<<

IBIMA Publishing International Journal of Interactive Worlds http://www.ibimapublishing.com/journals/IJIW/ijiw.html Vol. 2012 (2012), Article ID 418638, 19 pages DOI: 10.5171/2012.418638

Game Engines Selection Framework for High-Fidelity Serious Applications

Panagiotis Petridis 1, Ian Dunwell 1, David Panzoli 2, Sylvester Arnab 1, Aristidis Protopsaltis 1, Maurice Hendrix 1 and Sara de Freitas 1

1Serious Games Institute, Coventry University Technology Park, Coventry, West Midlands, UK

2IRIT-UT1C, Université Toulouse I Capitole, France ______

Abstract

Serious games represent the state-of-the-art in the convergence of electronic gaming technologies with instructional design principles and pedagogies. Despite the value of high-fidelity content in engaging learners and providing realistic training environments, building games which deliver high levels of visual and functional realism is a complex, time consuming and expensive process. Therefore, commercial game engines, which provide a development environment and resources to more rapidly create high-fidelity virtual worlds, are increasingly used for serious as well as for entertainment applications. Towards this intention, the authors propose a new framework for the selection of game engines for serious applications and sets out five elements for analysis of engines in order to create a benchmarking approach to the validation of selection. Selection criteria for game engines and the choice of platform for Serious Games are substantially different from entertainment games, as Serious Games have very different objectives, emphases and technical requirements. In particular, the convergence of training simulators with serious games, made possible by increasing hardware rendering capacity is enabling the creation of high-fidelity serious games, which challenge existing instructional approaches. This paper overviews several game engines that are suitable for high-fidelity serious games, using the proposed framework.

Keywords : High Fidelity, Serious Games, Game Engines, Visualization. ______

Introduction Council 1997). The of using games for simulation has been realized in a number of The potential for using game engines for military training- oriented games,including serious games has been recognized for years America’ s Army and Full Spectrum (Seung Seok Noh 2006). In 1997, the Command (Wray 2004). As a result, there has National Research Council cited some of the been a trend towards the development of objectives of multi-player games for military more complex, serious games which are simulation: “The underlying technologies informed by both pedagogic and game that support these objectives address similar elements. Whilst it is true that the technical requirements: networking, low-cost graphics state-of-the-art in serious games mirrors that hardware, human modeling, and computer of leisure games (Anderson 2009), the generated characters.” (National Research technical requirements of serious games are

Copyright © 2012 Panagiotis Petridis, Ian Dunwell, David Panzoli, Sylvester Arnab, Aristidis Protopsaltis, Maurice Hendrix and Sara de Freitas. This is an open access article distributed under the Creative Commons Attribution License unported 3.0, which permits unrestricted use, distribution, and reproduction in any medium, provided that original work is properly cited. Contact author: Panagiotis Petridis E-mail: [email protected] International Journal of Interactive Worlds 2

frequently more diverse and wide-ranging fidelity is best linked to learning than their entertainment counterparts. requirements; more is not necessarily always Serious game developers frequently resort to better, particularly from a cognitivist bespoke and proprietary development due to perspective (cognitive overload (Warburton their unique requirements, such as the use of 2008)); therefore, the ability to constrain and Blitz Games’ in-house engine for developing control fidelity can be seen as desirable, the serious game: Triage Trainer (Jarvis particularly with respect to composable 2009). Such examples of bespoke content. development and the limited use of off-the- shelf game engines for serious applications Hence, whilst visual fidelity is one highlight the difficulties that exist for engine component of an immersive environment, designers seeking to understand and serious games often seek to use abstraction comprehensively support the needs of or non-realistic visual elements. Knez and instructional design. Niedenthal (Knez 2008) present an interesting factor for consideration through a The convergence of high fidelity simulators study linking changes in the affect of users to with serious games represents an area with lighting within a virtual scene. Given that the increasing potential to fulfil cognitive affect of learners has a noted impact on the learning requirements with a high degree of efficacy of learning (Bryan 1996), the efficacy, whilst leveraging the advantages of capacity of a game engine to handle game-based content to motivate and engage sophisticated lighting approaches is of users. Modern games are frequently concern. An increasing recent trend in game developed on game engines, which can be engines to transfer lighting calculations to deployed on personal computers, game the GPU has led to a significant variation consoles, pocket PCs and mobile devices. The between engines in their support for popularity of video games, especially among models and in turn their capacity to perform younger people, results in them frequently sophisticated light and shading effects. being perceived as an ideal medium for instructional programmes aimed at hard-to- In the next section, the researchers consider reach demographics (Malone 1987); in more depth the motivation behind the however, studies have also shown this creation of high-fidelity serious games. demographic responds poorly to low-fidelity Through an analysis of background games (de Freitas 2009). literature, the researchers are able to highlight common pedagogic elements and The value and impact of realism on learning hence recognise the subsequent technical purely through simulation is well- implications. documented (de Freitas 2006; Jarvis 2009). However, when game elements are Background introduced, although they have been shown to improve learning in comparison to pure As seen before, the technical state-of-the-art simulation (Mautone 2008), the relationship in serious games mirrors that of leisure between simulation and game is less simple games (Anderson 2009), however, the to define. Visual fidelity, whilst being a technical requirements of serious games are common objective in simulations, is frequently more diverse and wide ranging considerably less valued in gaming; Jarvis than their entertainment counterparts. and de Freitas (Jarvis 2009) demonstrate Serious game developers frequently resort to excessive fidelity to be detrimental to bespoke and proprietary development due to learning and the approaches put forward by their unique requirements, such as (Playgen learning theorists such as Vygotsky avoid 2010) and (PIXELearning 2010), and creating a facsimile of reality in favour of difficulties exist for game engine developers internally-consistent abstractions. Visual 3 International Journal of Interactive Worlds

