Chapter 8: Commercial Animation Software

8.1 Introduction

The following text was excerpted from an article by David Sturman, from the ACM SIGGRAPH Retrospective series published in the SIGGRAPH newsletter, Vol.32 No.1 February 1998. The entire article can be read at
https://ohiostate.pressbooks.pub/app/uploads/sites/45/2017/09/sturman-animation.pdf

Perhaps one of the earliest pioneers of computer animation was Lee Harrison III. In the early 1960s, he experimented with animating figures using analog circuits and a cathode ray tube. Ahead of his time, he rigged up a body suit with potentiometers and created the first working motion capture rig, animating 3D figures in real-time on his CRT screen. He made several short films with this system, called ANIMAC. This evolved into SCANIMATE which he commercialized to great success in 1969. SCANIMATE allowed interactive control (scaling, rotation, translation), recording and playback of video overlay elements to generate 2D animations and flying logos for television. Most of the 2D flying logos and graphics elements for television advertising in the 1970s were produced using SCANIMATE systems. In 1972 Harrison won an Emmy award for his technical achievements [25]. As computer graphics systems became more powerful in the 1980s, Harrison’s analog systems began to be superseded by digital CG rendered keyframe animation, and now are no longer used in production.

The next widespread system was the GRAphics Symbiosis System (GRASS) developed by Tom DeFanti [at Ohio State University] for his 1974 Ph.D. thesis. GRASS was a language for specifying 2D object animation and although not interactive, it was the first freely available system that could be mastered by the non-technical user. With GRASS, people could script scaling, translation, rotation and color changes of 2D objects over time. It quickly became a great hit with the artistic community who were experimenting with the new medium of CG. In 1978 it was updated to work in 3D with solid areas and volumes and ran on a Bally home computer. This version was called ZGRASS, and also was important in bringing computer graphics and animation to the artistic community on affordable computing platforms [6].

Also in 1974, Nestor Burtnyk and Marcelli Wein at the National Film Board of Canada developed an experimental computer animation system that allowed artists to animate 2D line drawings entered from a data tablet. Animation was performed by point-by-point interpolation of corresponding lines in a series of key frames. The system was used for 1974 classic short film Hunger whose graceful melding of lines from one figure to the next won it an Academy Award nomination.

The New York Institute of Technology Computer Graphics Lab (NYIT), then under the direction of Ed Catmull, extended this idea, producing a commercial animation system called TWEEN. As with the National Film Board system, TWEEN was a 2D system that allowed the animator to draw key frames, and the computer interpolated corresponding line segments between the keys. TWEEN automated the process of producing in-between frames (sometimes called tweening), but still required the talents of a trained artist/animator for the keyframes. Although this method sped up the hand-animation process, animations produced this way had an overly distinctive fluid look and the method was not widely adopted for commercial animation.

The first complete 3D animation systems were typically in-house tools developed for use in particular academic environments or production companies. They could be categorized into two types of systems, scripted or programmed systems, and interactive keyframe systems. The first type were exemplified by ANIMA-II (Ohio State) [11], ASAS [23], and MIRA [16]. All three used a programming language to describe a time sequence of events and functions. When evaluated over time and a “snapshot” rendered at each animation frame, they produced the desired animation. ASAS is noteworthy since many of the CG sequences in the 1982 film TRON were animated with it. These systems were powerful in that almost anything could be done if it could be programmed, but limited in that programming skills were required to master them.

The keyframe systems were more amenable to animation artists. Based on the keyframe approach of traditional animation, these systems allowed the user to interactively position objects and figures in the scene, save these positions as keyframes and let the computer calculate the in-between frames to produce the final animation. GRAMPS [19] and BBOP [28] were examples of this type of system. Both relied on the real-time interactivity of the, then state-of-the-art, Evans & Sutherland Multi-Picture System, an excellent vector-graphics display system that worked from a display list allowing instantaneous updates of the on-screen graphics.

GRAMPS was developed for visualization of chemical structures although O’Donnell does give examples of how it could be used to animate a human figure. Ostensibly an interpreted script system, GRAMPS allowed script variables to be connected to dials for interactive manipulations.

