Internet Archaeology Archive: Publishing multimedia in archaeology

Published in EVA '97 London Conference Conference Proceedings: Electronic imaging and the visual arts, 1996, 5:32-37.

Publishing multimedia in archaeology

Alan Vince & Sandra Garside-Neville

Department of Archaeology,
University of York,
The King's Manor,
York
YO1 2EP

Introduction

In this paper we will describe the way in which an electronic journal, Internet Archaeology, has been using electronic visualisation and some of the issues that we have had to face in trying to make innovative use of the World Wide Web, but in a context where we aim to have the content online in perpetuity. We therefore have to balance innovation and experimentation against maintainability.

Internet Archaeology - An electronic journal of record

Internet Archaeology was conceived early in 1995 and Alan Vince was appointed as its first Managing Editor in August of that year (Heyworth et al 1996). The journal is funded by the Joint Information Systems Committee (JISC) as part of eLib, the Electronic Libraries Programme and is managed by a consortium of UK universities (York, Oxford, Durham, Southampton and Glasgow), the British Academy and the Council for British Archaeology. We have now published two issues, comprising ten papers, and have number of papers in preparation for our third issue, scheduled for August 1997.

Presenting multimedia in an e-journal

World Wide Web browsers support a wide variety of file types and by the use of "plug-ins" it is possible to increase this variety still further. There is therefore little technical difficulty involved in publishing static images, moving images or sound. However, many of the formats used to serve these images are proprietary, or experimental, or subject to periodic revision (eg GIF87 vs GIF89, the various TIFF formats, JPEG formats and so on) and on the basis of previous experience with computer software we can be certain that within a matter of years many of these formats will not be supported by the latest generation of web browsers. For most webmasters this hardly matters, since they would expect to be periodically updating their sites. In the case of an e-journal, or any sort of electronic archive, this is more of a problem and for e-journals the problem is acute. Every paper is the product of a different author, with access to varying sets of software and hardware. We have authors using Windows PCs, Macs, Linux boxes and Unix workstations. In most cases, the work they are doing uses computers but with a view to a printed end product. The range of tools required to filter, translate and manipulate multimedia are not normally present on their systems and it is beyond our capabilities to advise them unless their systems are similar to those used at York. Macs, for example, are a complete and utter mystery and we have wasted hours dealing with Mac/DOS/Unix compatability problems. Like most publishers (except those producing for the mass market with budgets for graphics) we compromise on every paper, trying to achieve the best quality for the smallest file size and using the closest thing we can get to established standards.

Static images

Most of the images in Internet Archaeology come to us as scanned colour photographs or monochrome line art, despite the widespread availability of painting and drawing packages for personal computers. We suspect this is partly a legacy of tradition. Archaeologists use colour slides in lectures and give a short, snappy verbal presentation for each slide but then publish their work with very few illustrations and with minimal explanation of the image, perhaps simply a one-sentence caption and a brief mention in the text. This tradition is certain reflected in the majority of the papers submitted to the journal and except to encourage people to be more adventurous and sometimes to add a dash of colour to a submitted monochrome diagram, there is little that a web publisher can do to change this.

Data Standards

There are now a large number of image formats, all trying to achieve similar goals: high compression; minimal loss of information; the ability to zoom in to examine detail; accurate colour reproduction and so on. In addition, we also need to know that the format can be used without copyright problems and is truly cross-platform. There are, of course, strong pressures to simply abandon the 10-20% of Internet users who do not use Microsoft operating systems and go with the flow. There are some interesting developments here, such as the Fractal Imaging format (FIF) marketed by Iterative Systems and Kodak's new flashpics format which allows a single image file to be displayed at a number of different resolutions, thus cutting down transfer time. However, essentially there are only two image formats in common use: GIF and JPEG. Early browsers could be relied upon to view GIF files but not JPEGs, and therefore, if for no other reason, GIFs are commonly used for inline images whereas JPEGs are reserved for external, larger images. There can be few browsers today which do not support both formats but the convention remains.

Artefacts

We have found some difficulty in getting good quality line art from our authors. This may be due partly to the inexpert use of scanners. Often, we find that producing a paper for IA is the first contact our authors have had with a scanner. However, it is obviously better to persevere with trying to get authors to produce these images rather than giving in and getting them to send the artwork for scanning in York. Most artefacts are drawn using the standard conventions for archive or print publication and would benefit from the use of colour and texture. For example, we have included some illustrations of copper alloy artefacts in which the presence of enamels is shown by shading. This is clearly ridiculous. Colour photographs, normally scanned as slides, are the next most popular method of illustrating artefacts. A big difference in quality exists between slides scanned using an attachment for a standard flatbed scanner and those produced from Photo-CD images. The use of colour images throughout a paper makes a big difference to its impact. In one case, for example, we have a glossary of terms used in discussion of Palaeolithic stone tools, from the raw material to finished objects. As a simple list this is pretty dull stuff and as a list illustrated by line art, the most one might expect in a printed journal, it is still rather mundane but when illustrated with colour photographs even this detail becomes pleasing to view, thus increasing the changes of people actually reading the glossary and absorbing the content.

