digital character development

Digital Character Development:Theory and Practice

This page inten ionally left blankt

Digital Character Development:Theory and Practice

RobONeill

Pratt Institute,Digital Arts Research Laboratory

AMSTERDAM BOSTON HEIDELBERG LONDONNEW YORK OXFORD PARIS SAN DIEGO

SAN FRANCISCO SINGAPORE SYDNEY TOKYOMorgan Kaufmann Publishers is an imprint of Elsevier

Morgan Kaufmann Publishers is an imprint of Elsevier.30 Corporate Drive, Suite 400, Burlington, MA 01803, USA

This book is printed on acid-free paper.

Copyright 2008 by Elsevier Inc. All rights reserved.

Designations used by companies to distinguish their products are often claimed as trademarks orregistered trademarks. In all instances in which Morgan Kaufmann Publishers is aware of a claim, theproduct names appear in initial capital or all capital letters. All trademarks that appear or are otherwisereferred to in this work belong to their respective owners. Neither Morgan Kaufmann Publishers nor theauthors and other contributors of this work have any relationship or afliation with such trademarkowners nor do such trademark owners conrm, endorse or approve the contents of this work. Readers,however, should contact the appropriate companies for more information regarding trademarks and anyrelated registrations.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form orby any meanselectronic, mechanical, photocopying, scanning, or otherwisewithout prior writtenpermission of the publisher.

Permissions may be sought directly from Elseviers Science & Technology Rights Department in Oxford,UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333, E-mail: [email protected]. You mayalso complete your request online via the Elsevier homepage (http://elsevier.com), by selecting Support& Contact then Copyright and Permission and then Obtaining Permissions.

Library of Congress Cataloging-in-Publication DataApplication Submitted

ISBN: 978-0-12-372561-5

For information on all Morgan Kaufmann publications,visit our Web site at www.mkp.com or www.elsevierdirect.com

Printed in the United States of America08 09 10 11 12 5 4 3 2 1

Contents

List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvForeword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviiAcknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xixIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi

PART 1 AN INTRODUCTION TO DIGITAL CHARACTERS 1

CHAPTER 1 Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.1 Overview of this Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Dening Digital Characters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.3 Roles in the Creation of Digital Characters . . . . . . . . . . . . . . . . . . . . . 7

1.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

CHAPTER 2 Contemporary Issues Related to Digital Characters . . . . . . . . . . . . . . . . . . . . . . 92.1 Viewer Perception and the Uncanny Valley. . . . . . . . . . . . . . . . . . . 9

2.2 How Do Style and Intended Platform Dictate CharacterSetup? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3 Case Studies in the Psychology of Digital Characters . . . . . . . . . . 15

2.4 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

CHAPTER 3 History of Digital Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.2 The Evolution of Digital Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.3 History of Digital Characters in Film . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.4 Overview of Digital Characters in Interactive Media . . . . . . . . . . . 40

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

CHAPTER 4 Character Technology and Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494.1 Commonalities Between Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

4.2 Programming and Scripting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 v

vi Contents

4.3 Algorithm Presentation in This Book. . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

PART 2 CHARACTER TECHNOLOGY 61

CHAPTER 5 Introduction to Character Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635.1 Nomenclature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

5.2 The Pipeline for Digital Character Creation andMaintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

5.3 Geometry Mesh Surface Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

5.4 Modeling Concerns for Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

5.5 Modeling Concerns for Video Games and Real-TimeEngines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

5.6 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

CHAPTER 6 Motion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 796.1 Character Motion Systems and Anatomy . . . . . . . . . . . . . . . . . . . . . . . 79

6.2 Anatomical Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

6.3 Anatomical Terms of Motion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

6.4 Joint Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

6.5 Comparative Anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

6.6 Matrices and Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

6.7 Joint Rotations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

6.8 Rotation Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

6.9 Euler Versus Quaternions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

6.10 Joint Naming and Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

6.11 Joint Hierarchies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

6.12 Anatomy Driven Hierarchies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

6.13 Constraints and High-Level Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

6.14 Forward and Inverse Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

6.15 Dynamics and Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

6.16 User Interface and Visualizing the Motion System . . . . . . . . . . . . . 115

6.17 Real-Time Engine Concerns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Contents vii

6.18 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

6.19 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

CHAPTER 7 Deformation Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277.1 Physiology of Muscles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

7.2 The Polygonal Model as Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

7.3 Deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

7.4 Skinning and Enveloping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

7.5 The Deformation Rig and Pipeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

7.6 Deformers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

7.7 Layered Deformation Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

7.8 Shape Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147


7.10 Spring Mesh Deformation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

7.11 Deformation Order. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153


7.13 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

7.14 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

CHAPTER 8 Face Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1638.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

8.2 Anatomy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

8.3 Psychology Behind Facial Expressions. . . . . . . . . . . . . . . . . . . . . . . . . . . 166

8.4 Face Shape Library. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

8.5 Emotions Through Face Shapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

8.6 Visemes and Lip Syncing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

8.7 Eyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

8.8 Interfaces for Facial Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177



8.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

viii Contents

8.12 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

CHAPTER 9 Rig Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1879.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

9.2 The Rig in a Scene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

9.3 Motion System Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

9.4 Deformation System Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

9.5 Face System Inventory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

9.6 Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197


9.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

9.9 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

PART 3 ANIMATION TECHNOLOGY 203

CHAPTER 10 Introduction to Animation Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20510.1 Denitions of Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

10.2 Integration of Animation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

10.3 Interfaces for Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

CHAPTER 11 Traditional Animation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21111.1 Classic Principles of Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

11.2 Curves and Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

11.3 Driving Versus Driven Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

11.4 Clip-Based Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

11.5 Sound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221


11.7 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

Contents ix

CHAPTER 12 Motion Capture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23112.1 Marker-Based Motion Capture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

12.2 Motion Capture Data Cleanup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

12.3 Skeletal Solvers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

12.4 Pipelines for Motion Capture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

12.5 Motion Retargeting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

12.6 Markerless Motion Capture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241

12.7 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244

CHAPTER 13 Procedural Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24913.1 Functions to Control Movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

13.2 Scripted Animation and Procedural Actors . . . . . . . . . . . . . . . . . . . . . 251

13.3 Physics to Control Movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

13.4 Behavioral Animation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

13.5 Articial Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

13.6 Crowds and Variation Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

13.7 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

CHAPTER 14 Case Studies in Character Interactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27114.1 AlphaWolf by the Synthetic Characters Group

at the MIT Media Lab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

14.2 Tartarus by Alan Price . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

14.3 MovieSandbox by Friedrich Kirschner . . . . . . . . . . . . . . . . . . . . . . . . 279

Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282

Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

PART 4 CONCLUSIONS 289

CHAPTER 15 The Future of Digital Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300

x Contents

CHAPTER 16 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

Bibliography 307

Index 311

List of Figures

1.1 Screen capture of a character motion and deformation system. . . . . . . . . . . . . . . 4

1.2 Character rig and model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.1 Abstraction of the human face by Scott McCloud [25] . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2 The Uncanny Valley as described by Masahiro Mori . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3.1 Bit from Tron (1982). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3.2 Jar Jar Binks from Star Wars Episode I - The Phantom Menace (1999). . . . 35

