USAGE OF TRADITIONAL ANIMATION
PRINCIPLES IN THE VIRTUAL
ENVIRONMENT TOONTOWN
Ruud Knieriem Ralph van der Meulen Joost van Wollingen
ABSTRACT
3D environments are being used more and more. However, there is no definitive methodology that supports the development process of these Virtual Environments. Some methodologies have been proposed, such as the methodology by [Fencott 1999] that focuses on Perceptual Opportunities. This paper aims to give a overview of some of these development methodologies, including using traditional animation techniques. Following on this a qualitative study of the Toontown environment is described, aimed at finding similarities between the traditional animation techniques and the animation techniques used in this virtual environment to define the importance of animation and relations to other processes.
KEYWORDS: animation techniques, virtual environment, Toontown.
1. INTRODUCTION
In the last few years many new multi-user online virtual environments have been created. These environments range from full-fledged online games like World of Warcraft [Ducheneaut 2006] to less gaming-oriented creations such as Second Life [Rymaszewski 2006]. The design process of Virtual Environments (VE) is however not an official standardized process accepted by developers.
[Fencott 1999] describes the term of designing an effective Virtual Environments (VEs) as two distinct interpretations, namely the aesthetic notion of designing something to create the desired perceptual responses and secondly the engineering notion of design as the creation of plans and models from which to test and build the desired artifact. [Fencott 1999] created a model, which will be discussed later on, that should support the design of VEs in both of these fields.
As discussed by [Bink 2007], animation has been around for a while. Drawing classes set up by Disney [Thomas 1981] for their animators at the Chouinard Art Institute in Los Angeles eventually led to the development of animation procedures to use as rules of the trade for new artists. Rules of the trade were gradually developed by animators when lessons of drawing classes had to be better communicated to new artists. These fundamental animation principles will be used in this paper and further discussed in the following section.
The goal of this paper is to take a closer look at animation used in Disney’s VE called Toontown Online [Mine 2003]. The animation principles designed by Disney animators themselves (!) will act as guidelines to test the quality of animation used in Toontown Online. By analysing the use of the animation principles in a VE the authors of this paper hope to find the relationship between the design of animations in the conceptual modeling and other processes defined in the process model mentioned in [Fencott 1999].
2. RELATED WORK
In this section several existing literature is described to give an overview of first the Toontown application and secondly related theories in the field of animation in virtual environments. The part about Toontown describes design issues that are considered when designing the application and gives an overview of the main aspects of the game. The theoretical part is giving a description of some major theories in the field of animation, varying from application in embodied agents to more general principles.
2.1. Toontown
Disney’s Toontown Online [Mine 2003] is an online 3D massive multiplayer game for children of ages seven and older. When designing such a massive multiplayer game lots of design issues have to be considered. Mine et al. described in their 2003 study how they handled some of these issues.
First of all the designers wanted the game to be positive, more or less in contrast to classical multiplayer games like World of Warcraft [Ducheneaut 2006], which can be a turn-off for the mass market due to its violent character. To achieve positivity, cartoony characters and colorful environments were developed. Also a lot of gags were implemented in the game to increase humor.
The animation issues concerning this design encompass both the environment and the actions and
appearances of the characters themselves. As is the Disney tradition, seen in their theme parks all over the world, the environment should consist of a lot of colors and nice shapes. Humor can be supported by the animated gags in the game.
Another design aspect focused on was the communication between players in the game. To ensure a child-friendly environment, global chat was kept safe by only allowing chatting with pre-defined phrases. This chat-system is named SpeedChat by the developers. The phrases can be chosen from an easy to use in-game menu. Only when real-life friends are met within the in-game (identified by using password-sharing outside of the VE), regular chat is available. Another part of the game to give the participants a feeling of safety was the humorous name generator, with which an appropriate character name can be generated. Own names are checked by game moderators to prevent abuse. These issues have to be supported by animation by showing emotion and speech of other players. Also the visibility and recognizability of other players in the game is an important issue, because this will inform users that communication is possible. Besides the possibility to use chat, other social activities that are available within Toontown Online are fishing, mini-games and battles against Cogs. Grouping and teaming up is made very easy by the developers, a battle or game can be joined by just walking up to it. The focus of the designers was on cooperative play, both new and more experienced players are stimulated to play together. Several mechanisms are implemented to protect fair play.