One of the worst problems for archaeological publishers is the archaeological plan. Conventions have grown up about the what ought to be shown on a published plan and these mean that there is almost always a conflict between what is wanted and what is practicable. For example, plans are usually required to be at standard scales, for ease of comparison, they distinguish between what was found and what was surmised and between areas which have no information (because they were not investigated or were destroyed by later activity) and areas in which nothing was found because nothing was ever there. This is a lot to ask of a single, monochrome image and there has been pressure upon print publishers to do two things: firstly to allow at least one other spot colour on their plans and secondly to allow foldouts or even separately boxed loose leaf sheets. Both of these options add to the cost of publication but enhance the usefulness of the finished product. We have yet to see any archaeological plan on the Web which is satisfactory. Most, admittedly, are actually scanned in from an A4 printed original and should not therefore be used to damn the concept. However, it is difficult to see how we can get around the problem of uniform scale when an image can appear at different sizes even on one computer monitor, depending on screen resolution. On the positive side, however, it is possible to use clickable imagemaps to link one plan to another, or the plan of a structure to its description, or a finds catalogue or photographs of the excavation.

Much of what we would like to do with archaeological plans is in fact already possible with CAD drawings. Vector graphics can be viewed at any scale, depending only on the accuracy of the original digitisation, and the drawing includes information about the drawn objects in the form of a "layer" or "attribute". These layers and attributes could be used to control viewing and linking across the web, should the appropriate software be available. There are several plug-ins available for DXF format vector files and several online archives of such drawings, though none, so far as we know, dealing with archaeological drawings. This ability to link drawn object with an underlying database is taken further in geographical information systems (GIS) and there are numerous attempts to make these work across the web. The most sophisticated of these projects allow the entire map to be customised and for a complicated archaeological site this could be an excellent way of presenting plan data, especially if the drawing was linked to a sophisticated database. It would, for example, be possible to send the following request by means of a form:

Display all features in existence during Period 5 and plot the frequency of pottery and animal bone as small histograms, pie charts or variably-sized symbols.

For a small number of variables and periods it is possible to achieve this already, by precompiling the relevant data as a series of GIF images and using either Java or a CGI script to serve up the appropriate map. The amount of investment in developing such a system that would allow us to use archaeological data in its native formats (relational databases and vector drawings) is high, especially since there is little standardisation in the use of software packages by archaeological projects. However, there is a strong likelihood that the producers of GIS software will themselves produce products that will carry out this task, since GIS is becoming embedded into many commercial and governmental organisations where remote access through the Web is seen as an advantage. It would then be the author's responsibility to obtain a copy of the GIS software and to import their data into it.

Distribution maps

A specialised aspect of the presentation of geographical data in archaeology is the distribution map and Internet Archaeology includes several papers (three in Issue One, one in Issue Two) which use distribution maps as an interface to archaeological data. All those used so far have a simple style, with grey coastlines and rivers, green land and blue sea but in future papers we will be including colour shading based on a public domain digital terrain model (DEM). Those used to date have little of the functionality of even the simplest GIS system in terms of control over the appearance and content of the map. The most complex is that accompanying a paper by Paul Tyers on the distribution of types of Roman amphora in the British Isles and western Europe since this includes two different base maps and a dozen or so different overlays. By contrast, the most sophisticated GIS or map viewer we have seen on a CD-ROM uses a series of zoomable vertical aerial photographs of London to build up a highly realistic map of the capital, an ideal starting point for any study of the modern city or its archaeology and perhaps the shape of things to come. However, it will be a long time before such software becomes commonplace on the Web.