in accurately understanding and supporting Visual Fidelity the needs of instructional design. High-fidelity in serious games is typically Although many serious games have limited seen as desirable in situations where there is visual interactivity, immersion and fidelity, a need to transfer process knowledge learnt there is an increasing motivation to create within the game to real world situations; and serious games that intend to support thus the closer the similarity between real situative (social and peer-driven) and and virtual space, the more effective the experiential pedagogies, partially because learning transfer is likely to be. Although a behaviourist approaches have been shown to link between learning transfer and be limited (e.g. people learn to play the game, verisimilitude of learning activity has been not address learning requirements), whilst observed particularly in training contexts cognitive approaches struggle to impart (Park 2005; Janet L. Grady 2008; Davidovitch deeper learning in the areas of affect and 2009)), this link does not necessarily hold motivation (Egenfeldt-Nielsen 2005). true in all game-based learning scenarios; Furthermore, recent work by Mautone Jarvis and de Freitas (Jarvis 2009) suggest (Mautone 2008) demonstrates enhanced that the of fidelity required must be learning when introducing game elements to mapped onto learning objectives. a standard flight simulator. Consequently, re- Furthermore, an over-emphasis upon visual evaluation of simulator approaches to fidelity can mask the complexities of incorporate game and game-like elements producing verifiable and replicable learning places an increasing demand for serious activities and experiences. game developers to deliver high-fidelity solutions. Given this motivation to create The need to engage the learner, and immersive, high fidelity serious games, an specifically, immersion through high-fidelity obvious development choice is to utilise content is one such mechanism through game engines, which provide ‘out of the box’ which such engagement can be achieved. The support for state-of-the-art desktop GPU concept of immersion is a common one in rendering and physics. In the remainder of serious games, although the components that this section, the researchers discuss a range constitute an immersive experience can be of key considerations when selecting the more difficult to define. The capacity to technology behind serious games, supporting immerse learners is a significant effective pedagogy, the learner and their consideration, although the means for context. achieving this immersion can be diverse from highly visual content to less technical Fritch (Fritsch 2004) et al compared approaches such as narrative immersion different games engines for large-scale (Mott 2006). Robertson et al (Robertson visualization of outdoor environments 1997) looked specifically at the relationship focusing mainly on the issues of between user and standard desktop PC as an composability. Similarly, Shiratuddin interface for , and compared it compared various game engines for to head-tracked systems. Robertson claims visualizing large architectural scenes mainly “immersion should not be equated with the focusing on accessibility and the availability use of head-mounted displays: mental and of game engines (Shiratuddin 2007). emotional immersion does take place, Following the analysis of all these potential independent of visual or perceptual factors, the methodology used here is immersion”, an opinion reinforced by defined to include areas, such as audiovisual Csikszentmihalyi and Kubey fidelity, functional fidelity, composability, (Csikszentmihalyi 1981). Thus, a further availability and accessibility, networking and discretisation of the concept of immersion heterogeneity. between psychological and perceptual levels is identified. The role consistently is one of International Journal of Interactive Worlds 4

‘drawing in’ the user, such that they Accessibility experience a perceptual shift between simply viewing the screen and existing within the Unlike leisure games, target demographics environment. Breaks in consistency, such as for serious games are often non-game those induced by low frame rate or players, with little interest in technology or discontinuities in world content are shown to knowledge of user interfaces. Furthermore, have a significant negative impact on the developers of serious games may be immersion (Csikszentmihalyi 1981). instructional designers wishing to explore a new medium, rather than traditional game With respect to all three of these aspects, developers seeking to develop instructional there are a number of dimensions in which content. Therefore, the capacity of the engine fidelity must be considered. At a high level, to support both developers and users with there are many aspects of a game that can be limited expertise is of relevance. Previous represented with differing levels of fidelity; Research (Jarvis 2009) has shown that the narrative, the depth of visual and conventional keyboard and mouse auditory content, the interaction medium interaction in a world with multiple degrees and the behaviour of characters and objects of freedom can prove initially overwhelming within the game world. Whilst all these for non-gamers. As such, serious games can concerns must be reflected upon when often require interfaces that deviate from designing a serious game, in terms of engine those common to entertainment games, and selection, a clear distinction exists between seek to simplify interactions based upon an affordances for audiovisual and functional understanding of learning requirements. fidelity. It is possible to engineer a world, which appears realistic but does not behave Heterogeneity in a realistic fashion. Although increasing audiovisual fidelity often implies increased Zyda identified as early as 1995 (Zyda, 1995) functional fidelity, as a virtual room fills out three fundamental challenges in multiuser with furniture to become more visually virtual environment design, which he defines realistic, players start expecting furniture to as composability, scalability and function as it would in the real world. For heterogeneity. These remain substantial example, placing a virtual telephone on top of challenges within current virtual a desk can bring with it a host of potential environment research. With respect to questions from users expecting to be able to serious games, heterogeneity is of particular dial out. concern, since target demographics are frequently ‘non-gamer’, and thus platforms The concept of immersion is a common one capable of deployment across a wide range of in serious games, although its definition, and hardware and software platforms are specifically, the components that create an significantly advantageous. immersive experience can be more difficult to define. The capacity to immerse learners is Composability a significant consideration, although the means for achieving this immersion can be Serious games often seek to model real world diverse: from visual fidelity to functional locales and situations, or adapt real world fidelity as well as less technical approaches data for use in games with minimal such as narrative immersion. A simple out-of- development overheads. Composability in place texture or inappropriate sound can this context is used to describe both the have catastrophic effects for believability, as reusability of content created within a game metrics of immersion such as the engine, and also its capability to import and performance indications and cognitive use data from common or proprietary surveys applied by Pausch et al. (Pausch sources. 1997). 5 International Journal of Interactive Worlds

