Reverse-Engineering AR

This section seeks to locate AR’s position within a wider context.

There are three media that converge in Mobile AR: Virtual Reality; the Internet and the mobile telephone, with other, subsidiary technologies as enablers to this end. Assessing the three of these in turn, we can glean knowledge of these highly influential media forms and their impact, the findings of which can be built into a model for the commercial diffusion and societal impact that Mobile AR might enjoy.

Virtual Reality is first up. Check out my next post in the series!

Web Discoveries for May 27th

These are my del.icio.us links for May 27th

Constructing A Methodology

Mobile AR is still highly prototypical, and has not received much previous academic attention thus far. A deep analysis of this technology and its implications requires a specially developed methodology, a methodology which acknowledges the pre-release status of the technology, and recognises that Mobile AR represents a fusion of a number of different media technologies. Given that there is no fixed point of entry for analysis, I look to academics writing on the subject of other radical and emergent technologies that (at the time of publishing) were yet to reach the mainstream.

At this early stage of the product cycle there is an interesting interplay between fields. This interplay is assessed in Digital Play: The Interaction of Technology, Culture, and Marketing by Kline et al. (2003), and it raises some arguments useful to the further analysis of Mobile AR as a cultural artefact. One of its lessons is that assessing new technologies is fraught with a long-standing academic and cultural issue: the problem of technological determinism. The theory is that new technologies drive social, political and cultural changes, and that the perceived linearity of technical progression is somehow representative of humanity’s own progression, parallel trajectories dependent on the other’s existence. The weaknesses arise when one assigns these same values to their own assessments, which isolate the subject technology from its wider context. Indeed, it is often forgotten that in order to achieve these innovations, social, economic, political and cultural forces have all worked in collusion. Digital Play (Kline et al., 2003), quotes Leiss (1990) in an especially provoking summarisation:

“Strictly speaking, there are no imperatives in technology. The chief mistake … is to isolate one aspect (technology) of a dense network of social interactions, to consider it in abstraction from all the rest, and then relate it back to that network as an allegedly independent actor.”

Leiss (1990: 2) in Kline et al. (2003:8)

Leiss’ point is that academic enquiry should seek to observe its subjects in the light of their true context. He highlights the importance of the “network” as the source of each technology, denying the idea that modern culture is ‘Under Technology’s Thumb’ (Leiss, 1990). A personal observation is that within the “network” also lie the forerunning technologies that gave rise to the newest developments, and the means to develop them further. This idea recalls McLuhan, who to the detriment of Leiss’ argument, was sometimes known as “the most famous media technological determinist” (Straubhaar & LaRose, 2005: 51) who, in reference to man’s “perpetually … modifying his technology” McLuhan (1964: 46) states that “man becomes, as it were, the sex organs of the machine world, as the bee of the plant world, enabling it to fecundate and to evolve ever new forms”. He hereby suggests a hidden complexity to human-technologic interactions, a complexity I return to later in this work. For now though, these thinkers’ opposing perspectives make a further analysis rather difficult. I recognise McLuhan’s view that there are forces at work within the “network” that need to be addressed, but accept Leiss’ view that I should view the web of interactions as a whole, in order that technological determinism cannot skew my findings. I must reconcile these perspectives in my own approach. Seeking to refrain from any dangerously deterministic hyperbole, I continue the assessment of Mobile AR as an emergent and potentially “network” enhancing new medium, but from which determinist-proof methodology?

Digital Play, though referring mainly to the digital games industry, looks into the complex dynamics between developer, distributor, market and economy. Its critical evaluation of this medium adopts a methodology that suits my own AR enquiries:

“The story of the emergence of interactive play and of its uncertain crisis-filled transformation into one of the premier industries of digital globalized capital is both exciting and revelatory. Historical perspective is vital to critical understanding. We strongly agree with Williams that it is impossible to diagnose the cultural impact of a new medium until the specific institutional circumstances of its development are understood. Moreover, critical media analysis requires historical perspective in order to argue against the deterministic view that technology “is a self-acting force which creates new ways of life [Williams, 1992: 8]”.

Kline et al. (2003:79)

Now, since Mobile AR is such a new and radical technology, it is best considered in terms of other, previously radical technologies and their own timely impact. This approach recognises the emergent nature of Mobile AR technology and the lack of current research in the field, but also allows the opportunity to reflect on the implications of this technology in a relatively safe way: that is, through the lessons learned from full-fledged media. I propose that a useful lens through which to view Mobile AR is offered by an historical analysis of Mobile AR’s better established component media, with a view to producing an understanding of the implications AR holds for society. This approach allows AR to be considered as a product of a convergence of paths: technologic, academic, social and economic, providing the basis for deeper analysis as a consequence.