An interesting aside: 2D and 3D animation.

BBOP was developed at the New York Institute of Technology’s Computer Graphics Lab (NYIT) by Garland Stern expressly for character animation and was used extensively by NYIT in six years of commercial production. In BBOP, animators could interactively control joint transformations in a 3D hierarchy, saving poses in keyframes which the computer could interpolate to produce smooth animation. The system was very responsive and easy to use, and conventionally trained animators produced some remarkably good animation with it. Examples include a CG football player for Monday Night Football promotions, Rebecca Allen’s CG dancer for Twyla Tharp’s dance The Catherine Wheel, Susan Van Bearle’s short film Dancers and numerous SIGGRAPH film show shorts featuring the digital characters User Friendly, Dot Matrix and User Abuser. These last three were some of the first CG characters to have expressive personalities that engaged the audience and brought CG to life. Much of this was due to an interactive keyframe system that gave the animator the control and freedom to manipulate the figures visually, in keeping with his training and experience.

Most modern commercial keyframe systems are based on the simple BBOP interactive keyframe approach to animation with added features that ease the animation process. At their core, they all have features of BBOP (some copied, some developed independently), including hierarchical skeleton structures, real-time interactive update of transformation values, interpolation of keyframes in channels so that different joints can have different keys in different frames, choice of interpolation functions such as linear, cubic, ease-in and ease-out, immediate playback and an interpolation editor.

In general, however, scripted systems still are best for repeated or easily describable movements, but require programming skills beyond the capabilities of most artists, especially as movements become more complex. Scripting expressive characters, for example, is extremely difficult, not to mention unnatural for an artist. Interactive keyframe systems are just the opposite. They allow artists to interact directly with the objects and figures within a familiar conceptual framework. But they become inefficient or tedious to use for mechanical or complex algorithmic motion. Because they are more easily used by artists, the interactive keyframe approach has won in the commercial software market. Curiously enough, as animators are becoming more sophisticated in their use of computer animation, scripting capabilities are beginning to reappear in keyframe systems. The newest version of Wavefront|Alias’ MAYA animation system has a built-in scripting capability that allows animators to tie actions to events, define movement as functions of other movements, create macros and more.

Early 3D animation systems mostly dealt with simple forward kinematics of jointed bodies, however inverse kinematics can also be an important element in an animation toolkit. By moving just a hand or a foot, the animator can position an entire limb. Michael Girard [at Ohio State University] built a sophisticated inverse kinematic animation system for his Ph.D. thesis [9] which was used for producing very graceful human body movement in his 1989 film Eurhythmy. He later commercialized his system as a 3D Studio MAX plug-in, Biped (part of the Character Studio package), where legged locomotion such as walks, runs, jumps and skips can be animated by placing footprints. His inverse kinematic algorithms compute the motions of the figure that cause it to follow the footprints.

When Softimage was first released, it was the first commercial system to feature an inverse kinematics capability (although in a simplified form). That feature helped greatly in selling the new system. Now, almost all 3D animation systems have some form of inverse kinematic capabilities.

Dynamics is also an important tool for realistic animation. Jane Wilhelms was one of the first to demonstrate the use of dynamics to control an animated character [31]. Since then, James K. Hahn (Ohio State), David Baraff and Michael McKenna [12, 2, 17] have all described robust dynamics for computer animation. Yet it is only in the past few years that the major commercial systems are incorporating dynamics into their software. The problems they are facing are how to integrate dynamic controls, inverse kinematic controls, and forward kinematic controls within the same system, and presenting and resolving clearly the potentially conflicting constraints each puts on the animated elements.

Kinematics and dynamics deal with jointed skeletal structures. However, not all animation is skeletal. A face, for example, is a single surface with complex deformations. Fred Parke was the first to attack this problem [20] with a parametric facial model. Using parameters to describe key aspects of facial form, such as mouth shape, eye shape and cheek height, and then animating these parameters, he was able to simulate the motions of a human face as a single surface. The system was used by NYIT in a music video for the group Kraftwerk, but never commercialized.