Technology evolution often requires the insufficient information for the game engine; recreation of game content in ever-increasing often the case when models are developed levels of visual and functional fidelity.. If the without adequate information on materials, scene consists of many high fidelity objects at textures, bump maps or levels-of-detail. various distances, it may be possible to adopt a low level of-detail approach (Engel) and Secondly, occlusion culling may only be use less complex geometry, or even dummy performed if information on the visibility of objects (Akenine-Moller T. 2008), to polygons is computationally less expensive to approximate distant objects (Sander 2006). obtain than rendering them. Early game Alternatively, if only a small sub-section of engines such as the seeks to the world or object is in sight at any one time, achieve visibility data prior to run time it may be possible to hold only these visible through pre-processing a visibility matrix parts in memory and then replace them as and binary partitioning of virtual space. new parts come into view by applying some Thirdly, mesh decomposition must be form of spatial partitioning (Crassin 2009). performed in concert with analysis and Another issue that the designer/developer scaling of texture level of detail to provide a has to consider arises when attempting to satisfactory visual outcome. import a high-fidelity model into a game engine, which may not support the geometry Networking or texture formats within the model, or require specific optimisations such as Multi-user elements are often specified at reduction to be rendered in real- early-stage designs of serious games, since time. they often affect the nature of the game and its role within a training programme as a Two principal solutions exist: the first is to whole. User interaction within the game itself address the problems on a large-scale is often used to address the difficulties in through an algorithmic approach that seeks automating the behaviour of non-player to provide automated conversion between characters in a believable and coherent formats and import the model as a whole into fashion. In this context, instructors play the the game engine. The second is to select role of virtual characters in order to converse specific components of the model for and interact with learners in a realistic and conversion and progressively convert and adaptive manner. Whilst this can be an integrate them into the game engine by hand. effective way of creating believable virtual However, algorithms intending to support scenarios, it suffers from limited scalability such an approach face a number of due to the availability of instructors and challenges: decomposition of a mesh into practical limits on pupil-tutor ratios. multiple levels-of-detail is difficult to optimise since the perceived visual fidelity of Support for larger-scale communities and the resultant lower-resolution meshes is social elements are gaining increasing linked to the perception characteristics of the recognition within the leisure gaming user, and hence a solution is not only community as a mechanism for increasing mathematically complex, but must also uptake and long-term play. As an example, consider how humans perceive and integrate the recent Guitar Hero [35] game wraps a features of a three-dimensional scene cognitively simple task within a socio- (Treisman 1980). Therefore, developers are culturally motivating setting, and has commonly forced to select specific consequently proved to be highly successful components of the model for conversion, and commercially. Within serious games, social progressively convert and integrate them elements often take the form of online into the game engine by hand, a task communities, and the convergence of games necessary when conversion tools are with social networking technologies remains inadequate or the original data format has an area of interest. International Journal of Interactive Worlds 6

Summary of Requirements for Serious game engine is any tool or a collection of Games tools that creates an abstraction of hardware, and/or software, for the purpose of As seen in section 0, the key elements are simplifying common game development fidelity, consistency and support for tools for tasks. Many computer game developers creating immersion and flow. Fidelity can be support modification of their game further subdivided into visual and functional environments by releasing level editors, for fidelity. Consistency relates to technical example to modify the game environment features in this area including the need to and tools to edit the game behaviour. This load between areas or stream geometry, or allows the reuse of the underlining game whether the engine is standalone or web engine technology, including 3D rendering, based. Finally, corresponding features for 2D drawing, sound, user input and world tools creating immersion and flow include physics/dynamics (Lewis M. 2002). For game scripting tools, especially when example, users can create new levels, maps narrative is a non-linearity (i.e. implementing and characters, adding them to the game, some form of or known as partial conversion or they can artificial life). This ties in with immersion, create entirely new games by altering the but the term immersion within the game engine, known as total framework is avoided due to the current lack conversion (Trenholme 2008). Modern game of consensus on its definition (Slater 1993), engines have a modular structure, so that (Slater 2003). Instead the researchers focus they can be reused into different games on the elements that contribute to immersive (Jacobson J. 2003; Trenholme 2008). Analysis experiences. of a range of current game engines suggests the following modules are common: the Additional technical elements include Graphics Module, the Physics Module, the heterogeneity (on which platforms can the Collision Detection Module, the I/O Module, engine be deployed ; what hardware Sound Module, the AI Module and the requirements; can it scale automatically), Network Module. The graphics module is accessibility (support for non-standard responsible for the generation of the 2D/3D interfaces and devices; as well as support for graphics in the environment, including standardised interfaces e.g. WASD) and libraries for texture mapping, shadowing, multiuser support, beneficial since in the lighting and shader effects. The Physics absence of sophisticated AI, human Module ensures that objects behave instructors often play a role in virtual according to physical laws, for example learning experiences, and similarly socio- objects fall under gravity and glass breaks. cultural elements can be key motivators as The Collision Detection Module is used to mentioned above. ensure that certain actions will occur when two objects collide. Furthermore, the Overview of Modern Game Engines input/output module is responsible for the input and output device which can be Modern game engines combine several integrated into the 3D engine. technologies from the area of computer science such as: graphics, artificial One of the most important elements of the intelligence, network programming, creation of serious games is the visual languages and algorithms. Modern computer representation of these environments. game engines are robust and extensively Although serious games have design goals tested (Lepouras G. 2005), in terms of the that are different from those of pure usability and performance, work on off-the- entertainment video games, they can still shelf systems (Robillard G. 2003) and can be make use of the wide variety of graphical easily disseminated, for example via online features and effects that have been communities (Burkhard C. Wünsche 2005). A developed in recent years. Most game 7 International Journal of Interactive Worlds