3.3 Second Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

3.4 The Movies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.1 Class and data diagram for joint rotate constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

4.2 Method for a joint rotate constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.1 Character technology pipeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

5.2 Surface type sphere primitives (polygon, NURBS, subdivision surface) . . . . . 69

5.3 Character model in T-Pose and A-Pose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

5.4 Knee model straight and exed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

5.5 Shoulder model straight and bent down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

5.6 Low- and high-resolution polygon mesh for character fromGears of War. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

6.1 Anatomical Position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

6.2 Human dog comparison of anatomical direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

6.3 Body planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

6.4 Basic human skeleton element names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

6.5 Human anatomical elements in motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

6.6 Comparison of insect, pterosaur, bird, and bat wings . . . . . . . . . . . . . . . . . . . . . . . . . 91

6.7 Class and data diagram for joint object . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

6.8 Left- and right-handed coordinate systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

6.9 Rotation around the x-, y-, and z-axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

6.10 Rotation orders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

6.11 Example of articulating one versus two Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

6.12 Method for a curve-based joint placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

6.13 Human shoulder girdle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

6.14 Class, data, and method diagram for position constraint . . . . . . . . . . . . . . . . . . . . . 105

6.15 Class and data diagram for compensating rotation constraint . . . . . . . . . . . . . . . 107 xi

xii List of Figures

6.16 Method for a compensating rotation constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

6.17 Class and data diagram for simple inverse kinematics constraint . . . . . . . . . . . . 109

6.18 Method for simple inverse kinematics constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

6.19 Class and data diagram for spline inverse kinematics constraint . . . . . . . . . . . . . 112

6.20 Method for spline inverse kinematics constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

6.21 Class and data diagram for tail dynamics constraint . . . . . . . . . . . . . . . . . . . . . . . . . . 113

6.22 Method for tail dynamics constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

6.23 Deforming model versus non-deforming proxy models . . . . . . . . . . . . . . . . . . . . . . 117

7.1 Biceps muscle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

7.2 Front and back of a muscle model setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

7.3 Abe sapien muscle setup from Hellboy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

7.4 Front and back of a character model with dened muscles . . . . . . . . . . . . . . . . . . . 131

7.5 Close-up model views . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

7.6 Undeformed, bent, stretched, and sheared polygon plane. . . . . . . . . . . . . . . . . . . . 133

7.7 Smooth (weighted) versus rigid (uniform) skinning . . . . . . . . . . . . . . . . . . . . . . . . . . 135

7.8 Class, data diagram, and method for stretching joints in a straight line. . . . . 137

7.9 Painting interface for editing skin weights for the left knee joint . . . . . . . . . . . . 138

7.10 Sphere modied by a lattice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

7.11 Surface acting as an abstraction of biceps muscle surface . . . . . . . . . . . . . . . . . . . . . 144

7.12 Layered deformation setup from Clovereld (2008) . . . . . . . . . . . . . . . . . . . . . . . . 145

7.13 One low-resolution surface controlling two subsequently high-resolutionsurfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

7.14 Class and data diagram for pose space interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . 148

7.15 Method for a pose-space interpolation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

7.16 Class and data diagram for jiggle deformer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

7.17 Method for a jiggle deformer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

7.18 Sample deformation order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

8.1 Major face muscles responsible for motion (platysma not shown) . . . . . . . . . . 165

8.2 Neutral face model with blend shape inputs results in an output facewith a wide range of emotion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

8.3 Basic happy expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

8.4 Basic sad expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

8.5 Basic angry expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

8.6 Basic disgust expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

List of Figures xiii

8.7 Basic fear expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

8.8 Basic surprise expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

8.9 Face visemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

8.10 Eye anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

8.11 Eye surfaces and eye model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

8.12 Possible face interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

8.13 Yoda from Star Wars Episode III - Revenge of the Sith (2005) . . . . . . . . . . . . . 181

9.1 Top-level nodes are often visualized as a series of curves that canbe selected. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189

9.2 Silly as they look putting the rig through a series of extreme-poseexercises is an ideal way to test deformations while in development . . . . . . . . 193

9.3 The neck is a complicated region at the boundary of the bodydeformation and face systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

9.4 The rig exported to the game engine includes a stripped-down set ofjoints representing the body and the face . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

11.1 Curve with stepped, linear, and spline interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . 217

11.2 Class, data, and method for a blending positions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

11.3 Audio waveform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

12.1 Motion capture performer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232

12.2 Sample motion capture marker placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

12.3 Blended input on a node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

12.4 Markerless motion capture by Organic Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241

13.1 Sine and cosine graph. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

13.2 Emotional curves for a procedural system by Actor Machine . . . . . . . . . . . . . . . 252

13.3 Prototype procedural character engine by Actor Machine . . . . . . . . . . . . . . . . . . . 254

13.4 Method for a A* pathnding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

13.5 Basic decision tree for a crowd character engaging in a ght sequence . . . . . . 263

14.1 AlphaWolf by the synthetic characters group at the MIT media lab. . . . . . . . 276

14.2 Tartarus by Alan Price. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

14.3 Empire by Alan Price . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279

14.4 The Journey (2004) by Friedrich Kirschner. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280

14.5 MovieSandbox by Friedrich Kirschner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

16.1 King Kong (1933) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304

List of Tables3.1 Notable Projects Featuring Digital Characters (Pre-1990) . . . . . . . . . . . . . . . . . . . . 32

3.2 Notable Projects Featuring Digital Characters (Early 1990s) . . . . . . . . . . . . . . . . . 34

3.3 Notable Projects Featuring Digital Characters (Late 1990s) . . . . . . . . . . . . . . . . . . 36

3.4 Notable Projects Featuring Digital Characters (20002001) . . . . . . . . . . . . . . . . . . 39

3.5 Notable Projects Featuring Digital Characters (20022004). . . . . . . . . . . . . . . . . . 41



6.1 Anatomical Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

6.2 Naming Convention Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

8.1 Phoneme to Viseme Mapping. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

9.1 Sample Attributes on an Animation Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

14.1 Summary of Five Major Differences Between Linear and InteractiveAnimation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272

xv

Foreword

Animation involves breathing life into things. The word animation comes fromthe Proto-Indo-European root ane-, meaning to breathe, via the Latin anima,meaning living spirit. The spirit of animation lies in the charactersentitiesthat people feel are alive (and possibly even sentient) when they engage withthem.

