JISC 3D Visualisation in the Arts Network

3DVisA Bulletin, Issue 1, September 2006

ISSN 1751-8962 (Print)
ISSN 1751-8970 (Online)

3D RESOURCES: Computer Reconstruction. TEMPLE SITE at PHIMAI, Thailand. A Review by Michael Greenhalgh, Australian National University, Canberra, ACT Australia

Introduction

The days when computing was text-only are long past, and a combination of computer and network speed, colour, WWW and graphics software has considerably expanded our views onto the world. For some, the computer monitor is indeed the window into a different world; one which may have few connections with that in which we all live. Of course, those worlds are different for all kinds of reasons, and the salient point about worlds inside the computer is that they must somehow be created with software, and are not necessarily intended to be a simulacrum of our real world. Indeed, assuming we want and need that computer world to be 'real', then the software must create the illusion of reality, some of the features of which include perception of depth, texture and even atmosphere.

What might a visitor expect to learn from a site such as Phimai? In basic information, perhaps details of the UNESCO submission for listing, something about the Phimai Historical Park, and links to the many other Khmer monuments available across the web. Remote sensing has proved very useful in the area, with an interesting report to UNESCO (PDF format) by Surat Lertlim, and perhaps information on heritage management in Asia, and comparative data from GAP - the Greater Angkor Project.

Description of the Website

Completed in 2001, the website is the work of Richard M. Levy, Associate Professor of Planning and Urban Design, and Director of Computing in the Faculty of Environmental Design at the University of Calgary. It is divided into a suite of HTML pages designed by Peng Peng: an Introduction gives a summary, and a project description; The Site gives access to photographs of various kinds, to a QuickTime movie, and to a plan of the Phimai site; it is spoiled by a line of seven thumbnail locations five of which change, two of which do not, and none of which is clickable and hence enlargeable. The Computer Model offers animation sequences, a QuickTime Movie, and access to a series of enlargeable thumbnails on Constructing the Model. There are also contact details, awards and exhibitions details.

Quality and Extent of the Content

The quality of the site design is fine, but the range of the content (see above) could be expanded. So also could the size of the images: all of three historic photos, and the twelve air photos are only 450 pixels wide. Some of the 'air photos' are in fact reconstructions, which is (probably unintentionally) deceptive. Any site dealing with both photographs and models should keep very clear boundaries between the two.

The Photo Gallery offers sixteen images, again all too small, but at least keyed via a red dot to their location on the adjacent site plan; but the plan is dead; you cannot click on a section of it and bring up the relevant image. Initially the heart leaps as the siteplan on the first page is indeed clickable, but all this does is bring up a larger version. ALL plans are in Thai script, which might be a bit of a steep learning curve for some of us. A difficult job in PhotoShop, scrubbing one script and labelling it in English...

As we have now come to expect, the Computer Model again offers small images only. There is no zooming, so it is impossible to get in close enough to see the detail if any. All the images are 'dead' - the user has no control over detail - and can only move from one image to the next.

The Purposes of the Model

The Introduction offers a somewhat politically correct explanation for the model: 'Reconstruction of the temple site in Phimai serves as a case study highlighting the potential of computer visualization as a tool in heritage resource management. Besides offering archaeologists, historians and museum curators a non-evasive [invasive? MG] environment for testing reconstruction scenarios, virtual worlds offer the public access to important historic monuments without the wear of excessive visitation'. This provokes several questions:

The notion of 'testing reconstruction scenarios' is misleading for monuments such as Phimai, where the setup is so structured and well known from many comparanda that the reconstruction is never in doubt. Such 'reconstruction scenarios' are indeed a use for such models, because they necessarily operate where the archaeologists have little idea of detail, and just want to block things out - which is conveniently just what such models are good at (but nothing much beyond such basics!). If virtual worlds can indeed 'offer the public access to important historic monuments', then surely setups such as this should be much richer in materials.

A final comment on the aims might be that the Freudian slip of evasive is that this is exactly what such models do: they evade the twin requirements of any scholarly work, which are detail and accuracy.

The Quality of the Model

The models provided are as good as the technology allows. Constructing the Model is the most interesting page, in that it demonstrates clearly the problems inherent in the technology - namely the simplification of forms and of textures, the repetition of both, and the overall unworldly look-and-feel (best seen in the comparison between the computer interior view and the photograph). The difficulties should not surprise us, since the process involves stripping down the real world to its computer-understandable components, and then rebuilding it in the machine, which is far from simple, and tedious, time-consuming and expensive to do to any level of accuracy. But to repeat, any shortcomings are not Professor Levy's fault, but endemic to such modelling. This might explain why the computer models on these (and plenty of other) pages are shown at such low resolution: seen in close-up the reconstructions would reveal themselves even more clearly to be lacking in both detail and accuracy.

Will Computer Modelling get Better?

The problems inherent in the technology seem incapable of solution. VRML and its descendants have been around for several years now, but have never lived up to the hype which is endemic to computing. The killer question is to ask what a computer model of any such site as Phimai can yield to the student/researcher than a good collection of photographs cannot? The answer must surely be 'very little', since the models are extrapolated from photographs, in the first place, gaily cutting corners (I know, I've done it) along the way. Even when money can be thrown at laser measuring and a huge turntable (as with the Digital Michelangelo Project), the result is no doubt one of micro-accuracy contour-wise, but with a surface as dead as old putty - the textures are still any enormous problem. And in any case, Phimai has insides as well as outsides, and lasering it for computer modelling would be for rich people with time on their hands.

I live in hope that all the doubts about the technology expressed above are now ill-founded. All any protester has to do is to email me with details of any reasonably complicated inside-and-outside model (no more Renaissance statues, thank you, far too simple), and I shall eat this HTML file and re-embark on building accurate and detailed computer models which may fearlessly be shown in a web browser at high resolution. Until I get that email, I shall stick to photographs, panoramas, zooming devices and the rest - and in large enough quantities to give some satisfaction to researchers.

© 3DVisA and Michael Greenhalgh, 2006.