engines provide support for texture mapping, Playstation controller provide 6 Degrees Of shadowing, lighting and shader effects in Freedom interaction that could enhance the their graphics model. Additionally, modern user interaction with the environment and engines frequently include a selection of such increase the immersion of the user. effects, which can include more traditional image processing, such as colour correction, With such a broad definition, what is referred film-grain, glow or edge-enhancement, as to as a game engine can vary among well as techniques that require additional developers and where a game engine ends scene information, such as depth of field and and a game begins is not always a clearly motion blur (Akenine-Moller T. 2008). defined line (Trenholme 2008). Game engines should be distinguished from An Artificial Intelligence module, often used graphics engines that come only with to create objects or "Non Playing Characters" rendering capabilities, and also from (NPCs), is able to interact "intelligently" with Software Developer Kits (SDKs) that aid the player. An important aspect in the game development. The reason for this creation of realistic scenes is to create in the differentiation is that graphics engines game environment intelligent behaviours for impose limitations as to what and how things the inhabitants of the virtual world, which is can be included in a game, whilst SDKs are achieved using Artificial Intelligence (AI) much more flexible but with narrower focus. techniques. However, it is important to For example, is a very flexible understand that when the researchers refer proprietary renderer but has no collision to the AI of virtual entities in game engines, it detection or physics capabilities, unlike is not truly AI – at least not in the Havoc, which is solely a physics engine. conventional sense (McCarthy 2007). The Similar include Criterion’s techniques applied to computer games are Renderware and Speedtree. usually a mixture of AI-related methods whose main concern is the creation of a A Framework for Game Engine Selection believable illusion of intelligence (Scott 2002), e.g. the behaviour of virtual entities In short, key elements are defined arising only needs to be believable enough to convey from the background as fidelity subdivided the presence of intelligence and to immerse into visual and functional fidelity, the human participant in the virtual world. consistency, (technical features in this area The Network Module is responsible for the would include need to load between areas or multiplayer implementation of the game. As a stream geometry, whether the engine is result, players could cooperate in exploring standalone or web based – basically things an area or exchange opinions about certain that impact on immersion), and support for aspects of a virtual environment, while being other tools for creating immersion and flow located in different areas of the world. such as narrative (corresponding features include: game scripting tools), particularly The IO module provides support for different narrative which supports non-linearity input/output devices. This module provides (artificial intelligence, artificial life). This ties tools that allow the user to communicate and into immersion; the researchers deliberately interact with the game environment. Most avoid “immersion” and the factors that affect game engines provide support for standard immersion within the framework and input devices such as, joysticks, gamepads because it is decided to focus on the elements and keyboard. Technological improvements that contribute to immersive experiences. and cost reduction in computing power, display and sensor technology have resulted Additional technical elements include in a widespread use of 3D Input Devices heterogeneity (on which platforms can the (Fröhlich 2000; Petridis 2005; Mourkoussis engine be deployed ; what hardware 2006). Devices, such as the Nintendo and requirements; can it scale automatically), International Journal of Interactive Worlds 8

accessibility (support for non-standard instructors often play a role in virtual interfaces and devices; as well as support for learning experiences, and similarly socio- standardised interfaces (e.g. WASD) and cultural elements can be key motivators as multiuser support, beneficial since in the mentioned above. absence of sophisticated AI, human

Table 1: Framework for Comparing Engines in SG

Audiovisual Fidelity Rendering Animation Sound Functional Fidelity Scripting Supported AI Techniques Physics Composability Import/ Export Content Developer Toolkits Accessibility Learning Curve Documentation and Support Licensing Cost Networking Client Server/ Peer –to - peer Heterogeneity Multiplatform Support

Current Game Engines Compared Using the engines over others is to meet several Framework identified criteria including: wide usage of engine, availability, modularity and At the current time, the researchers have innovative features. identified over 100 games engines available on the market (WikiPedia 2010). It was The first game engine under comparative decided to compare a subset of this number analysis here is the CryENGINE 3. The engine with the framework in order to gain some is available for the PC, PS3 and Xbox360. baseline data for validating the model. The CryENGINE 3 supports development in rationale for the selection of these games Direct X 9,10 and 11.