Computer animation specically involves breathing life into things bymeans of electronic systems that perform millions of calculations each sec-ond. Computers have the capability to present complex characters and theintricate worlds they inhabit in ways that draw people in and help us seepast the vast arrays of binary digits inside the box. The only challenge is thatsomeone has to tell the computer how to twiddle all those bits.

Breathing life into computer characters is a tricky blend of artistry andtechnology. The artistic side of this process entails a subtle understandingof both the characters in the animation and the audience to whom thosecharacters are being presented. Great animation requires intuition and nesse,and certain effortlessness that facilitates the audiences suspension of disbelief.Underlying this artistry is an array of elds ranging from graphic design andcinematography all the way to psychology, anatomy, and physics.

Being a skilled computer animator also entails a rigorous understand-ing of a wide range of arcane technological skills. Terms like quaternions,matrices, and inverse kinematics are tossed around like grenades. Tomake the challenge even greater, the hardware and software underlyingcomputer animation change so rapidly that it is hard to keep pace. Thus,computer animation requires a simultaneous grasp on both the esoteric andthe cutting-edge.

Rob ONeill is one among a small group of people in the world who canaddress both the artistry and technology of animation with equal skill. Inthis book, Rob provides a comprehensive treatment of several disparate yetinterconnected parts, covering the overlap of anatomy, artistry, technology,mathematics, and many other elements that are critical to the creation ofdigital characters. Rob skillfully illuminates the complex interdependenciesamong such seemingly distinct aspects of animation and helps the readerunderstand why it is important to treat all of these issues together. You havechosen a wise guide through the lush jungle of computer animation.

xvii

xviii Foreword

This book is relevant to practicing animators, animation students,researchers, and anyone else who may have interest to know how animationworks. It is not just a technical manual and not just a theoretical explorationof the creative process; it is both at once. As such, it gets to the heart of com-puter animationan incrediblyprecise andexacting craft that seeks toproducecharacters with elegance, subtlety, love, fear, anguish, humor, and compassion.

Bill TomlinsonUniversity of California, Irvine

Acknowledgments

This book would not be possible without all the hard work, innovation, andcreative engineering by those who developed the ideas summarized within.The thoughtful reviews byBill Tomlinson andTJGalda added valuable insightsand critical thoughts thatmade this a better book. I cannot thank themenoughfor the time they invested in reviewing rough versions of the manuscript.

I owe a huge debt to all the members, past and present, of the CharacterTechnical Direction Crew at PDI/Dreamworks Animation for their friendship,camaraderie, and ever-present desire to educate at every level. Special thanksto modeler/rigger extraordinaire Lee Wolland for the use of his immaculatehuman model for the creation of model screen shots and to Mark Snoswell,President of The Computer Graphics Society, for his generous donation of thecgHuman for the creation of muscle images in this book. His model set is anamazing resource for study and use in production.

I owe a huge debt to Dr. Bonnie Gustav of Brooklyn College for gettingme hooked on biological anthropology and the evolutionary anatomy thatgoes with it and to Anezka Sebek of Parsons School of Design for helping meharness that into a career in animation. The support of my colleagues andstudents in the Department of Digital Arts at Pratt Institute provided much ofthe inspiration to write this text. In particular, thanks to Provost Peter Barna,Chair Peter Patchen, and Assistant Chair Melissa Barrett Lundquist for theirsupport. Extra special thanks go to my Graduate Research Assistants at thePratt Digital Arts Research Lab: Paris Mavroidis and George Smaragdis whosehard work is reected in many of the images in this book and all the otheramazing work that has been produced in the Lab.

Thanks to the editorial staff at Morgan Kaufmann past and present fortheir guidance, support, and patience, in particular: Tim Cox, Jessie Evans,Georgia Kennedy, Laura Lewin, Chris Simpson and Anais Wheeler. Thank youto Claudia Pisano and Beth Millett for the help in securing image rights.

The book was made so much better by the inclusion of interviews with thetalented and generous Aaron Holly, Milana Huang, David Hunt, Steve Mann,Paul Marino, Ken Perlin, Nico Scapel, Jason Schleifer, and Todd Taylor. Theirviews are their own and they in no way endorse any of the other material inthe book, but I hope they like it.

xix

xx Acknowledgments

As a continual source of inspiration are my partners in the creation ofKickstand Animation Research + Design: Daniel Dawson, Greg Elshoff, andPhil McNagny. Keep an eye on the horizon for some good things in the nearfuture.

Thanks to my parents, grandparents, brother and sister for fostering mysense of imagination and humor, two critical skills for working in animationproduction. Thanks to the support of my friends, new and old, who keepme sane. Finally my wife, Kate, whose support and enthusiasm for the bookreally carried me through to the end. From allowing me to take drafts of themanuscript on vacation, to being a great copy editor, she was the backboneof the creation of this book. Words can not express my appreciation for theeffort she made.

Epic Games, the Epic Games logo, the Crimson Omen logo and Gears ofWar are either registered trademarks or trademarks of Epic Games, Inc. in theUnited States and/or elsewhere.

Introduction

Digital characters are a driving force in the entertainment industry today.Every animated lm and video game production spends a large percentageof its resources and time on advancing the quality of the digital charactersinhabiting the world that they are creating. We have entered an era whendigital characters have reached a level of sophistication that has promptedsome critics to question if a digital actor can win an Academy Award foracting. As articial intelligence and behavioral animation become more inte-grated with hand-animated entities, we will see a dramatic increase in therealism and interactivity of these characters. Practitioners of the subject willalso require a deeper understanding of the underlying conceptual foundationas the complexity of the technology increases. The eld of character technol-ogy has matured into a topic that spans the realms of anatomy, animation,computer science, performance, and behavioral psychology. The contempo-rary uses of digital characters are varied and range from purely entertainmentto biomedical, industrial simulation, and beyond. This book is an overview ofthe history, theory, and methods for creating digital characters. Many bookscover the step-by-step creation of digital characters using a particular piece ofsoftware. This book forgoes existing software and deals with the concepts froma software-agnostic point of view.

Recently, characters such as Gollum from The Lord of the Rings series(20012004) and Davy Jones from the Pirates of the Caribbean series(20062007) have both been discussed with regard to earning awards forachievement in acting. Almost more compelling, a recent panel entitled TheBiology of KingKong, part of the 2006 Tribeca FilmFestival, included a discus-sion on how King Kong (2005) was incredibly true to life and believable asa real gorilla. Panel member, Roger Fouts, Co-Director of the Chimpanzee andHuman Communication Institute, discussed how pleased he was that the risein technology and artistry has allowed for digital doubles and replacementsfor roles that were usually reserved for trained animals. While most critics arewaiting for a believable human replacement, there is no better complimentfor the team that created this digital character and no better indicator for thepotential of character technology.

Rob ONeill

xxi

Part1An Introduction to Digital

Characters

Chapter 1Overview

1.1 OVERVIEW OF THIS BOOK

This book is about character technology. Character tech-nology is the merging of character animation and three-dimensional (3D) computer graphics technology into a seriesof concepts and methodologies resulting in a character ableto be animated.

