(Excerpt from 'Introduction', A.K. Bowman and J.D. Thomas, The Vindolanda Writing Tablets Volume III, London: British Museum Press 2003, 14)
The tablets were originally photographed at Vindolanda by Alison
Rutherford, using the infra-red technique established for the earlier
finds 1 . Once
again, we should emphasise these photographs are a remarkable achievement
and provide, in most cases, an excellent basis for reading and reconstructing
the ink tablets. In addition, we have been able to achieve further
improvement in legibility because of the availability in the latter
part of the 1990s of improved equipment and techniques for digital
scanning of the ink tablets with infra-red filtration. Digital scans
of the tablets offered a range of quality in the images which was
in most but not all cases an improvement on the conventional black
and white infra-red photographs 2 .
The differences or the inconsistency of the level of improvement
to be achieved can generally be ascribed to the the state of preservation
of the originals. In 1996 a preliminary scanning programme on the
ink tablets excavated in the 1990s confirmed the potential of digital
image capture. Following technical improvements in the scanning
equipment, new images were captured in 2000 and 2001 directly from
the tablets. Over 2500 scans have been made of all the unpublished
and published ink tablets. In this recent programme images were
obtained using a PhaseOne PowerPhase digital scanning camera back.
This was mounted on a studio camera body (medium-format Fuji GX680
II), taking the place of the film holder. The scanning back was
controlled using PhaseOne software (version 3.1.1) from a laptop
(Macintosh Powerbook G3). Tungsten lighting and a Kodak Wratten
87C infra-red filter were used. For archiving purposes grey scale
TIFF images were created at 850 dpi.
The scanning element (CCD or 'charge coupled device') within the camera back is made up of many photo-sensitive cells. Its very high sensitivity to colour differences makes it an appropriate tool for detecting the very faint traces of ink on the Vindolanda tablets. The software also offers the opportunity to enhance subtle contrasts in images. While scanning, the camera back compiles more information about the colour values for each pixel than can be stored. This additional information is discarded on conversion to the TIFF file format, but the conversion process can be manipulated in order to retain more information in the part of the colour range which is of most interest. The effect of these operations is to identify and enhance subtle differences in shade between the often very faded ink and the wood of the tablets, or between ink and dirt, in order to make the writing more clearly visible. An additional benefit of the high resolution scanning is the possibility to examine individual letters at a high degree of magnification on the screen, at several times their actual size. The image can be further manipulated by image processing software (e.g. Adobe Photoshop), although this is no substitute for an initial image of as high as possible a quality.
Like conventional photography of the tablets, the capture of digital images proceeded by trial and error until the best result was obtained. Unlike photography however, the results can be assessed almost immediately. The largest images, those of complete diptychs, for example, took only 6-7 minutes to capture. Further images could then be taken if necessary.
The archive images are filed for long-term storage in TIFF format on the University of Oxford hierarchical file server. They can be very large, for example between 30 and 40 MB for complete diptychs. For day-to-day use in the preparation of this volume CDs were produced of JPEG images generated from the TIFFs. Lower resolution JPEG versions of all the images will also be made available online on this website, hosted by the Centre for the Study of Ancient Documents. The website will initially offer images of tablets published in Volumes I and II; those in the present volume will be made available after publication of Volume III
The images captured can be viewed on the computer screen using software developed for handling image files. The most versatile of these is Adobe Photoshop which offers the advantage of sophisticated tools for adjusting greyscale, brightness, contrast etc. which can be applied differentially to different tablet images or to different parts of the same tablet to bring out the writing more clearly. The result of this is that we have, we believe, been able to read the tablets from the 1990s more successfully than would otherwise have been the case, although we are painfully aware that there is much which have failed to read or to interpret satisfactorily. Some confirmation of this belief was derived from the fact that the schedule of the scanning programme dictated that the tablets published in Tab.Vindol.I and II were processed after the new tablets from the 1990s and that our systematic examination of the older material has enabled us to make numerous improvements and corrections, many but by no means all of them of a minor kind.
1.See VRR II, 103-6.
2.The technique for scanning and enhancing stilus tablets is quite different because the writing is incised and therefore 3-dimensional. For this see Bowman and Tomlin, "Wooden Stilus Tablets from Roman Britain", in A.K. Bowman, M. Brady (eds.) Images and Artefacts of the Ancient World (British Academy Occasional Paper no. 4, Oxford 2005), 7-14; M. Brady, Xiao-Bo Pan et al., "Shadow Stereo, Image Filtering, and Constraint Propagation", ibid., 15-30.
The technical details which follow have been supplied by Dr. John
3.The work of preparing the images for the website and for this volume has been carried out by Dr. John Pearce and Dr. Frédérique Landuyt with the support of grants from the AHRB and the Andrew W.Mellon Foundation.