9 International Journal of Interactive Worlds

Figure 1: Screenshot from the CryENGINE3

The CryENGINE supports a number of provides a number of features such as features that are useful for creating character animation, advanced AI, real-world immersive and realistic serious games, such physics, shader-based rendering and a highly as real-time world editor, bump mapping, extensible development environment. The dynamic lights, an integrated multi-threaded level editor that is used to produce levels for physics engine, , shadow support, games using the Source Engine is the Valve multi-core support, character animation Hammer Editor, commonly referred to as system, pathfinding, dynamic sounds and Hammer. The editor uses brushes, called interactive music. The engine supports all primitives, to a level. The engine currently available hardware and it is also supports particle effects, volumetric updated with further hardware support smoke and environmental effects such as fog when it becomes available. or rain. Typical games that use the Source Engine are Half Life 2(Valve 2010), Dark The second engine under comparison is Messiah of Might and Magic ( 2010). Valve’s Source Engine. Source Engine

International Journal of Interactive Worlds 10

Figure 2: Screenshot of Half Life 2

The third engine is the 3 effects include a for particles (Games 2009), which has a complete game composed of sprites, meshes, lines and development methodology for next- beams. The particle system supports various generation consoles and DirectX9-equipped lifetime, texture, movement and collision PC's, providing the vast range of core options. All particle features can be technologies, content creation tools and manipulated in real-time in the editor. The support infrastructure required by top game texturing features of the engine include a developers. The engine supports high material system that supports alpha- performance rendering, advanced animation blending, e.g. transparency and blending of and high-quality dynamic lighting. Supported multiple layers of textures.

11 International Journal of Interactive Worlds

Figure 3: Screenshot of 3

The final engine under comparison is the latest technologies of as engine (Unity 2009), which is a game can be seen from Table 2. Using these development tool that allows the developer features such as texturing, lighting, shadows to create games for different platforms, such and special effects, the developer has the as the iPhone, Nintendo Wii, Mac and PC. potential to deliver the same perceptual quality, as the user was present in the real CryEngine 3, Unity, Valve’s Source Engine scene. Thus, the player can feel actually and the Unreal engines provide high-fidelity present in the real scene being depicted. standards and provide support for all the

International Journal of Interactive Worlds 12

Table 2: Audiovisual Fidelity

CryEngine Source Engine Unreal Unity Basic, Multi - Basic, Multi - Basic, Multi - Basic, texturing, Bump texturing, Bump texturing, Bump Bumpmapping, mapping mapping mapping, Procedural

g Texturin Procedural Per -vertex, Per - Per -vertex, Per - Per -vertex, Per - Per -vertex, Per pixel, pixel, pixel, Gloss/ Pixel Lightmapping, Lightmapping, Specular Mapping Gloss Radiosity, Gloss Light mapping

Lighting map,Anisotropic maps Shadow Volume Shadow Mapping, Shadow Mapping, Projected planar Projected, Shadow

Volume ws Shado Environmental Environmental Environmental Environmental Mapping Mapping Mapping Mapping

Particle Systems, Particle Systems Particle Systems Particle Systems Bill Boarding Bill Boarding Bill Boarding Bill Boarding

Rendering Lens Flares Lens Flares Lens Flares Lens Flares Effects Special Forward Skeletal Forward Forward Kinematics, Animation, Kinematics, Kinematics Keyframe Morphing, Facial Keyframe Keyframe Animation, Animation, Animation, Animation, Skeletal Animation Skeletal Skeletal Animation Animation, Blending Animation, Morphing, Morphing, Morphing, Animation Animation Animation Blending

Animation Blending Blending 2D Sound, 3D 2D Sound, 3D 2D Sound, 3D 2D Sound, 3D Sound Sound Sound, Streaming Sound, Streaming

d Soun Sound Sound:

Another major challenge in the selection of detection and path finding. Additionally, all game engines for serious games is functional game engines have integrated their own fidelity. Functional fidelity is closely related physics engines, and furthermore, each of the to the AI, physics and scripting. All the selected game engines has support for selected game engines provide support for scripting languages. various AI techniques, such as collision

13 International Journal of Interactive Worlds

Table 3: Functional Fidelity

CryEngine Source Unre Unity al Scripting Script Yes Yes Yes Yes

Object Model Yes No No No

Supported AI Collision Detection Yes Yes Yes Yes Techniques Path Finding Yes No Yes Yes

Decision Making Yes No Yes Yes

Physics Basic Physics Yes Yes Yes Yes

rigid body Yes Yes Yes Yes

vehicle dynamics Yes Yes Yes Yes

The reusability of content created within a comparison of the game engines, the game engine and the capability of importing researchers identified that importing a 3D and using data from common sources should model from CAD software into the supported be considered before selecting a game format of the game engine is a major issue engine. As budgets for creating serious and requires the developer to select specific games are usually very limited compared to components of the model for conversion and commercial entertainment games, this progressively convert and integrate them becomes an especially important into the game engine by hand. Once this aspect(Protopsaltis 2010; Protopsaltis 2011). obstacle has been overcome, the game Additionally, in serious games, the engines can create high-fidelity indoor and developers need to have access to the SDK, outdoor scenes occupied with non-player GDK in order to add different peripheral characters in real-time. Table 4 compares the devices or connect the game engine with selected engines according to the suggested learning management systems or with other categories. software . However, from the

Table 4: Composability

Metrics CryEngine Source Unreal Unity Import/ CAD Platforms 3ds max, 3ds max, 3ds max, 3ds max, Export supported maya maya maya, maya, Content Import Export No No Yes No Limitations Content Small Large Large Medium Availability Developer SDK/GDK Yes Yes Yes No Toolkits