Years later, Philippe Bergeron and Pierre Lachapelle digitized plaster models of several dozen expressions of a face, and created a system to interpolate between several of these target expressions at once for their 1985 short film Tony de Peltrie [3]. The result was a rubbery-faced character with a wide range of human expression. Rudimentary implementations of this technique of 3D object interpolation (or 3D target morphing) were incorporated into Softimage and Alias|Wavefront systems a few years ago, and are being improved for the latest versions of their software. 3D target morphing is also the basis of Medialab’s real-time character performance animation system.

Keith Waters developed an even more sophisticated facial animation system based on muscle activation influencing regions of the face model [30]. This system produces very realistic facial motion and can be controlled by high-level commands to the muscle groups. His methods have not been commercialized, but simpler versions are used in some optical facial motion capture systems.

There follow a whole host of bits and pieces to animate particular effects. Some of these have been integrated into commercial animation systems. Others are used exclusively by the companies that developed them, while still others have just seen proof of concept and await a plug-in or incorporation into a more complete system.

The most influential of these (and perhaps not really in the bits and pieces category) is Bill Reeve’s particle systems [22]. Reeves developed a method of using controlled random streams of particles to simulate fire, grass, sparks, fluids and a whole host of other natural phenomena. First used in the movie Star Trek II, particle systems are easy to implement and quickly appeared in many amateur, academic and professional CG animations, most notably Particle Dreams in 1988 by Karl Sims. Commercial animation systems took a little longer to incorporate the technique into their established structures, but today everyone has it in some form or another.

Other animation techniques for specific effects in the literature include (but by no means are limited to) automated gaits (walking, running, jumping, etc.) [5, 13], flocking behaviors [24, 1], fluid flow [14], waves [7, 21], smoke [27], sand [15], flexible objects [29], snakes [18], cloth [29] and many more.

As was already mentioned, the most difficult animation is character animation, particularly human character animation. In a quest for more realistic motion, people have looked towards directly recording the motions of a human performer. Lee Harrison III in the 1960s was only the first of many to use this concept. In 1983 Ginsberg and Maxwell [8] presented an animation system using a series of flashing LEDs attached to a performer. A set of cameras triangulated the LEDs’ positions, returning a set of 3D points in real time. The system was used soon after to animate a CG television host in Japan. However motion capture systems and graphics computers were just not fast enough then for the real-time demands of performance animation.

When they did begin to become fast enough, around 1988 with the introduction of the Silicon Graphics 4D workstations, deGraf/Wahrman and Pacific Data Images both developed mechanical controllers (also known as waldos) to drive CG animated characters — deGraff/Wahrman for CG facial animation for a special SIGGRAPH presentation and for the film Robocop II, and PDI for a CG character for a Jim Henson television series and several other projects. For various reasons the technology and market were not ready and the systems were rarely exploited after their initial use.

Then, in the early 90s, SimGraphics, Medialab (Paris) and Brad deGraf (with Colossal Pictures and later Protozoa) all independently developed systems that allowed live performers to control the actions of a CG character in real time. These systems allowed characters to be animated live, as well as for later rendering. The results, particularly with Medialab’s system, are characters that have very lifelike and believable movements. Animation can be generated quickly by actors and puppeteers under the control of a director who has immediate feedback from the real-time version of the character. All three systems have survived their initial versions and applications, and continue to be successfully used in commercial projects.

At first, these systems existed on their own and were not integrated into other commercial CG systems. Animation done in a keyframe system could not easily be mixed with animation performed in a real-time system. As time has passed, both the real-time systems and the keyframe systems have evolved, and now many keyframe systems have provisions for real-time input and the real-time systems import and export keyframe animation curves.

Performance animation has become very popular recently and at the SIGGRAPH 97 trade show, no less than seven companies demonstrated performance animation systems.