This book is intended to provide an introductory overviewof the theory and practice of digital character development,often called character rigging, from a software-agnostic pointof view. As a starting-off point for more complicated math-ematical discussions and more artistic exploration, the textstraddles the line between the arts and sciences that arerequired for success in the creation of compelling digital char-acters. As this subject is a moving target, updates, errata, andadditional resources are compiled at:

http://www.charactertechnology.com

Starting with the evolution and history of digital charac-ters, which is critical to the understanding of how charactersdeveloped alongside the techniques and technologies for cre-ating them, we then begin the process of building charactersfrom the inside out. Character setup starts by dening theanatomical considerations required andhow they relate to the

3

4 CHAPTER 1 Overview

Motion System Deformation System

FIGURE 1.1 Screen capture of a character motion and deformation system. Character by George Smaragdis and Rob ONeill.

structure of the character motion system (Figure 1.1) or themechanical architecture that drives the model. At this point,we will deal with the control structure for the motion systemthat provides the interface to animation. From there, the lookof the character in motion is dened by the deformation sys-tem,which is the planning and implementationof techniquesthat sculpt the characters shape as the motion system drivesthe manipulation of the model. Much of the life and emotionof a digital character is read through the face, and while facialsetup is a mix of motion and deformation systems, the issuesintrinsic to this aspect of character development require spe-cic attention. In the areas of motion systems, deformationsystems, and facial setup, we cover techniques currently inuse by mixing descriptive text with algorithms and graphicalrepresentations of code for potential implementation. In thecourse of this discussion, digital characters created for bothgames and lm are addressed.

Digital characters would be nothing without a means ofmoving them. In the animation technology section, we coverkeyframe strategies and curve interpolation, with an empha-sis on performance through traditional techniques. This isfollowed up by motion capture and how human performersare stepping into the shoes of digital characters. Furthermore,code often also drives motion, so procedural animation, arti-cial intelligence, and multi-character systems used for crowdsimulations are a necessary component of modern character

1.2 Dening Digital Characters 5

development. Finally, interactive characters are consideredwith a focuson the setup and limitations for characters createdfor game and non-game related interactive media.

The future of digital characters is where we conclude witha look into research questions that remain outstanding andsome challenges for work beyond them.

1.2 DEFINING DIGITAL CHARACTERS

The word animation basically means physically or graphi-cally moving, derived from latin anima which refers to livingbeings, spirit, and feeling and the Greek anemos for wind andbreath. An animated character could be said to have had thedivine wind (animus) breathed into it. A digital character canbe dened as an animated entity that has been brought to lifeusing a computer. These characters are often found in lms,television, interactive media projects, and most prevalentlyin video games. The creation of these entities can employoff-the-shelf software or custom programming to attain thedesired results. Nomenclature for this digital species is diverseand continually evolving. Digital actors, virtual actors,vactors, and synthespians have all been used, but theseall imply a connection to acting, and thus humans. The termavatar has also been used as a catchall phrase, but thisterm has evolved into a representation of a player or humaninhabitant in an interactive system. For the purposes of thisbook, we will refer to computer-generated entities as digi-tal characters to encompass all imaginable forms and actingabilities, with an emphasis on 3D animation media. An inter-esting distinction has been made in the lm industry betweendigital characters and digital creatures [26], the main differ-entiation being that of performance. Where a digital beastattacking a group of tourists would be described as a digi-tal creature, while a compelling gorilla who falls in love withthe lms starlet would be described as a digital character.This distinction comes from the amount of work requiredin creating the character, the amount of screen-time, andthe viewers connection with the character on screen. This


FIGURE 1.2 Character rig and model. Character by George Smaragdis and Rob ONeill.

book ignores the distinction between digital characters anddigital creatures as the underlying functionality and setup isvery similar.

Creating a digital character with the ability to produce arich performance requires the interaction of a number of sub-jects and pathways of thinking. In a digital space, this processalways walks the line between aspects of art and science, usingtechnology as a bridge. This book covers all facets of digitalcharacter development but focuses primarily on the realmof what has been traditionally called character rigging orcharacter setup. This is the process of creating the architec-ture, called the rig (Figure 1.2), andmethodologies requiredfor animating, by means of user-set keyframes or proceduralcontrol, the motion of the character. This is a crucial role thatrequires knowledge of anatomy, as well as animation. Withthe increasing complexity of characters needed for variousforms of media, a knowledge of programming to reinvent orextend the available tools is crucial.

1.3 Roles in the Creation of Digital Characters 7

1.3 ROLES IN THE CREATION OF DIGITALCHARACTERS

There are many tasks when it comes to digital characterdevelopment. While job titles are constantly evolving andtechnological growth empowers the individual to accomplishmore on their own, it is worthwhile to discuss some of theroles associated with the creation of digital characters. At thestart of production, we have the concept artists and characterdesigners who plan the overall look of the characters. Char-acter modelers then take the character designs developed bythe art team and build the 3D sculpture based on produc-tion specications for size, proportion, style, and geometricsurface type. Once approved, this model is handed off to theposition most beneting from this book, the character tech-nical director. Often called rigger, character setup artist, orcreature developer, the character technical directors title atsome studios may even be condensed down to something asnon-descript as technical director, but the role remains thesame. Character technical directors are responsible for takingthe 3D model and adding all the bones and muscles thatallow an animator tomove the character in a predictableman-ner and deform those surfaces in amanner consistentwith theaesthetics of the production. Character technical directors arecharged with designing the motion system which providesthe architecture and control structures for how the rig is ani-mated as well as the deformation system which denes howthe model is affected by the motion system and the anima-tion that is fed to it.Other components such as dynamic cloth,hair, and accessories associated with the characters are oftenalso the domain of this role.

From here, the character is in the hands of the animatorsand the technicians responsible for attaching motion to therig, but the character technical director will continue to over-see the maintenance of the rig until the end of production.At the early stages of production, character technical directorsspend time doing research and development on new tech-niques for improving their character pipeline in addition tothe specic challenges of future productions.


1.4 CONCLUSION

Digital characters are developed using various techniques,collectively called character technology. This technology isimplemented by the character technical director based onthe needs of the production. This book serves as an overviewof the concepts and the technology behind digital charac-ters to foster the development of innovative techniques andcompelling performances.

FURTHER READING

Tim McLaughlin. Taxonomy of digital creatures: Interpret-ing character designs as computer graphics techniques. InSIGGRAPH 05: ACM SIGGRAPH 2005 Courses, page 1,ACM Press, New York, NY, 2005.