International Journal of Interactive Worlds 14

Another major challenge in the selection of are also a number of community websites game engines for serious games is that provide discussion forums and tutorials accessibility; that is, how easy it is to retrieve (Planet Unreal 2010; Unreal Wiki 2010). supporting information about the game engine (see Table 5). CryENGINE provides all Documentation for the Source SDK is the necessary development tools that can be provided on the Valve Developer Community accessed from games that use the engine (i.e. wiki (Valve's Wiki 2010), which provides the Crysis, FarCry). The engine offers the most comprehensive guide on using the Sandbox editor, which allows the user to edit Source Engine. The Source SDK contains all levels in real-time. Partial source code and the necessary tools in order to create a documentation is included with a freely and develop game content. The Source SDK downloadable SDK. Documentations for the also provides a Create a Mod option, which Unreal Engine are available through the copies the necessary source code and Epic’s Unreal Developer Network (UDN) resources to a working directory. A large ( 2010), which is the official number of other websites provide their own support site for licensees and mod discussion forums and tutorials, which range (modifications) developers, providing from general introductions to the Source SDK technical documentation, as well as tutorials, tools to more specific tasks (Interpolers.net ; for the Unreal engine and UnrealEd. There EditLife 2010).

Table 5: Accesibility

Metrics CryEngine Source Unreal Unity Learning Medium Medium Medium Medium Curve Docs and Docs SDK Official Subset of Docs and Support Quality includes documentat official Tutorials CryENGINE ion documentat available modding available on ion and from the guide and Valve tutorials official FAQ, as well Developer available on website as guides Community the Unreal for other Developer tools. Network (UDN) Technica Yes Yes Yes Yes l Support Commun Yes Yes Yes Yes ity Support Licensing Game Game UnrealScrip Indie and source code source code t game Pro version available, available, source code available. comes with comes with available CryENGINE Source SDK from UDN. MOD SDK.

The next step in the selection process of the game engine is to focus on the heterogeneity of the engines and their network support (see Table 6). All the selected engines support client-server architectures. However, if the serious game is going to support a large virtual world with hundreds of users, a network supported layer has to be built. 15 International Journal of Interactive Worlds

Table 6: Networking and Heterogeneity

CryEngine Source Unreal Unity Client - Yes Yes Yes Yes

Server, Peer -to - No No No No

Networ Peerking Multiplatfo Yes Yes Yes Yes rm Het ero

Our analysis of four leading entertainment which underpin serious and leisure games is game engines demonstrates that whilst all a complicated process that requires time, four provide a sound basis for serious game resources and teamwork. As serious games development, none give specific regard to become more complex, so do the engineering serious applications. Unsurprisingly perhaps, challenges that arise during the development as none of these games engines have been of the game. Hence, the early-stage selection developed with serious applications in mind; of the optimal engine for development is which in turn may have something to do with crucial. This paper presents a selection the relatively small budgets often found in framework, allowing the developer to select serious game development. Rather, the the ideal engine based on the technical emphasis is upon the designer to integrate requirements of the serious game. technology with instructional design, and to this end, the game engine plays the role of an This is the first framework for serious game implementation tool, rather than fully engine selection currently proposed and supporting the process of serious game tested, and is intended as a starting point for development. Future viability of game ongoing benchmarking and metrics for engines for serious applications will depend supporting serious game engine selection. on how capable they become at supporting However, whilst our framework relates the concepts and core dissimilarities overarching technical requirements to a between serious and leisure applications, range of modern engines, more research, which includes supporting participation and testing and validation must still be done to learner involvement in the development relate learning requirements and process, integration into intelligent tutoring instructional design principles to these systems (Dunwell 2011) and support for technical features. Ultimately, the design and metadata (Hendrix 2012) in the form of implementation of effective serious games learning objects and repositories. must be grounded in pedagogy, as well as Furthermore, user interactions with serious technology, and therefore future work should applications are often diverse and reflected address the many issues surrounding the in a wide range of levels of user expertise and equation of learning requirements to these expectation, and therefore support for identified technical features. Towards this interfaces outside of those common to end, future studies will focus upon the entertainment games is advantageous. analysis of the impact of the various engines and their functionalities on targeted learner Conclusions and Future Work groups.

The creation of a serious game is a complex Finally, whilst the researchers sought to engineering project that requires technical identify and evaluate the most widespread expertise, as well as a careful balance of game subset of game engines within the context of design principles with instructional content. our framework, given the rapid evolution of Similarly, the development of generic engines technology in this area, comparison of International Journal of Interactive Worlds 16

additional and emerging game engines may De Freitas, S. & Neumann, T. (2009). "The Use also offer opportunities to further refine and of ‘Exploratory Learning’ for Supporting validate the framework. Thus, future work Immersive Learning in Virtual will upon and further test game engines Environments," Computers and Education using this framework method of validation. 52(2): 343-352.