Similar to performance animation, but without the real-time feedback, are motion capture systems. These are generally optical systems that use reflective markers on the human performer. During the performance, multiple cameras calculate the 3D positions of each marker, tracking it through space and time. An off-line process matches these markers to positions on a CG skeleton, duplicating the performed motion. Although there are problems with losing markers due to temporary occlusions and the animation matching process can be very labor-intensive, motion capture permits an accurate rendering of human body motion, particularly when trying to simulate the motion of a particular performer as Digital Domain did with Michael Jackson’s 1997 music video, Ghosts.

1. Amkraut, Susan and Michael Girard. “Eurhythmy: Concept and Process,” Journal of Visualization and Computer Animation, v.1, n.1, John Wiley & Sons, West Sussex, England, August 1990, pp. 15-17.

2. Baraff, David. “Analytical Methods for Dynamic Simulation of Non-penetrating Rigid Bodies,” Computer Graphics, proceedings of SIGGRAPH 89, ACM SIGGRAPH, New York, NY, pp. 223-232.

3. Bergeron, Philippe and Pierre Lachapelle. “Controlling Facial Expressions and Body Movements in the Computer-Generated Animated Short Tony De Peltrie ,” SIGGRAPH 85 Advanced Computer Animation seminar notes, ACM, New York, 1985.

4. Blumberg, Bruce M. and Tinsley A. Galyean. “Multi-Level Direction of Autonomous Creatures for Real-Time Virtual Environments,” Computer Graphics, proceedings of SIGGRAPH 95, ACM SIGGRAPH, New York, NY, pp. 47-54.

5. Bruderlin, Armin and Thomas W. Calvert. “Goal-Directed, Dynamic Animation of Human Walking,” Computer Graphics, proceedings of SIGGRAPH 89 Proceedings, ACM SIGGRAPH, New York, NY, pp. 233-242.

6. DeFanti, T., J. Fenton and N. Donato, “BASIC Zgrass — A sophisticated graphics language for the Bally home computer,” Computer Graphics, proceedings of SIGGRAPH 78, ACM SIGGRAPH, New York, NY, pp. 33-37

7. Fournier, Alain and William T. Reeves. “A Simple Model of Ocean Waves,” Computer Graphics, proceedings of SIGGRAPH 86, ACM SIGGRAPH, New York, NY, pp. 75-84.

8. Ginsberg, Carol M. and Delle Maxwell. “Graphical marionette,” Proceedings ACM SIGGRAPH/SIGART Workshop on Motion (abstract), Toronto, Canada, April 1983, pp. 172-179.

9. Girard, Michael and Anthony A. Maciejewski. “Computational Modeling for the Computer Animation of Legged Figures,” Computer Graphics, proceedings of SIGGRAPH 85, ACM SIGGRAPH, New York, NY, pp. 263-270.

10. Grzeszczuk, Radek and Demetri Terzopoulos. “Automated Learning of Muscle-Actuated Locomotion Through Control Abstraction,” Computer Graphics, proceedings of SIGGRAPH 95, ACM SIGGRAPH, New York, NY, pp. 63-70.

11. Hackathorn, Ronald J. “Anima II: a 3-D Color Animation System,” Computer Graphics, proceedings of SIGGRAPH 77, ACM SIGGRAPH, New York, NY, pp. 54-64.

12. Hahn, James K. “Realistic Animation of Rigid Bodies,” Computer Graphics, proceedings of SIGGRAPH 88, ACM SIGGRAPH, New York, NY, pp. 299-308.

13. Hodgins, Jessica K., Wayne L. Wooten, David C. Brogan and James F. O’Brien, “Animating Human Athletics,” Computer Graphics, proceedings of SIGGRAPH 95, ACM SIGGRAPH, New York, NY, pp. 71-78.

14. Kass, Michael and Gavin Miller. “Rapid, Stable Fluid Dynamics for Computer Graphics,” Computer Graphics, proceedings of SIGGRAPH 90, ACM SIGGRAPH, New York, NY, pp. 49-57.

15. Li, Xin and J. Michael Moshell. “Modeling Soil: Realtime Dynamic Models for Soil Slippage and Manipulation,” Computer Graphics, proceedings of SIGGRAPH 93, ACM SIGGRAPH, New York, NY, pp. 361-368.