Gameware: A Case-Study in AR Development

I have been aided in this series by a connection with Gameware Development Limited, a Cambridge-based commercial enterprise working in the entertainment industry. Gameware was formed in May 2003 from Creature Labs Ltd, developing for the PC games market which produced the market leading game in Artificial Intelligence (AI), Creatures. When Gameware was formed, a strategic decision was made to move away from retail products and into the provision of technical services. They now work within the Broadcasting and Mobile Telephony space in addition to the traditional PC market. I use this business as a platform to launch into a discussion of the developments current and past that could see AR become a part of contemporary life, and just why AR is such a promising technology.

Gameware’s first explorations into AR came when they were commissioned by the BBC to develop an AR engine and software toolkit for a television show to be aired on the CBBC channel. The toolkit lets children build virtual creatures or zooks at home on their PCs which are uploaded back to the BBC and assessed:

 

A typical Zook, screenshot taken from Gameware's Zook Kit which lets children build virtual creatures
A typical Zook, screenshot taken from Gameware's Zook Kit which lets children build virtual creatures

 

The children with the best designs are then invited to the BAMZOOKi studio to have their virtual creatures compete against each other in a purpose-built arena comprised of real and digital elements. The zooks themselves are not real, of course, but the children can see silhouettes of digital action projected onto the arena in front of them. Each camera has an auxiliary camera pointed at AR markers on the studio ceiling, meaning each camera’s exact location in relation to the simulated events can be processed in real time. The digital creatures are stitched into the footage, and are then navigable and zoomable as if they were real studio elements. No post-production is necessary. BAMZOOKi is currently in its fourth series, with repeats aired daily:

 

BAMZOOKi, BBC's AR game show where children’s zooks compete in a studio environment
BAMZOOKi, BBC's AR game show where children’s zooks compete in a studio environment

 

BAMZOOKi has earned Childrens BBC some of its highest viewing figures (up to 1.2 million for the Monday shows on BBC1 and around 100,000 for each of the 20 episodes shown on digital Children’s BBC), which represents a massive milestone for AR and its emergence as a mainstream media technology. The evidence shows that there is a willing audience already receptive to contemporary AR applications. Further to the viewing figures the commercial arm of the BBC, BBC Worldwide, is in talks to distribute the BAMZOOKi format across the world, with its AR engine as its biggest USP. Gameware hold the rights required to further develop their BAMZOOKi intellectual property (IP), and are currently working on a stripped down version of their complex AR engine for the mobile telephony market.

I argue, however, that Broadcast AR is not the central application of AR technologies, merely an enabler for its wider applicability in other, more potent forms of media. Mobile AR offers a new channel of distribution for a variety of media forms, and it is its flexibility as a platform that could see it become a mainstream medium. Its successful deployment and reception is reliant on a number of cooperating factors; the innovation of its developers and the quality of the actual product being just part of the overall success the imminent release.

As well as their AR research, Gameware creates innovative digital games based on their Creatures AI engine. They recently produced Creebies; a digital game for Nokia Corp. Creebies is one of the first 3D games which incorporates AI for mobile phones. Gameware’s relationship with Nokia was strengthened when Nokia named them Pro-Developers. This is a title that grants Gameware a certain advantage: access to prototype mobile devices, hardware specifications, programming tools and their own Symbian operating system (Symbian OS) for mobile platforms. It was this development in combination with their experiences with BAMZOOKi and a long-standing collaboration with Cambridge University which has led to the idea for their HARVEE project. HARVEE stands for Handheld Augmented Reality Virtual Entertainment Engine.

Their product allows full 3D virtual objects to co-exist with real objects in physical space, viewed through the AR Device, which are animated, interactive and navigable, meaning the software can make changes to the objects as required, providing much space for interesting digital content. The applications of such a tool range from simple toy products; advertising outlets; tourist information or multiplayer game applications; to complex visualisations of weather movements; collaborating on engineering or architectural problems; or even implementing massive city-wide databases of knowledge where users might ‘tag’ buildings with their own graphical labels that might be useful to other AR users. There is rich potential here.

In HARVEE, Gameware attempt to surmount the limitations of current AR hardware in order to deliver the latest in interactive reality imaging to a new and potentially huge user base. Indeed, Nokia’s own market research suggests that AR-capable Smartphones will be owned by 25% of all consumers by 2009 (Nokia Research Centre Cambridge, non-public document). Mobile AR of the type HARVEE hopes to achieve represents not only a significant technical challenge, but also a potentially revolutionary step in mobile telephony technologies and the entertainment industry.