References Dunwell, I., Petridis, P., Arnab, S., Protopsaltis, A., Hendrix, M. & de Freitas, S. Anderson, E. F., McLoughlin, L., Liarokapis, F., (2011). "Blended Game-Based Learning Peters, C., Petridis, P. & de Freitas, S. (2009). Environments: Extending a Serious Game "Serious Games in Cultural Heritage," 10th into a Learning Content Management VAST International Symposium on Virtual System," ALICE 2011: International Reality, Archaeology and Cultural Heritage Workshop on Adaptive Learning via (VAST '09), VAST-STAR, Short and Project Interactive, Collaborative and Emotional Proceedings, Eurographics, Malta, 22-25. approaches, at INCOS 2011: Third International Conference on Networking and Bryan, T., Mathur, S. & Sullivan, K. (1996). Collaborative Systems "The Impact of Positive Mood on Learning," Learning Disability Quarterly 19(3): 153-162. EditLife. (2010). Online Resources for Half Life, Retrieved 21/06/2010, from Burkhard C. Wünsche, Blazej Kot, Andrew http://www.editlife.net/ Gits, Robert Amor & John Hosking. (2005). "A Framework for Game Engine Based Egenfeldt-Nielsen, S. (2005). "Beyond Visualisations," Proceedings of Image and Edutainment: Exploring the Educational Vision Computing New Zealand . Potential of Computer Games," Phd thesis, IT- University Copenhagen . Crassin, C., Neyret, F., Lefebvre, S. & Eisemann, E. (2009). "Gigavoxels: Ray- Engel, W., Hoxley, J., Kornmann, R. & Suni Guided Streaming for Efficient and Detailed Zink J. Programming Vertex, Geometry, and Voxel Rendering," I3D ’09:Proceedings of the Pixel Shaders, line book available at: 2009 symposium on Interactive 3D graphics http://wiki.gamedev.net/, 2008. 9, 11." and games (2009). 9: 15-22. Epic Games. (2010). UDN Network Csikszentmihalyi, M. & Kubey, R. (1981). Retrieved 21/06/2010, from "Television and the Rest of Life: A Systematic http://udn.epicgames.com/Main/WebHome. Comparison of Subjective Experience," Public html. Opinion Quarterly 45(3): 17-328. Fritsch, D. & Kada, M. (2004). "Visualization Davidovitch, L., Parush, A. & Shtub, A. (2009). Using Game Engines," ISPRS commission 5 . "The Impact of Functional Fidelity in Istanbul, Turkey,: 621-625. Simulator-based Learning of Project Management," International Journal of Fröhlich, B., Plate, J., Wind, J., Wesche, G. & Engineering Education 25(2): 333-340. Gobel, M. (2000). "Cubic-Mouse-Based Interaction in Virtual Environments," IEEE De Freitas, S. & Jarvis, S. (2006). "A Computer Graphics and Applications 20(4): Framework for Developing Serious Games to 12-15. Meet Learner Needs," The Interservice/Industry Training, Simulation Games, E. (2009). Retrieved 06/11/2009, and Education Conference. Florida,USA. from http://www.unrealtechnology.com/.

17 International Journal of Interactive Worlds

Hendrix, M., Protopsaltis, A., Rolland, C., Conative and affective process analyses. F. J. Dunwell, I., de Freitas, S., Arnab, S., Petridis, Snow R. E. Erlbaum: 223-253. P. & Llanas, J. (2012). "Defining a Metadata Schema for Serious Games as Learning Mautone, T., Spiker, V. A. & Karp, M. R. Objects," International Conference on Mobile, (2008). "Using Serious Game Technology to Hybrid, and On-line Learning (eL&mL) IARIA. Improve Aircrew Training," In Proc. of I/ITSEC 2008. Interpolers.net. Resources Available for Source Engine. Retrieved 21/05/2010, from McCarthy, J. (2007). "What is Artificial http://www.interlopers.net/. Intelligence," Available At http://www- formal.stanford.edu/jmc/whatisai/whatisai. Jacobson J. (2003). "Using ‘‘CaveUT’’ to Build html Immersive Displays with the Unreal Tournament Engine and a PC Cluster ,"ACM Moller, T. A., Haines, E. & Hoffman, N. (2008). symposium on interactive 3D graphics , ACM Real-Time Rendering, A. K. Peters . Press . Mott, B. W., McQuiggan, S. W., Lee, S., Lee, S. Y. Janet L. Grady, Rosemary G. Kehrer, Carole E. & Lester, J. C. (2006). "Narrative-Centered Trusty, Eileen B. Entin, Elliot E. Entin & Tad Environments for Guided Exploratory T. Brunye (2008). "Learning Nursing Learning," Proceedings of the Agent Based Procedures: The Influence of Simulator Systems for Human Learning Workshop at Fidelity and Student Gender on Teaching the 5th International Joint Conference on Effectiveness," Journal of Nursing Education Autonomous Agents and Multiagent Systems 47(9). (ABSHL-2006). Hakodate, Japan.

Jarvis, S. & de Freitas, S. (2009). "Evaluation Mourkoussis, N., Mania, K., Petridis, P., White, of an Immersive Learning Programme to M., Rivera, F. M., Pletinckx, D., Troscianko, T. support Triage Training," Proceedings of the & Hawkes, R. (2006). "An Analysis of the 1st IEEE International Conference in Games Effect of Technological Fidelity on Perceptual and Virtual Worlds for Serious Applications, Fidelity," To appear in proceedings of the IEA IEEE Computer Society ,. Coventry,UK: 117- 2006 (International Ergonomics 122. Association), 16th World Congress on Ergonomics, The Hague, Netherlands. Knez, I., & Niedenthal, S. (2008). "Lighting in Digital Game Worlds: Effects on Affect and National Research Council (1997). "Modeling Play Performanc," Cyberpsychology & and Simulation: Linking Entertainment & Behavior (11): 129-137. Defense," National Academy Press .