Moving images

Many of our authors and potential authors have access to video cameras but few have access to the equipment needed to produce video of formats commonly supported on the web (Quicktime (.mov files), Microsoft's Video for Windows (.avi files) or MPEG, for example). It is not particularly surprising therefore to find that we have not been offered any video clips. Furthermore, as any regular web surfer will confirm, there are severe problems with most of these formats, in that they produce huge files which do not start to play on the users computer until the file is completely downloaded. More recently, We have seen VivoActive video which streams onto the users computer and starts to play after only a few seconds downloading. Within months, this technology, if combined with an increase in bandwidth/reliability, will probably solve the technical problems of publishing video. We then need the widespread availability of good quality analogue-to-digital video filters and for VivoActive (or whoever it is by then) to make their video writers, as opposed to viewers, to be cheaply available. Until then, for serious web use video is still going to be a novelty. After all, even video on CD-ROM leaves a lot to be desired when compared with television. There are, however, two areas where moving images have been used to advantage in Internet Archaeology, animated GIFs and VRML.

Animated GIFs

If a video sequence of a few tens of images is required, this can be achieved using animated GIF images, for which there are now a number of construction kits available. Most animated GIFs on the web, however, are used to produce icons and adverts rather than serious content. Nevertheless, two of our authors, Christian Beardah and Mike Baxter, used an animated GIF to produce a 360-degree view of a complex 3D graph. We would also like to see people experimenting with the use of animated GIFs to produce images of simple artefacts, such as pots, which need two or three shots to indicate their form to the viewer. One drawback of these images, apparently, is that unless the number of cycles is stipulated, they continue to download all the time they are being viewed, and thus add to the waste of bandwidth on the Internet.

VRML

We have published one paper, by Glyn Goodrick and Mark Gillings, which exploited VRML to visualise and study landscapes (as well as a rather jolly Neolithic pot). Two case studies were published in their paper, one of a site on Hadrian's wall (Peel Gap) and the other, the site of a medieval dwelling. Even on a Silicon Graphics workstation these files take a long time to load. Furthermore, we had difficulty in getting them to run using different VRML browsers, in one case due to the rapidity in the change of specifications for the VRML language. Clearly, it would be irresponsible to adopt VRML as a regular means of portraying 3D information to the public at present. Despite this reservation, it is nevertheless also clear that VRML or a similar web-based standard will soon be mature enough to use and that when it is, it has tremendous potential for presenting the results of archaeological excavation and fieldwork. There will probably always be situations where more formal plans and elevations are of more use but for the visualisation of an archaeological site or landscape three dimensions are definitely better than two and the ability to explore a site from any vantage point is better than having a single fixed viewpoint.

Sound

There are few cases in archaeological publication where the use of sound is crucial. One of these is in the study of early musical instruments and their experimental reproduction. Another might be in the study of the history of archaeology, where the reminiscences of some of the pioneers of the discipline would be valuable. However, even with the development of streamed audio it is clear that at present the medium for such information is television or film, followed soon perhaps by DVD.

Challenges for the future

It seems to me that there are already some areas in which the Web is a mature medium, in that it can deliver a fast, consistent product. These areas are: text, hypertext and small images (still and moving). Much of what we collectively term "Multimedia" on the other hand can only be served across the Internet in a haphazard way. It is difficult to believe that a specialist discipline like archaeology will invest time and money in either production or consumption of information in these media until they are more mature. Indeed, except in the case of flagship projects they may always be too expensive for us.

Our limited experience and information of the authoring process for multimedia as realised on local computers and CD-ROMs suggests that whilst there will always be people willing to put large amounts of effort into the production of multimedia presentations for their own amusement, or for the mass market (eg teaching aids, as with the TLTP consortium, or as part of multimedia encyclopaedias, such as Encarta), it is unlikely that it will become a standard method of archaeological publication until cheap authoring packages and standardised templates become readily available.

The World Wide Web cannot, at present, do justice to true multimedia presentations, in particular because of the lack of control over timing and the lack of options for transition from one image or screen to another. These things will come, but whether the Web is the right medium for extended presentation anyway is doubtful. In any event, so long as IA is based in an academic institution and connected to the Internet through JANET we are barred from including advertising on the web site, and much multimedia seems to consist of elaborate credits and logos rather than content.

There are several media available now which are suitable for multimedia presentations and the advent of DVD ought to give these formats a new lease of life. For archaeological publications, these should mean that anyone wanting every photograph and plan taken on the site can do so whilst we can publish on the Internet with enough detail to satisfy most readers and whet the appetite of anyone who might want a fuller presentation.

Bibliography

Grace, R (1997) The `chaîne opératoire' approach to lithic analysis Internet Archaeol 2
( http://intarch.ac.uk/journal/issue2/grace_index.html)

Heyworth, M, Richards, J, Ross, S and Vince, A (1996) Internet Archaeology: An international electronic journal for archaeology. Archeologia e Calcolatori 7, 1195-1206

 HOME   ABOUT   FOR AUTHORS