanced Study Institute on "Fundamental Algorithms for Computer Graphics" held in Italy,
England, in 1985, a Co-Chair of the BCS/ACM International Summer Institute on "State of the Art in Computer
Graphics" held in Scotland in 1986, and a Director of of the
NATO Advanced Study Institute on "Theoretical Foundations of Computer Graphics and CAD" held in Italy in
1987. He is a member of ACM, IEEE, CGS, EG, and a Fellow of the British Computer Society.

Norman Wiseman is Northern Area Computer Services
Manager for the Natural Environment Research Council.
His special interests are in the application of visualization
in physical and biological sciences of the environment; education and training of scientists in the use of graphical techniques; and raster print technology. He has been a systems
consultant for a number of years and has been involved in
several graphics hardware, software and training initiatives
in the UK Academic and Research Council communities.
Prior to this he has worked on a number of software projects involving the acquisition, storage, display and analysis
of seismic and borehole log data, primarily for use by scientists in the British Geological Survey. He is a member of
Eurographics.

Contents
Acknowledgements .................
Disclaimer . . . . . . . . . . . . . . . . . . . . . . . ..
Copyright Material .................
Trademarks ........................

Part I

XIII
XIV
XIV
XV

Basics of Scientific Visualization

Chapter 1 Introduction and Background
1.1
1.2

Introduction. . . . . . . . . . . . . . . . . .
Background ..................

3
3

Chapter 2 What Scientific Visualization
can do!
2.1
2.2
2.3
2.3.1
2.3.2
2.3.3
2.3.4
2.3.5
2.3.6
2.3.7
2.3.8

What is Scientific Visualization?
How to do Scientific Visualization
Some Examples of Scientific
Visualization .................
The March of Napoleon's Army
Cholera Outbreak .............
Weather Maps from Meteorology
Molecular Modeling ...........
Pelvic Reconstruction ..........
Oil Exploration ...............
Designing Ship Propellors ......
Visualization of Forest Growth

5
7
8
8
9
10
12
12
14
16
18

Chapter 3 Explanation of Scientific
Visualization Terminology
3.1
3.2
3.3
3.3.1
3.3.2
3.3.3
3.3.4
3.4

Techniques...................
Volume Visualization ..........
Data Types ...................
Overview of Facilities. . . . . . . . . .
HDF ........................
NetCDF .....................
Databases ....................
Current Application Areas .....

20
25
27
27
29
30
31
31

Contents

3.4.1
3.4.2
3.4.3
3.4.4
3.4.5
3.4.6
3.4.7
3.4.8

Cartography ..................
Statistics .....................
Remote Sensing ...............
Archeological Reconstruction ...
Molecular Modeling ...........
Medical Science ...............
Oceanography ................
Computational Fluid Dynamics.

31
32
32
32
33
33
34
34

Chapter 4 Facilities for Scientific Visualization
4.1 Visualization Software Categories
4.1.1 Graphics Libraries and
Presentation Packages ..........
4.1.2 Turnkey Visualization
Applications ..................
4.1.3 Application Builders ...........
4.1.4 Choosing a Package ...........
4.2 Software Costs ................
4.2.1 Subroutine Libraries and
Presentation Packages ..........
4.2.2 Turnkey Visualization Systems ..
4.2.3 Application Builders ...........
4.3 Hardware Considerations
(including Hardcopy) ..........
4.4 Vendor Systems Versus Public
Domain Systems ..............
4.5 Summary ....................

35
35
36
37
37
38
38
39
39
39
40
4~

Chapter 5 Outputting Results
5.1
5.2
5.3

Hardcopy. . . . . . . . . . . . . . . . . . . .
Video. . . . . . . . . . . . . . . . . . . . . . . .
Other Media .................

44
45
46

Chapter 6 Current Developments
and Activities
6.1
6.2
6.3

USA.........................

UK..........................
Europe. . . . . . . . . . . . . . . . . . . . . . .

47
49
50

Contents

Part II

Overview of Current Systems
and Developments

Chapter 7 Current Vendor Systems in Use
7.1
7.2
7.3
7.4
7.5
7.6
7.6.1
7.6.2
7.6.3
7.6.4
7.7
7.7.1
7.8
7.9
7.9.1
7.9.2

7.10
7.11
7.12

Wavefront Technologies, Inc. ... . 53
UNIRAS A.S. ................ 58
Precision Visuals, Inc. . . . . . . . . . . 63
Stardent Computer, Inc. ... . . . . . 67
Silicon Graphics, Inc. .......... 72
Sun Microsystems, Inc. .... . . . . . 79
SunVision - Sun's Visualization
Software Package .............. 79
SunVision Programming
Interfaces .. . . . . . . . . . . . . . . . . . . . 79
SunVision Window-based Tools
80
The VX and MVX - Sun's
Visualization Accelerators ...... 82
Sterling Federal Systems, Inc. ... 88
FAST (Flow Analysis Software
Toolkit) ...................... 88
Dynamic Graphics Ltd. ........ 89
Spyglass, Inc. ................. 96
Spyglass Transform ............ 96
Spyglass Dicer ................ 97
LightWork Design Ltd. . . . . . . . . . 98
Ricoh Company Ltd. .......... 100
Vital Images, Inc. ............. 105

Chapter 8 Current Public Domain Systems
in Use
8.1 Khoros ......................
8.1.1 Overview ....................
8.1.2 Subsystem Component
Descriptions ..................
8.1.3 Current Status of Khoros ......
8.2 apE: A Dataflow Toolkit
for Scientific Visualization . . . . ..

106
106
107
112
117

Contents

XII

8.3
8.4

8.4.1
8.4.2
8.4.3
8.4.4
8.5
8.6
8.6.1
8.6.2
8.6.3
8.6.4
8.7
8.8

National Center for Supercomputing Applications (NCSA)
GPLOT, DRAWCGM, P3D
(Pittsburgh Supercomputer
Center) ......................
The GPLOT CGM Interpreter ..
The DrawCGM Graphics
Subroutine Library ............
The P3D Three-Dimensional
Metafile Project ...............
Software Availability ...........
RAYSHADE. . . . . . . . . . . . . . . ..
NASA Ames Software .........
PLOT3D .....................
SURF .......................
Graphics Animation System
(GAS) .......................
Applications in Computational
Fluid Dynamics (CFD) ........
Irisplot ......................
ISVAS. . . . . . . . . . . . . .. . . . . . . ..

127

128
129
131
133
135
135
137
137
137
138
138
140
140

Chapter 9 Other Uses of Visualization Tools
9.1 Art and Design ............... 142
9.2 The 5th Dimension Animation
System....................... 142
9.3 Multimedia Environments
146
Chapter 10 Conclusions
10.1 Strategic Importance of Scientific
Visualization .................
10.2 Current Developments .........
10.3 More User-Friendly Facilities ...
10.4 Further Information ...........
10.5 What to do next? .............

149
150
150
151
151

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 152
Sources of Figures ............................ 155

Acknowledgements

XIII

Acknowledgements
Phil Andrews (Pittsburgh Supercomputer Center), Mike
Bundred (UNlRAS Ltd.), Tat-Seng Chua (National University of Singapore), D. Scott Dyer (Ohio Supercomputer
Center), Basem EI-Haddadeh (University of Leeds), Todd
Elvins (San Diego Supercomputer Center), Mark Goossens
(Silicon Graphics Ltd.), Chris Green (British Geological
Survey), Simon Hansford (Precision Visuals), Dee Holmes
(Stardent Computer Ltd.), Peter Irwin (Dynamic Graphics
Ltd.), Teruaki Ito (Ricoh Company Ltd.), Tosiyasu L. Kunii
(University of Tokyo), Hideko S. Kunii (Ricoh Company
Ltd.), Chris Little (Meteorological Office), Donna McMillan (Sun Microsystems Inc.), Nadia Magnenat-Thalmann
(University of Geneva), Eihachiro Nakamae (Hiroshima
University), Gordon Oliver (LightWork Design Ltd.), Aidan O'Neill (Ricoh Company Ltd.), Peter Quarendon (IBM
UK Scientific Centre), John Rasure (University of New
Mexico), David F. Rogers (US Naval Academy), Peter Stothart (Wavefront Technologies Ltd.), Yasuhito Suenaga
(NTT Human Interface Laboratories), Daniel Thalmann
(Swiss Federal Institute of Technology), Hiroshi Toriya (Ricoh Company Ltd.), Craig Upson (Silicon Graphics Inc.),
Joel Welling (Pittsburgh Supercomputer Center), Jane
Wheelwright (Dynamic Graphics Ltd.), Michael Wood
(University of Aberdeen), Brian Wyvill (University of Calgary), Geoff Wyvill (University of Otago).
The contributions of members of the the AGOCG
Workshop on Scientific Visualization held in the UK,
22-25 February 1991, are gratefully acknowledged. We particularly appreciated the comments of the following on a
first draft of this guide: Ken Brodlie, Lesley Carpenter, Kate
Crennell, Todd Elvins, Hilary Hearnshaw, Roger Hubbold,
Chris Little, Anne Mumford, Howard Watkins, and Mike
Wood. However, responsibility for the final text remains
with the authors.

Many people have
supplied information on their uses
and applications
of visualization
systems. Many designers and implementors have supplied details of
their systems and
also illustrations.
Others have supplied details of aspects of visualization, as well as slides. We express
our thanks and
appreCiation to:

Disclaimer I Copyright Material

XIV

Some companies were unable to supply information or
illustrations of their products, despite being invited to do so.
They have therefore been omitted from the information on
current vendor systems. The list of vendor systems is not
therefore claimed to cover all the systems in the market
place at the time of writing. Those that are covered are the
ones where information was obtainable.

Disclaimer
We are indebted to
the following for
the use of copyright material and
illustrations:

The views expressed by the contributors of information on
products is believed to be accurate and given in good faith.
However, authors and publisher do not hold themselves responsible for the views expressed in this volume in connection with vendor products or public domain products. In
addition, the authors and publisher do not hold themselves
responsible for the accuracy or otherwise of data extracted
from vendor specifications.

Copyright Material
Chris Little, UK Meterological Office (UK Government),
Peter Quarendon, IBM UK Scientific Centre, David F. Rogers, US Naval Academy, Regional Geophysics Research
Group, British Geological Survey, Peter Stothart, Wavefront
Technologies Ltd., Mike Bundred, UNIRAS Ltd., Precision
Visuals Ltd., Stardent Computer Ltd., Silicon Graphics Inc.,
Spyglass Inc., Ricoh Company Ltd., San Diego Supercomputer Center, Donna McMillan, Sun Microsystems Inc.,
Gordon Oliver, LightWork Design Ltd., Nadia MagnenatThalmann, University of Geneva, Daniel Thalmann, Swiss
Federal Institute of Technology.

Trademarks

Trademarks
UNIX is a trademark of AT & T Inc., OPEN LOOK is a This is an aggretrademark of AT & T Inc., X Window is a trademark of gated list of regisMassachusetts Institute of Technology, Xll is a trademark tered trademarks
of Massachusetts Institute of Technology, Motif is a trade- and trademarks
mark of the Open Software Foundation Inc., PostScript is used in the vola registered trademark of Adobe Systems Inc., Ethernet is a ume. In order to
trademark of Xerox Corporation, MS-DOS is a trademark identify products
of Microsoft Corporation, Stardent is a trademark of Star- unambiguously it
dent Computer Inc., AVS is a trademark of Stardent Com- is necessary to
puter Inc., DORE is a trademark of Stardent Computer use these terms.
Inc., Silicon Graphics is registered trademark of Silicon The following are
Graphics Inc., IRIS is a registered trademark of Silicon known to be tradeGraphics Inc., POWER series is a trademark of Silicon marks or regisGraphics Inc., Graphics Library is a trademark of Silicon tered trademarks
Graphics Inc., Image Vision Library is a trademark of Sili- of the companies
con Graphics Inc., IL is a trademark of Silicon Graphics concerned. We
Inc., GL is a trademark of Silicon Graphics Inc., Live Video trust that others
Digitizer is a trademark of Silicon Graphics Inc., StereoView that may not be
is a trademark of Silicon Graphics Inc., Explorer is a trade- noted are known
mark of Silicon Graphics Inc .. , Personal Visualizer is a to readers and are
trademark of Wavefront Technologies Inc., Data Visualizer referenced in a
is a trademark of Wavefront Technologies Inc., Advanced manner acceptVisualizer is a trademark of Wavefront Technologies Inc., able to the comSunVision is a trademark of Sun Microsystems Inc., Sun- panies concerned.
View is a registered trademark of Sun Microsystems, OpenWindows is a trademark of Sun Microsystems Inc., XDR is
a trademark of Sun Microsystems Inc., XGL is a trademark
of Sun Microsystems Inc., SPARC is a registered trademark
of SPARC International Inc., SPARCstation is a trademark
of SPARC International Inc., IBM, PC, PS/2 are trademarks
of IBM Corporation, PV-WAVE is a trademark of Precision
Visuals Inc., NAG is a registered trademark of Numerical
Algorithms Group Ltd. and, Numerical Algorithms Group,
Inc., RenderMan is a registered trademark of PIXAR, RIB
is a trademark of Pixar, Spyglass is a trademark of Spyglass
Inc., MacIntosh is a trademark of Apple Computer Inc.,

XVI

Trademarks
Laserwriter is a trademark of Apple Computer Inc.,
LightWorks is a trademark of LightWork Design Ltd.,
NeXT Cube is a trademark of NeXT Computers, SpaceBall
is a trademark of Spatial Systems Inc., DataGlove is a trademark of VPL, 3D Polhemus Digitizer is a trademark of
Polhemus, EyePhone is a trademark of VPL Research Inc.,
VoxelView is a trademark of Vital Images Inc., VoxelLab is
a trademark of Vital Images Inc.

Part I

Basics of
Scientific
Visualization

Chapter 1
Introduction and Background

1.1 Introduction
This guide seeks to answer the following questions:
What is scientific visualization?
What can it do?
What do the technical terms and the jargon really mean?
What products are currently available?
What kind of hardware do I need?
What are the costs?
What do I get?
Where do I go next to find out more, or to explore current possibilities?
• What are the prospects for the future?

•
•
•
•
•
•
•
•

The first part of this volume is concerned with introducing
the topic, definitions, terminology, techniques, methodology, and equipment. The second part contains an overview
of current systems and developments.

1.2 Background
The area encompassed by scientific visualization is defined
in this guide, with the range of possible applications, and
the potential for the future. Considerable advances have
been made in the USA by dissemination of information and
by coordinated initiatives from industry and professional
organisations such as the Association for Computing Machinery (ACM).

Information

Introduction and Background
Objectives

Facilities

Summary

The objective of this guide is to inform the general reader
about scientific visualization, what it offers, and what it can
do. It should be useful to scientists and engineers who are
not specialists in computing matters, but nonetheless wish
to use effective computer-based tools to further research objectives.
Developments and initiatives in the USA are summarized. These demonstrate the relevance and importance of
scientific visualization.
Software products are outlined and summarised - for
purposes of general information. These are indicative of the
kind of products available in the market place, and that are
supported on a variety of platforms. However, this is not intended to be exhaustive, and in certain application areas a
wide variety of software has been developed.
Current developments in animation are summarized because it is likely to become increasingly important for scientific visualization.
This guide provides an overall summary of the benefits
that accrue from scientific visualization and the methods,
tools and strategies that comprise its domain.

Chapter 2
What Scientific Visualization Can Do!

2.1 What is Scientific Visualization?
"The purpose of computing is insight, not numbers" wrote
the much-cited Richard Hamming in Numerical Methods for
Scientists and Engineers (McGraw-Hill, 1962). Scientific visualization is an amalgam of tools and techniques that seeks
to promote new dimensions of insight into problem-solving
using current technology.
Scientific visualization is concerned with exploring data
and information graphically - as a means of gaining understanding and insight into the data. Scientific visualization is
a graphical process analogous to numerical analysis, and is
often referred to as visual data analysis. Scientific visualization systems are combinations of hardware and software systems and techniques.
By displaying multi-dimensional data in an easily understandable form on a 2D screen, it enables insights into 3D
and higher-dimensional data and data sets that were not formerly possible.
Often data sets are very large, and this gives rise to problems of scale and of finding correlations and relationships
between different parts of the data.
Visualization is also a means of gaining a quick understanding of processes. This could be done in more classical
ways, but might take much longer.
The difference between scientific visualization and presentation graphics is that the latter is primarily concerned
with the communication of information and results that are
already understood. In scientific visualization we are seeking to understand the data.

Insight not
numbers

What is it?

Multidimensional

Large volumes

Speed

Not presentation
graphics

What Scientific Visualization Can Do!
Lots of tools

Help for
applications

Examples

Simulating nature

Interactive steering

Visualization involves aspects in the areas of computer
graphics, user interface, cognitive science, image processing,
design, and signal processing. Formerly these were independent fields, but convergence is being brought about by the
use of analogous techniques in the different areas. Visualization is thus an additional tool for scientific research and investigation.
Visualization highlights applications and application areas because it is concerned with providing the means for a
user to achieve greater exploitation of computing tools now
available. In a number of instances visualization has been
used to analyze and display large volumes of multi-dimensional data in such a way as to allow the user to extract significant features and results quickly and easily. Tools and
techniques in this area are therefore concerned with data
analysis and data display, perhaps with provision for the display of data changes with respect to time.
Non-destructive and non-invasive examination of the internal structures of living organisms (e.g., reconstructions
from brain scan data), turbulence effects in fluid flow, and
genetic engineering are all examples that have caught the
public attention, and where scientific visualization has
brought substantial benefits. However, this is but one aspect
of the whole field, as indicated above.
Visualization fits into the overall process of numerical
simulation as indicated in Figure 2.1 below.
In the computational sciences the main goal is to understand the workings of nature. In order to accomplish this,
the scientist proceeds through a number of steps from observing a natural event or phenomena to analyzing the results of a simulation of the phenomena. Visual representation of this data is often indispensible in gaining an understanding of the processes involved.
Visualization systems can be used for the interactive
steering of computations. The user observes the progress of
the computation visually and alters parameter values accordingly. These in turn determine the future computation.

How to do Scientific Visualization

~I--_r-l--,,- - - Observations

Fig, 2,1
~_~--- Physical Laws

l""".=~=="" '"

Mathematica l
Model
Approx. to
Mathematical

Mathematical
Formu lation
of Laws

Simulating Nature,
Simulating natural
phenomena: the
boxes represent
processes, the
circle's size
indicates the

Simula tion
Specification

relative volume of
information

Simulation
SOlution

passing between
each pair of
Images

processes

2.2 How to do Scientific Visualization
Visualization tools benefit from the availability of modern
workstations with good performance, large amounts of
memory and disk, and with powerful graphics facilities in terms of range of colors available, resolution, and speed
of display by the workstation. This close coupling of graphics and raw computation power is a powerful combination
for those areas where visual insight is an important part of
the problem-solving capability.
Such workstations now offer substantial computation
power coupled with high-speed 3D graphics. These facilities
can be exploited to significant advantage in application areas
such as modeling, simulation, and animation. Real-time dynamical simulation can involve the processing and display
of large amounts of data, and often the only effective analysis of the performance or validity of the model is through
visual observation.
Leading-edge applications will tend to require the most
powerful systems available. The vendors listed later in this

Computing power

3 dimensions

What Scientific Visualization Can Do!

Supercomputers

Output

volume provide a range of systems to match a wide variety
of applications, and are continuously improving computation power and graphics capabilities.
Such workstations provide the computation power to
process the data, and the high-speed graphics pipeline can
transform this into graphical images, often in real time. In
those cases where additional computational resources are required, the calculation can be off-loaded onto a supercomputer, or other advanced workstations with spare capacity,
and the resulting image down-loaded for viewing (and perhaps even interaction) when it is ready.
Output is often most useful as information on the workstation screen, especially when the process is interactive.
More permanent copies of screen images can be printed in
full color on A4 to AO plotters and printers and on video
or slides. However, this requires additional hardware and
software. Generally, the higher the cost the greater the variety of colors and quality of the final images.

2.3 Some Examples of Scientific
Visualization
2.3.1 The March of Napoleon's Army

,-,

----_.'.-"
,
Fig. 2.2

March of Napo/eon's Army

Some Examples of Scientific Visualization

The classic map/chart of Napoleon's march in Russia, and
the retreat of 1812, drawn by Charles Joseph Minard.
This is a good example of visualization which clearly
pre-dates scientific visualization!

2.3.2 Cholera Outbreak
The work of Dr. John Snow (Gilbert 1958) provides an early example of cartographic visualization in problem analysis. While investigating the 1853-54 cholera outbreak in
London he identified what he called a "cholera-field" in the
Soho area. He had plotted the homes of the 500 victims
who had died in the first 10 days of September 1854 and this
simple visualization (of quite a large and complex data set)
drew his attention to the previously unsuspected link between water supply and the disease. All victims had drunk
from the Broad Street pump, in the middle of the "field",

Fig. 2.3
Dr. John Snow's
map (1855) of
deaths from
cholera in the
Broad Street area
of London in
September 1854

What Scientific Visualization Can Do!

which, it was later established, was being polluted by a leaking cesspool. The "link" was confirmed by noting that a virtually disease-free area (a large workhouse) within this zone
had its own clean water supply.

2.3.3 Weather Maps from Meteorology

Fig. 2.4
Weather Map 1

The following three maps illustrate various aspects of weather patterns.
This first map (Figure 2.4) shows the raw numerical forecast data with contour lines showing the pressure. This is a
distillation of the information contained in thousands of
numbers! The second map (Figure 2.5) shows where the
fronts have been positioned. This is an interpretation of the
above map, and presents the data in a form that people can
understand more easily. The third map (Figure 2.6) shows
a tailored short-hand form of the second map, and is the
kind used by aircraft pilots.

What Scientific Visualization Can Do!

2.3.4 Molecular Modeling
This example shows a molecular model of liver alcohol dehydrogenase - calculated by computer and then displayed.
The mauve molecule fitting into the enzyme shows the
structure of the underlying molecule.

Fig. 2.7
Molecular Model

2.3.5 Pelvic Reconstruction
Karen Frankel
(1989) reported the

following case

A young man in his late twenties suffered a crushed pelvis
in an auto accident. His orthopedists said that the fracture
was too complicated to operate on and elected to treat him
conservatively; he would be in traction for a few months.
The doctors were certain that the young man would be permanently crippled.
Luckily the man's father, also a physician, knew of research in 3D rendering of computed tomography (CT) scan
data. He sent his son's CT scan studies to the researchers,
a radiologist, and orthopaedic surgeon, and a computer
graphics expert, who studied the volumetric rendering of
the pelvis that was created with specially designed hardware

Some Examples of Scientific Visualization

and software. Able to see it from all angles, they determined
the extent of the fracture and locations of several key fragments. The pelvis was operable and the next day the surgeons set the fragments. Three months later the patient returned for a check up and demonstrated full-range hip monon.

Fig. 2.8
Fractured Pelvis.
Volume renderings
of the broken
pelvis using CT
scan data by
Professor Elliot
K. Fishman of
Johns Hopkins
University Hospital, Dept of
Radiology. The
extent of the fracture and location
of the fragments
are clearly visible.
Although
radiologists have
been using CT
data for almost
20 years, volume
renderings of CT
offer

a new way of

interpreting such
data

What Scientific Visualization Can Do!

This case coupled great medicine and great computer science. The technique of volume rendering changed the
course of treatment by providing the physicians with more
data. This data ultimately gave them the confidence to operate and thereby improve the patient's quality of life. While
volume rendering helped manage the medical complexities,
this case also represents departures from tradition for both
disciplines.
Text: courtesy of ACM

2.3.6 Oil Exploration
This example shows the use of visualization in oil exploration. The volumetric data was produced as part of a simulation of a method for recovering oil from the tar sands of
northern Alberta, Canada.
This process was simulated by the Alberta Oil Sands
Technology Research Authority and the visualization was
computed by Geoff Wyvill and Brian Wyvill.
Courtesy of Geoff Wyvill and Brian Wyvill.