Another main design issue when developing Toontown Online was discouragement of griefing. Griefing is behavior of players who try to ruin the experience of others, verbally and by abuse of game features. In contrast to this, it has been made very easy to find friends and other previously met players. A final design issue can be summarized as keeping the design and controls as simple as possible, using graphically represented information.
An important issue for the designers of Toontown Online was whether they would develop their own engine or use an existing one. They chose to create an engine from the ground, because of several reasons. One of them was the amount control it would give over the features of the programme, which gave the designers the possibility to control the way in which animations were designed. The 3D engine that was developed was called Panda-3D. The architecture of Panda was designed to support the design issues described in the previous section. An important technical aspect that was considered, is the use of
downloading is a technique with enables an application to start running as fast as possible and then keeps downloading the rest of the content that is not directly needed. Other advantages are the open source nature of the engine, the possibility to use standard programming languages and the rapid prototyping. The latter means that developers can make changes to their code without recompiling, making fast
development and changes possible.
Because Toontown is an online game for the mass market it should be playable over narrow-band connections. Therefore some animations have to be compressed using a lossy method, resulting in some minor loss of information. Some are left decompressed, because otherwise unacceptable artifacts occur.
2.2. Embodied agents
The next three sections describe some theories about the application of animation in virtual environments. A very specific one is found within the domain of embodied agents. Embodied agents are human
representations in a virtual environment, which are often used to assist or help the user in achieving tasks like learning and training. Essential factors for the successfulness of such an embodied agents are found in a 1997 study of Johnson & Rickel [Johnson 1997]. They showed that such an application can be a good solution to support users in learning both simple and complex tasks. Several issues are concerned, but what they found was that a full human body compared to only parts is preferred. Animation should be the facilitator of such an embodiment. Another thing concerning animation they concluded was that an agent who provided also non-verbal communication to the user was better in reaching the goal.
The results from this short examination of the use of animation in embodied agent applications might be useful in defining whether animation has been applied in the right way in the Toontown game.
2.3. Design of Virtual Environments
Besides applications of animation in virtual environments as described in the previous section, others described in a more general way issues about designing animation. But even though the use of Virtual Reality (VR) and 3D environments is growing vigorously, the actual development process of Virtual Environments (VE) is not well documented or researched to its full extent. In order to design an effective Virtual Environment two major problems have to be overcome. First, the designer has to determine how to create something that is aesthetically pleasing to the user and there is the engineering side of designing the models from which to test and build [Fencott 1999].
In general it can be said that the designer should find an optimal balance in the tension between aesthetics design and engineering design. The structural part of the modeling can be supported through software engineering practice, but this does not hold for the perceptual modeling. Also, the structural part cannot be modeled before the perceptual modeling takes place or vice versa. Both have to be done concurrently and interaction has to take place during this process [Fencott 1999].
Based on Kaur’s 5 step methodology Fencott has designed a model, specifically tailored to combine the structural and perceptual modeling of a VE (Figure 1). Fencott’s model consists out of almost the same five phases Kaur identified, but presents these in a non-linear fashion.
The perceptual modeling phase which relates to Kaur’s point 5 is concerned with the act of modeling the perceptual opportunities and their inter-relationships [Fencott 1999]. This means that the intended user experience is being modeled in this phase.
The structural modeling phase relates to Kaur’s third point and runs alongside the conceptual and perceptual modeling. Use cases and UML(Unified Modeling Language) can be used here to create the layout of the code structure of the VE [Fencott 1999].
Figure 1:Informal model of Fencott’s VE design methodology
Fencott hopes the proposed model helps focus further research and discussion on the whole process of developing and implementing virtual environments.
Fencott describes in another paper [Fencott 1999a] a perceptual mapping technique that models perceptual opportunities within the VE. Perceptual opportunities can be anything within the VE, any imaginable object or concept. The perceptual opportunity (PO) is assigned to a certain category, which indicates the effect it is supposed to instigate in the user. Therefore PO’s can be seen as the basic units of meaning of the virtual content within the VE. There are three main principal forms PO’s can take: sureties, suprises and shocks.