Gameware’s HARVEE project is essentially the creation of an SDK (Software Development Kit) which will allow developers to create content deliverable via their own Mobile AR applications. The SDK is written with the developer in mind, and does the difficult work of augmenting images and information related to the content. This simple yet flexible approach opens up a space for various types of AR content created at low cost for developers and end-users. I see Mobile AR’s visibility on the open market the only impediment to its success, and I believe that its simplicity of concept could see it become a participatory mass-medium of user-generated and mainstream commercial content.

What is AR and What is it Capable Of?

Presently, most AR research is concerned with live video imagery and it’s processing, which allows the addition of live-rendered 3D digital images. This new augmented reality is viewable through a suitably equipped device, which incorporates a camera, a screen and a CPU capable of running specially developed software. This software is written by specialist software programmers, with knowledge of optics, 3D-image rendering, screen design and human interfaces. The work is time consuming and difficult, but since there is little competition in this field, the rare breakthroughs that do occur are as a result of capital investment: something not willingly given to developers of such a nascent technology.

What is exciting about AR research is that once the work is done, its potential is immediately seen, since in essence it is a very simple concept. All that is required from the user is their AR device and a real world target. The target is an object in the real world environment that the software is trained to identify. Typically, these are specially designed black and white cards known as markers:

An AR marker, this one relates to a 3D model of Doctor Who's Tardis in Gameware's HARVEE kit
An AR marker, this one relates to a 3D model of Doctor Who's Tardis in Gameware's HARVEE kit

These assist the recognition software in judging viewing altitude, distance and angle. Upon identification of a marker, the software will project or superimpose a virtual object or graphical overlay above the target, which becomes viewable on the screen of the AR device. As the device moves, the digital object orients in relation to the target in real-time:

armarker2
Augmented Reality in action, multiple markers in use on the HARVEE system on a Nokia N73

The goal of some AR research is to free devices from markers, to teach AR devices to make judgements about spatial movements without fixed reference points. This is the cutting edge of AR research: markerless tracking. Most contemporary research, however, uses either marker-based or GPS information to process an environment.

Marker-based tracking is suited to local AR on a small scale, such as the Invisible Train Project (Wagner et al., 2005) in which players collaboratively keep virtual trains from colliding on a real world toy train track, making changes using their touch-screen handheld computers:

crw_80271
The Invisible Train Project (Wagner et al., 2005)

GPS tracking is best applied to large scale AR projects, such as ARQuake (Thomas et al, 2000), which exploits a scale virtual model of the University of Adelaide and a modified Quake engine to place on-campus players into a ‘first-person-shooter’. This application employs use of a headset, wearable computer, and a digital compass, which offer the effect that enemies appear to walk the corridors and ‘hide’ around corners. Players shoot with a motion-sensing arcade gun, but the overall effect is quite crude:

100-0007_img_21
ARQuake (Thomas et al, 2000)

More data input would make the game run smoother and would provide a more immersive player experience. The best applications of AR will exploit multiple data inputs, so that large-scale applications might have the precision of marker-based applications whilst remaining location-aware.

Readers of this blog will be aware that AR’s flexibility as a platform lends applicability to a huge range of fields:

  • Current academic work uses AR to treat neurological conditions: AR-enabled projections have successfully cured cockroach phobia in some patients (Botella et al., 2005);
  • There are a wide range of civic and architectural uses: Roberts et al. (2002) have developed AR software that enables engineers to observe the locations of underground pipes and wires in situ, without the need schematics
  • AR offers a potentially rich resource to the tourism industry: the Virtuoso project (Wagner et al., 2005) is a handheld computer program that guides visitors around an AR enabled gallery, providing additional aural and visual information suited to each artefact;

The first commercial work in the AR space was far more playful, however: AR development in media presentations for television has led to such primetime projects as Time Commanders (Lion TV for BBC2, 2003-2005) in which contestants oversee an AR-enabled battlefield, and strategise to defeat the opposing army, and FightBox (Bomb Productions for BBC2, 2003) in which players build avatars to compete in an AR ‘beat-em-up’ that is filmed in front of a live audience; T-Immersion (2003- ) produce interactive visual installations for theme parks and trade expositions; other work is much more simple, in one case the BBC commissioned an AR remote-control virtual Dalek meant for mobile phones, due for free download from BBC Online:

A Dalek, screenshot taken from HARVEE's development platform (work in progress)
A Dalek, screenshot taken from HARVEE's development platform (work in progress)

The next entry in this series is a case study in AR development. If you haven’t already done so, please follow me on Twitter or grab an RSS feed to be alerted when my series continues.