Some Examples of Scientific Visualization

Fig. 2.9
Oil Exploration 1.
Iso-temperature
contour surfaces
from Volumetric
Data produced by
simulation. The
surfaces are tiled
using the "Soft
Object" algorithm.
The red pipe
represents the
"injection well"
which pumps
superheated
steam into the
rocks

Fig. 2.10
Oil Exploration 2.
As the simulation
progresses the
surface changes
shape as the
rocks are heated.
After a period of
time, oil is
precipitated into
the production
pipe

What Scientific Visualization Can Do!

2.3.7 Designing Ship Propellors
Sculptured Surface Fitting and Fairing
Shape design

Visualizing the
surface

The fairness of a sculptured surface is important for design
applications as diverse as modern artistic sculptures and aircraft or automobiles. The requirement for fairness can be
based on either aesthetic or technical considerations. Currently there is no universally accepted mathematical definition of fairness. One technique that aids in evaluating surface fairness is to look at the Gaussian curvature of the surface. The Gaussian curvature of a surface is visualized by using the values of Gaussian curvature to color encode the surface. If the surface is fair, then the color hue smoothly transitions across the surface. U nfairnesses show up as splotches
or lines of color within the surface.
The three accompanying images show one side of the
surface and the fillet for a single blade of a ship propeller.
The data supplied by the designer is shown in Fig. 2.11 visualized with a dynamic three-dimensional rational B-spline
surface design program called Rbssd developed by Professor
David F. Rogers.
The data set is a combination of two independent data
sets comprising the surface and the fillet. Visualized in this
way it is obvious that the lines of data for the fillet and the
surface are not aligned. This has implications when a rational B-spline surface is fit to the data.
Figure 2.12 shows the rational B-spline surface generated
by the defining polygon net. The surface appears reasonably
smooth.
Figure 2.13 illustrates the color coded Gaussian curvature
surface. This shows that the surface is unfair. Here the green
indicates zero Gaussian curvature and the yellow positive
Gaussian curvature. The areas of yellow indicate ridges or
bumps or hollows in the otherwise developable surface. Notice that many of these are concentrated in the area where
the fillet and the main surface data were joined. Visualization makes it obvious.

Contributed by Professor David E Rogers,
US. Naval Academy.

Some Examples of Scientific Visualization

Fig. 2.11
Designer's Data
Visualized with
Rbssd

Fig. 2.12
Rational a-spline
Surface

Fig. 2.13
Color Coded
Gaussian
Curvature Surface

What Scientific Visualization Can Do!

Supporting References
Rogers, D. F. and Adlum, L.: Dynamic Rational B-spline
Surfaces. Computer Aided Design Journal, invited paper
in the commemorative issue honoring Pierre Bezier on
his 80th birthday, Computer Aided Design Journal,
Vo1.22, pp.609-616, 1990.
Dill, ].c., and Rogers, D.F.: Color Graphics and Ship Hull
Surface Curvature. Proceedings of International Conference on Computer Application in the Automation of
Shipyard Operation and Ship Design IV (ICCAS '82),
7-10 June 1982, Annapolis, Maryland, pp.197-205,
North-Holland.

2.3.8 Visualization of Forest Growth
Growth of forest

Choice of forest to
model

Physical properties

An interactive tree model, FOREST, has been developed at
the University of Tokyo to enable the processes of forest formation to be visualized. The model includes a parallel algorithm of individual tree growth which considers both the
differences between the species and also the time-dependent
interactions among the trees through mutual shading.
A tropical rain forest in the equatorial zone is chosen as
a typical case because other types of forest can be derived
from it by imposing a set of constraints, such as diminished
rain fall and lower tempratures, which slow down the speed
of growth processes. The results of the visualization of algorithmically animating a few hundred years of forest growth
processes using this model have been validated against the
data obtained in experimental observations in Pasoh on the
Malaysian peninsula.
The model of individual trees considers the internal
properties of trees such as the rate of the light/photosynthesis relationship, the death rate of the branches, and the proportion of the foliage active in photosynthesis.
Figures 2.14 and 2.15 show two frames from a sequence
which show the entire life history of the forest. The first
shows the initial growth period (1-60 years) and the second
shows the higher layer formation processes starting at age
60.

Fig. 2.14
Forest Visualization.
Age 60 years

Fig. 2.15
Forest Visualization.
Age 250 years

Understanding a typical rain forest and its ecosystem is
thus expected to lead to understanding of other types of forests and their ecosystems.
Such models and their visualization can be used to increase our understanding of the growth processes in nature
and the way these processes can be affected by apparently
minor disturbances in the environment. For example, forestation affects the percentage of carbon dioxide in the atmosphere, which in turn affects global warming, which in turn
affects the total area of deserts. Understanding these effects
and the relationship between the variables is the key to understanding how to influence the future of the planet. Visualization can playa significant role in furthering this understanding.

In/ormation supplied by Professor T. L. Kunii,
University 0/ Tokyo.

Growth
processes
in nature

Global modeling

Chapter 3
Explanation of Scientific Visualization
Terminology

3.1 Techniques
The schema in Fig. 3.1 outlines some of the current representation techniques.

3D
•S!

11<:
....

Fig. 3.1

The visualization

Mapping Space:
the mapping from
the computational
domain into the
visualization
domain

"5
.~

Solids Modellog

,....... ~.................................................... !............." ......................VolulII~ ReDderiDi

.....1..................

ioldo

HoI ....
Pseudo Color

Images

.• Llnel
CUr:flS

Contour Maps

........................ ContourM..,.
Field Vectors
Space C,unes

C~11es

......Stacked
Tlied Surr....
Textures
Rlbboas

3DVKtor Nets
.. .. .. .. .. .. .. .. .. .. "Field Vectors

lcoa,

Attribute

Maprog

................ AUribute
M.,pplog

.................... .............. .
~

HedgeHop
Rlbbo ...

.S!

.5
~

............................ Scatter Plots

Partldes
DotSurfKeI

Scatter Plots
Particle Tracers
Dot Sur:faces

2D
3D
Dimensionality of the Computational Domain

The ongm of several visualization techniques can be
traced back to line-based two-dimensional contour maps.
These extensions result in higher-dimensional or more continuous representations as shown in Fig. 3.2.
This chapter outlines techniques that can be used by a
scientist confronted with a vast amount of data generated by
computer models, remote sensing devices, and automated recording equipment.

3D Cell
Rendering

'"=

)

--g

3DVoxel
Rendering

~ 2.50 Continuous Tone

2D Continuous Tone

.!l

7"'' ' ' '

Contour Maps

3D Polygonal
Surface Tilings

3D Vector Nets

~
~
Q~------r------------------------r------------2D Contour Maps

Dimensionality of the Domain

Fig. 3.2
Three-dimensional
extensions of
contouring

On the following pages are some examples of the methods that are chosen to represent quantities in the visualization domain.

Visual Picture
This picture shows a drug molecule reacting with a large enzyme and illustrates the underlying protein structure.

Representation
Method

Fig 3.3
Dot Surface.
Quick method for
small objects.
Time-consuming
and counterproductive for
larger objects

Explanation of Scientific Visualization Terminology

3D wire frame
representing the
shape of a
surface.
Fig. 3.4
Vector Net

Fig. 3.5
Polygonal Surface
Surface
represented by
polygons

Perspective view of a calculated gravity field represented
as a 3D smooth shaded polygonal surface, with lighting and
shading effects generated from an implementation of the
PHIGS PLUS model.

Techniques

This picture shows part of the North Atlantic where the
gulf stream is flowing. The sections show temperature at
three different depths with plumes of water demonstrating
mixing taking place in the vertical plane.

Fig. 3.6

Stacked Contour
Map.
Overlaying of 20
cross·sections to
represent 3D
volumes

Fig. 3.7

Ribbons and
Streamers

This shows airflow round a wing. Color shows one parameter; x, y shows direction; and twisting shows vorticity.

Explanation of Scientific Visualization Terminology

Fig. 3.8
Hedgehogs

A method for showing a direction relative to the surface
(hence the term hedgehog spine!). We can of course show a
third variable (by the color of the vector).

Fig. 3.9
Shaded Contours

This picture shows the use of color to identify areas in
the plot between upper and lower threshold values specified
by the boundaries of the areas.

Volume Visualization

Fig. 3.10
Attribute Mapping

Overlaying an additional data set on an existing 3D one.
Height represents axial velocity; shade represents radial velocity. As a turbine pushes fluid through an opening, scientists can observe the density of a particular slice - shown
as shading, with red being the most dense.

Representation
Method

3.2 Volume Visualization
Volume rendering is used to view 3D data without the usual
intermediate step of deriving a geometric representation
which is then rendered. The volume representation uses voxels (volume elements) to determine visual properties, such
as opacity, color, and shading at each point in the computational domain. Several images are created by slicing the volume perpendicular to the viewing axis at a regular interval
and compositing together the contributing images from
back to front, thus summing voxel opacities and colors at
each pixel. By rapidly changing the color and opacity transfer functions, various structures are interactively revealed in
the spatial domain.

Voxe/s not
geometry

Explanation of Scientific Visualization Terminology
Applications

Fig. 3.11
Rendered
Isosurfaces from
Slice Contours

"nerve cell"

Fig. 3.12
Cell Rendered
Volumetric Image

A number of projects in the USA have demonstrated the
benefits to medical and surgical planning from these new
techniques. Further information may be found in Frenkel
(1989), Kaufman (1990), and Upson (1991).

Data Types

Figure 3.11 shows paired helical filaments (orange) PHF
cracking a cell nucleus (blue). This is used in the study of
Altzheimer's disease. The digitized slices are hand contoured; MOVIE· BYU mosaic connects the contours. The
"Marching Cubes" algorithm is used for PHF.

Fig. 3.13
Volume Rendering
of CT Data
"Dolphin Head"
91 slices

3.3 Data Types
3.3.1 Overview of Facilities
As scientific visualization is often concerned with large
amounts of data, it is inevitable that methods for organizing
it, transferring it, manipulating it, and storing it are of great
importance. There are also a wide variety of data formats
and utilities for translating between them.
Application data is concerned with information at the
application level. Data formats are often developed in association with particular application areas. Examples of such
formats are Hierarchical Data Format (HDF) developed by
the National Center for Supercomputing Applications at
the University of Illinois, and Network Common Data
Form (netCDF) developed at NSSDC and NASA. Further

Handling data

Application
aspects

Explanation of Scientific Visualization Terminology

Graphics formats

Image data

Networking
implications

Remote sensing

Multi-media

information on these formats is contained in Sections 3.3.2
and 3.3.3.
Graphics data is comprised of information output by the
graphics system (e.g., vectors, polygons) which is then converted into appropriate image information. Examples of
graphics formats are Computer Graphics Metafile (CGM)
and Postscript.
Image data is the information corresponding to the image on the graphics display screen. For a display with 1000
by 1000 points on the screen we would need 1 million bits
to store the information, just for a simple black/white display. With a wide range of colors this amount of information increases, since we need to store a value representing the
color for each pixel. There are a number of formats for representing image data, including GPF (Graphical Pixmap
Format), TIFF (Tagged Image File Format), Group 3 and
Group 4 Fax, VIFF (Visualization Image File Format for
the Khoros visualization software), PICT (MacIntosh format), and PCX (IBM format).
In order to be able to transfer image data effectively (particularly over networks) it is important to reduce the size
of the image data set to manageable proportions without
losing essential information. The Joint Photographic Experts Group GPEG) and the Moving Picture Experts Group
(MPEG) are formulating proposals for standards for single
images and multiple frames respectively. JPEG and MPEG
are proposals for standards in this area. In addition, fractal
compression techniques are being used with considerable
success.
Remote sensed image data contains real information
which may be extracted by image processing techniques.
This is a well-establised field.
Multi-media systems combine software with facilities for
sound, images, graphics, video, and animation to create powerful communication tools. Interest in the area is due to the
ability to incorporate data from many sources, and the benefits arising from this. Multi-media products are widely avail-

Data Types

able on personal computers and are moving into the UNIX
workstation environment.
In view of the wide variety of data formats currently in
use, potential purchasers of scientific visualization systems
are advised to check that a given system will handle the data
formats required, and also have sufficient capability to handle the volumes of data required.

Choice of system

3.3.2 HDF
The Hierarchical Data Format (HDF) was developed by the
National Centre for Supercomputing Application (NCSA)
and is available via anonymous ftp.
Hierarchical Data Format (HDF) is a multi-object file
format for the transfer of graphical and floating-point data
between different hardware platforms. FORTRAN and C
calling interfaces for storing and retrieving 8-bit and 24-bit
raster images, palettes (color tables), scientific data and accompanying annotations have been developed. HDF allows
for the self-definition of data content and aims to be extensible, thereby allowing for the inclusion of future enhancements or compatibility with other standard formats.
HDF provides a general purpose file structure that encompasses the following:
- makes it possible for the programs to obtain information
about the data directly from the file, rather than from another source (e.g. look-up table),
- enables the storage of arbitrary mixtures of data and related information in different files, even when the files are
processed by the same application program,
standardizes the formats and descriptions of many types
of commonly used datasets, such as raster images and scientific data,
encourages the use of a common data format by all machines and programs that produce files containing a specific dataset,
can be adapted to accommodate virtually any kind of data by defining new tags or a new combination of tags.

File format

Extensible

Facilities

Explanation of Scientific Visualization Terminology

HDF currently supports sharing data across machines and
systems such as CRAY (UNICOS), Silicon Graphics
(UNIX), Alliant (CONCENTRIX), Sun (UNIX), VAX
(UNIX), Macintosh (MacOS), and IBM PC (MS-DOS).

3.3.3 NetCDF
National
cooperation

Storage and
retrieval of data

Architecture

Network
transparency

Availability

The Network Common Data Form (netCDF) was developed as part of U nidata - a U. S. national effort sponsered
by the Division of Atmospheric Sciences of NSF. The initiative is managed by the University Corporation for Atmospheric Research. The software is available via anonymous
ftp.
NetCDF is a data abstraction for the storing and retrieval
of scientific data, in particular multi-dimensional data.
NetCDF is a distributed, machine-independant software library based upon this data abstraction which allows the creation, access and sharing of data in a form that is self-describing and network-transparent. Both C and FORTRAN
interfaces are supported.
NetCDF software utilizes the concept of an abstract data
type, which means that all operations to access and manipulate data in a netCDF file must be via a defined set of functions provided by the C library interface. As the actual representation of the data is hidden from the application, internal data representations can be changed without affecting
the program.
To achieve network transparency, netCDF is implemented on top of a layer of software for external data representation known as XDR. XDR is a nonproprietary standard for
describing and encoding data developed by Sun Microsystems, Inc.
The netCDF software provides common C and FORTRAN interfaces for applications and data. The C interface
library is available for many common computing platforms,
including UNIX, VMS, MSDOS, and MacOS environments. The FORTRAN interface is available on a smaller
set of environments (due to the lack of a standard for calling
C from FORTRAN).

Current Application Areas

XDR has been implemented on a variety of platforms,
including SUNs, VAXs, Apple Macintoshes, IBM-PCs, IBM
mainframes, and CRAYs.

3.3.4 Databases
The currently accepted storage method for most scientific
data is the Relational Database Management System. Many
commercial examples are available (e.g., Oracle, Ingres). Data can be extracted using Standard Query Language (SQL)
based commands. Some scientific visualization systems have
these command interfaces built in (e.g., UNIRAS).

Storage of data

3.4 Current Application Areas
This section provides an overview of application areas
where visualization techniques are being used on input data
from the real world, processed data, and computer-generated
data.

Input data

3.4.1 Cartography
Cartography is rapidly moving from a discipline concerned
with the presentation of data (the map) to one concerned
with the storage and analysis of spatial data via Geographic
Information Systems (GIS). These systems may be used to
to store large amounts of information in a database and allow retrieval and display based on user-specified criteria.
The visualization is used to select spatial features based on
their attributes, or to observe topological relationships with
other features. For example, the user may wish to make requests such as the the following:

Analysis of

• Show me all the regions where forests are adjacent to
lakes and which have access by road.
• Show me all the principal roads which have houses within 50 meters.

Interrogating

spatial data

the data

Explanation of Scientific Visualization Terminology

• Display all the houses which have not had their gas service supplies and electricity supplies renewed in the last
30 years.
Data relationships

Project planning

Obviously, there is more information stored in the database
than just the terrain. In particular, it illustrates how useful
it can be to have other information to do with the same territory available so that it can be interrogated and overlaid on
the terrain map.
The visualization can then be used to plan for work to
be done (e.g., by the service industries) in such a way as to
minimize costs. Equipment can be moved to an appropriate
point in the area and used to supply all the requirements for
the work to be done.

3.4.2 Statistics
Visual representation of statistical data is very useful for
providing insight and understanding into the data.

3.4.3 Remote Sensing

Increasing
dimensionality

Satellite and other imaging devices are producing large
amounts of data. Many two-dimensional processing methods
exist for analyzing this data. New methods are being developed to allow for increasing the dimensionality of the data
as the number of frequency bands increases. Visualization
methods allow horizontal (two dimensions of space at a given frequency value) and vertical (one dimension of space and
one of frequency) sections through the data. An example of
the kind of picture produced is shown in Fig. 7.7.

3.4.4 Archeological Reconstruction
Rebuilding history

Data from archeological excavations has been entered into
visualization systems to enable partial or major reconstruction to be done, and the resulting constructs to be viewed
interactively on a computer display screen. This enables the
archeologist to build up a picture of the original buildings,
objects, and their relationships.

Current Application Areas

3.4.5 Molecular Modeling
Chemists and biologists have been using physical models of
molecules for many years to enable the relationships between the various components to be understood. This is
now more easily and and effectively done by using a computer-based model and interacting with it on a graphics display screen. An example of such a molecule is shown in
Fig. 2.7. Such molecules can be rotated and viewed from various angles, as well as providing relevant quantitative information (e.g., potential energies, bond distances, etc.). Analyzing X-ray diffraction maps is greatly facilitated by visualization methods. The atomic positions in protein structures
can be adjusted by interaction until they best fit a given electron density map. In the design of new drugs, existing molecules can be modified by introducing new molecules into
the overall structure. Such systems often provide a stereo
view capability, where by means of special glasses the user
is able to view the molecule on the screen in full 3D. This
can provide further understanding and insight into the overall structure.

Physical models

Computer models

Studying the
molecule
Refining the model

Design of
new drugs

3.4.6 Medical Science
Historically, radiologists have looked at a series of two-dimensional cross-sections and built up mental pictures of
three dimensional structures. However, these are subjective
and can vary from one radiologist to another. In many cases
more detailed and accurate three-dimensional information
can be very useful when planning surgical procedures for
complex and/or intricate structures, or in radiation treatment planning. Volume visualization techniques are increasingly being used to provide three-dimensional information
from a series of two-dimensional slices. A case of pelvic reconstruction is outlined in Section 2.3.5, and volume methods are shown in Fig. 3.11-3.13.

From 2D to 3D

Accuracy in
planning
procedures
Visualizing
volumes

Explanation of Scientific Visualization Terminology

3.4.7 Oceanography
Large and
complex natural
systems
Simulating
ocean behavior

Protection of the
earth's resources

Modeling the behavior of oceans is increasingly being done
using visualization techniques. Often there are a large number of variables involved, such. as temperature, salinity,
depth, vorticity, etc. Representations are chosen to enable
this multi-dimensional data to be viewed on a display
screen. Internal structures can be shown and simulations
performed. This leads to an increase in overall understanding of ocean behavior. Areas such as these are becoming increasingly important as attention shifts from interplanetary
investigations to earth-environment matters such as global
warming and the ozone layer. Examples in the area of oceanography are shown in Figs. 3.6 and 7.21.

3.4.8 Computational Fluid Dynamics (CFD)
Air flow over
aircraft wings

Fluid flow

Visualization is being used to analyse complex flow systems.
Numerical simulations produce values of velocity, temperature, pressure, vorticity, and even tensor fields. A variety of
techniques exist for displaying this data. Particle tracers are
used in real-time to show aspects of the flow. Interactive
CFD is important for tracking and steering solutions. Figures 3.7 and 7.21 show examples.
For further information on application areas, readers are
referred to the more detailed reference work Scientific Visualization - Techniques and Applications edited by K. W
Brodlie, L.A. Carpenter, R.A. Earnshaw, J.R. Gallop, R.J.
Hubbold, A. M. Mumford, C. D. Osland, P. Quarendon,
also published by Springer-Verlag, 1992. This volume also
has more detailed reference information on data formats.

Chapter 4
Facilities for Scientific Visualization

4.1 Visualization Software Categories
Visualization software has evolved over a period of time and
three distinct categories can be identified which appeared in
succession. In general the older the category the less the
power, memory and storage required to run them. This
makes the software in the first category suitable for use in
PC or terminal-mainframe environments, and the most recent developments only suitable for the most modern
supercomputers or supercomputer workstations. At the
same time, because the first and second categories have been
around longer, more applications have been developed using
them and many products in the market fall into these classes. It is likely in the future that tools using the most modern
techniques will appear, but at the present time these techniques are very much in the experimental and developmental stage and will need some time to mature.

Evolution of
facilities

4.1.1 Graphics Libraries and Presentation
Packages
This is the traditional method for creating ways to view and
analyse data.
The libraries interface directly to graphics hardware or
provide graphics functionality in software. The user has to
supply nearly all the pieces of the application: the main program, the user interface, data handling and geometry mapping. The most basic libraries only supply an interface to
the graphics devices (e.g., PLOT10 for Tektronix terminals,
HCBS for Calcomp plotters) and some higher-level libraries

Libraries of
routines

Facilities for Scientific Visualization

Pros and cons

PC Graphics

handle more sophisticated graphic entities such as axes,
curve drawing, and so on. Typical examples in common use
are the UNIRAS subroutine libraries, DISSPLA, GL (from
. Silicon Graphics), GKS(2D), PHIGS, DORE (Stardent),
and NAG Graphical Supplement.
The advantage of this type of software is its flexibility
and direct control but it suffers from the disadvantage of the
large amount of time necessary to write and support code.
Many PC-based packages such as Harvard Graphics,
Slidewrite and CricketGraph have taken on board user-interface functions to provide friendlier software, but still require a great deal of user effort to achieve good results.

4.1.2 Turnkey Visualization Applications

Dedicated

applications

No programming
needed!

These offer a fixed functionality to solve a limited range of
specific problems. The user supplies the data and the computational instructions to the main program and possibly
some geometric mapping. The application supplies the
main program and rendering and usually has an attractive
user interface.
Many products in this category are extremely application-specific and examples in oil exploration, molecular
modeling, and architectural modeling are common but of
limited use in other fields. They are also very often only
available on a very few hardware platforms in common use
in these industries - for commercial reasons. More general
examples are the UNIRAS interactives, PV-Wave (PVI), Data
Visualiser (Wavefront), SunVision (SUN) and VoxelView
(Vital Images).
The user does not have to program these packages and
can obtain results very quickly. Their disadvantage is that
they have limited extensibility and therefore may often only
provide a part of the solution a user requires. They have all
reached a high level of maturity and many users applying visualization to their work will probably be using one of these packages.

Visualization Software Categories

4.1.3 Application Builders
These offer a series of modules linked by interfaces which
are connected interactively at runtime. It combines features
from both of the other two categories by providing turnkey
solutions for individual parts of the program and the flexibility to customise the final solution adopted. The supplied
modules can be replaced by user-written modules as required, providing they conform to the data input/output interface requirements, therefore giving greater extensibility.
In these systems virtually everything the user needs is
provided by the program. The user has only to direct the execution path of the program, provide the data, and optionally, their own computational modules if required.
Applications are constructed by a mouse-driven interface, manipulating icons on screens and linking them with
data paths. Once the required modules have been connected
and built the program can be executed. New applications
can be prototyped very quickly by connecting modules in
different ways but the user needs to know how to manage
the flow of data through the network, and how to extend
the module set.
Examples are AVS3 (Stardent), Explorer (Silicon Graphics Inc.), apE (Ohio Supercomputer Centre) and Khoros
(University of New Mexico). More advanced application
builders are currently under development and new advances
in visualization techniques and software will extend and improve the application builder functionality. At the present
time these products are not mature enough to have had substantial numbers of packages built around them, but these
products should be appearing in the future.