Chapter 2Contemporary Issues Related to Digital

Characters

The challenge of creating a compelling digital character thatboth looks and behaves in an appealing manner is a tremen-dous undertaking. Digital characters share the same difcul-ties inherent in traditional animation, but their trend towardrealism presents many additional challenges. In this section,we examine the design decisions that must be acknowledgedin the development of a digital character and the pitfalls thatcan arise along theway. Characters created for live-action inte-gration and those for fully animated productions share thesame challenges, and some of the best lessons come from theworlds of illustration and robotics. Also in this section, wewill look at howdigital characters are being used in importantareas, such as education and user survey, where the delicateconnection between the character and the human makes allthe difference in assessing the successfulness of the projectand the intended outcome.

2.1 VIEWER PERCEPTION AND THE UNCANNYVALLEY

Its all in the eyes...

The perception of digital characters is determined by the reac-tionof the viewer to the look andperformance of the characteron screen. While a poor performance by a human actor may 9

10 CHAPTER 2 Contemporary Issues Related to Digital Characters

result in laughter or snickering by the audience, the uncon-vincing performance by a digital character in a live-action lmhas a much more jarring effect, particulary when that char-acter is supposed to be realistic. Animated characters witha high degree of abstraction or a cartoon-like appearanceare generally more accepted by the audience whether theyare interacting in an animated or a live-action world. Thatbeing said, the stakes are higher for characters, which must beintegrated with live human actors.

Writer and comic book theorist Scott McCloud [25]has distilled the distinction between character realism andabstraction. An abstract character with a simplied face hasthe ability to emote more clearly via the process of ampli-cation through simplication. By eliminating details, we areallowed to focus on the specic details required to create a per-formance, and thus the expression is amplied (Figure 2.1).We have seen this in the incredible popularity of animatedcharacters, such as Mickey Mouse and Bugs Bunny, that withtheir simplied design have the ability to captivate a widerange of audiences.

FIGURE 2.1 Abstraction of the human face by Scott McCloud [25]. Copyright 1993, 1994Scott McCloud. Reprinted by permission of HarperCollins Publishers.

2.1 Viewer Perception and the UncannyValley 11

When it comes to digital characters, much of the discus-sion with regard to believability and the ability to forge aconnection with the human observer is rooted in the studyof robotics. Long before digital characters were at a level ofsophistication to even address this subject, engineers wereexperimenting with the visual representation of emotion inreal-world robots. Robots share many similarities with digi-tal characters, from the computational aspects of kinematicsto their perception by viewers. The connection between themcanbe encapsulatedby the observations of Japanese roboticistMasahiro Mori, who in 1970 raised the idea of the UncannyValley [28].

Moris concept states that as we approach a realistic visageof a human, the familiarity level plummets. The drop gener-ally takes place at around the 75% point toward realism, andwe experience a drop-off into a familiarity valley where theconnection to the entity becomes more akin to how we per-ceive zombies and prosthetics (Figure 2.2). This is a commonissue that, for themost part, cannot be expressed by the viewerbut is endemic to digital characters. We, as humans, are natu-rally accustomed to the nuances and idiosyncrasies of a living

human likeness

corpse

zombie

prosthetic hand

healthy person

stuffed animal

humanoid robot

industrial robot

Uncanny Valley

50% 100%

fam

ilia

rity

1

2

moving entitystill entity

FIGURE 2.2 The Uncanny Valley as described by Masahiro Mori.


person. What the Uncanny Valley describes is how we pickup on the small percentage of non-humanness in a char-acter and xate on that. The other details become irrelevantto us; our senses tell us something is unnatural about whatwe are seeing. We focus on the lack of skin translucency, thestillness of the eyes, and the slightly roboticmotion. The char-acter that was once appealing has given way to an animatedcorpse and our sense of familiarity plunges.

This small percentage is what animators and roboticistsare charged with overcoming when they are in the process ofcreating something that is supposed to be alive. This is nota simple feat. At some point, most of us have experienced anuneasy feeling when watching a digital character. Perhaps,in watching a lm with a character that you were supposedto connect with, you felt as if that character was lifeless, notin a stiff, poorly animated way, but in manner that can onlybe described as creepy. Films such as Final Fantasy (2001)and Polar Express (2004) are themost cited examples of thisphenomenon, and one might argue that for Final Fantasy,this hurt the lms box-ofce returns. Both lms incorporatedmotion capture, but therewere discrepancies in the charactersfaces that created the most unease, the so-called dead-eyesyndrome. The real/unreal notionwas even discussed by lmcritic Roger Ebert in his review of the lm [10]:

Not for an instant do we believe that Dr. Aki Ross, the hero-ine, is a real human. But we concede she is lifelike, whichis the whole point. She has an eerie presence that is at oncesubtly unreal and yet convincing.

The question then remains, how do we combat the lack ofsubtlety in these characters that causes viewers to have suchan unsettled reaction to them? Part of the answer lies in thedesign of the character and the style of the lm. As we haveseen, when possible, a more stylized character will communi-cate easily to its audience. The design of the character is basedon the style of the project and is informed by the intendeduse of the character and the media that it exists in.

2.2 How Do Style and Intended Platform Dictate Character Setup? 13

The essence of the character generally comes across mosteasily when stylistic choices are made that accentuate char-acter attributes such as the eyes. The eyes are one of themain conduits of communication to the viewer. A charac-ters expressions should read as well in a still pose as they doin motion. Motion, however, adds a level of complexity asthis is where you get into the behavior and the psychology ofthe character. The animator is responsible for what is goingon inside the characters head. This is accomplished solelythrough the subtle actions of the character. Behavioral cluesthat trigger a negative response come about from behaviorthat is categorized as abnormal. In the next sections, we raisequestions about how the style and use of a character informsetup in addition to taking a brief look at how we perceivepeople in day-to-day interactions and the notion of abnormalpsychology. To help look at this from the other perspective,we discuss how psychologists are using digital characters intheir own research as interview avatars. The psychology of adigital character allows us to think about how we perceive ourdigital counterparts.

2.2 HOW DO STYLE AND INTENDED PLATFORMDICTATE CHARACTER SETUP?

When planning the character setup for a project, the overallstyle must be analyzed and the manner of motion plannedout.Whilemost character setups tend to be complex, for high-resolution digital characters, the requirements of the setupare predicated on the style of motion required of that char-acter for the project. This is especially true for characters thatare expected to behave in a more stylized, cartoon-derivedmanner. Often the difference comes down to the nal defor-mations, which dictate the shape and prole of the characterin motion. Such decisions could illustrate the character ineither a realistic manner adhering to a realistic anatomyor a cartoon rubber-hose style. The rubber-hose style ofanimation was initiated with Oswald the Lucky Rabbit cre-ated by Ub Iwerks and Walt Disney and exemplied by the


Fleischer Brothers Olive Oyl character in the Popeyecartoon series. Other examples of this style of animationinclude early Mickey Mouse and Bugs Bunny. The differencebetween cartoon rubber-hose animation and that of morerealistic motion is the presence or lack of denable visibleanatomical structures, such as bones. Certainly, complex dig-ital characters created for integration with live-action featurelms often contain an underlying structure that mimics theskeletal and musculature needed if that character existed, nomatter how fantastical. For the readers with knowledge ofcharacter setup, this discussion of bones, versus the lackthereof, is referring to the visual anatomical and kinematicrepresentation and not the underlying methodology used tosetup the character. While it is assumed that a realistic char-acter with bones and muscles is more complicated than themalleability of a rubber-hose digital character, they bothhave their own unique challenges. These basic questions needto be addressed during pre-production of the project:

1. What is the overall visual style or genre of the character?2. Is this a realistic character whose anatomy reacts to

motion?3. Do the characters body and limbs need to squash and

stretch with motion?4. Does the style of motion or deformation change

throughout the project?