Shocks are unintended impulses for the user drawing their attention away from the VE. An example given by Fencott [Fencott 1999a] states users were shocked that they could see were the sea ended. Shocks break users away from the VE and have a negative effect on feelings of presence, and thus should be avoided. Sureties are mundane and highly predictable details. They are attractive to the user because of their predictability. For example, railings at the pier give an indication of the size of the virtual appearance of the user, but do not necessarily have to look like the railings on a real pier.
Finally, surprises are PO’s that emphasize certain aspects of the VE. They do not have to be plausible to be beneficial. They can be split up in three categories, which are: attractors, connectors and rewards. Attractors seek the attention of the user to a certain place or moment. Connectors help users to reach goals set for them through attractors. Rewards are specific experiences within the VE that are memorable to the user, also helping in keeping the attention of the user focused on the VE.
Fencott concludes that the Perceptual Opportunities model helps the designer in creating a sound PO’s, which can later on be used to create a full-fletched high-quality VE. The method should not be used with the intention to create large list of surprises and attractor graphs.
2.4. Disney’s animation techniques
1. Squash and Stretch
Defining the rigidity and mass of an object by distorting its shape during an action. 2. Timing
Spacing actions to define the weight and size of objects and the personality of characters. 3. Anticipation
The preparation for an action. 4. Staging
Presenting an idea so that it is unmistakably clear. 5. Follow Through and Overlapping Action
The termination of an action and establishing its relationship to the next action. 6. Straight Ahead Action and Pose-To-Pose Action
The two contrasting approaches to the creation of movement. 7. Slow In and Out
The spacing of the inbetween frames to achieve subtlety of timing and movement. 8. Arcs
The visual path of action for natural movement. 9. Exaggeration
Accentuating the essence of an idea via the design and the action. 10. Secondary Action
The action of an object resulting from another action 11. Appeal
Creating a design or an action that the audience enjoys watching.
A more elaborate description and application of the techniques in 3D animations can be found in [Lasseter 1987]. The use of these techniques for 3D user-interfaces is discussed by [Bink 2007]. Because Toontown Online is a 3D environment and not just a 3D animation, users can navigate freely in the VE. Focusing on all 3D user-interfaces, [Bink 2007] concludes that:
“When applying them (Disney’s animation techniques) to interfaces, something like Pose-To-Pose
becomes useless because they are meant for creating stage, where the stories are linear but the user’s interaction is not.”
By this, the authors mean that the use of the Pose-to-Pose technique is somewhat limited to linear stories, because when applying the technique important parts of the animation are designed beforehand, thus there is a need for a linear story to fill in the gaps. As the authors state, this make it difficult to apply this technique in user interfaces, because these mostly are highly interactive. Although this can be true for most user interfaces, this is however not entirely true for VEs. The user’s interaction may not be linear, this does not mean the animations that define the actions (such as movement) of a user’s virtual self is not. Actually by defining small animations that together form the total animation can be a form of hierarchical
modeling, which is a form of Pose-To-Pose action modeling according to Lasseter.
Having defined a theoretical basis for the application of animation in virtual environments, for the current research we now apply the techniques of Lasseter to the Toontown game. It could be argued that also the design issues of Fencott should be taken into account, but because these are stated for another phase of development, we chose to only test the classical animation principles. Further research can be done to also test whether the perceptual opportunities of Fencott have been applied.
3. ANALYSIS OF TOONTOWN ONLINE
3.1. Method
To define whether the classical animation principles as defined by Disney really are applied in the several parts of Toontown Online, a qualitative analysis of the application was conducted. The authors of the current study spent time in the environment investigating the possibilities of Toontown Online and find applications of each of the principles. To check whether this was the case, the authors tried to go over a lot of different aspects of the game, varying from mini games and Cog battles to just walking around the streets of Toontown.
Toontown Online has been designed to equally divide time spent by users into three major activities, shown in Figure 2 [Mine 2003]. The figure shows that the game is designed in three main parts: the playground where games can be found, the actual games in the minigames part of the game, and the streets where the user can walk around and meet others. Each of the parts has been analysed independently of the other parts. This is done so that the gathered results will be as representative as possible in relation to the experience of the users.