Select the options
you need

Constructing
applications

Some examples

4.1.4 Choosing a Package
Current work in scientific visualization tends to be done
with turnkey application tools - because of their functionality and their ability to process large data sets. However,
this does not mean that good visualization work cannot be

What should I
use?

Facilities for Scientific Visualization

PC graphics
survey

Data transfer
limitations

Access via X

done with software libraries and PC packages, but merely
reflects the good understanding of these systems that exists
in the graphics community. Application builders are still
primarily used as research tools, but it is anticipated that
more applications will utilize them.
PCs are more widely available in the academic/Research
Laboratory sphere than any other type of hardware and a
thorough review of currently available graphics software has
recently been completed by the UK Inter-University Software Committee Working Party and published as a Report.
It is probable that restrictions on internal data transfer
bandwidth and graphics display facilities would limit the
usefulness of the current PC products as hardware platforms
for the majority of applications.
For those without workstations, PCs can be utilized as
X-servers for such software running on superworkstations,
supercomputers, or mainframes, and connected via ethernet
or X25 OANET). Users of PCs therefore can have access to
visualization systems remotely and view the results at their
desks, albeit at lower graphics resolution and with a time
penalty introduced by current network performance.

4.2 Software Costs
4.2.1 Subroutine Libraries and Presentation
Packages
You get what
you pay for

Costs of subroutine libaries and packages are related to their
functionality and degree of sophistication. Those which only provide interfaces to graphics devices (HCBS, PLOT10,
HPGL, Xll) are bundled with the hardware. Packages with
a wide range of higher-level functionality, such as the UNIRAS subroutine libraries, can be fairly expensive. Packages
for PC-based operations reflect the prices that PC software
can command, and are typically in the range 100-500
pounds sterling. Much of this software is covered by educational deals and is available to the UK community and Re-

Hardware Considerations (including Hardcopy)

search Councils under bulk discount arrangements (e.g., via
the UK Combined Higher Educational Software Team CHEST).

4.2.2 Turnkey Visualization Systems
The cost of these systems reflects the complexity of the systerns and the comparative affluence of the targeted application areas! For example, volume rendering products for the
petroleum industry are very expensive, but the users can
usually afford to pay. At the other extreme, products bundled with, or designed for, particular hardware platforms are
very reasonable. Educational deals exist for some of these
products.

High-cost areas

4.2.3 Application Builders
At present these systems are either available free (or almost
free) in the public domain (e.g., Khoros, apE) or tend to be
expensive if they are mainly machine-dependent and bundled in with their hardware platforms (e.g., AVS). In the future, more sophisticated systems may follow this scenario or
may become more expensive, if unbundling takes place.

Public domain,
or proprietary

4.3 Hardware Considerations
(including Hardcopy)
Useful visualization work has been performed on very inexpensive equipment, but there is a growing requirement to
perform more complex analysis of increasingly large volumes of data, using more and more sophisticated graphic
display techniques. When this is coupled with a demand for
much faster response it is obvious that hardware costs for
'ideal' systems could escalate. Although much valuable
work has been done on superworkstations attached to a supercomputer (e.g., CRAY) - this is a solution which is not
available to many people.

Functionality
needed

Facilities for Scientific Visualization
Costs of
components

Add-on extras

The amount of money in a hardware budget will determine what kind of work can be performed, or the speed of
the computation. However, we can identify components
which are important. Systems to perform visualization with
subroutine libraries and packages are fairly modest and terminal connection to a mainframe or a PC-based system costing less than 2 K pounds sterling is likely to be adequate.
Turnkey visualization systems usually require a UNIX
workstation, with color screen and hard disk, which will
cost in excess of 6 K pounds. Higher-powered processors,
better quality, faster displays, and substantial hard disk storage will improve throughput but can increase the cost beyond 20 K pounds. Application-builder software will only
work properly on high-performance workstations. The absolute minimum configuration will cost around 15K-20K
pounds, but usable systems are likely to exceed 30 K. Highpowered systems for complex analysis are likely to cost in
excess of 70 K pounds.
Hidden costs should not be overlooked! These could include items such as high-speed networking access, archive/
backup facilities, sophisticated I/O devices, and - more importantly - hardcopy. A3 color Postscript devices suitable
for use with most workstations, PC's, or on a network, are
available for 10-15K pounds through educational deals. If
it is desired to store graphic images in raster form, then large
amounts of disk space are required.
Costs for training and provision of expert advice and
support also need to be included.

4.4 Vendor Systems Versus Public Domain
Systems
Vendor systems

Users are advised to note that vendor systems usually have
software support (e.g., for sorting out software faults and
other difficulties). As part of the licence agreement the user
will usually pay an annual software maintenance fee (typically 5-10% of purchase price). Users also need to enquire

Vendor Systems Versus Public Domain Systems

whether such licence agreements entitle the user to receive
upgrades and new versions of the software as they become
available, if it is desired to continue with the same system
for a number of years. Users should note that usually upgrades are included in the original licence agreement, but
that major developments or major new versions will often
be the subject of separate new licence agreements, for which
users have to pay. This is because the vendor has decided it
is necessary to recoup research and development costs, and
therefore the new version has essentially been made into a
new item of software. If the user requires the additional new
functions in this software, then he or she is faced with a further licence charge.
In this area, as in others, users get what they pay for.
A variety of current vendor systems are detailed in
Chapter 7.
Public domain systems often have very limited support,
or even none at all. Occasionally the authors and originators indicate they will receive reports of bugs or difficulties,
but they cannot necessarily guarantee to provide remedies.
Also, future updates and revisions of the system may be more uncertain. However, informal help is sometime available
via email discussion lists, where users report their problems
and other users with the same software can provide help and
assistance. Khoros has such an email discussion list. This list
also has the advantage that the Khoros Group at the U niversity of New Mexico are also on the list and so can provide
expert technical advice and help as appropriate. However,
such help is done on a voluntary basis; there is no contractual commitment.
Public domain systems often offer state of the art facilities, and can be very useful for research and development
purposes. They are increasingly subjected to procedures
such as those applied to commercial vendor systems, for example, rigorous testing (e.g., Beta testing) before general release to the community, and detailed on-line and hardcopy
documentation. They are also essentially free, though occasionally a small distribution fee is charged. Software can

New versions may
mean further costs

Public domain
systems

Addressing
problems

Quality is
improving

Facilities for Scientific Visualization

usually be obtained directly by anonymous FTP across the
international communications network. For users with no
funds, such software can be very useful indeed.
However, users and potential users should be aware of
the following points before committing themselves to using
public domain software, especially for .large, on-going projects:
Check-list before
deciding to use
the software

Software support is usually fairly limited.
Software support is usually in the hands of a small number of experts (often the designers and originators of the
project).
• Future developments (if any) are in the hands of the current developers and the resources available to them.
• The current developers may decide to abandon the current software and work on something completely different, or their managers may move them to work on different projects.
• The software may be sold by the originators or their site
to a commercial vendor, who will then only release versions of it as for normal vendor software. The original
designers may no longer be involved, and the future directions the product will take become uncertain. An example of this is scenario is apE, which began as a public
system and has been recently acquired by a corporation.
Further information on this point is given in Chapter 8.
•

•

4.5 Summary
Comparing
software
capabilities

In addition to the obvious functional requirements that users need in the software to meet their specific application,
users should also bear in mind the following points when
considering how the system is to be used:
• Software support
• Availability of source code
• Range of hardware platforms on which the software is
available

Summary

• Is a library of graphical subroutines required?
• Can the software be distributed across a number of different platform types (e.g.; computational server and
workstation)?
• Any graphics interfaces that may be required by higher
level software
• X-support
• Data formats supported
• Import formats supported (for reading in information
from other sources)
• Export formats supported (for outputting information
to other systems)
Thus users of scientific visualization systems need to consider very carefully not just the functions of the system, but
also the environment in which is to be run, and the general
requirements associated with input of data and export of results to other systems.

List al/ your
requirements

Chapter 5
Outputting Results

5.1 Hardcopy
Neglected topic

Results

Hardcopy

Color Postscript

Stand-alone
or networked?

This topic is included because it is often neglected in the
evaluation of scientific visualization systems and because requirements for it only surface when the user has started to
use the system and applied it to his or her problem.
Users should think carefully about the ultimate destination of their results. Often this includes written papers and
reports, and also presentations to funding bodies and conferences.
Intermediate hardcopy is often required for the production of draft reports for circulation within research groups
or departments. Thus, although slides and video are usually
the preferred medium for the submission for publication, it
is often very useful to be able to produce color paper hardcopy for draft purposes.
Color Postscript printers and plotters are now widely
available in a variety of paper sizes from A4 to AO. Most visualization software contains drivers for producing color
Postscript, since it is a well accepted industry standard.
For a small research group, a dedicated printer connected
to the workstation is a feasible option. For a larger laboratory, or large groups of workstations and users, it is often
more economical to consider a larger printer connected via
the network. Users then have the capability of generating
output of larger size and often greater range of colors. However, it should be noted that large plotters are often expensive to maintain and run (a typical annual maintenance contract can be 7-10% of purchase price).

Areas such as geophysics and seismic applications applications often use AO-size maps as standard in the industry.
If these are required then a top of the range color electrostatic plotter will be required.

Specialized
applications

5.2 Video
Video is becoming an increasingly important medium for
storage and display of real-time simulations for publication
of research results and presentations at seminars and conferences. It is the only cost-effective medium for the publication of large amounts of color information (1 hour
= 100,000 frames), and which is cheap and easy to copy. It
has a natural interface to the TV technology domain and
provides a portable and easy-to-use medium.
The use of video to demonstrate time-varying and dynamic processes is becoming increasingly important. A sequence of still frames (e.g., on slides) may not convey the
full information relating to the process involved, in a way
that a sequence of moving images can.
The display of real-time simulations offers the user a dynamic analysis tool to supplement other visualization methods.
Video technology is coming within the range of the
workstation user by the increasing availability of low-cost
interface boards and also animation software. pes and
workstations can now be interfaced to a video recorder by
means of a video board, an animation controller, a PAL encoder, and a sync generator for around 7K pounds sterling.
An editing U matic VTR would cost a further 7 K pounds.
The quality of the finished product is proportional to
the cost of the system. If broadcast quality is required, then
the equipment required is currently expensive. However,
much more inexpensive systems (as the above) can produce
reasonable output.
Effective presentation of information via the video mediurn is a non-trivial task. In particular, almost all viewers of

Color and
real-time images

Costs of video

Quality of the
result

Viewers can be
stern critics

Outputting Results

Professional
advice

such information have become accustomed (unconsciously)
to a high level of presentation quality through the programs
presented on national and commercial television channels.
Those new to video need to take advice from those with
substantial experience in this area, e.g., graphic artists, TV
producers, video editors, etc. A close association with those
with experience in this area is likely to produce substantial
benefits in the quality of the work produced.

5.3 Other Media
CD ROM and
laserdisk

Other media include CD ROM and laserdisk. The typical
cost of a laserdisk is 700 pounds. This is capable of storing
large amounts of picture information, but is expensive unless many copies are required.

Chapter 6
Current Developments and Activities

6.1 USA
Initial impetus for scientific visualization was provided in
1987 by a National Science Foundation (NSF) Panel Report
of a Workshop on "'Visualization in Scientific Computing»
(McCormick et al. 1987).
The principal recommendations of the McCormick Report were that national funding should be provided for
short and long term provision of tools and environments to
support scientific visualization, and to make these available
to the scientific and engineering community at large. Such
provision was considered to be essential if the enabling tools
were to be effectively harnessed by current and future scientists and engineers.
Such tools often require access to significant computation resources. A natural focal point for these developments
has been the funding of Supercomputer Centers - to provide both the facilities and access to them by the community.
An example of this at the San Diego Supercomputer visualization Center is the development of network-based
general tools purpose visualization tools. These are accessed
by 2800 users with 350 different applications. Such users access the facility by a variety of different routes including dial-in lines, national networks, and dedicated high-speed
links. In addition to this broad range of provision there are
also more specialized tools for high-end applications (e.g.,
molecular modeling, computational fluid dynamics).
Similar provision has also been made at other Supercomputer Centers at Cornell, Pittsburgh, and the University of
Illinois at Urbana-Champaign.

NSF Report

Recommendations

Supercomputer
facilities

Network-based
visualization tools

Specialized
applications
Other centers

Current Developments and Activities
Workshops

Visualization
laboratories

Visualization
conference

NSF support

Network support

Workshops on scientific visualization have been established by ACM SIGGRAPH and IEEE to address specific aspects such as data facilities (to facilitate ease of use and transfer of information), and volume visualization (to enable representation of real 3 D information and to give inside views).
Representatives from the Department of Defense and the Department of Energy have initiated a Working Group to define a Visualization Reference Model. A conference of CG
International was held at MIT, Media Laboratory, in June
1991 with the theme "Scientific Visualization of Physical Phenomena ~ The proceedings have been published as a book by
Springer-Verlag (see Chapter 11 - References).
A large number of major universities are establishing visualization laboratories, and often such installations receive
supplementary funding for further proposals in specific application areas. Funding is provided by such bodies as NSF,
DARPA, and NASA. State supercomputers and associated
visualization facilities exist in Ohio, North Carolina, Minnesota, Utah, Alaska, and Florida.
To provide a forum for the presentation and discussion
of the latest advances in scientific visualization, the IEEE
Technical Committee on Computer Graphics has established an international visualization conference, which is
held on an annual basis.
In addition, the National Science Foundation is providing funds for the support and promotion of educational initiatives in scientific visualization by means of institutes,
workshops, and summer schools.
Fast networks are required for distributed and remote visualization. Developments in networking infrastructure are
planned to provide faster communication, interconnection,
and the ability to aggregate computing resources at different
locations on to one particular problem. For example, the
CASA test bed project is funded by the NSF to develop a
1 Gbit/sec network link between Los Alamos National Laboratory, the California Institute of Technology, and San Diego Supercomputer Center, to enable all three resources to be
concentrated on one application simultaneously.

A multi-million-dollar grant has recently been awarded
by NSF to California Institute of Technology, Brown U niversity, University of Utah, Cornell University, and the
University of North Carolina at Chapel Hill, to explore the
foundations of computer graphics and visualization.

49
Foundation
aspects

6.2 UK
A number of centers in UK academic institutions are concerned with application areas such as molecular modeling
and computational fluid dynamics (CFD). There are a number of collaborative projects between academia and industry
in the areas of parallel processing and scientific visualization. One example, GRASPARC, a Graphical Environment
for Supporting Parallel Computing, is a joint project between NAG Ltd., the University of Leeds (School of Computer Studies), and Quintek Ltd. The major objective of the
work is to improve the interaction between the computational scientist and the parallel computer through the development of interactive visualization software.
Vis Lab at Sheffield University is engaged in five projects:
extending surface reconstruction to irregularly sampled
fields; rendering vector and tensor fields; building radiotherapy planning tools; reconstructing cerebral blood vessels
from a pair of x-ray projections; and issues surrounding perception.
The IBM UK Scientific Centre in Winchester is primarily concerned with scientific visualization and has a Visualization Group, a European Visualization Group, a Medical
Imaging Group, and a Parallel Programming and Visualization Group. There are a number of collaborative projects
with academia and industry in the areas of parallel processing, user-interface aspects, and medical informatics.
Natural Environment Research Council (NERC) has a
Visualization Advisory Group concerned with evaluating
products for the areas of geological surveys and oceanography. Science and Engineering Research Council Engineer-

Application areas

VisLab

Industry and
academia

Research councils

Current Developments and Activities

Video facility

University of Leeds

Other projects

ing Board has evaluated superworkstations in the areas of
hardware and software. The present AGOCG Scientific Visualization Workshop which initiated this guide and the
Status Report arose out of proposals by the UK Universities
funding body for computing (the Computer Board) and the
Advisory Group on Computer Graphics (AGOCG).
The Rutherford Appleton Laboratory of the SERC,
Central Computing Division, has developed a video facility
for use by the academic and research community in the UK,
and is involved in projects in the areas of oceanography, atmospheric physics, laser design, mechanical engineering,
ecological simulation, and CFD.
The University of Leeds has an interdisciplinary Scientific Visualisation Group and promotes a wide range of software on state of the art hardware platforms to support a variety of applications. (Reference: ACM SIGGRAPH Computer Graphics, June 1992)
There are numerous other projects underway in this
field - the above is only an indication of the range of work
being done.

6.3 Europe
European centers

Volume
visualization
Workshops

IBM has a number of European centers actively involved in
projects involving Scientific Visualization. These include
the European Petroleum Applications Centre (EPAC) in
Stavanger, the Paris Scientific Centre which is involved in
visualization in the medical area, and the European Scientific Centre in Rome which is involved in engineering and
modelling turbulent flow. IBM also has a joint project with
the Centre of Competence in Visualization at the University of Aix-Marseilles.
FhG-AGD in Darmstadt is working on a number of areas, including tools for volume visualization on a variety of
platforms, and handling different kinds of data sets.
Eurographics arranged a Workshop on Scientific Visualization in April 1990. The proceedings will be available
from Springer-Verlag. Further workshops are planned.

Part II

Overview of
Current Systems
and Developments

Chapter 7
Current Vendor Systems in Use

Readers are recommended to read Chapter 4 for a classification and categorisation of scientific visualization systems
before reading this chapter. Chapter 4 sets out the overall
framework into which the products outlined in Chapters 7
and 8 fit.
Contact addresses are provided for each product at the
end of each section.

7.1 Wavefront Technologies, Inc.
Wavefront was founded in 1984 in California and provides
graphics products for use on a wide range of UNIX workstations including Silicon Graphics, IBM, Hewlett Packard,
DEC and SUN. The company has a well established worldwide sales and support network with its own offices in all
the key European countries, including the UK.
Wavefront's Visualizer software is designed to help engineers, scientists, designers and graphic artists use the power
of today's 3D graphics workstations. There are three products providing professional visual communication to aid
rapid understanding.
- The Data Visualizer is designed to speed up the analysis
of large volumes of 3D data on any type of grid. It displays many variables at once and uses color and dynamics to create easily understandable images or animation
sequences of of complex processes.
- The Personal Visualizer is an easy- to-use image renderer
for CAD geometry. It provides photo-realistic images for
product designers, engineers and marketing groups.

Founded
in 1984

Visual
communication

Summary of
Visualization
Products

Current Vendor Systems in Use

The Advanced Visualizer includes interactive surface
modeling, animation and dynamics, and the ability to
render motion sequences and record them on tape or
film for presentation. It has interfaces to a wide range of
CAD/CAE and dynamics packages and is fully compatible with the other visualizer products.
The Data Visualizer
Data
Visualizer
Graphical interface

Interactive analysis

Animation

Analysis of fluids
and structures

Objects and flows
round them

The Data Visualizer is a graphic analysis toolkit for 3 D scalar and vector data on any type of grid - including regular,
irregular, and unstructured grids.
It has a mouse-driven point-and-click interface that allows a wide variety of graphic tools including cutting planes, iso surfaces and iso volumes, particle systems, ribbons,
and sheets to be positioned and turned ON and OFF at the
click of a button. There is no limit to the number of tools
that can be created, grouped, and rendered simultaneously.
The Data Visualizer is designed for the interactive analysis of large 3 D volumes where many tools are required for
simultaneous analysis, and their rapid combination is a key
productivity factor. The user interface therefore provides refined management of tools and screen layout and lends itself
well to an environment in which many users require volume throughput.
The data and all the graphic tools can be animated over
time, and there are also common image file formats such as
color Postscript for use with DTP printers.
The ability to handle unstructured grids gives the Data
Visualizer a strong capability in the next major area of advancement in data visualization. Adaptive unstructured
grids allow complex shapes to be defined by the user with
relatively simple and intuitive tools. This technology will
make numerical analysis of fluids and structures more accessible to the non-expert engineer or designer.
One of the unique features of the Data Visualizer is its
ability to use the same analysis tools in the unstructured environment and for multi-block data where individual grids

Wavefront Technologies, Inc.

have been created around different components in a flow
field. This makes it possible to see a complex assembly such
as an entire aircraft, together with the flow conditions
around it and its various parts.
Finite Element Analysis applications run exclusively on
unstructured grids, and the Data Visualizer provides an assortment of tools for viewing data on the surface of these
grids. These include clipping planes and boxes that cut away
portions of data volumes and iso-surfaces to view interior
detail, while maintaining the exterior view.
There is also a command language for users who wish to
build combinations of tools and results externally - for example from within their solver - and data reader source code is provided to allow on-line network transfer of results
from a host directly to the screen of the graphics workstation.
Figure 7.1 shows a transparent isosurface of rainwater
density in a cloud together with wind velocity flow ribbons
that are themselves mapped with a scalar density value. The
user interface includes interactive color map editing.

Viewing data on
unstructured grids

Command
language for tool
building

Fig. 7.1
Transparent
Isosurface

Current Vendor Systems in Use

Fig. Z2
Fighter Fuselage

Figure 7.2 shows computational fluid dynamics analysis
of a fighter fuselage. The grid is irregular and contains multiple data blocks and approximately 250000 nodes. The picture shows air pressure on the fuselage with particle traces
mapped with mach number, spiralling around a vortex.

The Personal Visualizer
Personal Visualizer

This product was developed for the casual user who needs
to postprocess computer-aided design models. It interfaces
directly with many leading CAD packages and allows interactive control of lights, cameras, and surface materials to create highly realistic images. The Personal Visualizer provides
texture mapping, bump mapping, transparency, refraction
and ray tracing, and has a library of over 700 prepared materials as well as its own surface material editor with which
new surface textures can be created.

Wavefront Technologies, Inc.

The Advanced Visualizer

This product is for those who wish to create realistic motion
sequences, for example an engine assembly in motion, or a
spacecraft docking sequence. It provides all the modeling
tools needed to create the geometry internally and can also
read geometric and stress data from external systems.
The Advanced Visualizer has sophisticated animation
tools and on-line motion channels that can be driven by
ASCII data in real time. This latter feature makes it a useful
tool in displaying the results of computed dynamic analysis,
such as in vehicle crash performance testing. The Advanced
Visualizer includes all of Wavefront's state-of-the-art image
rendering technology, allowing the user to create virtually
any effect that may be required.

Advanced
Visualizer

Sophisticated
animation facilities

Fig. Z3
Engine

Figure 7.3 is taken from an animated engine sequence
showing all the major parts in synchronised motion seen
through the transparent engine block.
Contributed by Peter Stothart, UK Managing Director,
Wavefront Technology Ltd.

Current Vendor Systems in Use

For further information contact:
US.A.
Wavefront Technologies, Inc.
530 East Montecito Street
Santa Barbara
CA 93103 US.A.
Tel: 805-962-8117
Fax: 805-963-0410

Europe
Wavefront Technologies
Guldenspoorstraat 21-23
B-9000 Gent Belgium
Tel: 32-91-254555
Fax: 32-91-234456

United Kingdom
Wavefront Technologies Ltd.
Oakridge House; Wellington Road
High Wycombe
Bucks HP12 3PR UK.
Tel: 0494-441273
Fax: 0494-464904

7.2 UNIRAS A. S.
High quality color
graphics

Raster based

Pioneering
developments

Real world
application of
visualization tools

UNIRAS was established in 1980 with the objective of satisfying the growing need among computer users for a range
of high-quality color graphics software. This need arose
partly from current requirements to analyse ever increasing
volumes of data and partly to utilize fully the high performance graphical output devices coming on to the market.
UNIRAS software uses raster techniques to deliver a
broader spectrum of colors, improved resolution, and greater throughput. UNIRAS therefore makes the most of new
hardware technology, while still effectively supporting the
traditional vector output devices of an earlier generation.
Key management and development people at UNIRAS
have worked with color raster graphics since the early 1970s
and were closely involved in the design of software for the
first inkjet plotters. The current UNIRAS product range
has evolved from this pioneering work.
Reflecting the rapid growth of the offshore oil industry
at that time, UNIRAS' first commercially available product
was a software package to aid exploration companies in their

UNIRAS A.S.

search for oil and gas. This was an early example of the use
of scientific visualization tools in real-world applications
with considerable strategic benefits. Since then UNIRAS
has taken its technology into a number of other application
areas, including automotive and aerospace manufacturing,
pharmaceutical industry, communications, defense, energy
generation and distribution, and environmental management.
UNIRAS software technology comes in two forms. Interactive, user-friendly packages permit non-specialist computer users - scientists, engineers, and managers - to learn
to use the extensive facilities quickly and easily; while the
range of subroutines provides a choice of powerful tools to
help the application programmer integrate high quality
graphics with new and existing applications. All UNIRAS
products are computer and device independent and comply
with accepted international standards.
Today UNIRAS is a truly international organization and
can list many famous companies, research institutions, and
universities among its hundreds of customers. UNIRAS has
its headquarters in Denmark, with wholly owned subsidiaries in the USA, UK, France, Germany and Italy, a sales office in Tokyo, and representatives in other parts of the
world. Research and development takes place in Denmark
and the USA. The major shareholders in UNIRAS are the
Danish financial services group Hafnia, the Dutch investment company Halder Holdings, and UNIRAS' own management.
UNIGRAPH + 2000 is a powerful, fully interactive data
visualization system which enables users to:
• retrieve their technical and scientific data from a file or
database,
• edit and operate on it in a variety of ways,
• analyse and visualize it quickly and in many forms, from
very simple charts to advanced multidimensional surfaces,
• give plots an extra touch of professional presentation
quality,
• present the information graphically as hardcopies of the
highest quality.