These questions come before the details and issues relatedto the construction of the actual character rig. Many ofthese questions can be answered by developing or consultingconcept illustrations of the character.

The intended use of a digital character also has a tremen-dous impact on the design and architecture of how it is built.Certainly, the difference between characters built for gamesand those for lm animation is very different, but there arelevels and distinctions within these two uses that will alsoinuence setup. The distinction between lead, secondary, andtertiary characters plays its part as does whether a character ina game is going to be controlled by the player or not. There are

2.3 Case Studies in the Psychology of Digital Characters 15

limitations to real-time game engines that will inuence howthe character is developed. Because lm characters are notinteracted with in their nal form, there can be a complex-ity distinction between what the animator interacts with andwhat is rendered. Thus, with a careful construction, there isno limitation to what can be built into a lm character. Somebasic questions can be asked of the production to determinehow character use will impact setup:

1. What system will the character be animated in?2. Is the character intended for rendered animation or a

game engine?3. If the character is to be implemented in a game engine,

what types of deformations does it support, and whatare the limitations with regard to number of joints?

The challenges that result in answering all these questionsare the topic of this book. Beyond form and function, we alsoneed to consider performance and the psychology of digitalcharacters to consider how we relate to them based on howthey relate to us.

2.3 CASE STUDIES IN THE PSYCHOLOGY OFDIGITAL CHARACTERS

A number of psychologists have begun to study digital charac-ters with research topics ranging from how they interact withhumans to how effective digital characters are as interviewersfor survey use. This research is another component as to howartists can transcend the Uncanny Valley in the creation ofthe digital characters. The depth of this subject is too great togo into detail here but worth a brief discussion as it is increas-ingly informing the way digital characters are designed andanimated.

Digital characters are static without the input of motion.We are arguablywired frombefore birth to instinctively recog-nize how a human acts down to the smallest details. Any cluethat something is not operating as we would expect causesus to assume that there is something out of the ordinary


about that individual. This is a steep challenge for the digitalcharacter and falls into the category of abnormal psychology.Abnormal psychology can be dened as an area of psychologythat studies people who are consistently unable to adapt andfunction effectively in a variety of conditions. Digital charac-ters are inherently abnormal. It takes a tremendous amountof nuance to overcome the smallest pathology to create acharacter whose actions do not detract from what would beconsidered normal. Character pathology is often discussedwith regard to pure character motion where popping kneesand an unnatural stance can make something appear injuredwhen in fact it is purely just a poorly animated motion.Pathology at the characters performance level canhave amuchmore damaging effects on the viewers experience as notedabove. Viewers often describe the motion and expressions ofdigital characters as too perfect. The implication here is thata little bit of idiosyncrasy is actually what makes us human.Striking the balance between character traits and characteraws is a delicate line. A deeper, sustained connection to dig-ital characters is required when it comes to interactive mediathat strives to teach or collect important information. Withthis in mind, we can look at situations where digital charac-ters are being tested in the realms of education and automatedinterview.

Educational software strives to teach lessons through newforms of interaction and connection with its users. The ideathat a digital character in the role of a teacher, adviser, or guidecould be compelling enough to create a strong connectionwith a student is a compelling and active subject of research.Preliminary systems are already inplace in childrens software,agent-based teachers for medical situations, military training,and even at the simple level as help agents for commercialsoftware. Research in education demonstrates that the use ofdigital characters has many benets, as discussed by James C.Lester and his group at North Carolina State University [22]where this group describes their work that tries to capitalizeon the emotional connection we, in particular children, feeltoward digital characters:

2.3 Case Studies in the Psychology of Digital Characters 17

Because of the immediate and deep afnity that childrenseem to develop for these characters, the potential pedagogi-cal benets they provide are perhaps even exceeded by theirmotivational benets. By creating the illusion of life, lifelikecomputer characters may signicantly increase the time thatchildren seek to spend with educational software, and recentadvances in affordable graphic hardware are beginning tomake the widespread distribution of realtime animationtechnology a reality. (p. 273)

Lester has, in other experiments [23], demonstrated a phe-nomenonhe termed the persona effect, which is the positiveresponse and sense of engagement shownbyuserswhen inter-acting with an animated agent. The conjecture is that (inparticular for children) animatedpedagogical agentswill havea higher success rate than similar software without animatedagents. It can be hypothesized that children are more accus-tomed to seeing animated characters. It is also possible thattheir ability to recognized nuanced behavior is not as tunedas an adults making them more forgiving of the shortcom-ings of current digital character and thus able to have moremeaningful interaction with them.

Winslow Burleson [4], during his work at MassachusettsInstitute of Technology (MIT), used sensors and input devicesto look at the subtle non-verbal cues that can make digitalcharacters more adept at social situations. He found that ininteractions with pedagogical agents, as in human-humaninteraction, responsiveness can be enhanced by non-verbalsocial mirroring, in a sense, the imitation of anothers non-verbal cues. By using a four-second delay, the charactersmirroring behavior is not consciously detected by users, yetthis is a short enough time for the mirroring to have a socialeffect. Animated agents in education is an ongoing researchtopic. Similarly, digital characters in the role of interviewagents is another application helping us distill the nuancedrelationship between digital characters and humans.

Collecting information via human-to-human interview isa time consuming and costly endeavor. The process is also ripe


with sources of error and can generally be viewed as incon-venient for the respondent (the person answering the ques-tions). Human interviewers have been shown to result in thehigher response rates and improved respondent comprehen-sion over questionnaires. Questionnaires, on the other hand,incur lower costs, increased privacy, and convenience andcontrol for respondents. The hope of using a digital char-acter as an interviewer is that they will actually reap thebenets of both the human interviewer and a questionnaire.The non-judgmental, anonymous aspect of digital charactersmay also have some benet. But what are the fallbacks ofthis forward-thinking technology? Frederick G. Conrad andhis group at the University of Michigan are developing dig-ital conversational agents for use in interviewing scenarios.Conrad and Michael Schober [9] of New School Universityheld workshops, entitled Envisioning the Survey Interviewof the Future, intended to expose survey methodologiststo upcoming technologies that might be applicable to sur-vey data collection. One of these technologies are videogame characters and the associated game engine as a vehiclefor survey data collection. To test the hypothesis that ani-mated agents are more successful than live humans, Conradand his group are conducting exploratory laboratory experi-ments [8]. Rather than developing the actual agent software,they are simulating animated agents using what they term aWizard-of-Oz technique, where respondents believe they areinteracting with a computer-generated agent when, in fact,they are interacting with an actual interviewer whose imagehas been rendered graphically in the user interface as if it wascomputer-generated. The overall goal of the proposed workis to determine when animated agents might help and whenthey might hurt quality of responses.

