The digitisation by the HUMI Project
(17 November - 4 December 2003)
Supported by the Japanese Ministry of Education, Culture, Sports, Science and Technology's Open Research Center (ORC) Project.
The digitisation of the Winchester Manuscript was undertaken by a team of nine Japanese experts, led by director Professor Toshiyuki Takamiya and technical director Professor Masaaki Kashimura.
The team measured the sizes of the text area, margins, manuscript pages and the thickness of the book, and also checked the binding. This enabled them to estimate the distance between the surface of the page and the camera, and to set the position of the specially designed book cradle.
The camera used for the digitisation of the Winchester Manuscript was the Kodak DCS Proback Plus + Mamiya RZ67 Pro II digital camera system (16M pixel (4072x4072), 36-bit RGB (12 -bit depth per channel) ), which can produce a 16-million pixel image. The lens used was a Mamiya-Sekor M 140mm F4.5 M/L-A, at F16.
A black curtain was suspended in the studio, and the team wore black clothing to minimise unwanted colour casts and reflections. Standard colour charts were regularly used to verify the colour accuracy.
The book cradle used was originally designed for the Gutenberg Bible Project, and modified to suit the Winchester Manuscript. It consists of three separately moveable parts, which can be gently manipulated so that the book is never open more than 110 degrees and yet can present a nearly flat image of each page to the camera. The combination of this adjustable book cradle and a square mirror allowed the team to ensure that the page surface was always parallel to the camera without placing the camera above the manuscript.
A specially made focusing device was used to bring out the best from the camera system. The team realised that the focal point on the focusing screen is slightly closer than the actual point. Hence they used the focusing device, which allowed correction to an accuracy of 1 mm. This was a three-dimensional wedge-shaped device placed on the book. On the top slope away from the book there was a graduated scale with Chinese characters meaning 'Keio'. The Chinese character indicating the sharpest focus showed the exact distance between that point and the surface of the book using a corresponding millimetre scale. This, together with the camera stand, also designed by the HUMI Project, allowed the operator to lower the camera by exactly the same distance and gave a reliably sharp focus.
The structure of the book moves gently as the leaves are turned, slowly increasing the distance between the book and the camera. This was corrected by checking the image size in pixels on a computer screen.
Once photographed, the image underwent primary processing in the camera and was then processed through a battery of three computers. It was important that as much of the processing, checking and backing up of the images took place immediately, for once the team and equipment returned to Tokyo it would not be possible to rectify any errors.
The manipulation of the camera position and the light setting allowed the team to capture some hitherto-veiled features such as watermarks, scribal corrections in multiple colours, dry-point glosses, and offsets of Caxton's printing ink discovered by Dr Lotte Hellinga in 1977.
The whole process took about 25 seconds for each image. It took the team 14 days to shoot the whole manuscript: 971 full-page-size images and 44 special shots in total.
Back in Tokyo the team processed the TIF archive images (2800 X 4072 pixels) to create sets of JPEG files of different sizes suitable for delivery via the web. The images provided for this preview site are sized 1400 X 2020 pixels, compressed to 'standard' quality and they can be viewed on the samples page.