Interactive
modules or library
routines

International
vendor and clients

UNIGRAPH

facilities

Current Vendor Systems in Use

60
Visualizing
datasets

Wide variety
of output
devices

Integrating
visualization and
presentation
Windowing
environments

Networked
facilities

Platforms

Benefits and
advantages

Datasets can be accessed, edited, and analyzed using mathematical, logical, or statistical operators. A comprehensive set
of interpolation techniques correctly handles such complexities as regions and barriers, allowing datasets to be visualized as 2 D, 3 D, or 4 D surfaces in color or monochrome.
The UNIGRAPH + 2000 hardcopy system produces
high-resolution hardcopies on a wide variety of output devices including raster and vector devices as well as the new
generation of Postscript printers. Pictures can also be saved
for later use or exported to other systems by the creation of
ISO Standard Computer Graphics Metafiles (CGM) or encapsulated Postscript files.
agX/TOOLMASTER is a suite of high-level graphics
tools which allows the software developer to easily integrate
visualization and presentation techniques into application
programs.
agX/TOOLMASTER programming tools are callable
from C and have been developed and optimized for use in
an X Window environment. The open systems architecture
of agX/TOOLMASTER allows it to be fully integrated
with the X Window, OSF/MOTIF and OPEN LOOK windowing environments.
With agX/TOOLMASTER the application programmer can combine high-level graphics functions with the best
features of X such as multiple windows and event handling
for interactivity, pixmap generation for animation, and client/server techniques for network computing.
agX/TOOLMASTER runs on all major UNIX workstations and supercomputers as well as the VAX/VMS environment, making applications portable across the network.
The virtual color system in agX/TOOLMASTER provides
X-server independence with its support of monochrome
and color displays and various bitplane depths.
The benefits of building an application with agX/
TOOLMASTER are:
• The amount of time and code required for building applications in the X environment is greatly reduced,

UNIRAS A.S.

• Code maintenance and support over the lifetime of an
application is also reduced,
• Productivity of programmers is increased by access to
high-level single function calls for visualization of numerical data,
• Hardware investments are protected.

Fig. Z4

agXJCONTOURS
example

Fig. Z5

agXJVOLUMES
example

Current Vendor Systems in Use
Commitment to
standards

Ongoing
developments

UNIRAS technology is based on standards, and in the
future UNIRAS will continue to develop products that offer flexibility together with computer and device independence. The further development and enhancement of both
the UNlRAS interactive packages and subroutine libraries
will continue to provide high-quality graphics solutions to
both end users and application programmers.
Research and development now takes place in both Europe and the USA in order to reinforce the global scope of
the company's products. Future developments will also take
place with the cooperation of both UNIRAS users and the
computer hardware vendors. The UNIRAS network of local subsidiaries in Europe, USA, and Japan will enable the
company to continuously strengthen local sales and support
activities.
Contributed by Mike Bundred,
General Manager UNlRAS Ltd.

For further information contact:
Denmark
UNIRAS A.S.
376 Gladsaxevej
DK-2860 Soborg
Denmark
Tel: 45-31-672288
Fax: 45-31-676045
United Kingdom
UNlRAS Ltd.
Ambassador House
181 Farnham Road
Slough SLl 4XP

u.K.

Tel: 0753-579293
Fax: 0753-821231

Germany
UNIRAS GmbH
Niederkasseler Lohweg 8
W-4000 Dusseldorf 11
Federal Republic of Germany
Tel: 0211-5961017
Fax: 0211-5961019

Precision Visuals, Inc.

7.3 Precision Visuals, Inc.
Engineers, scientists and researchers have a common need
for visual data analysis (VDA) in order to understand and
use their data. A common requirement is for large datasets
and fast graphics with analytical capabilities such as mathematics, statistics, signal processing, and image processing.
Precision Visuals Workstation Analysis and Visualization Environment (PV-WAVE) is a powerful software system
which lets users display, reduce, analyse and re-display large
multi-dimensional data sets.
As an example, Figure 7.6 exhibits temperature, carbon
monoxide, and sulfur dioxide contents in a city's air for one
year. The 3 D surface allows observation of how all the parameters interact, and integrates the three variables in one
plot. Thus relationships between parameters can be visualized. Menus also allow highlighting and magnification of a
specific dataset.
Visual Data Analysis and PV-WAVE improves upon traditional data analysis by allowing the user to control data

Visual data
analysis

Visualization
environment

User interaction

Fig. Z6
Test Engineering
- Air Quality Data

Current Vendor Systems in Use

analysis by user interaction with visual representations. The
features of PV-WAVE are:
• Reads and defines large, multi-dimensienal datasets
• Tools for fast manipulation and subsetting
• Quick graphical displays of immediate results
• Immediate user interaction
• Advanced graphic tools for animating and displaying
multidimensional data
Increase in
productivity

PV-WAVE

Features

VDA techniques are invaluable for scientific discovery and
engineering analysis. They offer impressive advantages, including increases in productivity and a means for visual
communication with colleagues.
PV-WAVE is interactive software for visualizing and analyzing technical data. It consists of a set of high-level, interpretive commands and procedures that provide:
•

•
•
•
•
Application areas

Data access, reduction and analysis
2 D and 3 D graphics
Dynamic graphics
Image processing and manipulation
Application development

PV-WAVE is being successfully applied in the following application areas:
Laboratory Science
to visualize and analyze data from analytical instruments
to develop instrument automation systems
to create custom laboratory information systems
Test Engineering
to visualize vibration, heat transfer, and emissions test
data
to compare prototype performance with theoretical expectations
- to implement quality control systems in manufacturing
environments

Precision Visuals, Inc.

Real-time Data Acquisition and Control
- to build automated remote sensing systems
to manage water quality and sewage systems
to monitor atmospheric conditions
to create simulations based on real data
Space Exploration and Astrophysics
- to reconstruct planetary and stellar environments
- to study geodynamics in planets and satellites
- to study seismology
to simulate astronomical events and objects
Computational Fluid Dynamics
to identify flow patterns such as shock waves, vortices,
and shear layers
to apply CFD research to aeronautics, automotive design, weather forecasting, and oceanography
- to analyze data from thermal dynamics, fluid dynamics
and nuclear reactions
Finite element modeling and analysis
- to assure quality and reliability in computations involving field equations
to apply finite element pre- and postprocessing methods
to such areas as aircraft design and stress analysis in building components
Imaging - medical and remote sensing
- to postprocess remote sensing data
- to display and analyse bioscience imagery, including
NMR/MRI, X-Ray, CAT, and electron microscopy
Earth Resources
to interpret seismic data
to analyze well logs for locating mineral deposits
- to make meteorological predictions
- to compile and combine raw data for mapping

Current Vendor Systems in Use

Fig. 7.7
Remote Sensing
- Landsat Image

In Figure 7.7. PV-WAVE Point and Click subsets this satellite image in multiple windows using the toolbox. The picture shows the Boulder Valley east of the Rocky Mountains.
The Boulder Reservoir is highlighted in the smaller windows.
Other products

Other members of the PV-WAVE family of visual data analysis software products include:

Numerical analysis

PV-WAVE: NAG - features the powerful numerical
analysis capabilities of the NAG Workstation Library with
the sophisticated visualization and data analysis functions
of PV-WAVE to create a single tightly integrated system, allowing access to 172 subroutines and functions from the
NAG library through a seamless link.
PV-WAVE Point and Click - combines the power and
functionality of PV-WAVE with an easy to use Point and
Click mouse-driven interface that allows technical professionals to access, analyse, and visualize their data, without
the need to program.

Programming
not needed

Stardent Computer, Inc.

For further information contact:
US.A
Precision Visuals, Inc.
Lookout Road
Boulder
CO 80301
US.A.
Tel: 303-530-9000
Fax: 303-530-9329

Germany
Precision Visuals
International GmbH
Lyoner Stern
Hahnstrasse 70
W 6000 F rankfurt/Main 71
Federal Republic of Germany
Tel: 49-69-6690150
Fax: 49-69-6666738

United Kingdom
Precision Visuals International, Inc.
Royal House
1-4 Vine Street
Uxbridge
Middlesex UB8 1XF
UK.
Tel: 0895-235131
Fax: 0895-272299

7.4 Stardent Computer, Inc.
AVS is an advanced interactive visualization environment
for scientists, engineers, and technical professionals. AVS
supports geometric, image and volume datasets - the user
can interactively select the appropriate menu option. No
programming is required.
For the more sophisticated user, the AVS Network Editor can be used to build processing networks into which user-developed modules and computational programs can be
integrated easily.
Modules can by dynamically added, connected, and deleted. Modules are only re-executed when new data is required or an input parameter is changed. Modules have control panels for interactive control of input parameters by on-

Interactive
visualization

Visual Network
Editor

Modular approach

Current Vendor Systems in Use

Building
applications

Integration of
user programs

Further modules

Filters

New filters
Mapping

Renderers

Image processing

screen sliders, file browsers, dials and buttons. The control
panel is automatically generated when a module is connected into the network.
The complete network can be saved with all the user defined interactive controls and layout specifications as a complete application. This can then be invoked directly, bypassing the standard AVS menus and the Network Editor.
User programs can be coupled into the network to allow
real-time visualization of dynamic simulations. This allows
the user to transform existing batch programs into interactive visual applications.
AVS has a wide range of data input, filter, mapper and
renderer modules. User-written programs or subroutines in
FORTRAN or C can be easily converted into AVS modules.
Filters transform data into data or geometry into geometry. Some filters convert the output data of widely-used applications into displayable form. AVS includes filters for applications in engineering analysis, computational fluid dynamics (CFD), chemistry and other fields. Other filters process commonly used data formats such as the Brookhaven
Protein Databank (PDB) molecular structure format, or
graphics formats such as MOVIE.BYU and Wavefront Technologies.
New filters can be developed by means of templates and
geometric conversion utilities.
Mappers transform data into geometry. Multiple visualization techniques can be selected to suit the problem being
studied. Examples of mappers include: isosurfaces of a 3 D
field; 2D slices of a 3D data volume; 3D meshes from 2D
elevation datasets.
Renderers display geometry, images, and volumes on
screen. AVS networks can incorporate multiple rendering
modules, including a fully-featured 3 D geometry renderer,
an image display renderer, and a range of volume renderers.
Graphic images may also be output to hardcopy devices or
video tape.
AVS provides a complete image display capability, including real-time pan and zoom, rotation and transforma-

Stardent Computer, Inc.

tion, flip book animation, and support for 8-bit, 24-bit, and
floating point images. Imaging filters include look-up table
operations such as contrast stretching, pseudo-coloring, and
histogram balancing, as well as data resizing operations such
as interpolation.
AVS takes image processing a step further by generalizing
these modules for 3 D volume imaging. AVS provides a variety of tools for rendering volume data; a real-time isosurface
generator; a unique transparent volume renderer which creates real-time, semi-transparent images with full rotational
and lighting control; generation of geometric objects such
as arbitrary slicing surfaces, dot surfaces and vector nets;
and VBUFFER, a unique, high-quality volume renderer.

Volume imaging

Fig. Z8
Stardent AVS

Current Vendor Systems in Use
Geometry
viewing

Different views

Hierarchies

Animated views
Platforms

The AVS Geometry Viewer gives full control with simpIe menu-driven parameter selections. It offers wireframe,
Gouraud, or Phong shading; 16 individually controlled colored light sources, selectable as point, directional or spot
lights; surface properties such as specularity and transparency; real-time texture mapping and anti-aliasing.
AVS allows creation of multiple windows with different
views of the same geometric object or simultaneous display
of multiple objects.
Scenes with hierarchies of objects can be created and manipulated individually or as one or more groups. Scenes can
be saved, with all viewing selections preserved for later redisplay. Sequences of images can be created and saved, and
the sequence cycled through to provide animated views of
dynamic behavior in real time.
AVS is designed for portability and multi-platform support, from desktop systems to supercomputers. Written in
C, AVS runs in a UNIX X-Window environment. The geometry renderer is designed to support a variety of graphics
display subsystems including the standards PHIGS and
PHIGS+, Stardent's advanced rendering and display environment, DORE (Dynamic Object Rendering Environment), and other display list or immediate mode graphics
interfaces.
AVS modules are a convenient means of exchanging new
computational and visualization software.
AVS is supplied free with every Stardent visualization
system. Licensing of the software is available for other platforms.
Figure 7.8 shows 3D terrain elevation mapping using
Stardent's AVS package. The area shown is Orange County,
southern California.
Figure 7.9 shows the visualization of electron orbital
within a hydrogen atom using Stardent's AVS package.
From information supplied by Stardent Computer Ltd.

Stardent Computer, Inc.

Fig. 7.9
Visualization of
Electron Orbit

For further information contact:
USA
Stardent Computer, Inc.
6 New England
Tech Center
521 Virginia Road
Concord MA 01742
U.S. A.
Tel: 508-287-0100
Fax: 508-371-7414

International Headquarters
Stardent Computer
Hagenauer Strasse 42
W-6200 Wiesbaden 1
Federal Republic of Germany
Tel: 49-611-22037
Fax: 49-611-260181

United Kingdom
Stardent Computer Ltd.
7 Huxley Road
The Surrey Research Park
Guildford Surrey GU2 5RE
U. K.

Tel: 0483-505388
Fax: 0438-505352

Current Vendor Systems in Use

7.5 Silicon Graphics, Inc.
Visual processing

Four levels
of products

Upper and
lower levels
Graphics
Library GL

Silicon Graphics is the world's leading manufacturer of visual processing systems. Visual processing allows staff and
researchers to work with data in a more natural, intuitive
way - graphically. In addition, Silicon Graphics designs,
manufactures and markets computational systems used by
engineers, scientists and animation professionals for design
and analysis of 3D objects and for general purpose technical
computing. By employing RIse technology and proprietary VLSI components to provide both high-performance
computing and high-performance graphics, Silicon Graphics continues to deliver amongst the most powerful systems
available for engineering and scientific applications.
As part of its visual processing software, Silicon Graphics
is offering four modular software products which fit into
different levels between application and source code. Figure
7.10 shows how these products relate to each other. Moving
up the chart shown in Fig. 7.10 away from the source code
and toward the application makes it easier for the user to use
a product without specialist knowledge, but this loses some
of the flexibility at the source level.
Often this difference between the upper and lower levels
is partly or wholly obscured by the proliferation of tools
and facilities currently available in the market place.
The products depicted in Fig. 7.10 are as follows: GL is
the Graphics Library that has been available from Silicon
Graphics for a number of years, and is now gaining acceptance with other hardware suppliers. This provides the basic
system calls required to address the 3D graphics hardware incorporated into every Silicon Graphics system. It is a set of
calls that can be included into source code, which may be
written in a programming language. Above this level is a
graphics toolkit which provides a library of standalone GL
calls that have been developed to address a wide range of
graphics problems and which is available to be incorporated
into software developers' products.

Silicon Graphics, Inc.

Application at User Level

Flexibility

Explorer

)

Image Vision Library

Ease of Use

GL Toolkit

GL (Graphics Library)
Fig. Z10

Source Code at Programmer Level
The more significant products are at the highest level.
Image Vision Library (IL) is an object-oriented, extensible
toolkit for creating, processing and displaying images on all
Silicon Image Graphics workstations. The IL toolkit provides image-processing application developers with a robust
framework for managing and manipulating images. The
toolkit is specifically designed to provide a constant software interface to hardware that may change underneath,
thus ensuring that applications can continue to run unchanged in the future.
The Image Vision Library consists of a shared library developed in C++, with interfaces for C and FORTRAN. It
has a core set of more than 70 image processing operators,
and is user-extensible for specific needs. Silicon Graphics
provides a set of data abstractions and access functions to
make it easy to augment the IL toolkit's image operators
and design new ones.
Image data sets have a wide variety of formats. The IL
toolkit allows new file formats to be integrated into the library. The toolkit currently supports three standard formats: SGI, and extended version of TIFF, and a simple tiled
format called FIT.

Levels of products
Higher level
products
Image Vision
Library

Image processing
operators

Image data set
formats

Current Vendor Systems in Use
Image
manipulation

Data processed
on demand

Visualization
facilities

Customised
application
building

Modular approach

The IL toolkit provides an efficient model for the manipulation of image data and image attributes. The toolkit's image model includes a configurable cache to allow access to,
and processing of, the very large images common in many
disciplines. IL provides a common interface for image manipulations, while requiring little or no programmer knowledge of the image's internal structure of format. IL implements a demand driven execution model, such that data is
processed only on demand. This model is based on the same
cached-image model as file images. This technique enables
an application to process just the area of interest, providing
significant benefits in terms of reduced I/O and improved
system performance.
The fourth product is called Explorer. This is a true applications developer package. It provides visualization and
analysis functionality for users whose needs are not met by
commercial software packages, or who want to extend existing systems with their own algorithms and techniques.
The software environment falls into the category of Application Builders - environments that consist of functional program pieces called modules which are visually connected together through a point-and-click user interface into
a data flow style network. This flexible and interactive environment of building application programs by choosing
from a suite of functional modules is the true power of the
system.
Modules are the building blocks of the visualization software and cover a wide range of functionality. Because the
software environment encompasses a distributed execution
model, modules may execute on the local workstation or on
other platforms on the network. Users can easily integrate
their existing algorithms into the system in the form of new
modules. Through point-and-click selections, a modelbuilding facility is provided which generates the code needed to make the user's algorithm into a visualization software
module. Modules generally fall into the following categones:

Silicon Graphics, Inc.

• Input
Modules that read data files
• Feature Extraction and Analysis
Modules that produce data from data (e.g., extract a planar slice from a volumetric dataset)
• Geometric Representation
Modules that produce geometry-based display lists from
data
• Renderers
Modules that produce images from geometry, volumes
or Images
• Output
Modules that write to disk
System Data Types are the data formats for passing data
through the system. The system data types are powerful and
abstracted, and each can represent an entire class of data as
well as a very specific instantiation of the data. Date types
are as follows:
• Parameter
This data type conveys widget interaction to the module.
Parameters are scalar quantities including long integer,
double precision, floating point, and character string.
• Lattice
This is the most widely used data type. It is essentially
a multi-dimensional array with two major components:
data stored at nodes of the lattice, and a coordinate mappmg.
• Pyramid
This data type combines lattices with connectivity in a
hierarchical structure. The depth of this structure is arbitrary.
• Geometry
This data type contains a hierarchical geometrical scene
description. The geometric description contains all information concerning geometric objects and their attributes, cameras, lights, etc.

Data types

Current Vendor Systems in Use

• Unknown
The unknown type is an uninterpreted array of bytes.
The organisation and interpretation is left to the programmer.
Widening the use
of visualization

IRIS Explorer is a key part of Silicon Graphics' new technical computation environment, aimed at making the company's industry-leading visualization technology more accessible to the broad range of workstation users. With IRIS Explorer, users view data and create applications by visually
connecting software modules into flow chart configurations
called module maps. Modules, the building blocks of IRIS
Explorer, perform specific program functions such as data
reading, data analysis, image processing, geometric and volume rendering, and many other tasks. Modules can be executed across heterogeneous platforms, delivering powerful,
resource-efficient distributed computing capabilities to application users and developers.
The following pictures were created by Silicon Graphics
using AVIRIS data, courtesy of the Jet Propulsion Laboratory. AVIRIS data is made up of 224 spectral bands. The images were created by traversing through the data in the spectral
dimension, and show spectral responses and histograms of
the data.
Figure 7.11 is a general view of the user-interface showing an image of San Francisco with several multi-spectral analyses being carried out. (Created by Silicon Graphics using
AVIRIS data, courtesy of the Jet Propulsion Laboratory.)

Silicon Graphics, Inc.

Fig. Z11
Multi-spectral
Analyses

Fig. Z12
Spectral
Signatures

Figure 7.12 contains an isometric view showing spectral
signatures for a number of pixels along a specific transect of
the original image. Vertical scale is the radiance, horizontal
scales are position and wavelength.

Current Vendor Systems in Use

Fig. Z13
Chair Diagram

Figure 7.13 is a chair diagram of multi-spectral image data. The horizontal plane shows an image at a particular
wavelength. The vertical planes are slices at a given X,Y location showing radiance at the spectral wavelengths which
have been recorded.
From information supplied by Mark Goossens, Education
and Research, Silicon Graphics Ltd.
For further information contact:
USA
Silicon Graphics, Inc.
2011 N. Shoreline Boulevard
P.O. Box 7311
Mountain View
CA 94039-7311
U.S.A.
Tel: 415-960-1980
Fax: 415-961-0595

Europe
Silicon Graphics
International
18 Avenue Louis Casai
CH-1209 Geneva
Switzerland
Tel: 41-22-7987525
Fax: 41-22-7988230

United Kingdom
Silicon Graphics Ltd.
Forum 1 Station Road
Theale Reading
Berks RG7 4RA
U.K.
Tel: 0734-306222
Fax: 0734-302550

Sun Microsystems, Inc.

7.6 Sun Microsystems, Inc.
7.6.1 SunVision Sun's Visualization Software Package
SunVision is a software platform with which application developers and sophisticated end users can develop visualization applications. It includes two programming interfaces
(for image processing and high-quality rendering), and
Open Windows-based tools for image processing, volume
rendering, manipulation of 3 D geometric data, high-quality
rendering and movie loop display. The tools and libraries
are highly integrated so that data and images can be shared
among them.
SunVision 1.1 runs on any 8-bit color SPARCstation/Sun-4 Workstation and on Sun workstations equipped
with the true color VX/MVX visualization accelerators.
No special purpose hardware is required. If a VX or
VX + MVX visualization accelerator is present in the system, the SunVision applications are transparently accelerated.