A very good overview of a psychological approach todesigning game characters is collected by Katherine Isbister[14]. In it, Isbister covers the principles required for gamecharacters to create powerful social and emotional connec-tions with users interacting with current and next-generationgaming environments. By tackling issues of culture and

2.4 Exercise 19

gender alongside the tools we use to communicate, suchas the body, face, and voice, Isbisters text challenges us tomake conscious decisions about every aspect of designing andanimating a digital character.

The research steps outlined in the case studies above mayseem small, but the results of them will make our interac-tionswith digital charactersmore effective, seamless, and darewe say natural. All of this research is relevant to charactersin the entertainment industry. As we put people face-to-facewith digital characters and break down our understandingof human behavior into pieces that can be analyzed andrecreated individually, it is only a matter of time before thisknowledge will make its way into video game and lms. Thecritical topic of behavior will be readdressed when we discussfacial animation, procedural motion, and articial intelli-gence later. The historical and evolutionary steps that haveoccurred are outlined next and will help inform the technicalpractice of building digital characters.

2.4 EXERCISE

Start designing a character you would like to use throughoutthis book. If it is a human, then think about Moris UncannyValley.What level of stylizationwill give you the right place onthat continuum so as to not present you with the challenge ofcreating something almost real but not quite? What mediumwill this character exist in? What aspects of the characters facecan you accentuate to make it appealing even if grotesqueor evil?

2.1 Start creating a character bible which includes descrip-tions and sketches of your character. It should alsoinclude references for details of the character that youcollect over time. These references canbe images, quotes,or pieces of fabric that dene a texture. Whether thesethings make their way into your character formally orremain strictly as inspiration is something that onlytime will tell; so do not hesitate to collect anything thatstrikes a chord.


2.2 Get feedback from others about the bible. Does thedesign of the character tell part of its story throughthe use of details such as period clothing or culturalelements?

2.3 Incorporate the feedback you receive to rene the char-acter bible.

FURTHER READINGFor a solidoverviewof thepsychologybehind character designrefer to:

Katherine Isbister. Better Game Characters by Design: A Psycho-logical Approach (The Morgan Kaufmann Series in Interactive3D Technology). Morgan Kaufmann Publishers Inc., SanFrancisco, CA, 2006.

An interesting panel conversation about Moris Uncanny Val-ley took place at Siggraph 2007 in San Diego California. Theabstract is below but video transcripts are available:

Thierry Chaminade, Jessica K. Hodgins, Joe Letteri, andKarl F. Mac-Dorman. The uncanny valley of eeriness. InSIGGRAPH 07: ACM SIGGRAPH 2007 panels, page 1, ACMPress, New York, NY, 2007.

Scott McClouds Understanding Comics is requiredreading:

Scott McCloud. Understanding Comics: The Invisible Art.Perennial Currents, New York, NY, 1994.

INTERVIEW: MILANA HUANG, CHARACTERTECHNICAL DIRECTION SUPERVISOR,PDI/DREAMWORKS ANIMATION

Milana Huang, Character Technical Direction Supervisor,PDI/DreamWorks Animation

BIOMilana Huang is currently the Character TD Supervisor onDreamWorks Animations anticipated fall release, Madagas-car: Escape 2 Africa. Huangs responsibilities includeworkingwith the directors, production designers, and animators tointerpret the style and movement of characters in a 3Denvironment.

Huang joined PDI/DreamWorks in 1999 as a CharacterTD on the Academy Award-winning animated feature,Shrek. She was quickly promoted to Senior Character TDon Shrek 2 and then to her current supervisory role whileworking on Madagascar. Prior to joining PDI/DreamWorks,Huang spent three years as a Software Engineer at Mult-Gen Inc., writing functional specications and implementingdeformation tools for the software Creator, a real-time 3Dmodeling software package. She also participated as a VirtualReality Room (VROOM) exhibitor in SIGGRAPH1994 for herco-development in mathematical visualization and applica-tion for rapid prototyping.

Huang received her MS from the University of Illinois inChicago, writing her thesis on Untangling Knots with EnergyMinimizationbySimulatedAnnealing. She alsoholds aBS inArt and Computer Science from the University of Wisconsin.

Q&AQ) The components of character technology are diverse and spanthe artistic and technical spectrum. How did you get your startin the industry? Based on where you see the eld heading, whatexperience/training do you recommend people have?

A) My interest in computer animation started when I was inhigh school. My Dad would chat with me about what kind ofjob I might like when I was older. I realized that I enjoyed art 21


and science. We looked around and noticed this thing calledcomputer graphics was starting to happen. It was starting tobe used in commercials, movies, and print publications. Thatkind of work would need artistic and technical skills! At thetime I went to college, there were no computer graphic pro-grams, so I majored in art and computer science to learn thebasic skills that I hoped to combine. After earning my degreeat theUniversity ofWisconsinMadison, I went to theUniver-sity of IllinoisChicago for graduate school. At the universitysElectronic Visualization Laboratory, my studies and researchinvolved computer graphics and virtual reality. From there,I soon started working at PDI/DreamWorks as a CharacterTechnical Director.

I recommend that folks interested in this kindofwork learnthe artistic and technical skills of character setup and anima-tion. I would emphasize learning the technical fundamentals,including computer programming and the mathematics usedin computer graphics. These are important tools in ones tool-box of character setup skills so that one can easily innovateon top of evolving software and paradigms.

Q) Creating compelling digital characters is part art and partscience. What is the greatest challenge you encounter on a dailybasis? What aspect of character technology keeps you up thinking atnight?

A) Our characters are built to have a large range of motionand to look great when viewed from all angles. Ensuringthat the characters are developed and then supported in amanageable way to achieve these goals is a daily excitingchallenge.

There are so many interesting character technology issuesto think about. An animator may need differing types ofcontrols on the same character depending on the type ofacting required. Building a character that allows the anima-tor to easily switch and blend between differing systems isa challenge. Some example systems are forward kinematicverses inverse kinematic control, physical simulation verses

Interview 23

hand animated, and bipedal verses quadrupedal movement.Another issue we think about is speed: how can we improvecharacter setup tools so that we can prototype character solu-tions more quickly, and how can we make the character runfaster so animators can work more efciently.

Q) What are the ingredients for a successful digital character?