16. Magnenat-Thalmann, N. and D. Thalmann. “The Use of High-Level 3-{D} Graphical Types in the Mira Animation System,” IEEE Computer Graphics and Applications, v.3, 1983, pp.9-16.

17. McKenna, Michael and David Zeltzer. “Dynamic Simulation of Autonomous Legged Locomotion,” Computer Graphics, proceedings of SIGGRAPH 90, ACM SIGGRAPH, New York, NY, pp. 29-38.

18. Miller, Gavin. “The Motion Dynamics of Snakes and Worms,” Computer Graphics, proceedings of SIGGRAPH 88, ACM SIGGRAPH, New York, NY, pp. 169-178.

19. O’Donnell, T.J. and A. J. Olson. “GRAMPS — A graphics language interpreter for real-time, interactive, three- dimensional picture editing and animation,” Computer Graphics, proceedings of SIGGRAPH 81, ACM SIGGRAPH, New York, NY, pp. 133-142.

20. Parke, Frederic I. “Computer Generated Animation of Faces,” Proceedings ACM annual conference, August 1972.

21. Peachey, Darwyn R. “Modeling Waves and Surf,” Computer Graphics, proceedings of SIGGRAPH 86, ACM SIGGRAPH, New York, NY, pp. 65-74.

22. Reeves, W. T. “Particle Systems — a Technique for Modeling a Class of Fuzzy Objects,” ACM Trans. Graphics, v.2, April 1983, pp. 91-108.

23. Reynolds, C. W. “Computer Animation with Scripts and Actors,” Computer Graphics, proceedings of SIGGRAPH 82, ACM SIGGRAPH, New York, NY, pp. 289-296.

24. Reynolds, Craig W. “Flocks, Herds, and Schools: {A} Distributed Behavioral Model,” Computer Graphics, proceedings of SIGGRAPH 87, ACM SIGGRAPH, New York, NY, pp. 25-34.

25. Schier, Jeff. “Early Scan Processors– ANIMAC/SCANIMATE,” in Pioneers of Electronic Art, Exhibition catalog, Ars Electronica 1992, ed. David Dunn, Linz, Austria, 1992, pp. 94-95.

26. Sims, Karl. “Evolving Virtual Creatures,” Computer Graphics, proceedings of SIGGRAPH 94, ACM SIGGRAPH, New York, NY, pp. 15-22.

27. Stam, Jos and Eugene Fiume. “Turbulent Wind Fields for Gaseous Phenomena,” Computer Graphics, Proceedings of SIGGRAPH 93, ACM SIGGRAPH, New York, NY, pp. 369-376.

28. Stern, Garland. “Bbop – a Program for 3-Dimensional Animation,” Nicograph ’83, December 1983, pp. 403-404.

29. Terzopoulos, Demetri, John Platt, Alan Barr and Kurt Fleischer. “Elastically Deformable Models,” Computer Graphics, proceedings of SIGGRAPH 87, ACM SIGGRAPH, New York, NY, pp. 205-214.

30. Waters, Keith. “A Muscle Model for Animating Three- Dimensional Facial Expression,” Computer Graphics, proceedings of SIGGRAPH 87, ACM SIGGRAPH, New York, NY, pp. 17-24.

31. Wilhelms, J. and B. A. Barsky, “Using Dynamic Analysis to Animate Articulated Bodies such as Humans and Robots,” Graphics Interface ’85 Proceedings, Montreal, Quebec, Canada, May 1985, pp. 97-104.

For an historical overview of the development of animation systems in general, see the summary of Steve May’s 1998 PhD dissertation. https://ohiostate.pressbooks.pub/app/uploads/sites/45/2017/09/smay-animation.pdf

For the entire dissertation,
Encapsulated Models: Procedural Representations for Computer Animation.
Stephen F. May. PhD thesis, The Ohio State University, March 1998.
go to

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.128.1095&rep=rep1&type=pdf

 

License