Figure 2: Toontown trinity
3.2. Results
3.2.1. Squash and Stretch
Squash and stretch is visible when looking at any character in the game. When a Toon stands still it is breathing and its body stretches and squashes up and down to illustrate this. Also the use of the portable holes to go directly to another player has a scripted animation that includes the Toon using it to squash before the leap into the hole (Figure 3), and stretches himself out when jumping into it. This is a good example of squash and stretch without deforming the object.
Figure 3: Using the portable hole to teleport to someone
The exploding Cog (after defeat) is another (common) example of the use of squash and stretch. The Cog clearly first squashes parts of its body before an explosion sends all the parts flying.
It is however clear that a lot of animations do not use squash and stretch. Jumping with your Toon is an example of this which results in an unrealistic animation.
3.2.2. Timing
Most of the movement within the game is quite slow and easy to follow. This has probably been done on purpose, in order to make the game fitting for young children.
Small Toons seem to walk faster, due to their smaller size their feet touching the ground more often. This use of timing gives the impression that thesecharacters are fast but small. The use of timing here therefor defines the weight and size of objects.
The emotional state of the characters when being sad (after defeat in battle with a Cog) is also reflected in its movement speed.
3.2.3. Anticipation
There is little anticipation in actions that the user starts, such as walking around and jumping. This is probably done in order to keep the game responsive to the user actions, anticipation in these animations could be considered as slow or delayed responses. In animations user has no control of some anticipation occurs. When using a portable hole for example, the Toon bends his knees en bends over before jumping into the hole.
Another good example of anticipation in Toontown Online occurs when performing a gag in battle. The Toon takes a certain object, indicating that it will do something with it before actually performing the gag (Figure 4). After defeating a Cog, all Toons in the battle will protect themselves from the anticipated explosion of the destroyed Cog.
3.2.4. Staging
Toontown Online uses staging in battles between Toons and Cogs. The viewpoint of a user is changed so that it is obvious that a battle starts and takes place. Also when other users are battling the typical position of their characters gives an indication that a battle is going on, which you can join by just walking up to it. The turn-based character of the battles allows the attention of the user to always be directed to the action being performed and not missing important parts.
Staging occurs as well when the Toon engages in fishing. The Toon moves over to the small landing stage and the viewpoint changes. This makes it clear to the user that the Toon is getting ready to fish.
3.2.5. Follow Through and Overlapping Action
Follow through and overlapping action is missing from most of the animations that are started by the user. As mentioned for squash and stretch, at the end of a jump a Toon is instantly in the standing position as if never jumped just after the Toon hits the ground. In uncontrolled animation follow through is however present, as can be seen in the Cog battles (throwing movements etc).
3.2.6. Straight Ahead Action and Pose-To-Pose Action
The slowing down of movement is subtly visible in a walking Toon. Because the delay between the steps, more attention is pointed to the movement. Pose-To-Pose action is applied so that small animations can be used in different settings and re-used to combine more complex animations.
3.2.7. Slow In and Out
Slow in and out is hardly used within Toontown. Perhaps because all of the timing is already rather slow, slow in and slow out is hard to observe by a user, and missed in this analysis.
3.2.8. Arcs
Arcs are used a lot in Toontown Online. When Mickey Mouse (Toontown Central area) walks around for example he swings his arms following an arc pattern. Most of the character animation in Toontown show exaggerated arcs in their movement, making the animation appropriate for a cartoon-based VE.
An example of the use of arcs can be found in the application of gags to a Cog when in battle. When your character throws something (like a pie), the object moves following an arc-shaped route through the air. In some mini-games the use of arcs is also quite clear when launching your character in a catapult for example. In the mini-game where you use your toon as a ‘human’ cannonball, the arcs aren’t only visible but even an integral part of the game dynamics; the user has to predict the curve to be as successful as possible.
3.2.9. Exaggeration
Exaggeration is a very important point in Toontown Online, as this technique returns almost everywhere in the VE. Everything has been abstracted to its basic form and sometimes distorted a little to emphasize its original meaning. Houses are very colorful and the whole setting has a cartoony feel to it. The
(movements of the) Toons are exaggerated as well, their arms disproportionally long in comparison to the rest of their body and swinging around when they move. The contrast between Toons and Cogs is also exaggerated by the difference in movement, the Cogs having a very stiff and robotic movement.