A) Here are three main ingredients:

Animator controls that are easy to use

Beautiful deformations that result in a model thatrespects the design intent on every frame

Character setups that run at reasonable speed

Q) While the industry is competitive, it is ultimately respectful ofgreat work. Outside of the work you have been personally involvedin the production of, what animated characters have you beenimpressed or inspired by? What is the historical high-bar?

A) The Lord of the Rings trilogy comes to mind as anexcellent example of pushing the envelope further. The crowdcharacters created such amazing battle scenes andhad impres-sive abilities to intelligently react to other characters and theenvironment. Gollum created a wonderful performance.

Q) Where do you see digital characters in ten years? Where areanimated characters for lm and games heading?

A) Visual media that is authored for stereoscopic displaywill increase to give audiences a more immersive experience.This will affect digital characters as we determine not onlywhich two-dimensional tricks we use no longer apply butalso what new tricks we can apply to enhance our charactersperformance in a stereoscopic viewing.

There is also a great branching of character develop-ment where, on one hand, we see a push for hyper-realism,where it will be difcult to determine if leading characters


are real or digital. We also see a push for more cartoony,stylized characters. We will be better able to translate hand-drawn character designs into digital characters that defy thestandard three-dimensional physical rules but have their ownset of visually stylized rules.

Chapter 3History of Digital Characters

3.1 INTRODUCTION

This chapter deals with the evolution and history of digitalcharacters. Digital characters developed in response to theprojects that they were employed in. By using an evolutionarymodel to frame the historical development of digital charac-ters, we group innovations into ve stages leading us intocontemporary projects and hints at what is ahead. The his-tory of digital characters follows and paves the way for theprimary concepts in this book.

3.2 THE EVOLUTION OF DIGITAL CHARACTERS

The scale of history for digital characters is not a large one.Digital characters found their start and some might arguegreatest success in the realm of interactive video games. Theact of arranging pixels into images in the form of characters ofsome variety goes back to the rst days of computer imaging.The moving of these pixels in two dimensions (2D) over time(also known as animation) was not far behind. Alongsidedevelopments into character motion came character inter-activity, complex imagery, and eventually three-dimensional(3D) graphics. The artistry and the technology tended toadvance together, each pushing the other, and they continueto do so today. As graphics hardware provides systems with 25

26 CHAPTER 3 History of Digital Characters

more power for real-time graphic processing, the level ofinteractivity and visual complexity available increases. Thisis a constant give and take and in many ways sets the pacefor digital character technology development. Before we lookat the historical path of digital characters in various forms ofmedia, it is worth taking a look at the changing technologyand methods that have brought digital characters to wherethey are today and which may give us insight into where theywill go in the future.

If we were to frame the development of digital characters, ageneral framework for our discussion on the history of digitalcharacters may be to look at this history from an evolutionaryperspective. Let us use as a springboard the path proposedby artist and animator Matt Elson [11]. Elsons ve stagesof digital character evolution deal with the technology andmethods with which 3D character animation is carried out.Each stage is characterized by a technical advancement inter-related with a paradigm shift regarding motion.

Stage one in Elsons schema is typied by keyframe ani-mation or the process of positioning a character every fewframes then letting the system interpolate the in-betweenframes. Animation at this stage is modeled on the work-ow of 20th century cell animation, which included a mas-ter animator who drew the keyframes and a more juniorin-betweener who did what was expected of their title.A primary technical innovationof this phase is programmableexpressions, in which technical directors dene interactiverelationships in the 3D environment. These expressionsprovide the opportunity for complicated, non-hierarchicalrelationships within the character to add automatic sec-ondary motion for example. Another example of a pro-grammable expression is the exing of associated skin ormuscles when a joint is bent. Other noteworthy develop-ments of the keyframe animation era include restructuredhierarchical objects, 3D paint, and socking and skinning.Critically, keyframe animation is said by Elson to be cumber-some and requiring large crews and complex infrastructures

3.2 The Evolution of Digital Characters 27

with a high ratio of technical-to-creative talent. Control of thecharacters is generally achieved at the expense of productionefciency.

Stage two in Elsons schema is layered animation andsingle skins. While largely a technical stage, it involves thecreation of a re-architected workow that enables animators,programmers, developers, artists, and technical directors toworkon individualmodules that address their respective areasof production. These modules are contained and insulatedfrom the larger production until they can be plugged in withthe other components. These parallel work ows make it pos-sible for large teams of artists to work simultaneously on aproject. This restructuring of pipelines and process happensin part due to improved Application Programming Interfaces(API), which allow third-party developers to write object-oriented system extensions and plug-ins. There is also a moveaway from straightforward animation production to layerableanimation, allowing different motion types to be combinedon the same character.

The third stage described by Elson is one he terms scriptedmemory and behavior. This stage is typied by the dynamicanimation of crowds via scripted actions. As we know, crowddynamics such as these are already being implemented.Complex animation and the creation of realistic natural phe-nomenon using tools such as procedural systems for creatinghair, fur, cloth, clouds, smoke, and grass are implementedin this stage as are dynamics and physics tools for simulat-ing gravity, motion damping, and wind. On the charactertechnology side, anatomical systems for simulating muscle,bone, fat, and other underlying structures of the body in theform of layered deformations are included. Motion capturewill become merely another animation method rather than alast resort. Skeletons will become smarter with built-in func-tions for pose mirroring and position and motion memories.Animation clips or as Elson calls them performance blockswill be kept in a library and reused as needed, for example incrowds.

28 CHAPTER 3 History of Digital Characters

It is safe to say that we have achieved stage three of Elsonsschema. Stages four and ve are where Elsons predictionsturn toward the slightly more fantastic and futuristic. Keep inmind, as we will nd later, the notions encapsulated by thesestages are current topics in character technology research anddevelopment both in academia and industry.

Elsons fourth stage, character autonomy, is where truecharacter independence begins. In this particular stage,embedded physics systems in the character will enable it tomove about an environment with weight, mass, and improvi-sational behaviors giving the appearance, though not yet thereality, of intelligence. He provides a note for the fearful ani-mator here: Animators will still be in great demand to craftall of a characters behaviors, actions, and response [11].

The fth and nal stage of Elsons evolution of digi-tal characters is personality. Here, Elson departs from therealm of procedural animation to true AI. Elson describesthis stage as characters that have developed emotional andcognitive interior lives. He argues they will begin to thinkand interact with their environment, each other, the ani-mator, and end-user, employing knowledge structures fortheir basic decision-making. Characters will understand andrespond to human speech, so they will be able to trans-late, perform menial tasks, and sift and sort the digitalexpanses.

Elson concludes by pointing out that his stages representan ideal world and that future developmen

digital character development

Documents

overview of digital

history of digital characters

dening digital characters

creation of digital

digital characters92

character technology

morgan kaufmann publications

elsevier homepage http