Visualization
applications

No special
purpose hardware
add-ons required

7.6.2 SunVision Programming Interfaces
SunVision provides two libraries for visualization tasks (image processing and high-quality rendering), and one utility
library. Future releases will include an additional library interface for volume rendering. In addition, SunVision is designed to work in an integrated fashion with XGL, Sun's
3 D interactive graphics library.
SunIPLib is Sun's image processing library, providing extensive image processing functionality. It consists of C-callable functions for:
•
•
•
•
•

arithmetic and logical operations
spatial filtering
Fourier domain processing
image analysis
geometric operations

Libraries

Image processing

Current Vendor Systems in Use

High-quality
pictures

RenderMan

In addition, there are library functions to create and manipulate subimages and regions of interest.
Images can have multiple bands, and can have unsigned
byte, signed 16-bit short, or 32-bit floating point data types.
Additional imaging functions can be added to the library.
For high-quality rendering, SunVision provides a RenderMan function library and a RIB (Renderman Interface
Bytestream) protocol interpreter. The RenderMan interface
provides a way to describe geometry, scenes, the camera, and
lights so that computer images can be generated from this
information.
Users of the RenderMan interface specify a set of procedures that describe a scene. Object color and location, lighting, and viewer perspective can all be specified. The RenderMan shading language provides a way to create shaders specific to a given scene.
The SunVision RenderMan function library is compliant
with version 3.1 of the RenderMan Interface Specification.
It also supports the following optional features:
• solid modeling
• programmable shading
• displacements
• texture mapping
• environmental mapping (partial implementation)
• bump mapping
• volume shading (partial implementation)
The following surface shader options are also provided:
• general (the default)
• the 6 standard RenderMan surface shaders (plastic, painted plastic, metal, shiny metal, matte, and constant)
• the 2 standard RenderMan atmosphere shaders (depth
cue and fog)

7.6.3 SunVision Window-based Tools
Visualization tools

SunVision also provides Open Windows-based visualization
tools for image processing (SunIP), volume rendering (SunVoxel) , 3 D graphics manipulation (SunGV), high quality

Sun Microsystems, Inc.

rendering (SunART), and movie loop display (SunMovie).
Additionally, there is an interactive colormap editor that is
accessible by each tool. These tools can be used "as is" by
sophisticated end users, or as "application prototypes"
which developers can tailor to a specific application.
Each window-based tool is an independent program that
communicates with a shared parameter database program
(PMGR), which, in turn, communicates with a user interface management program (SunVIF). The user interface for
each tool can be changed at run time, with no programming. Additional programs can be easily added to the SunVision user interface.
SunIP and SunART are tools that implement the image
processing and RenderMan functions described above. The
source code for SunIP and SunART is provided in the form
of examples for using these libraries, along with SunVIF and
PMGR.
SunVoxel is an interactive tool for the generation of images from volume data. It consists of window-based rendering
and analysis functions. With SunVoxel, volume data can be
manipulated and viewed in two modes:

Shared information

Images from
volume data

(1) manipulate the entire volume, using orthogonal or oblique slicing planes where needed, and view internal
structures using ray-casting, and
(2) extract and view 2D slices of the volume data in a "light
box" mode. There is also a "cloud" mode for displaying
data stored in point cloud format.
SunVoxel supports unsigned byte data on uniform rectangular grids. Data filters are provided to convert TAAC-1
volume data to the SunVision data format.
SunGV is used to interactively view 3D geometric data.
It can also be used to edit scenes that can be input to SunART for final rendering. SunGV provides wire frame and
Gouraud shaded display of polygon and patch data types.
Scenes can be composed of a series of objects which are organized hierarchically in a tree structure. Editing functions
allow the user to select, copy, paste, cut, and delete objects

Viewing 3D
geometric data
interactively

Current Vendor Systems in Use

Movies

in the hierarchy. Objects can be transformed and assigned
attributes, such as color, opacity, specularity, texture, etc.
Functions are also provided for changing the viewing and
projection parameters, and for defining the lighting model,
which supports up to 32 light sources. The source code for
SunGV is provided as an example for using the XGL graphics library along with SunVIF and PMGR.
SunMovie is a tool for the display of image and movie
loop data. Images and sequences of images generated by other components of SunVision can be viewed using this tool.

7.6.4 The VX and MVX Sun's Visualization Accelerators
Accelerators

Applicatioh areas

Sun's new visualization workstations are powered by the
new VX and MVX accelerator boards. The VX accelerator
is a successor to the TAAC-l, incorporating added features,
including twice the memory, double the performance, multiple windows, and a lower price.
The MVX is a multi-processor accelerator that can be
added to a VX system, providing performance of more than
4-6 times the TAAC-1. Both accelerate SunVision's visualization tools and libraries, and XGL, Sun's graphics library.
VX and MVX systems are for developers who require a
combination of imaging and graphics to develop visualiza~
tion software. Target markets include medical imaging, remote sensing, earth resources, scientific visualization, and
AEC.
7.6.4.1 Features and Benefits

VX
Characteristics

• Accelerates SunVision and XGL.
• High-performance Intel i860 processor (40 MIPS, 80
MFLOPS) for high-speed integer and floating point computation required by visualization applications.
• Dual frame buffer architecture with 32-bit VX and 8-bit
GX accelerated frame buffers on one board; a digital keying technique is used to cleanly integrate the VX windows into the system display.

Sun Microsystems, Inc.

• Transparent integration of multiple VX windows into
the GX Open Windows environment; four independent
colormaps are available for use by the multiple VX windows, with a fifth colormap allocated for the GX.
• Reconfigurable VX frame buffer can be used to display
24-bit true color plus 8-bit alpha, or four independent
8-bit channels.
• Supports Sun's new 1280x1024 @ 67Hz format as well
as the existing 1152x900 @ 66 Hz format.
• Single 9U VME board.
MVX
• Provides additional acceleration of SunVision and XGL.
• Four Intel i860 processors, offering a total of 160 MIPs
and 320 peak single-precision MFLOPS (240 peak double-precision MFLOPS).
• Four Mbytes of memory per processor for fast data and
Image access.
• High-speed data bus for fast, smooth data and image
transfer between the MVX and VX.
• High-speed control bus for transfer of commands between the MVX and VX, eliminating VME overhead.

Software Included
• SunVision and XGL for the most integrated, easy-to-use
visualization environment, with the widest range of
graphics and imaging functionality, in the industry.
• Complete set of C-development tools, including compiler and debuggers, for application developers.
For further information, contact:
Doug Schiff
SunVision Product Manager
p. O. Box 13447
Research Triangle Park
NC 27709
U.S.A.
Tel: 919-469-8300
Email: doug.schiff@East.Sun.COM

Current Vendor Systems in Use

Sun's new VX visualization accelerator delivers high-performance across the full range of visualization techniques, including image processing, volume rendering, 3D graphics,
and high-quality rendering. A multiprocessor MVX board
can be added to the VX model to boost performance to 160
MIPS and 320 peak single precision MFLOPS. The VX and
MVX accelerate Sun's XGL and SunVision software libraries. Shown here are a Gouraud shaded teapot, a volumerendered air duct, and a 2D Landsat image, running in Sun's
Open Windows environment. (Figure 7.14)

Fig. Z14
VX and MVX
Overview

SunVision is a software toolkit for integrated, desktop visualization on any SPARC-based Sun Workstation with a
GX or VX accelerated color frame buffer. It delivers image
processing, volume rendering, interactive 3 D graphics viewing, and high-quality rendering in Sun's Open Windows environment. Figure 7.15 shows a wireframe model of a tea
set, a volume rendered CT head, SunVision's colormap editing tool, filtered images, and a photorealistic rendering of
geometric objects.

Sun Microsystems, Inc.

......

-----

Fig. 7.15
SunVision

Overview

A new volume rendering technique called "splat", developed by Lee Westover while at the University of North Carolina, was used to render different views of a 256x256x96
CT scan of a human head (Figure 7.16). The four images depict the skin and bone surfaces. The transparency of the
skin is changed in each picture. The upper left view shows
an intermediate step as the data is being rendered from back
to front. The algorithm is easy to parallelize; these images
were generated on a Sun VX +MVX visualization accelerator, with four processors doing the data shading and one
processor compositing the samples into the final image.
SunVision's volume rendering tool, SunVoxel, generates
images from 3D volumetric data. The data can be manipulated and viewed in one of several modes, including slice
mode, ray-casting, surface editing mode, point cloud mode,
and light table mode. In Figure 7.17, a 256x256x96 CT scan
of a head is rendered in ray-casting mode with semi-transparent substances. The user can easily assign color and opacity properties to ranges of data values that represent areas of
interest.

Current Vendor Systems in Use

Fig. 7.16
Four Views of CT
Head

Fig. 7.17
Semi-transparent
CT Head

Fig. 7.18
2 D slice from
CT Head

Fig. 7.19
Filtered 2 D slice
from CT Head

Sun Microsystems, Inc.

SunVision's volume rendering tool, SunVoxel, lets the user extract 2D slices from the 3D volume. In Figure 7.18 a
single slice from the 256x256x96 CT scan of a head (shown
in Fig.7.17) is selected and displayed. The image can be
scaled to any size, and regions can be marked and deleted.
The image can be saved to a file for future use.
Sun's image processing library, SunIPLib, provides software for analysis and manipulation of images. SunIPLib
provides arithmetic operations, logical operations, spatial
filtering, Fourier domain processing, morphological operations, geometric operations, statistics, and more. The image
saved in Fig.7.18 has been processed with morphological
functions to give Fig.7.19.
All Sun pictures are reproduced by permission of Sun
Microsystems, Inc.
From information supplied by Donna McMillan, Sun Microsystems, Inc.
For further information contact:
West US.A.
Sun Microsystems, Inc.
2550 Garcia Avenue
Mountain View
CA 94043
US.A.
Tel: 415-960-1300
Fax: 415-969-9131
United Kingdom
Sun Microsystems Europe
Bagshot Manor
Green Lane
Bagshot
Surrey GU19 5NL
UK.

East US.A.
Sun Microsystems, Inc.
P. O. Box 13447
Research Triangle Park
NC 27709
US.A.
Tel: 919-469-8300
Email: donna.mcmillan@
east.sun.com

Current Vendor Systems in Use

7.7 Sterling Federal Systems, Inc.
7.7.1 FAST (Flow Analysis Software Toolkit)
Aerodynamics
applications

Modular approach

FAST functions

Visualization of computational aerodynamics requires flexible, extensible, and adaptable software tools for performing
analysis tasks. Full scale, 3 D, unsteady, multi-zoned fluid
dynamics simulations are common features of typical problems at NASA Ames' Numerical Aerodynamic Simulation
(NAS). NAS scientists perform calculations on CRAY 2
and CRAY-YMP supercomputers and then graphically visualize the results on IRIS workstations. New developments
in the scientific computing environment warrant a new approach to the design and implementation of analysis tools
with multiple processor workstations available in the 2-8
Mflop range. FAST is a software environment for analysing
such computational fluid dynamics (CFD) data.
The FAST environment consists of a collection of separate programs (modules) that run simultaneously. Using these modules, the NAS CFD scientist can efficiently examine
the results of numerical simulations.
FAST provides functions which include:
• Loading data files,
• Performing calculations on the data,
• Constructing scenes of 3 D graphical objects that may be
animated and recorded.

Superior to
earlier approaches

While these capabilities existed to some extent, they were
spread across many specialized programs such as RIP, SURF,
GAS and PLOT3D with overlapping functionality. These
specialized programs were problematic because the data was
only partially compatible, and user interfaces varied widely.
The approach used in FAST solves these problems. FAST
creates an environment of compatible modules, each with
its own purpose and functionality. In addition, each module
has a consistent, easy-to-use, highly interactive user interface
(using the Panel Library developed by David Tristram, NASA, Code RNR). A programmer can add a new module to

Dynamic Graphics Ltd.

FAST by making use of the data in shared memory, the
PANEL LIBRARY interface, and the NAS (input/output)
module. With these features, the FAST team has worked to
make the environment as extensible as possible.
Complex fluid dynamic simulation problems created a
need for new visualization techniques not possible with the
existing software programs. These techniques will change as
the supercomputing environment (and hence the scientific
methods employed) evolve even further. Flexibility means
the ability to handle a diverse range of problems. Extensibility means the ability to interact at all levels of the software hierarchy, either through existing built-in functionality or through the implementation of custom 'plug-in'
modules. Adaptability means the ability to adapt to new
software and hardware configurations through the use of
modular structured programming methods, a graphics library standard, and common network communication protocols (like UNIX sockets) for distributed processing.

Enhanced
techniques

Flexibility
Extensibility

Adaptability

For further information contact:
Sterling Federal Systems, Inc.
1121 San Antonio Avenue
Palo Alto, CA 94303, U.S.A.
Tel: 415-964-9900

7.8 Dynamic Graphics Ltd.
Interactive Volume Modeling (IVM) is a product from Dynamic Graphics, Inc., which models, displays, and interactively manipulates measured property values P in three dimensions located by X, Y, and Z.
The modeling procedure in IVM takes scattered data
points of a physical property value (e.g., porosity, temperature, salinity, chemical concentration) and calculates a threedimensional grid. This grid represents the modeled distribution of the property in three-dimensional space.

Multi-dimensional
modeling

Data pOints

Current Vendor Systems in Use
Calculation
methods

3D gridding
technique

New methods

Displaying the
information

Display
requirements

The property model can be calculated by one of three
methods. The first method calculates the grid throughout
the volume defined by the input data distribution, or by the
user. The second method restricts the calculation laterally to
the area enclosed within a predefined polygon, i.e. limits in
X and Y, but not in Z. The third method allows the user
to specify previously calculated faulted or unfaulted two-dimensional structural surfaces as hard boundaries which limit the modeling process in X, Y, and Z. This third method
enables the user to calculate, for example, a model of porosity or permeability within a zone while not allowing the
model to be influenced by measured values in underlying or
overlying layers. This procedure provides a much more realistic model of property variation within a zone.
The procedures for calculating property models are
based on a three-dimensional extension of the robust two-dimensional gridding techniques used by Dynamic Graphics'
Interactive Surface Modeling (ISM) program. These routines
utilise a variation of the minimum tensions surface algorithm.
A newly developed strategy is currently being tested.
This procedure is designed to enhance lateral continuity
within layers. The user is given control over the degree of
horizontal continuity that the gridder tried to establish.
The calculated property model can be made to conform to
the shape of either an underlying or overlying structural
surface. The results so far have been promising, especially
with thin laterally continuous or discontinuous beds.
A three-dimensional grid is of limited value without display techniques that allow the user to rotate, slice, peel, and
otherwise manipulate the model in real time. This rapid interactive ability is vitally important because no single view
can adequately reveal the complex geometric relationships
contained within any model. IVM provides these capabilities.
Before any manipulation can occur the user must build
a display file from the three-dimensional grid. This display
file is in essence the three-dimensional equivalent of a two-

Dynamic Graphics Ltd.

dimensional contour map. This file contains three-dimensional isovalue surfaces drawn at selected intervals throughout the modeled volume. These surfaces are displayed as
smooth color-filled Gouraud-shaded bodies.
Once built, the user can manipulate the display file in
numerous ways to better understand the internal relationships contained within the model. The model can be sliced
along the X, Y or 2 axes at specified intervals. The model
can also be sliced first along Y, then X, then 2, or any combination thereof to produce color-filled sections along all
three sides. The model may be rotated to any combination
of user specified azimuth and inclination, or may be "grabbed" and turned to any desired orientation.
While slicing and rotating, the user can select a particular range of isovalue surfaces to be displayed (or not displayed). For example, the user may decide to display only
those porosity values between 6% and 9%, or to display all
porosity values except those between 6% and 9%.
Alternatively the user can use the "chair mode" mode
display which cuts out only a piece of the model parallel to
the X, Y, and 2 axes. The chair void is bounded by vertical
walls (X and Y axes) and a horizontal floor (2 axis) on
which are displayed color-filled sections of the property distribution. The user can interactively adjust the width,
depth, and height of the void along any of the axes. Also,
the user can elect to display, within the chair void, a range
of isovalue surfaces. These surfaces give the appearance of
being extruded into the missing volume.
The user can also rapidly flicker between two different
property models which gives the effect of superimposing
the two models. The user loads one model and selects the
proper orientation that best displays the property distribution. Then the other model is loaded with the same display
parameters. With the touch of a single key, the user flicks
between each display as rapidly or slowly as desired. This is
very useful for comparing distributions of such important
properties as porosity and permeability.

Understanding
aspects of
the model

Slicing and
rotating

Chair mode

Superimposing
models

Current Vendor Systems in Use
Choice of colors
Analysis

Calculating
volumes

Trend grids

Grid operations

Data extraction

The Color Table Editor allows any combination of colors to be selected.
IVM contains a full complement of the extensive analysis capabilities found within Dynamic Graphics' two-dimensional mapping package, ISM. .
The user can calculate volumes in a variety of ways. Volume can be determined for the entire model, between isovalue surfaces, within a surface polygon, between two-dimensional structural surfaces, above and below specific depths,
or any combination of these possibilities. A typical problem
could be: "Calculate the pore volume between 6% and 9%
porosity within Zone B inside of Lease J above the oil/water
contact W".
Three-dimensional trend gridding is available for polynomial surfaces between 0 and 14th order. These trend grids
can be subtracted from property grids to calculate three-dimensional residual surfaces. These residual grids can be used
for display or volumetric calculations, if desired.
IVM contains three-dimensional extensions of nearly all
of the standard grid operations found in two-dimensional
mapping packages. A property grid can be modified
through such mathematical operations as addition, subtraction, multiplication, or division by a constant or another
property grid. Other grid functions are also available.
Data can be extracted from the three-dimensional grid in
the form of individual data points or as two-dimensional
grids. For individual points the user must supply X, Y and
Z coordinates for the locations at which values will be back
interpolated from the property model. Two-dimensional
grids can be extracted either along a slice through the body,
or at node locations defined by a previously calculated twodimensional grid. The extracted two-dimensional grid values can be either the discrete back interpolated values or the
column averages between two structural surfaces. The data
extracted from three-dimensional grids can be used for simulation purposes, additional analyses, or for standard two-dimensional computer mapping.

Dynamic Graphics Ltd.

IVM has been successfully used to display, verify, and edit three-dimensional seismic velocity files. IVM is being
used with reservoir simulators both as a front-end processor
(build and verify geometric relationships) and as a back-end
processor (display and manipulate results from time steps).
IVM is being used to monitor and evaluate various thermal
and chemical enhanced oil recovery projects, and is being
used to better understand reservoir geometries and properties.
One key area that is being studied is the distribution of
permeability within a reservoir. Within certain reservoirs
tilted permeability barriers have been well defined and studied. This knowledge can lead to changes in established drilling programs, and hopefully, significantly increased oil production. IVM is also used to monitor distribution and
movement of pollutants within aquifers, and to assist in clean-up efforts. Concentrations of ozone within the Earth's
atmosphere are being studied with IVM, as well as salinity
and temperature distributions within oceans.

93
Applications

Studies in
the environment

Distribution and
concentration levels of PCE organic
solvent contamination below a
WW11 airbase site.
Data from test
boreholes.

Fig. Z20
PCE Plume

Current Vendor Systems in Use

IVM modeled and
displayed temperature data collected
for 11 years to reveal the higher
temperature water
from the Gulf
Stream does not
pass into the Arctic Ocean.

Fig. Z21
Fram Strait Connecting the
Atlantic and Arctic
Oceans
IVM was used to
depict the simulated concentration
and dispersion of
pollutants in a
plume from a municipal garbage incinerator stack in
Minneapolis.

Fig. Z22
Simulated Plume
from Smoke Stack

Dynamic Graphics Ltd.

Geoscientists can now study and analyse three-dimensional geometric and property relationships in ways that
previously have been impossible with two-dimensional
mapping techniques. IVM's modeling, display, manipulation, and analysis capabilities are indeed applicable to a large
number of geoscience problems.

95
Summary
Advantages of 3D

IVM was used to
model and display
porosity data from
borehole readings
within discrete
lithologic units
which were then
combined into a
single model. This
provided a unique
visualization of porosity distribution
and enabled improved volumetric
calculations and
recovery techniques to be employed.
Fig. 7.23
Porosity Modeling
for Oil Recovery

Porosity variations
in a particular reservoir. Isoporosity
shells are peeled
back to reveal
those areas with
porosity greater
then 6%.
Fig. 7.24
Porosity Variations

Current Vendor Systems in Use

From information supplied by Peter Irwin, Dynamic
Graphics Ltd.
For further information contact:
US.A.
Dynamic Graphics, Inc.
1015 Atlantic Avenue
Alameda, CA 94501, US.A.
Tel: 415-522-0700
Fax: 415-522-5670

United Kingdom
Dynamic Graphics Ltd.
Addison Wesley Building
Finchampstead Road
Wokingham
Berks RG 11 2NZ, UK.
Tel: 0734-774755
Fax: 0734-774721

7.9 Spyglass, Inc.
7.9.1 Spyglass Transform
Analysis and
visualization

Data input

Output of results

Transform is a comprehensive tool for analysis and visualization of two-dimensional data on the Macintosh. It can
generate contour plots, surface plots, vector plots, line
graphs, polar images, animations, overlays, and raster images
(assigning colors to data values).
The Import command reads and converts 2D and 3D
HDF datasets, 2D and 3D generic datasets (byte, integer,
long integer, float), HDF image files, PICT files, TIFF files,
FITS files, 2 D ASCII data, and X-Y column data not already
in array form. Transform reads 3 D data one slice at a time.
Every image, plot, or dataset can be printed to any color
or black-and-white Postscript printer, or any Macintoshcompatible color printer. Exporting graphics to other Macintosh applications, or to produce 35 mm slides is easy. Data
and images created in Transform can be exported as PICT
or HDF files, or via the Clipboard.

Spyglass, Inc.

7.9.2 Spyglass Dicer
Dicer is a comprehensive tool for visualizing volumetric data on the Macintosh. It can perform 3 D blocks, slices in
three planes, 3 D cubes and cutouts, variable orientation, data re-sampling, variable color maps, and animation sequences.
Dicer reads 3 D HDF (float and byte), netCDF, and generic (byte, short and long integer, single- and double-precision floating point) file formats directly. A utility can convert and import folders of 2D files in ASCII byte, integer,
and floating-point formats, and construct 3 D datasets from
the 2D files. The utility also imports 3D ASCII data files,
as well as PICT, TIFF, FITS, and HDF image files.
Dicer offers a menu of over 20 color tables, or user tables
can be imported. Interactive tools enable any color to be
made transparent to enable the user to 'see through' the volume it previously occupied, or to substitute colors in selected regions.
Any configuration created on screen can be saved with
or without corresponding data, and snapshot images can be
saved alone as PICT or HDF image files. Dicer images can
be exported as PICT files using the Macintosh Clipboard,
and 2 D slices can be exported to Spyglass Transform for further manipulation. Data sets can be re-sampled and saved as
either HDF or netCDF. Any image created on screen can
be printed to color (or black and white) Postscript printers,
or to Macintosh-compatible color printers.
Dicer also offers two ways to create and save animations.
After defining parameters, the user can generate and save sequences of 3 D frames, or a sequence of 2 D slices from 3 D
frames created in Dicer. The sequenced images are saved in
folders and can be viewed as animations using Spyglass
View.
Information courtesy of Digital Studio.

Visualizing
volume data

Data input

Color options

Output

Animations

Current Vendor Systems in Use

For further information contact:
US.A.
Spyglass, Inc.
701 Devonshire Drive, C-17
Champaign
IL 61820, US.A.
Tel: 217-355-1665
Fax: 217-398-0413

United Kingdom
Spyglass, clo Digital Studio
Clifton Mews
Saffron Walden
Essex CB10 lEE, UK.
Tel: 0799-513773
Fax: 0799-513454

7.10 LightWork Design Ltd.
High-quality
images

Natural
phenomena

Applications

User interface

LightWorks is a powerful new image generation system
from LightWork Design Ltd. that enables users of modeling
software to create high-quality images of their models,
showing accurate surface finishes and lighting effects. It is
optimised for interactive operation in computer-aided design environments.
Images of photographic quality are produced by simulating natural phenomena such as reflective surfaces, transparency and shadows from any number of light sources. Arbitrarily complex material characteristics can be defined to
create realistic finishes such as marble, wood, brick, chrome
and steel. A wide range of geometric modeling primitives
are supported on to which scanned images can be mapped.
LightWorks can be used in many application areas. Examples are visualization for product designers; visualizing
simulation data for scientists and engineers; building designs for architects; accurate simulations of color and lighting schemes for interior designers; print and animation for
graphic designers.
The system has an interactive mouse and window-based
scene editor module which presents an easy-to-use yet powerful user interface with which the visual characteristics of
a model can be controlled. Incremental rendering techniques enable many color and lighting combinations to be
compared in a short time.

LightWork Design Ltd.

Fig. Z25
Engineering
Component

This engineering component shows different kinds of
metal surfaces - the machined chrome, the cast red, and the
threads. The "bumps" on the background and also on the
red surface are done by displacements. Lighting and shadows are also present.

Fig. Z26
Architectural
Application

An example of an architectural application. This shows
different finishes, soft shadows, and a perspective view.