3.2.10. Secondary Action
There are no real apparent secondary actions within Toontown Online, presumably because this causes extra complexity. Adding this extra level of complexity might cause users with older computers not being able to use the VE.
3.2.11. Appeal
The colorful design of Toontown Online is designed to be appealing to its target audience. The cartoony look and feel is apparent everywhere in the VE. Twins are avoided effectively. The basic standing position of a Toon has one foot slightly more forward and the arms move independently in the breathing animation. The clear distinction between the bright colored Toons and the dark colored Cogs and their grey buildings have extra appeal to the roles that characters have within the VE.
4. DISCUSSION
After this study some issues remain for discussion: in the mini games part of Toontown Online, most of the animation techniques are not used. Possible explanation for this can be the high system effort multi user game play would require. Another explication can be found in that the designers wanted to make it relatively easy to add more games to the game without having to add additional animations. Because it is necessary to keep users interested, updating and changing games can be a good strategy to achieve this. All the choices the designers made have to be done with three main things into account: complexity of the application, quality of the animation and the user experience. These things sometimes do not go in line, so a trade-off between different aspects sometimes has to be made. An example of this is that increased complexity of the mini games can lead to a decrease in user experience, so there should be thought carefully about this.
5. CONCLUSION
Many of the animation techniques were found in different parts of the Toontown Online world. Table 1 shows the level of usage of the individual techniques.
Table 1: Animation techniques used in Toontown Online
Animation technique Used in Toontown Online
Squash and Stretch High
Timing High
Anticipation High
Staging High
Follow Through and Overlapping Action Low/None Straight Ahead Action and Pose-To-Pose Action High
Slow In and Out Low/None
Arcs High
Exaggeration High
Secondary Action Low/None
Disney has clearly designed their Virtual Environment product with their own animation techniques in mind. Toontown Online however has also been designed to also be used by low-end systems. It appears that this design requirement has limited Disney in their technological possibilities which add support to more complexity of animations within the environment.
As mentioned by [Bink 2007], user behaviour is not linear. This results in the fact that a user’s action cannot be animated as well as the animators would have liked, keeping all the classical animation techniques in mind. The multi-user identity of this VE also makes it hard to create decent animations, without letting the users lose the feeling of being in control of their actions.
Animation techniques in VEs are thus closely related to technological possibilities, and intended usage of the VE. It should be clear that the coordination between the different modeling processes defined by Fencott [Fencott 1999] should be handled with care, even when focussing solely on the animation techniques used in the conceptual modelling process.
Also the trade-off between the quality of animation and the usability of a VE has to be determined in the development of a VE. Without the feeling of control a virtual world can be as beautiful as an animated movie, but will probably remain uninhabited.
6. REFERENCES
Bink, Thijmen;Pennings, Sebastiaan E.(2007). How classic animation techniques can be applied to 3D
user-interfaces and data representation. Paper presented at the course Animation, Utrecht, The
Netherlands.
Fencott, C. (1999). Towards a design methodology for virtual environments. Paper presented at the Workshop user centered design and implementation of virtual environments, York, England. Ducheneaut, Nicolas (2006). Building an MMO With Mass Appeal. A look at Gameplay in World of
Warcraft. Games and Culture, Volume 1 No.4. Sage Publications. Pages 281-317.
Goslin, Mike;Mine, Mark R. (2004) The Panda3D Graphics Engine. Computer, Volume 37, No. 10 IEEE Computer Society. Pages 112-114
Johnson, Lewis W. & Rickel, J. (1997). Steve: An Animated Pedagogical Agent for Procedural Training in Virtual Environments. SIGART bulletin, Fall 1997.
Kaur, Kulwinder. Designing Virtual Environments for Usability. PhD Thesis, City University, London, 1998.
Mine, Mark R.; Shochet, Joe; Hughston, Roger (2003). Building a Massively Multiplayer Game for the Million: Disney’s Toontown Online. ACM Computers in Entertainment, Volume 1, No.1. ACM Press, New York.
Rymaszewski, M., Au W.J., Wallace, M., Winters, C., Ondrejka, C., Batstone-Cunningham, B. (2006).
Second Life – The official guide. San Fransisco: Sybex.
