Internet Archaeology 10 - Jeffrey. Conclusion

6.0 Conclusion

In the examples presented within this article, hopefully it is clear that this technique, while not offering all of the advantages of full 3D models incorporating DTM/DEMs, does provide an efficient, accessible and accurate method of presenting reconstructed objects as if they were in context. In relation to the constraints mentioned in the introduction, this approach offers the benefits of 3D modelling of the artefact while taking advantage of QTVR panoramas to represent landscapes accurately. These advantages, while applicable to other types of monuments and artefacts, are particularly important in presenting models of early medieval sculptured stone. This is because these monuments cannot easily be modelled using only QTVR object movie techniques, nor can their relationship with their landscapes be adequately explored using DTMs generically rendered and without reference to the horizon. The lack of interactivity in comparison with VRML that would diminish a model of a building or settlement does not have such a profound effect when dealing with a single object that can easily be viewed from all sides. This is especially true when alternative landscape viewpoints are offered via linked QTVR panoramas.

The interactivity offered by the resulting QTVR object still allows the user to rotate the image and zoom in and out. Indeed, it would also be simple to use the spare vertical angled image slots to appear to change the angle of lighting on the monument, although in the case of the example in this article the photogrammetric model of the cross does not have enough surface detail for this to be worthwhile. Additionally, the object movie of the artefact can be embedded within a sequence of linked panoramas as discussed earlier. This would allow the user limited movement around a landscape, but with far better detail and realism than the more interactive VRML.

One of the uses that is most likely to be made of QTVR object movies generated in this way will be as an interactive front end to a tree of further information. The reason for this is that although Internet deliverable models of about 1Mb give a good impression of the landscape, and the 3D shape and location of the monument, the image resolution is poor. A high resolution version of the QTVR object movie used as an example here (Fig. 9 a & b) is around 50Mb in size. This object movie can be run quite adequately on a local disk and gives image quality as high as the original photographs, but is obviously not yet suited for Internet delivery. A low resolution object movie acting as a clickable interactive technique for selecting high resolution 2D images could be a reasonable compromise.

This new approach to presenting models of archaeological objects in photorealistic landscape contexts offers many of the benefits granted by both QTVR and three-dimensional modelling. It is unlikely that this technique would ever replace full three-dimensional landscape modelling where there is the information, hardware, software, time and money available to use it. But in situations where one or all of the above factors is an issue, then it could be seen as a viable alternative, especially where Internet delivery is crucial. The non-reliance on large geographical datasets also sits well with a move away from sophisticated and expensive VR methodologies, striving for a near perfect digital copy of the world, towards a general critical acceptance of VR models as being as much laden with layers of interpretation as any other form of representation. As with all modes of representation, this technique should be approached critically by its intended audience and the originators of such representations should be as explicit as possible about both the methodologies used and how they expect their representation to be perceived. For example, as Gillings and Pollard state in relation to VR representations 'far from treating the models as deficient surrogates a more productive approach is to characterise them as manufactured intensities, emphasising their hyperreality and inauthenticity and encouraging archaeologists to explore their true potential for the study of issues bound intimately with engagement and process' (1998, 148)