Lepouras, G. & Vassilakis, C. (2005). "Virtual Park, G. D., Allen, R. Wade, Rosenthal, Museums for all: Employing Game Theodore J. & Fiorentino Dary. (2005). Technology for Edutainment," Virtual Real . 8: "Training Effectiveness: How Does Driving 96-106. Simulator Fidelity Influence Driver Performance?," Human Factors and Lewis M. & Jacobson, J. (2002). "Game Ergonomics Society Annual Meeting Engines in Scientific Research," Commun ACM Proceedings, Training . 5: 2201-2205. 45(1): 27-31. Pausch, R., Proffitt, D., & Williams, G. (1997). Malone, T. W. & Lepper, M. R. (1987). "Quantifying Immersion in Virtual Reality," "Making learning Fun:A Taxonomy of Proceedings of the 24th annual conference Intrinsic Motivations for Learning," In on Computer graphics and interactive Aptitude, learning and instruction : III. techniques, ACM Press : 13-18. International Journal of Interactive Worlds 18

Petridis, P., White, M., Mourkousis, N., Sander, P. V. & Mitchell, J. L. (2006). 'Out-of- Liarokapis, F., Sifiniotis, M. Basu, A. & Core Rendering of Large Meshes with Gatzidis, C. (2005). "Exploring and Progressive Buffers,' ACM SIGGRAPH 2006: Interacting with Virtual Museums," CAA Proceedings of the conference on SIGGRAPH 2005: The World in your eyes, 2006 course notes (2006). 9: 1-18. Tomar,Portugal. Scott, B. (2002). The Illusion of Intelligence, PIXELearning. (2010). Pixel Learning: AI Game Programming Wisdom , Charles River Serious Games and Immersive Simulation for Media: 16-20. Learning and Development, Retrieved 21/06/2010, from Seung Seok Noh, Sung Dea Hong & Jin Wan http://www.pixelearning.com/. Park (2006). "Using a Game Engine Technique to Produce 3D Entertainment lanet Unreal. (2010). Planet Unreal Resource, Contents," Proceedings of the 16th 21/06/2010, from International Conference on Artificial Reality http://planetunreal.gamespy.com/. and Telexistence--Workshops (ICAT'06).

Playgen. (2010). "Make Games and Shiratuddin, M. F. & Fletcher, D. (2007). Simulations with PlayGen," Retrieved "Utilizing 3D Games Development Tool for 21/06//2010, from http://playgen.com/. Architectural Design in a Virtual Environment," 7th International Conference Protopsaltis, A., Auneau, L., Dunwell, I., de on Construction Applications of Virtual Freitas, S., Petridis, P., Arnab, S., Scarle, S. & Reality. Hendrix, M. (2011). "Scenario-Based Serious Games Repurposing," ACM SIGDOC 29th Slater, M. (2003). "A Note on Presence International Conference on Design of Terminology," from www.presence- Communication, ACM connect.com.

Protopsaltis, A., Panzoli, D., Dunwell, I. & de Slater, M. & Ushoh, M.(1993). Freitas, S. (2010). "Repurposing Serious 'Representation Systems, Perceptual Games in Health Care Education," 12th Position, and Presence in Immersive Virtual Mediterranean Conference on Medical and Environments,' Presence: Teleoperators and Biological Engineering and Computing Virtual Environments 2(3): 221-233. (MEDICON 2010) Treisman, A. M. & Gelade, G. (1980). "A Robertson, G., Czerwinski, M. & Van Dantzich, Feature-Integration Theory of Attention," M. (1997). "Immersion in Desktop Virtual Cognitive Psychology 12(1): 97-136. Reality," Proceedings of the 10th annual ACM symposium on User interface software and Trenholme, D. & Smith, S. P. (2008). technology, ACM Press .: 11-19. "Computer Game Engines for Developing First-Person Virtual Environments," Virtual Robillard, G. Bouchard, S. Fournier, T. & Reality , 12(3): 181-187. Renaud, P. (2003). "Anxiety and Presence Using VR Immersion: A Comparative Study of Ubisoft. (2010). Retrieved 21/06/2010, the Reactions of Phobic and Non-Phobic 2010, from Participants in Therapeutic Virtual http://darkmessiahgame.uk.ubi.com/. Environments Derived from Computer Games." CyberPsychol Behav 6(5): 467-475.

19 International Journal of Interactive Worlds

Unity. (2009). Retrieved 04/11/2009, from http://unity3d.com/.

Unreal Wiki. (2010). Unreal Wiki. Retrieved 21/06/2010, from http://wiki.beyondunreal.com/.

Valve. (2010). Retrieved 21/06/2010, 2010, from http://orange.half-life2.com/.

Valve's Wiki. (2010). "Valve's Source Developer Wiki," Retrieved 21/06/2010, from http://developer.valvesoftware.com/wiki/S DK_Docs.

Warburton, S. (2008). 'Defining a Framework for Teaching Practices Inside Virtual Immersive Environments: the Tension Between Control and Pedagogical Approach,' Proceedings of RELive ’08 Conference .

WikiPedia. (2010). "Game Engines," Retrieved 21/06/2010, from http://en.wikipedia.org/wiki/List_of_game_e ngines.

Wray, R. E. Laird, J. E. Nuxoll, A. Stokes, D. & Kerfoot, A. (2004). "Synthetic Adversaries for Urban Combat Training," Proceedings of the 2004 Innovative Applications of Artificial Intelligence Conference, San Jose, CA.