100

Current Vendor Systems in Use

Written in C, LightWorks has been designed to be easily
ported across a wide variety of computer platforms including UNIX workstations, MS-DOS PCs, and cost-effective
parallel processing computers. LightWorks is supported on
Sun-4/SPARC, IBM RISC System/6000, Silicon Graphics
IRIS/4D, Hewlett Packard 9000 300/400/700/800, Sony
NEWS, and 80386 and 80486 IBM PC compatibles running
MS-DOS with Phar Lap's 386/Dos-Extender or Microsoft's
Windows 3.0.
Information supplied by Dr R. Gordon Oliver,
LightWork Design Ltd.
For further information contact:
Lightwork Design Ltd.
Sheffield Science Park
Arundel Street
Sheffield S1 2NS, U. K.
Tel: 0742-724126
Fax: 0742-720379

7.11 Ricoh Company Ltd.
Use of Visualization in Modeling
and CAD/CAM
Solid modelers, designed to represent 3 D shapes as solids,
3D shapes have become essential tools in computer-aided design and
manufacturing (CAD/CAM) systems. The benefits of solid
modelers vary from designing to molding, and from structural analysis to robot simulations. Unlike surface modelers,
however, existing solid modelers have had limitations in the
representation of complex surfaces.
3D modeling
DESIGNBASE is a 3 D solid modeling system developed
by Ricoh Company Ltd. This is designed to represent complex free-form surfaces in UNIX workstation environments. Using a surface representation method called the
"Gregory patch", DESIGNBASE enables free-form surfaces
to be smoothly connected and locally operated.

Representing

Ricoh Company Ltd.

101

Fig. Z27
Scene produced
by DESIGNBASE

DESIGNBASE provides various surface representation
methods: the Gregory patch, the rational boundary Gregory patch, natural quadric surfaces (spheres, cylinders, cones),
and n-th degree rational Bezier patches. Featured with these
surface representation methods and translation libraries,
DESIGNBASE is capable of bi-directional data exchange
with other CAD/CAM systems. DESIGNBASE translates
its surface data to non-uniform rational B-splines (NURBS),
the de facto standard surface representation method, and
sends the data through Initial Graphic Exchange Specification (IGES) to other CAD/CAM systems.
The Gregory patch is the more suitable surface representation method for the smooth connection of free-form surfaces compared to the Bezier patch. A bicubic Gregory
patch, which is defined by 20 control points, is an extension
of the bicubic Bezier patch, which has 16 control points.

Surface
representation

Connecting
surfaces smoothly

102

Current Vendor Systems in Use

Fig. Z28
Free-form Surface
represented by
Gregory Patch

The extra 4 internal control points of the Gregory patch are
used in the smooth connection of surfaces and enable the
interpolation of irregular meshes.
Fig. Z29
Irregular Mesh
represented by
Rational Boundary
Gregory Patch-1
Fig. Z30
Irregular Mesh
represented by
Rational Boundary
Gregory Patch-2

Fillets

Figure 7.29 shows the control points of the rational
boundary Gregory patch. A six-sided irregular mesh is interpolated by 6 RBG patches.
In Figure 7.30 the contour lines show the smoothness of
this surface representation method.
Filleting is a crucial capability for CAD software for designing products such as automobile engines or electric appliances.

Ricoh Company Ltd.

Boolean operations (union, difference, and intersection)
are critical in the representation of complex shapes. However, most solid modeling systems approximate the shape of
the free-form surfaces with facets when executing these operations.

103
Boolean
operations

Fig. 7.31
Engineering
Component
created by
applying Boolean
Operations

When natural quadric surfaces are intersecting,
DESIGNBASE uses a special library for high-speed surfaceto-surface intersection calculations.
This shows an example of a body generated by applying
Boolean operations. The geometric accuracy of the resultant
solid based on these operations is within the margin of
10- 6 on intersection vertices and 10-3 on intersection
curves.
To support the trial and error implicit in the design process, DESIGNBASE provides high-speed Undo, Redo and
Re-execution commands. Each command of DESIGNBASE
is subdivided internally into primitive operations, and each
primitive operation has a reverse operation. By executing
the reverse operations, previous shapes can be generated by
tracing the tree-type history backwards (Undo) and forward
(Redo). In addition, shapes input previously can be modified easily by giving different parameter values (Re-execution); this facility is useful for designing analogous shapes.

Trial and error
in the design
process

104

Current Vendor Systems in Use
Summary

Applications

Computer-aided design (CAD) began with 2D systems
which replaced drafting instruments. However, total
CAD/CAM or computer-aided engineering (CAE) systems
require the facility to process 3 D data, and it must be possible to freely exchange this data between systems.
Solid modeling systems are assuming a more significant
role as industry introduces CAD/CAM/CAE systems.
High powered workstations enable such systems to perform
efficiently and also cost-effectively.
DESIGNBASE has been applied to the automobile and
electricity industries as the basis of CAD/CAM/CAE requirements. It has also been used for rendering, stereo lithography, and pre-processing for the production of finite
element meshes.

Reference
Chiyokura H.: Solid Modeling with DESIGNBASE: Theory
and Implementation. Addison Wesley, Reading, MA,
1988.
Information supplied by T. Ito, A. O'Neill, and H. Toriya,
Ricoh Company Ltd.
For further information contact:
Ricoh Company Ltd.
1-1-17 Koishikawa
Bunkyo-ku
Tokyo 112
Japan
Tel: 81-3-3815-7261
Fax: 81-3-3818-0348

Vital Images, Inc.

105

7.12 Vital Images, Inc.
Vital Images began developing software as part of a research
project at Maharishi International University to visualize laser-scan confocal microscope data of living nerve cells. This
research was funded by Iowa Department of Economic Development and the National Science Foundation.
VoxelView is a general purpose, high-performance volume rendering package. Data values of voxels can be mapped to corresponding opacity values. This enables the user
to view faint or small details inside the volume. It is also
possible to do thresholding to remove voxels, or redistribute
voxel values from one range to another. Sequences of renderings can be stored and then played back in real time. The
following features are also included:

VoxelView

graphical data base system,
full surface shading with lighting,
user control of animation parameters,
gradient operations to extract and selectively display
nested surfaces within the volume,
• autoconfiguring for multiprocessor systems.
•
•
•
•

VoxelLab is an entry-level version of VoxelView and is available on the Silicon Graphics workstation. It enables beginning users to appreciate the power and potential of volume
rendering systems.
For further information contact:
Vital Images, Inc.
P.O. Box 551
Fairfield
10 52556
U.S.A.
Tel: 515-472-7726
Fax: 515-472-1661

VoxelLab

Chapter 8
Current Public Domain Systems
in Use

Readers are recommended to read Chapter 4 for a classification and categorisation of scientific visualization systems
before reading this present chapter. Chapter 4 sets out the
overall framework into which the products outlined in
Chapters 7 and 8 fit.
In addition, readers should note the points on public domain systems that are set out in Chapter 4 (Section 4.4).
Contact addresses are provided for each product at the
end of each section.

8.1 Khoros
Khoros is an open environment for data processing, visualization, and software development. This summary describes
how the Khoros software system can be utilized as a foundation or platform to improve productivity and promote software reuse in data processing applications. First, a high-level
description of Khoros is given, then the current status of
Khoros is discussed.

8.1.1 Overview
Tool for research
and development

The Khoros system integrates multiple user interface modes,
code generators, instructional aids, data visualization, and
information processing. The result is a comprehensive tool
for computational research and development. The Khoros
infrastructure consists of five major subsystems:
• a general visual language,
• a user-interface development system (UIDS),

107

Khoros

• an interoperable data exchange format (vif£),
• application-specific data processing libraries,
• interactive data display/manipulation programs and a visualization toolkit.
The software structure that embodies this system provides
for extensibility and portability, and allows for easy tailoring to target specific application domains and processing environments. Khoros is a successful example of how research
programming, end-user applications programming, information processing, data visualization, instruction, documentation, and maintenance can be integrated to build a
state-of-the-art software environment.

8.1.2 Subsystem Component Descriptions
a) X Windows Applications
The interactive graphical user interface programs are based
on MIT XllR4 and the Athena widgets. They are all designed to have a simple and consistent look and feel.
Program Name

Description

cantata

extensible visual programming
language
interactive image display and
manipulation
interactive image sequence display
comprehensive 2D and 3D
plotting packages
surface visualization (imagery
draped over elevation data)
image registration and warping
distributed user interface controller

editimage
animate
xprism2 and xprism3
vlewlmage
warplmage
concert

Portable and
extensible

108

Current Public Domain Systems in Use

b) Visual Language
Dataflow graphs

Many applications

The visual language of Khoros, cantata, is a graphically expressed, dataflow-oriented language. The user builds a cantata application program by connecting processing nodes to
form a dataflow graph. Nodes are selected from an application specific library of routines created using the Khoros
UIDS, and may have arbitrary granularity, from fine to large
grain. Control nodes and a parser extend the functionality
of the underlying data flow methodology. Visual procedures, representing a hierarchy of subgraphs, add structure
to the visual language and help to manage the complexity
often associated with visual programming. A dynamic execution scheduler allows the user to interactively execute the
entire flow graph across a heterogeneous computer network.
The execution can be set to either a demand-driven or datadriven model depending on the application and desired level
of interactivity.
Cantata has been targeted at a variety of application domains: a visual language interface has been completed for
the LINPACK/EISPACK libraries, an image processing library, a digital signal processing library and a remote sensing/geographical information system. The design of cantata
promotes code reuse and modular design of libraries.

c) User Interface Development System

Dialog

The Khoros system combines interactive graphical user interface specification/editing and code generation to give the
user a programmer's assistant. This UIDS can be used to create general X Windows applications or to extend cantata.
The central component of the UIDS is a high-level user
interface specification that represents a formal description
of the dialog between the user and the application, independent of the user interface mode. The specification is used directly to generate the code for either a graphical or command-line user interface. When the user interface specification is combined with a formal Khoros program specifica-

109

Khoros

tion, the entire application (documentation and code) can
be maintained via a set of automated source configuration
tools.
The software development tools that are provided allow an
end-user to act as a developer to extend the system.
Program Name

Description

preVIew
composer

graphical user interface display tool
interactive graphical user interface
editor
code generation tool for a graphical
user interface
code generation tool for a command
line user interface
source configuration and management tool

conductor
ghostwriter
kinstall

The user interfaces created with the Khoros UIDS all utilize
the same layered libraries. This provides features common to
all applications, such as:
•
•
•
•

journal recording and playback,
distributed user interface,
reconfigurability without recompiling,
consistency of use.

Perhaps the most powerful and innovative item in the list is
the distributed user interface or groupware capability. All of
the graphical user interfaces created within the Khoros system allow for multiple user interfaces to be running on different machines. This allows a group of researchers (either
as master and slaves or as all masters) to simultaneously interact across a network using the same data and application
software. This groupware capability motivates users to share
resources.

User interfaces

110

Current Public Domain Systems in Use

The UillS also promotes consistent structured programming methodology and styles as well as code reuse. The hope is that as Khoros evolves, many reusable libraries can be
provided in various languages for various applications that
have a consistent and powerful user interface.
d) Interoperable Data Exchange
Data formats

Standard formats

The Khoros data structure or visualization model supports
general geometric objects, multidimensional data, and a robust mapping scheme. Storage type conversion between different architecture platforms is automatically performed by
the read/write utilities, i.e., DEC VAX floating-point data is
automatically converted to IEEE format if read on a SUN
computer. The consistent use of the Khoros data structure
promotes an algorithm library that can be used in many disciplines and supports data sharing.
It is important to state that as Khoros expands, there will
be a need to support a set of "standard" file formats. Currently, Khoros provides for data interchange with other systems via file format converters. Khoros supports the following file formats: TIFF, pbm, BIG, DEM, DLG, ELAS, FITS,
Matlab, sun raster, TGA, and xbm.
e) Data Processing Libraries

Functions

Interface levels

Khoros includes a library of programs that can operate on
point data, one-dimensional data, two-dimensional data and
multi band or vector data. These operators are designed to be
polymorphic, i.e., they function on bit, byte, short, integer,
float and complex data types. This also implies that the
functions will operate differently depending on the dimensionality of the data.
There are two interface levels defined in the library levels
functions in Khoros; the program or process interface and
the function call or procedure interface. The program interface is determined completely by the high-level user interface specification described above. The procedure interface
is currently not as well defined, but allows the procedures

111

Khoros

(functions) to be easily combined into a single program. Visual programs are built by executing a set of programs using
the program interface.
The library contains over 260 programs, in the following
categories: arithmetic, classification, color conversion, data
conversion, file format conversion, feature extraction, frequency filtering, spatial filtering, morphology filtering, geometric manipulation, histogram manipulation, statistics,
signal generation, linear operations, segmentation, spectral
estimation, subregion, and transforms.

Library programs

f) Visualization Toolkit

A visualization tool is of limited utility if it cannot be modified to view and process data in a new user-specified way.
This will only be possible if the user can modify and extend
the software system. A scientist should not be required to
modify a large C program to do this; a high-level language
or specification should be provided as in the Khoros UIDS.
Khoros includes generic interactive X Windows applications for image (2D data) visualization and three-dimensional surface rendering. But more important are the highlevel graphics and display libraries that are accessible from
the UIDS to build custom visualization programs. The following libraries act as a visualization toolkit that is layered
on top of Xlib, Xtk and the Athena Widgets.
Library Name

Description

forms

hierarchical user interface based on
forms and panes
browsers, error reporting, pop-up
lists, and help
2D and 3D drawing library; supports
X, Postscript, HPGL
image display and editing; manages
color allocation and editing

utils
graphics
display

Adapt and extend

High-level tools

112

Current Public Domain Systems in Use

8.1.3 Current Status of Khoros
Many users

Help

Platforms

The Khoros user community is applying Khoros primarily
to image and signal processing research and development
projects. In addition, sites are retargeting Khoros to application domains such as, three-dimensional volume rendering,
relational databases, and telecommunications. Khoros is being used as a teaching tool at several universities.
The documentation for the system is a combination of
on-line help and printed manuals. The manual comprises
2200 pages in three separate volumes: User's Manual, Programmer's Manual, and Reference Guide. Also, journal
playback files are provided to give the new user "live" demonstrations of the various applications.
Khoros currently runs on SUN, DEC, APOLLO/HP,
SGI, IBM, NeXT, and CRAY platforms and there are porting efforts for 386/486 and Apple platforms. Khoros is now
available via anonymous ftp at no charge, or a tape and
printed documentation can be ordered for $250.00. Hundreds of Khoros users participate in a mail user's group,
email khoros-request@chama.eece.unm.edu for more information. The software can be obtained on tape by mail or
by ftp. Email khoros-request for an order form, or mail the
request to the address below. Orders can also be faxed to
505-277-1439.
The ftp address for the software
(129.24.24.10).
Login: anonymous (or ftp)
Password: user~ame@machine
cd /pub/khoros

pprg.eece.unm.edu

For users in Germany, the ftp is ftp.uni-erlangen.de
Login: anonymous
Password: user~ame@machine
cd /cyber/khoros
The documentation can also be printed by usmg the
"prnmanual" program.

Khoros

113

In Figure 8.1, the cantata visual language is being used
for two simple applications: blending two images and then
pseudo-coloring (top), enhancing a magnetic resonance image of a human spine (bottom).
In Figure 8.2 the cantata visual language program is used
to synthesize, filter, and display a one-dimensional signal.

Fig. 8.1
Cantata Visual
Language
C

1\1\~""""'

-_-u.u---'111 ..."
I ~~- II

jr"...

II
II

_ _ _ "'_

CO"

II .. ·
II ..

'V"

II
II

I
.. - I

Fig. 8.2
Filter and Display

Current Public Domain Systems in Use

114

The left xprism2 plots shows the signal before and after filtering and the right xprism2 plot shows the filter response.
The three-dimensional scientific data plotting package
xprism3 can be used to interactively visualize surfaces, contours. and meshes (Fie:ure 8.3).

Fig. 8.3
3D Capabilities

The Khoros system can be used to integrate satellite imagery with ground elevation and map data to produce a
three-dimensional scene of the earth's surface. The viewimage program is used to interactively change the perspective
view of the surface, and the animate application can be used
to create a 'fly-by' sequence (Figure 8.4).

Khoros

115

Fig. 8.4
Integration
Methods

Warp image is an interactive application for registering
and then warping images to produce integrated data sets
(Figure 8.5).

Fig. 8.5
Image Warping

Current Public Domain Systems in Use

116

The cantata visual language is being used to remove shot
noise from an image of the moon. This is done by using a
count loop containing a median filter (Figure 8.6).
(

-...

~ ...... II

.-_----,

I. ~:".- .,,.,,.
'''I'I· '" .•,.. II

'.14~"'_---"''''';:==---.

1'1.

~
~

Fig. 8.6

Noise Removal

Contributed by Dr. John Rasure, University of New Mexico.

For further information contact:
The Khoros Group
Department of Electrical and Computer Engineering
University of New Mexico
Albuquerque
NM 87131
U.S.A.
Fax: 505-277-1439.
Email queries: khoros-request@chama.eece.unm.edu

apE: A Dataflow Toolkit for Scientific Visualization

117

8.2 apE: A Dataflow Toolkit for Scientific
Visualization
In 1984, the Ohio State University competed with institutions across the United States to host a National Supercomputer Center. While its proposal was highly ranked, Ohio
State lost its bid for National Science Foundation funding
to obtain a center. However, the highly motivated group of
computational chemists who spearheaded Ohio State's
chemistry efforts then received help from the state. In 1987,
the legislature appropriated funds for the Ohio Board of Regents' supercomputer initiative to create a center serving academic and industrial users in the state of Ohio. In June
1987, a Cray X-MP was installed at the Ohio Supercomputer Center, followed by a Cray Y-MP in June 1989.
One of the early supporters of the Ohio Supercomputer
Center was Professor Charles Csuri, a pioneer in the field
of computer graphics and Director of the Advanced Computing Center for the Arts and Design at Ohio State. He
foresaw the rise of scientific visualization in the early 1980s
and built a significant graphics research component into the
base of the then-fledgling Ohio Supercomputer Center.
Thus in late 1987, the newly formed Ohio Supercomputer
Graphics Project set out to design an effective software system for visualization, apE.
Rather than dictate to the scientific community a particular methodology, extensive time was spent with potential
users to understand the real needs of scientific research. The
apE group listened to users of all kinds of current graphics
software and hardware, discovered the realities of fixed budgets that permit only modest hardware acquisition and the
effects of slow network connections on high speed computing. In short, they tried to face the real world, and to design
and build a product that would outlast current hardware
platforms while providing a high degree of flexibility to today's users.

Background

Computational
chemistry

Developments in
computer graphics

Tools for
supercomputers
What do potential
users really need
from computers?

Hardware
independence and
flexibility

118

Current Public Domain Systems in Use
Dataflow model

Steps in
understanding
scientific data

Feedback

Utilizing
networked
computers

Local and
centralized
computers

Dataflow language

Efforts in the mid-1980s at the Computer Graphics Research Group (now known as the Advanced Computing
Center for the Arts and Design) at the Ohio State University led to selection of a dataflow model for the apE system.
A dataflow system maps very well to the general steps followed in visualizing scientific data. Most researchers follow
a five-step process, beginning with a computational or experimental simulation, and concluding with interpretation.
Intermediate steps include preparation, mapping, and rendering (the preparation stage is occasionally omitted or
merged with the map stage). Ideally, the results of interpretation can be fed back into the original experiment or simulation. This kind of feedback is known as steering, and has
been used with great success in some limited applications.
New software technology, beyond the reach of apE, is needed to investigate the steering issue completely.
The dataflow abstraction is ideal for remote execution
and parallel operation. Network computing environments
are commonplace, and distributed computation is a requirement for maximum resource utilization. Dataflow systems
can naturally distribute each execution element on a separate machine or processor. The apE dataflow is data-driven
and not demand-driven, which offers the benefit of parallel
execution for time-dependent or multi-frame data sets. Each
element operates not under the control of some central authority but instead only as input data, frame boundaries,
and other local conditions dictate. Successive elements in a
visualization pipeline can be operating on separate groups of
data, all in concert, without any additional user interaction.
This notion of distributed computing maps well to the realities encountered among researchers, as they are often located
at distant sites, far from their supercomputing resource, but
may have some local computing power available.
Once the group was firmly committed to the data flow
concept, they examined the requirements for a data language. Incompatible binary formats are common in a heterogeneous network environment. While transmission of data
as text files would mitigate this problem, the operational

apE: A Dataflow Toolkit for Scientific Visualization

overhead for such transmission was out of the question in
an interactive system. Thus a dataflow language was born,
designed to represent not only common data elements from
the scientific domain (such as grids and variables) but also
common graphical forms, such as objects, images, and geometries.
There is a great difference between writing a small piece
of personal software and constructing a large software environment. Additional complexity appears when consideration is given to machine and device independence and portability. Still more demands come from software which is to
be distributed not as a closed system but as source code, to
be modified, improved, and extended as required. The
group tried to build as portable a graphics environment as
is possible using existing software and hardware technology.
Clearly, the analysis of this requirement results in a different
answer today than it did in 1987 when this project was begun. However, many of the design decisions faced then are
also faced today by large scale developers. These decisions
can be summarized as three primary turning points: the selection of an operating system, the selection of a graphics library, and the selection of a user interface.
The apE system is built on the UNIX platform. The
mid-1980s saw an explosive growth in a new breed of personal computer known as the "workstation". Performance,
power, and software resources that were once only part of
large mainframe systems rapidly became available on the
desktop, and the UNIX operating system quickly became
the de facto standard operating system. Manufacturers that
did not respond to this trend saw their sales diminish.
The group chose not to embed any graphics library in
the basis of the system. They were criticized for not building their software upon a graphics software layer such as
CORE, GKS, PHIGS, PHIGS+, PEX, or others. In late
1987, when this decision was made, the number of competing standards was large, and no clear winner had emerged.
None of the standards available then were really sufficient
for scientific visualization. Constructing the software on

119

Representing data
elements and
graphical forms

Software
engineering
principles
Extensibility
by users
Portability

Design decisions

UNIX platform

De facto standard
Independent of
graphics libraries
Proliferating
standards
Visualization
requirements

120

Current Public Domain Systems in Use

Add-on costs for
graphics libraries

Independent of
window systems

Motif and
Openlook
Developing
the software

User interface
Look and feel

First version

Data language
extensions

such a platform would be a tacit endorsement of one of these standards and would require users to obtain the necessary
licenses to actually program within apE. Most workstation
vendors do not currently ship a PHIGS product, for e~am
pIe, as no-cost, bundled software with their systems, so additional cost is incurred in purchasing, installing, and maintaining a graphics library in addition to apE. All that is
needed to run apE is apE.
apE incorporates a new interface layer on top of existing
"standards". The group chose not to build upon one of the
existing window systems. Clearly today the only "standard"
window system is the X Window System. However, in 1987,
a number of alternative threatened to steal the glory from
X. Despite the claims, the intense battle between such competing higher-level standards as Motif and Openlook will
continue this uncertainty.
With source code control, interface, and (lack of) a
graphics library in hand, the group was ready to implement
the application software. This phase of the development was
divided into three logical elements: the construction of the
libraries, the construction of the individual dataflow elements (or modules), and finally the construction of the
tools and interface that would comprise the look and feel of
apE to the average user. The library implementation was
done in phases, with the UNIX-level hiding functions completed first. The data language, usel"_ interface library, and
graphics functions were done in preliminary test forms prior to full implementation. The resulting test software was
released (as version 1.1) and used to help motivate the full
implementation of apE version 2.0.
Version 1.1 of apE, released in early 1989, had many of
the important aspects of the dataflow design, but lacked the
full implementation needed to make a truly useful software
package. Many changes occurred in the 18 months from the
public release of apE 1.1 to the first glimpses of apE 2.0.
The data language "flow" that had been developed for
apE was enhanced, extended, and rechristened flux. Specifically designed to deal with large amounts of data in user-de-

apE: A Dataflow Toolkit for Scientific Visualization

signed grouping, flux is a powerful information management tool. All data entities, from images to variables to pipeline descriptions to icons, are represented in the flux.
The generic user interface, first presented in apE 1.1, was
expanded and renamed face. The face libraries provide a
complete, window-system-independent interface for program development. Face elements include most of the standard interface items, such as buttons, menus, sliders, scanners, and text entry boxes. On top of this layer more complex elements are provided as well, such as alerts, browsers
(for selecting a text element from a list), and collectors (for
selecting several text elements from a list). Face provides a
generic, application-based interface for interactive tool design which allows a single application to execute under SunView, X, and GL without significant source code changes.
The operational tools provided in the first release were
also significantly reworked. The pipeline construction tool
has been reworked to increase interactivity and to handle
different connection methods between the elements (apE
1.1 used UNIX pipes to connect the dataflow elements; apE
2.0 uses both UNIX pipes and sockets for connections). A
central console provides an outlet for error messages and access to documentation. An interactive image viewer allows
manipulation of single and multiple images and real-time
"playback" of image sequences. A geometry viewer allows
interactive viewing of geometries.
While apE 1.1 was limited to nearly-linear pipelines, apE
2.0 is designed to allow complex pipeline configurations, including multi-input and multi-output and cyclic graphics.
This cyclic capability provides apE 2.0 users with the ability
to investigate connections between graphics and supercomputer simulations, and to attempt to "steer" a simulation
through visual feedback. These additions are all natural extensions of apE 1.1.
Finally, the filters/modules have been extended to include three-dimensional elements as well as the traditional
two-dimensional elements found in the first release of apE.

121

Flux

Advances in the
user interface
Face elements

Interworking with
SunView X, GL
Operational tools
Pipelines

Errors and help
Playback of
sequences
Complex pipelines

Supercomputer
simulations

3 D element filters

Current Public Domain Systems in Use

122
Visualization
techniques
Volume methods

Extensions
Distribution policy

Criteria
for success

Release of code

Modifications
Changes

Understanding

Source code
availability

Visualization techniques include carpet and contour plots,
surface detection, terrain generation, and all forms of rendering from scanline polygonal techniques to ray tracing. A
volumetric rendering system based on methods methods developed by Levoy is also included. Particle tracing, advection, and surface feature detection (such as stream lines) are
also included. In addition, full prototypes are provided to allow extension of the system by the addition of new filters,
data types, tools, and interface elements.
One of the real keys to the success of the apE software
effort has been the distribution policy. While a corporation
must be concerned about profits, competition, market analysis, and other factors, the group concentrated solely on
providing the best tools for the research community, knowing that success would be judged by user productivity, not
the corporate bottom line. The best and only result hoped
for was widespread usage and increased productivity among
Ohio's researchers.
The first version of apE was released in binary form only. For many users this was insufficient, because it prevented
them from fully utilizing the software. First, many people
needed to modify the code to suit particular needs or demands in a particular application or field of interest. Some
needed to make changes to suit local equipment or configurations. Finally, for many, not being able to see the source
code caused a lack of confidence in the final results. Even
if the code is not modified, it is of great value to examine
sections to understand how a particular function is implemented or why an unexpected result is seen. University environments typically enjoy source code for most applications for precisely this reason.
The group can now distribute the second version of the
software in source code form. All of the apE system, including window system layers, program development layers, data format layers, and all existing filters and tools will be released in source code form with the software. Academic institutions can request this software (with manuals) at no
charge, although a license (prohibiting redistribution) must
be signed.

apE: A Dataflow Toolkit for Scientific Visualization

For commercial and non-profit users, the university has
chosen to pursue commercialization of the software with
the TaraVisual Corporation in Columbus, Ohio. This company offers maintenance, installation, and consulting services related to apE. It is not affiliated in any way with the
Ohio State University or the developers of apE and thus the
version of apE offered through TaraVisual is expected to diverge from that offered by Osu. While this may seem
counter to some of the philosophies of apE, it was a decision
made by the university and is not representative of the general feelings of the developers.
The apE system does not represent a breakthrough in
computer graphics. Most of the technology that has been
harnessed to construct apE has been in existence for a number of years, and precious little of it could in any way be
considered to be state of the art. However, the apE system
does represent a significant new step in placing sophisticated
tools into the hands of users. The project has helped to push
industry toward a greater realization of the nature of the scientific visualization problem. The potential of visual methods for data analysis is enormous, and we need to recognize
that the grand challenge that faces us today is not in making
faster silicon but in finding new ways to improve the productivity of our research community.
apE Runs on Convex C-l, C-2, Cray, SGI, SUN, HP,
NeXT, DEC, Stardent, and IBM (RS6000, AIX).
apE supports data manipulation, data mapping, image
rendering, and animation. Data manipulation includes a data flow language (flux), creation and editing of polygonal data, image format conversion, image processing, and image labelling. Data mapping includes isosurface construction of
volumetric data, support for RGB, HSV, HLS colorspace
mapping, color palette editing, data in uniform and nonuniform grids, rectangular, polar, spherical and geocentric
coordinate systems in 2, 3, and n dimensions.
The user interface supports SunView, X Window System,
and Silicon Graphics GL. It has a visual language paradigm,
and fully indexed on-line user documentation. It also allows
distributed processing over a computer network.

123

Commercial users

Aggregation of
existing tools and
methodologies

Tools for real users
Power of visual
methods

Platforms

Summary of
facilities

User interfaces

124

Current Public Domain Systems in Use

Image rendering

Reference
information

Image rendering includes lD frequency plots, 2D contour line images, 2D continuous-color contour images, 2D
colour contours with gradient shading (bump mapping),
carpet plots, photorealistic rendering of polygonal data, volume rendering, particle animation of vector fields, and
"glyph" rendering.
"A Dataflow Toolkit for Visualization" by Scott Dyer in
IEEE Computer Graphics and Applications, Vol. 10, No.4,
1990, pp.60-69, gives further information on apE .

mCb'O'

Fig.B.7
Isosurface
Rendering

•

Isosurface rendering using apE. A pipeline has been created to examine a 3 D volumetric dataset consisting of MRI
data from a human subject. Two isosurface values have been
selected - the outer one has been made transparent.

apE: A Dataflow Toolkit for Scientific Visualization

11111",,,

.1.,

, ....

~.I.I'

125

~I'III ...
(_.-t

"""-

...-'.-1
... _11-1

_,,_I

,.,... "-"-I
.111'."'" I

r=====~:::::::::::=~

Volumetric rendering using apE. A pipeline has been created to examine a 3 D dataset of temperature in the Atlantic
Ocean and has been rendered using a ray-tracing technique.

Fig. B.B
Volume Rendering

126

Current Public Domain Systems in Use

Acknowl-

edgements

Support

This project has been possible only because of the dedication and commitment of the members of the Ohio Supercomputer Graphics Project. The listed authors would like to
thank Barb Dean, Manager of the Ohio Visualization Laboratory, and Michelle Messenger, our Project Coordinator. In
addition, the participation of the Advanced Computing
Center for the Arts and Design, in the form of graphics specialists Steve Spencer and Jeff Light, has added features and
capabilities to apE that would have otherwise been absent.
The authors also thank Prof. Charles Csuri, who marshalled the resources to bring this project into being and
supported it throughout its history, Dr. William McCurdy,
for providing the resources and being the catalyst for many
of apE's scientific concepts, and Dr. Charlie Bender, Director of the Ohio Supercomputer Center, for believing in the
project, and continuing to supporting it, during its most
crucial hours.
This work was supported by the Ohio Board of Regents,
through the Ohio Supercomputer Center, and by the Ohio
State University. This work was supported in part by a grant
from Cray Research, Inc., by an equipment grant from Apple Computer, Inc., and by an equipment loan from Silicon
Graphics, Inc.
Contributed by the Ohio Supercomputer Graphics Project:
Scott Dyer, Project Leader
Steve Anderson
John Berton
Pete Carswell
John Donkin
Jeff Faust
Jill Kempf
Robert Marshall
Since this information was submitted, apE has been taken
over by TaraVisual Corporation, who are now responsible
for its distribution and support. Although it was initially
made available as a public domain product from Ohio Su-

National Center for Supercomputing Applications (NCSA)

127

percomputer Center, this is no longer the case at the time
of writing. However, discounts on the software are available
for academic use. All enquiries on apE should now be directed to the address below, and not to Ohio Supercomputer
Center.
TaraVisual Corporation
929 Harrison Avenue
Columbus
OH 43215
U.S.A.
Tel: (800) 458-8731
Tel: 614-291-2912
Fax: 614-291-2867

8.3 National Center for Supercomputing
Applications (NCSA)
The NCSA Tools for the Macintosh

The National Center for Supercomputing Applications
(NCSA) offers a number of tools that are available in the
public domain. NCSA Distributed DataScope is an interactive data analysis tool that displays 32-bit scientific data values in spreadsheet form or as simple scaled, interpolated, or
polar color raster images. NCSA Image is a color imaging
and analysis application that permits manipulation of twoand three-dimensional image data sets. Distributed capabilities across TCP/IP network connections allow image processing on powerful computers such as a CRAY 1. Specific
data manipulation features in NCSA Image include histogram equalizations, contrast enhancements, and useful utility functions. NCSA Layout is a presentation tool that allows the and user to display and annotate two-dimensional
data annotation images so that users can photograph their
Macintosh screen display with a 35 mm camera and produce
presentation-quality slides. NCSA Telnet provides a link be-

Interactive data
analysis

Color imaging
and analysis

Layout and
annotation

128

Current Public Domain Systems in Use

File transfer

tween the Macintosh and the TCP/IP networks. It includes
a standard file transfer server (FTP), which allows file sharing with other machines.
All these tools are available free via the Internet. The
software and documentation are also available for purchase
through the NCSA Technical Resources Catalog.
Contact:
NCSA Documentation Orders
152 Computing Applications Building
605 East Springfield Avenue
Champaign, IL 61820, U.S.A.
Tel: 217-244-0072.
(This information is supplied courtesy of visualization
Technology: an Introduction", by Anne Kaplan-Neher, in
Syllabus, Summer 1991, Number 17, P.O. Box 2716, Sunnyvale, CA 94087- 0716, AppleLink: SYLLABUS, Internet:
SYLLABUS@APPLELINK.APPLE.COM; Phone and
Fax: 408-773-0670. It was initially obtained from the "Articles database of CCNEWS, the Electronic Forum for Campus Computing Newsletter Editors, a BITNET-based service of EDUCOM").

8.4 GPLOT, DRAWCGM, P3D

(Pittsburgh Supercomputer Center)
GPLOT can interpret CGM metafiles and can run animation hardware. DRAWCGM produces rasters from 2D arrays of integers or reals. P3D creates and views 3 D models.
Such models can be viewed on SUNs and SGIs. It can build
isosurface models and molecular models, and create animations directly on video tape. It is available via anonymous
FTP from calpe.psc.edu. Further information is given below.

GPLOT, DRAWCGM, P3D

129

8.4.1 The GPLOT CGM Interpreter
The Pittsburgh Supercomputing Center began in 1986 as
one of several sites charged by the U. S. National Science
Foundation with providing supercomputer (and other) capabilities to NSF researchers. Unlike other such sites, the user base (over 2000 researchers) is very diverse, both geographically and technologically, with almost all users physically remote.
Providing a graphical capability to these users was problematic; but being a new center provided the option to design systems from scratch. It was decided to standardise
completely on the Computer Graphics Metafile (CGM) format for two-dimensional image storage. Accordingly only
graphics packages that could produce CGM files were purchased. Each of these packages came with a CGM translator,
but it was found that these translators could only reliably
translate CGM files from the corresponding graphics package!
It was unacceptable to distinguish between types of
CGM files and it was decided to write a CGM translator:
GPLOT. In addition to homogenizing the CGM file population there were several other advantages to using this software. Firstly, GPLOT could be freely distributed to remote
sites and users were encouraged to produce their CGM files
at the PSC and ship them home for viewing with a local copy of GPLOT. This improved response for the users dramatically compared to viewing the CGM files across their network connection and also reduced the load on the network
connections. Secondly it allowed the rapid addition of new
output devices and facilities.
GPLOT was more successful than expected; there are
now over 275 sites on the list of users, including most of the
major universities and research laboratories in the USA and
several sites in other countries. GPLOT now supports output to many different output devices with three different user interfaces.

Supercomputer
center

Remote access
Graphics
standards

CGM

Which CGM?

GPLOT
Distributed
software

Reduced
network load

275 sites

130

Current Public Domain Systems in Use
Video

Frame checking

Animation
sequences

Cost savings

Random access

Graphical user
interface
Object oriented

Port to
Apple Macintosh

GPLOT was also used to greatly facilitate the creation of
videos by remote users. Users are encouraged to create CGM
files with many (possibly very many) frames. They can examine individual frames at their home site using their copy
of GPLOT and when satisfied with their "look" submit
them for animation. A local copy of GPLOT at the PSC is
then used to create a full animation either on a Sony U matic
recorder or on a Sony laser disk recorder. It can then be
dubbed to VHS tape and mailed to the user, with a turnaround time of a few days. In fact this is the only output
that PSC mails to users. In this manner several minutes of
animation can be produced, possibly spanning several individual animations, a night. Presently there are three parallel
animation systems, two U matics and one laser disk recorder.
In addition to purely remote use, three of the heaviest animation users decided to purchase their own hardware, and
used GPLOT software to produce their own animation systems. Since all necessary software was provided, the cost to
these users was much less than a commercial system.
The size of the CGM files required for animations (frequently close to a gigabyte) motivated extension of the
CGM standard to include a random access capability. This
was done in cooperation with the SLATEC supercomputing community and it has worked very well to date.
The original GPLOT system was written entirely in the
C programming language, compilable under either the
UNIX or VMS operating systems. As the move began towards sophisticated graphical user interfaces and users began
to request more capabilities for GPLOT (including onscreen animations for workstations) it was decided to completely rewrite GPLOT in an object-oriented fashion using
the C+ + language. This greatly simplified the addition of
some features, including on-screen animations using the XWindow system and a Motif user interface.
This C+ + version also simplified the port to the Apple
Macintosh operating system. The great majority of the code
is common with two Macintosh specific modules, one for

GPLOT, DRAWCGM, P3D

the user interface and one for the Quickdraw imaging system.
The object-oriented design was also intended to allow
the easy integration of all of GPLOT's capabilities into other packages. Work is underway to perform this in combination with a documentation system to allow true text-graphics integration in a single system.

131

Access to
other packages

8.4.2 The DrawCGM Graphics
Subroutine Library
During the development of GPLOT, it became necessary to
generate CGM files for test purposes. A simple library called
CGMGen was written to do this. It had the ability to produce indexed or direct color CGM files, the interface being
set up so that a single call generally produced a single CGM
element.
At the same time it was becoming obvious that graphics
packages available then, like Disspla, DI-3000, and the
NCAR Graphics Library, lacked features needed to do some
of the graphics required in a supercomputing environment.
In particular, it is very common for a supercomputer user
to wish to produce a color image from a two-dimensional
regular array of data. This is done simply by mapping the
data into a range of integer values, and drawing the image
using those values as indices into a color map. It was required to provide this functionality to the users in a simple
way.
This led to the development of the DrawCGM graphics
library, which is particularly well suited to the generation of
raster images. The package provides facilities for manipulating color maps, scaling and quantizing rasters of reals, and
drawing multiple images within a single CGM frame. The
ability to produce other CGM primitives was added, such
as markers, lines, polygons, and text as well, because these
functions were readily available in CGMGen. Utilities to
easily draw color bars and labels were also included. Because
the primary task of DrawCGM was to handle color mapped

Test facilities

Graphics from
supercomputers

Raster images

Current Public Domain Systems in Use

132

Interface

GPLOT

Applications

images, only the indexed color facilities in CGMGen
were used. DrawCGM is written in FORTRAN, while
CGMGen is written in C.
Eventually confidence in the interpretation of the CGM
standard increased so that support for the independent
CGM generator in CGMGen could be dropped. This CGM
generator was replaced with a direct interface to the GPLOT
device driver library, making it possible to use the
CGMGen interface to do graphics interactively on any device supported by GPLOT. Since GPLOT also supports device drivers which create binary or clear CGM, the ability
to produce metafiles from DrawCGM was not lost. Thus
DrawCGM became an interactive package, supporting a
wide range of devices.
DrawCGM and CGMGen are distributed with the
GPLOT library, and are now quite widely used. DrawCGM
has been particularly successful for producing animations,
for example of hydrodynamic systems. Users will preview
animations either interactively or via a CGM metafile and
GPLOT, and will pass a CGM file containing the entire animation for recording when their results are satisfactory.
DrawCGM provides the ability to have multiple images
with distinct color maps on screen simultaneously, which
can greatly improve the information content of this sort of
animation. The underlying CGMGen layer still supports
the ability to use direct color, so it has been used to interface
a number of 24-bit color applications to the GPLOT device
driver library.
One feature which DrawCGM does not currently support is the ability to have multiple output devices open simultaneously. It is hoped to correct this when CGMGen is
recoded to interface to the new object-oriented version of
GPLOT.

GPLOT, DRAWCGM, P3D

133

8.4.3 The P3D Three-Dimensional
Metafile Project
Success with a metafile-based environment for 2D graphics
led to the consideration of a similar system for three-dimensional models. The goal would be to produce a format
which all programs generating 3 D models at the PSC would
produce, and which could be rendered in three dimensions
on a wide variety of platforms. The existence of this format
would allow models to be transferred between the central site and user sites, and would simplify the software support
situation analogous to the simplification provided by CGM.
This is not to belittle the importance of interactive 3 D
graphics; it is simply believed that a complete environment
requires both interactive and metafile forms and that it is appropriate to investigate the metafile approach.
Unfortunately, unlike the 2D case in which the standard
CGM format already existed, there was and is no standard
format for 3 D scientific graphical models. The available
non-standard formats were examined and none of them
were found very satisfactory for current needs. Therefore a
further format was developed, called P3D - the P denoting
'Programmable'.
P3D is based on a subset of the Common Lisp language,
with a small number of extensions to describe geometry.
This makes P3D a complete programming language in itself,
allowing it the same flexibility which programmability provides to the Postscript page description language. Like Postscript, it is never necessary for a user to actually write a program in P3D. A model generating program (for example a
molecular modeler) produces a P3D model, and a P3D
viewer translates the model into images which the user can
view. The P3D viewer includes a locally written Lisp interpreter; there is no need for the site using P3D to license an
interpreter from a third party. Because the full 3 D structure
of the model is stored in the P3D file, the model can be
viewed from any direction or incorporated into more complex models.

2D to 3D

Format for 3D

3D language

134

Current Public Domain Systems in Use

Needs of scientific
visualization

Rendering options

Animation

Exchange of 3D
mode/ data

P3D was designed to support the needs of scientific visualization, rather than those of photorealism or, for example,
computer aided design. The current implementation supports ten geometrical primitives, fairly complete lighting
and camera information, arbitrary transformations, and a
very extensible attribute structure in a hierarchical model
environment.
P3D models can be viewed on seven different renderers,
with more under development. These range from a simple
mouse-driven renderer for the X Window System environment to renderers for solid modeling workstations and a ray
tracer. A number of generators for P3D models now exist,
including translators for molecular dynamics output,
marching cubes algorithms, a general purpose subroutine library, and a simple tool for generating fly-bys of P3D models. The programmability of P3D makes it easy to modify
existing codes to produce models in P3D format, since the
metafile can essentially be tailored to the needs of the code
rather than the other way round.
As with CGM, it is possible to produce animation from
a P3D model file in which a number of views are specified.
This allows animations to be previewed on a user's workstation, and then nicely rendered (possibly ray-traced) and recorded as high quality video animation. The animation system uses the same hardware and some of the same software
as the CGM-based system.
It would be useful for the P3D format to become a common medium of exchange for 3D models. There are currently about 40 sites on the mailing list of those using or interested in P3D, so some progress is being made toward this
goal. Current work is on designing interfaces which will allow general interactive visualization packages like Stardent's
AVS and the Ohio Supercomputer Center's apE to read and
write P3D models, and to incorporate additional renderer
interfaces such as Pixar's Renderman. Other development
projects include a translator to generate P3D from finite element models, and general modifications to improve the
functioning of the P3D renderers.

135

RAYSHADE

8.4.4 Software Availability
Software developed at the Pittsburgh Supercomputing Center, including GPLOT, DrawCGM, and the P3D software
suite, is available free of charge by anonymous FTP from
the machine ftp.psc.edu. DrawCGM is included in the
GPLOT package; the P3D software is a separate distribution
but (depending on the configuration chosen) may require
the GPLOT software. If you take GPLOT, please send mail
to Anjana Kar (kar@psc.edu) to be added to the appropriate
mailing list. Advanced questions regarding GPLOT, as opposed to simply taking and installing the software, can be
directed to Phil Andrews (andrews@psc.edu). If you take
P3D, please send mail to Joel Welling (welling@psc.edu) to
be added to that mailing list.

Free of charge

Information

Contributed by Joel Welling and Phil Andrews.

Further information from:
Dr. Joel Welling
Pittsburgh Supercomputer Center
4400 Fifth Avenue
Pittsburgh, PA 15213, U. S. A.
Tel: 412-268-6352
Email: welling@psc.edu

8.5 RAYS HADE
This is an excellent ray-tracing program for scene rendering.
It handles many different kinds of object. It is available from
the University of Yale and the University of Utah.
Rayshade reads a multi-line ASCII file describing a scene
to be rendered and produces a Utah Raster RLE format file
of the raytraced image.

Scene rendering

Format

Current Public Domain Systems in Use

136

Facilities

Features include:
• Primitives:
boxes
cones
cylinders
height fields
planes
polygons
spheres
triangles (flat- or Phong-shaded),
• Composite objects,
• Point, directional, and extended (area) light sources,
• Solid texturing and bump mapping of primitives, objects, and individual instances of objects,
• Antialiasing through adaptive supersampling or "jittered" sampling,
• Arbitrary linear transformations of primitives, instances
of objects, and texture/bump maps,
• Use of uniform spatial subdivision and/or hierarchy of
bounding volumes to speed rendering,
• Options to facilitate rendering of stereo pairs,
• Support for the Linda parallel programming language.

C language

Platforms

Getting

a copy

An awk script is provided to translate NFF format scripts
to rayshade format.
Rayshade is written in C with parsing support provided
through lex and yacc. The C, lex and yacc files comprise approximately 8000 lines of code. Sites without lex and yacc
can make use of the C source files produced by lex and yacc
which are included in this distribution.
Rayshade has been tested on a number of UNIX-based
machines, including Vaxes, Sun Workstations, Iris 4D Workstations, Encore Multimax, AT&T 3B2/310, CRAY XMP,
and IBM RTs. In addition, support is provided for the Amiga using the Aztec C compiler.
Rayshade makes use of the Utah Raster toolkit, a package consisting of a large number of useful image manipulation programs, test images, and a library to read and write

NASA Ames Software

137

images written using the toolkit's RLE format. The toolkit
is available via anonymous FTP from cs.utah.edu or from
weedeater.math.yale.edu.
Those sites that cannot or do not want to use the Utah
Raster toolkit can make use of a compile-time option to
produce images written using a generic file format identical
to that used in Mark Van de Wettering's "MTV" raytracer.
Rayshade is copyrighted in a "Gnu-like" manner.
Rayshade is available via anonymous ftp from weedeater.math.yale.edu (192.26.88.42) in pub/Rayshade.2.21.tar.Z.
The Utah Raster toolkit is available in pub/UtahToolkit.tar.Z.

8.6 NASA Ames Software
8.6.1 PLOT3D
PLOT3D is a computer graphics program designed to visualize the grids and solutions of computational fluid dynamics.
Eighty-five functions are available, and versions are available
for many systems. PLOT3D can handle multiple grids with
many grid points, and can produce varieties of model renderings, such as wireframe or flat shaded. Output from
PLOT3D can be used in animation programs.
PLOT3D User's Manual and PLOT3D software can be
distributed free of charge and without copyright to any institution or business in the USA.

Computational
fluid dynamics

Availability

Contact:
Workstation Applications Office
NASA Ames Research Center, MS258-2
Moffett Field, CA 94035, U. S. A.

8.6.2 SURF
A further program, SURF, allows the user to input
PLOT3D grid and solution files and interactively create wireframe, shaded, and function mapped parts to view, and

Interactive viewing
of PLOT3D files

138

Current Public Domain Systems in Use

Further facilities

then output to ARCGraph files which can be animated using the Graphics Animation System (GAS). Shaded parts
are created based on user specified lightsources (at least 20),
viewpoint, and the ambient light level. The function mapped parts can have their color spectrum adjusted interactively. Legends can be created to show the correlation of color
and normalised function values (i.e., pressure, density, temperature, and mach number). Also, function-mapped parts
can be clipped so that they only show contours within a
specified range of function values (e.g., normalised pressure
between 1 and 2).
Other features of SURF include the ability to work with
several grids and solutions, grid/solution deletion, support
of multi-grid files, input/output for colormaps, matrices,
light sources, function extrema, screen dump pixel input/output, display of current grid and part attribute data,
and a UNIX shell escape.

8.6.3 Graphics Animation System (GAS)
Animation system

GAS is a graphics animation software system that is menudriven and provides fast, simple viewing capability as well
as more complex rendering and animation features. It is
used to display two- and three-dimensional objects along
with computed data, and also to record animation sequences on video digital disks, videotape, and 16 mm film.

8.6.4 Applications in Computational
Fluid Dynamics (CFD)
Illustrations of use

Getting

a solution

Some example applications have been the following: pressure distribution inside the space shuttle main engine, vortex flows over the wing/strake surface of F-16 aircraft, pressure distribution on an oscillating F-5 wing, simulation of
turbine engine rotor-stator interaction, particle traces over
the space shuttle orbiter, and pressure distributions over the
high-speed National Aerospace Plane.
The CFD software analysis cycle begins with the design
of a test geometry 'grid' (e.g., forward-swept wing model

NASA Ames Software

with surrounding airspace), specification of simulation conditions (e.g., angle of attack, mach number, reynolds number) and coding and execution on supercomputers of 'flow
solver' programs that solve the higher-order mathematical
equations governing the flight characteristics. Then the numerical solution data is collected and converted to graphics
images of fluid flows, pressure distributions, shock waves,
and particle traces using workstations running PLOT3D
and other specialised graphics programs. Then SURF can be
used to add shadinglcoloring enhancements to the images.
Finally, animation sequences are generated and recorded
with GAS.
The results produced are animated 16 mm films and videotapes showing solid, pressure mapped aircraft models
with wing/body vortices, particle traces, temperature distributions and shock waves. The ability to display the physical
properties in aerodynamic flight is a tremendous aid in understanding and designing aircraft geometries for specific
flight characteristics. By using interactive computer graphics
in the aerodynamics study, the critical areas (e.g., high turbulence, high temperature, reverse flow) immediately become obvious so they can be studied more closely using a
finer grid, more particle traces, and higher-resolution mapping of pressure or temperature contours. For example,
graphical studies revealed high turbulence and pressure inside the Space Shuttle Orbiter main engine hot gas manifold, and further analysis led to a redesign of the engine
with reduced decisions internal pressure and turbulence.
(Courtesy of PLOT3D User's Manual by P. P. Walatka and
P. G. Buning, SURF User's Guide by T. Plessel, and GAS
User's Manual by T. Plessel).
Contact:
NASA Ames Research Center
Mail Stop 258-2
Moffett Field, CA 94035, U. S. A.
Tel: 415-694-4052

139

Use of computer
graphics

Animation

Visualization aids
understanding

Studying physical
properties

Design decisions

140

Current Public Domain Systems in Use

8.7 IRISPLOT
Display of surfaces

Viewing the
surface

IRISPLOT is an extended version of GNUpiot for the Silicon Graphics workstation that allows algebraic surfaces to
be displayed, shaded, etc.
IRISPLOT allows the user to define some of the graphical objects built from surfaces and curves, which in turn are
defined from mathematical functions, discrete maps, differential equations and data files. The user has full control of
the graphical attributes, which includes viewing, orthogonal
or perspective projection, object transformation, object slicing, 8 different light sources with different color and location, and different material properties for each object in
plot. It also allows contouring on the surfaces.
For more information, contact system@math.arizona.edu

8.8 ISVAS
Interactive
visualization

Finite element
data

FhG-AGD in Darmstadt provides a tool called ISVAS, an interactive volume visualizer for SUN (Xll/0SF Motif) and
SGI (GL/Motif). ISVAS stands for Interactive Software for
Visual AnalysiS of fracture mechanics. The system has been
developed for the visualization of the results of three-dimensional finite element simulations in fracture mechanics and
other application areas.
The main aim in the development of ISVAS is interact ivity. Finite element analysis produces large amounts of data
and the graphical presentation of the data is a computer-intensive process. Therefore there is a need for presenting data
at speed, but with low amounts of detail. The tools allow
a rough preview of the data and interactive specification of
parameters, such as viewpoint and cut planes. The user can
thus produce a quick picture of what the data looks like,
specify the parameters of the required image, then render it
for a higher-quality presentation.

ISVAS

The Xll version, which is running on SUN, DEC, and
other machines, has already been installed at the Technical
University of Munich, the University of Lisbon, and the
University of Mexico. Sun (SPARe) code is available, and also source code of the data filters for adaption to the user's
own volume data. Data filters for several FE-data types are
provided. Version 1.2 of ISVAS allows the user to visualize
scalar volume data. The next version will also allow vector
data to be displayed.
ISVAS is built upon X Windows and OSF/Motif toolkit
to be portable to any UNIX colour workstation.
Further information:
Dr. Martin Goebel
Fraunhofer-Arbeitsgruppe
fur Graphische Datenverarbeitung
Wilhelminenstrasse 7
W-6100 Darmstadt
Federal Republic of Germany
Tel: 49-6151-155123
Fax: 49-6151-155199
Email: goebel@de.fhg.agd

141

Portability

Chapter 9
Other Uses of Visualization Tools

9.1 Art and Design
Creative and
artistic
applications

Artists and
designers

Tools such as those described in this guide can be also used
by those whose primary interest is not in the scientific content of the information presented, but rather the creative or
aesthetic value.
Barlow et al. (1990) outlines how artists create effects and
explores issues at the interface between art and science.
Artists and sculptors have been using computer-assisted
tools for a number of years (Lansdown, 1989) and these
tools often promote new and unexpected ways of creating
and developing images and objects. Architects, designers and
engineers also use these methods. Thus visualization tools
are not confined to scientific visualization but can be used
in all areas where the user is seeking to create and manipulate information via visual means.

9.2 The 5th Dimension Animation System
3D animation and
visualization

Applications

The 5th Dimension Project is a large research project in
three-dimensional animation and visualization. The main
objective of the project is the animation of synthetic actors
in their environment, which involves a number of related
areas of computer animation and scientific visualization. In
particular, the following applications are being developed:
• animation of articulated bodies based on mechanical
laws,
• vision-based behavioural animation,

The 5th Dimension Animation System

•
•
•
•
•
•
•

143

hair rendering and animation,
object grasping,
facial animation,
personification in walking models,
synchronization in task-level animation,
deformation of flexible and elastic objects,
cloth animation with detection of collision.

To coordinate efforts and allow good communication between the various applications, a toolkit of high-level dynamic classes, both two- and three-dimensional, has been
constructed. This toolkit, called the 5th Dimension Toolkit,
uses a uniformly object-oriented design for all its data structures, resulting in a high degree of integration between various applications.
The 5th Dimension animation system is intended to offer to the animator a full 3 D interaction including the possibility of entering into the virtual world and communicating
with the synthetic actors. The hardware used consists of 21
Silicon Graphics IRIS Workstations including three Powervision (VGX) models. Most 5th Dimension applications
take advantage of visual 3 D interfaces using the various 3 D
devices available in the laboratories: two datagloves, several
SpaceBalls, an EyePhone, a 3-D Polhemus digitizer, a Live
Video Digitizer, a StereoView station, and a synthesizer keyboard controlled by a NeXT Cube workstation.
In the current version, six applications provide a user interface based on 3 D devices:
• the sculpting program SURFMAN,
• the Muscle and Expression editor in the SMILE Facial
Animation system,
• the cloth design software,
• the hand gesture recording system GESTURE LAB,
• the program to create 3 D paths for cameras, objects and
light sources,
• a communication program animator-actor (in development).

High-level toolkit

Object oriented

3D interaction
Virtual worlds

3D devices

144

Other Uses of Visualization Tools
3D input

The first three programs are mainly based on the ball and
mouse metaphor. SURFMAN may also take advantage of
StereoView and the 3 D Polhemus digitizer. Hand gestures
are recorded using the DataGlove and 3 D paths are mainly
generated using the SpaceBall. We are developing a way of
creating camera paths based on the EyePhone. The communication program animator-actor uses the Living Video Digitizer to capture the animator face.
Other applications in the 5th Dimension system are only based on mouse interaction. They include:
• an interactive system to design individual walking,
• the BODY-MOVING human keyframe animation system,
• a hair modelling and rendering program.
Submitted by Nadia Magnenat Thalmann, University of
Geneva, and Daniel Thalmann, Swiss Federal Institute of
Technology.

Fig. 9.1
Cloth Animation

Figure 9.1 shows cloth animation from the film Flashback,
by B. Lafleur, N.M. and D. Thalmann, University of Geneva and Swiss Federal Institute of Technology.

The 5th Dimension Animation System

145

Fig. 9.2
Human Walking

Frame from the film Still Walking by A. Paouri, R. Boulie, N. M. and D. Thalmann, University of Geneva and
Swiss Federal Institute of Technology.

Fig. 9.3
Hair Rendering

Hair rendering by A. leBlanc, A. Paouri, N.M. and D.
Thalmann, University of Geneva and Swiss Federal Institute
of Technology.

146

Other Uses of Visualization Tools

9.3 Multimedia Environments
Project KICK
Multi-media

Hierarchy

Video and image

Training

KICK is an interactive multimedia environment designed to
support industrial training applications (Serra et al. 1991).
The main component of KICK is an authoring environment for the designers to organize the contents and orchestrate the presentation of multimedia information. The medium types supported include text, image, 3 D graphics, and
video.
The effective organization of multimedia information
for interactive access is of major concern. In KICK, multimedia information is organized primarily using the natural
hierarchy of the physical objects to be modeled. In order to
permit associative access to information as advocated in hypermedia, auxiliary access paths are provided by means of
other media, such as text, image, and video. The resulting
information structure thus facilitates both structured and
associative access to information.
In KICK, the same direct manipulation interface is used
to manipulate information of any medium type. To achieve
this for video and image media, techniques have been developed to allow the synchronization of 3 D graphics models
and animation of video sequences and images. In addition,
techniques to model object motions, constraints and relationships have also been developed.
The system developed is intended for training applications, where users can learn about the structure and operations of a complex mechanism by interactively:
- accessing its component hierarchy,
experimenting with how various components interact
with each other in dynamic operating conditions,
- studying the effects of various externally applied forces
on different parts of the mechanism.

Examples

To demonstrate the usability of system KICK, two applications for industrial training have been developed. One is ba-

147

Multimedia Environments

sed on an aircraft and the other on a 1/8 scale model car
with a 0.21 cc engine. Figure 9.4 shows an engine component of the aircraft application in four different media - 3 D
graphics, text, video and image. Figure 9.5 shows the engine
component of the model car application. Through the display, the user may interact with any medium type to retrieve further information or to view the operations of the
engine in context.
KICK is developed on a Silicon 4D/210 GTX workstation with a Live Video Digitizer. The video input is obtained from a Sony LDP-1S00 laser disk player. KICK is implemented using Starship, a frame language developed at ISS
(Loo 1991).

Fig. 9.4
Turbine Engine
from Aircraft
Application

The figure shows the concept of a turbine engine in four different media - 3 D graphics (top right), text (top left), video
(bottom right) and image (bottom left). Users may interact
directly with any medium type to retrieve further information about its subcomponents, view the video about the
assembly of the engine, or generate an animation of a 3 D
model.

Other Uses of Visualization Tools

148

Fig. 9.5
Concept Engine
from Car
Application

This figure shows the interface of KICK without the video window. The users may again interact with any medium
type directly to obtain further information.

References
Loo J.p.L. (1991) The Starship Manual (Version 2.0), ISS Internal Technical Report, TR91-54-1
Serra, L., Chua, T. S. and Teh, W. S. (1991) A Model for Integrating Multimedia Information around 3 D Graphics
Hierarchies. The Visual Computer (in press)
Information supplied by Luis Serra and Wei-Shoong Teh,
Institute of Systems Science, National University of Singapore, and Tat-Seng Chua, Department of Information Systems and Computer Science, National University of Singapore.

Chapter 10
Conclusions

10.1 Strategic Importance
of Scientific Visualization
In view of the believed strategic importance of scientific visualization it is timely to consider what tools and techniques should be provided in this area for the community,
and also what kind of initiatives and objectives should be
supported and promoted.
It is important that scientific visualization be developed
and promoted. Here are some of the strategic issues that
need addressing if scientific visualization tools are to be effectively utilised.
Forum for:
• Planning, discussion, and problem-solving,
• Coordinate developments that may be required in the
area,
• Exchange experiences in different application areas,
• Share common software tools, where available,
• Disseminate information,
• Assist with teaching materials,
• Education and training issues.
Plan for:
• Any new products that may be needed,
• Any general network developments that may be needed,
• Any general video facilities that may be needed,
• Providing visualization tools for the community,
• Supporting research into scientific visualization.

Tools and
techniques
needed

Promotion

Conclusions

150

10.2 Current Developments
More detailed
modeling

Interactive
3D design

Handling mUltidimensional data

New techniques
for analysis

These advances will allow mathematical models and simulations to become increasingly complex and detailed. This results in a closer approximation to reality, thus enhancing the
possibility of acquiring new knowledge and understanding.
Scientific visualization is concerned with methods of understanding large collections of numerical values containing a
great deal of information. The scientist has to be able to
make effective use of this information for analytic purposes.
A further aspect is that increases in computer performance allow 3D problems in simulation and design to be
done interactively. In addition, processes that formerly separated out simulation and design can now bring them together (e.g., in CAD, or in the design of new drugs). This in turn
moves the user into a new era of methods of design.
Control over fine simulations, interactivity, and computer performance mean that vast amounts of multidimensional data can be generated. Superworkstations allow this data
to be displayed in optimum ways. These features and capabilities are driving the current wave of interest in scientific
visualization.
Work is proceeding in evolving new techniques for data
display as the data is being analysed.

10.3 More User-Friendly Facilities
What you drive is
what you see
(WYDIWYS)

A further current trend is to make software tools for visualization more user-friendly and accessible to a wide variety
of application areas, thus increasing their potential and usability.
Improvements to the graphical user interface and the
way the user interacts with the model are likely to provide
more effective ways of communicating relationships and
other aspects of data to the user.

What to do next?

151

10.4 Further Information
The book by Nielson et aI. (1990) contains a wide variety
of current applications of scientific visualization and also an
excellent bibliography of scientific papers. A 2-hour video
tape is available with the book. This tape gives effective
demonstrations of the projects described in the book. The
format is NTSC but can be played on a dual-format player
(such as Panasonic J35) in the UK without any problem.
This player can take both PAL and NTSC formats.
Frenkel (1988) provides a general introduction to basic
visualization techniques.
Thalmann (1990) contains a number of papers in the areas of scientific visualization and graphical simulation.
For further details of any of the software mentioned earlier, please contact the local office of the vendor. In addition, Chapter 11 contains a list of references which may be
of further interest.
There are numerous electronic mail subscription lists
and bulletin boards on topics in scientific visualization and
also on specific items of software.

Applications

Introduction

10.5 What to do next?
The majority of users have some experience of tools for presentation graphics. Some users are currently using visualization tools of one kind or another, in that the use of such
tools facilitates the process of understanding more about the
data. However, very few users have access to visualization
systems of the kind described in this guide. It is expected
that access to such systems will become more common as
costs decrease and experience in application areas increases.
The developments in the U. S. A. outlined in Chapter 6
indicate the trends as of 1991, of course!
A more detailed reference volume on scientific visualization is available entitled Scientific Visualization - Techniques
and Applications edited by Brodlie et aI., Springer-Verlag,
1991.

Understanding
data

References

H. Barlow, C. Blakemore, M. Weston-Smith (eds.): Images
and Understanding: Thoughts about Images, Ideas and
Understanding. Cambridge University Press, 1990
Outlines how artists create effects and explores issues at the interface between art and science.

K. W Brodlie, L. A. Carpenter, R. A. Earnshaw, J. R. Gallop,
R.J. Hubbold, A.M. Mumford, c.D. Osland, P. Quarendon (eds.): Scientific Visualization - Techniques and Applications. Springer-Verlag, 1991
This volume represents a full consideration of the subject of scientific visualization and is intended to be a reference guide for the
community on the technical aspects of the subject. The topics
covered include: the framework, visualization techniques, data facilities, human computer interface, applications, products, a glossary of terms, enabling technologies, and an extensive bibliography.

D. Scott Dyer: A Dataflow Toolkit for Visualization. IEEE
Computer Graphics and Applications, 10:4, 60-69 Ouly
1990)
This article describes the design principles behind the apE visualization software (apE stands for Animation Production Environment).

E.J. Farrell (ed.): Visual Interpretation of Complex Data.
IBM Journal of Research and Development, 35:1/2 OanMarch 1991)
A special issue of the IBM Journal of Research and Development
on visualization. There are papers on visualization of volumetric
data, image display and interpretation, and animation for data interpretation. There is also a companion video "Understanding
Complex Data with Computer Animation".

References

K.A. Frenkel: The Art and Science of Visualizing Data.
Communications of the ACM, 31:2, 110-121 (1988)
An introductory paper which looks at a range of application areas
and the uses of visualization tools and techniques. A wide range
of pictures illustrate some of the techniques currently being used.

K.A. Frenkel: Volume Rendering. Communications of the
ACM, 32:4, 426-435 (1989)
Outline of volume visualization, with application areas.

A. Kaufman (ed.): Volume Visualisation. IEEE Press, 1990
A collection of most of the key papers in the area of volume visualization.

R.J. Lansdown, R.A. Earnshaw (eds.): Computers in Art,
Design and Animation. Springer-Verlag, 1989
Collection of papers on the area of creative uses of computer
graphics and associated tools and techniques.

B.H. McCormick, T.A. DeFanti, M.D. Brown (eds.): Visualization in Scientific Computing. ACM SIGGRAPH
Computer Graphics, 21:6 (November 1987)
The original Panel Report which outlines the political, economic,
educational, and technological aspects of scientific visualization as
an emerging discipline.

B.H. McCormick, T.A. DeFanti, M.D. Brown (eds.): Visualization in Scientific Computing. IEEE Computer, 23:8
(August 1989)
An updated version of the original McCormick (1987) report,
outlining current progress and advances in scientific visualization.

G.M. Nielson, B. Shriver, L. Rosenblum (eds.): Visualization in Scientific Computing. IEEE Press, 1990
A collection of papers in the areas of techniques and applications
of scientific visualization from a variety of academic, government,
and industrial organisations in the USA.

153

References

154

N. M. Patrikalakis (ed.): Scientific Visualization of Physical
Phenomena. Springer-Verlag, 1991
Proceedings of the 9th International Conference of the Computer
Graphics Society on the theme of scientific visualization. This
volume contains a number of key invited papers in the areas of
applications of scientific visualization, including engineering design, spacecraft exploration of the solar system, and remote subsea
exploration.

D. Thalmann (ed.): Scientific Visualisation and Graphics
Simulation. Wiley, 1990
Computational and graphical techniques that are necessary to visualize scientific experiments are surveyed in this volume, with a
number of case studies in particular application areas.

N. Magnenat-Thalmann, D. Thalmann (eds.): New Trends
in Animation and Visualization. Wiley, 1991
A collection of papers covering state-of-the-art topics in the areas
of scientific visualization, animation, graphical simulation, modeling, hypermedia, facial animation, natural phenomena, human
modeling, and applications.

E. R. Tufte: The Visual Display of Quantitative Information. Graphics Press, USA, 1983
E. R. Tufte: Envisioning Information. Graphics Press, USA,
1990
Two introductory texts with guidelines on displaying information
effectively.

C. Upson, T. Faulhaber, D. Kamins, D. Laidlaw, D. Schlegel, J. Vroom, R. Gurwitz, A. van Dam: The Application
Visualization System: A Computational Environment
for Scientific Visualization. IEEE Computer Graphics
and Applications, 9:4, 30-42 (1989)
This paper describes the design principles of AVS.

C. Upson: Volumetric Visualization Techniques. In: D.E
Rogers, R.A. Earnshaw (eds): State of the Art in Computer Graphics - Visualization and Modeling. SpringerVerlag, 1991
Description of volumetric techniques from one of the designers
of AVS and the designer of Explorer.

Sources of Figures

Rendered by DESIGNBASE

Figure 7.27

Reproduced by permission
of Dynamics Graphics Ltd.

Figure 7.20,
-7.24

Figure 2.7, 3.3

Figure 7.25, 7.26

Figure 3.10, 7.6,
7.7

Figure 3.5

Reproduced by permission
Ricoh Company Ltd.

Figure 7.27, 7.28,
7.29, 7.31

Figure 2.11-2.13

Reproduced by permission of
Stardent Computer Ltd.

Figure 3.12

Figure 7.14

Figure 7.15,
7.17-7.19

Figure 7.16

Data courtesy of Frank Bryan
at NCAR

Figure 8.8

156

Sources of Figures
Courtesy of Dr. B. EI-Haddadeh,
University of Leeds,
and UNIRAS software

Figure 3.9

Courtesy of Todd Elvins, San Diego
Supercomputer Center (SDSC)j
Data: Mark Ellisman, University of
California, San Diegoj Visualization:
Dave Hessler, SDSCj Software: SYNU
from SDSC

Figure 3.11

Courtesy of Todd Elvins, San Diego
Supercomputer Center (SDSC)j
Data: Ted Cranford, University of California, Santa Cruzj Visualization: Todd
Elvins, Phil Mercurio, SDSC, Software:
SUN Microsystems Voxvu

Figure 3.13

Courtesy of Prof. Elliot K. Fishman MD

Figure 2.8

Courtesy of Prof. T. L. Kunii

Figure 2.14, 2.15

Courtesy of the Lamont Doherty
Geologic Observatory

Figure 7.21

Courtesy of the Lawrence Livermore
National Laboratory

Figure 7.20

Courtesy of Silicon Graphics Ltd.

Figure 7.11-7.13

Courtesy of Simultec, Switzerland

Figure 7.22

Data courtesy of Dr. Michael Torello

Figure 8.7

Courtesy of Craig Upson

Figure 2.1, 3.1,
3.2, 3.4

Figure 9.1-9.3

Figure 2.4-2.6

Figure 7.4, 7.5

Figure 3.6-3.8,
7.1-7.3

Courtesy of Geoff Wyvill &
Brian Wyvill

Figure 2.9, 2.10

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.4
Linearized                      : No
Has XFA                         : No
Language                        : en
XMP Toolkit                     : Adobe XMP Core 4.2.1-c043 52.372728, 2009/01/18-15:08:04
Modify Date                     : 2012:12:10 16:03:47+08:00
Create Date                     : 2011:08:27 02:26:10+08:00
Metadata Date                   : 2012:12:10 16:03:47+08:00
Creator Tool                    : Adobe Acrobat 8.0
Format                          : application/pdf
Creator                         : 
Title                           : 
Document ID                     : uuid:a61ea4c3-ea1d-46b1-b6b2-f4530e81a873
Instance ID                     : uuid:f4a1b7cf-47ec-4bf1-b742-7abcd6f79184
Producer                        : Adobe Acrobat 9.46 Paper Capture Plug-in
Trapped                         : False
GTS PDFX Version                : PDF/X-3:2002
Page Layout                     : SinglePage
Page Mode                       : UseThumbs
Page Count                      : 169

EXIF Metadata provided by EXIF.tools

An Introductory Guide To Scientific Visualization

Navigation menu

Versions of this User Manual:

Views

Navigation