Staff Papers
Preparing Collections for Digitisation: the National Library's Digitisation Project
A paper presented by Lisa Jeong-Reuss and Chesley Engram, National Library of Australia at the AICCM Book and Paper Symposium, Melbourne, 17-19 April 2002
Abstract:
Digitisation is the current project buzzword in cultural institutions throughout the world and is spreading rapidly around Australia. The replacement of traditional photography with digital techniques in the production of preservation surrogate images of collections will have far-reaching implications on conservation and preservation departments. During 2001, the National Library of Australia (NLA) commenced a project to produce, from paper-based and photographic originals, around 40,000 digital images. This had a large impact on Preservation Services in terms of staffing, resources, and treatment needs. The paper will discuss this impact and hopefully provide some pointers in the process of preparing collections for digitisation.
Introduction
The National Library of Australia, in its Directions 2000-02 policy document stated that it would
digitise selected Australian original material and make them widely accessible through the Internet.
The subsequent Digitisation Policy 2002-04 further states that by digitising its collections the Library will
increase access to and help to preserve rare and fragile collections by providing digital surrogates and . produce publication-quality reproductions for Library users.
Preservation Services Branch consequently formulated the digital preservation surrogates strategy, which sought to enable the Librarys preservation copying program to move to a digital format. Many of the reasons behind this move were mentioned in the paper by Colin Webb and Deirdre Burgess at the last AICCM paper symposium .
Essentially, they include:
- Preservations continued involvement in all significant digitisation projects.
- The recognition that digitisation will not cancel the need to preserve original items, but will assist preservation goals by reducing the need for originals to be physically handled.
- The Librarys policy had been to use the traditional/analogue methods of microfilming and photography for preservation surrogates and access. Digitisation had been used mainly to create image files for catalogue access to Images 1 (now the Pictures Catalogue). A decision has now been made to use digitisation for both preservation and access.
Many institutions in Australia and overseas are still using analogue preservation copies to support digital copies. This is mainly because the framework to manage digital surrogates is not yet generally in place, however, despite this, the NLA decided to trial digitisation for both preservation and access. The main factor that influenced this choice was the Librarys commitment to a Digital Object Storage System (DOSS) and a Digital Object Management System (DOMS) as its archive. Preservation Services Branch will monitor and evaluate this approach over the next year.
For the financial year 2001-02, funds were allocated to provide staffing and equipment resources to implement a project to digitise at least 20,000 items from the Pictorial collection, 18,000 pages of sheet music, 3,000 Manuscript items and 1,000 maps.
The Library had previous experience with digitisation in projects such as the Rare Maps Digitisation Project (1998-99), Australian Cooperative Digitisation Project (1999), and the digitising of the Wills Diary (2000) and the Barton papers (2000-01). With the exception of Wills and Barton, most other projects have involved digitising from microforms or transparencies. As well, much of the Librarys digitisation of Pictorial items had been outsourced and produced from transparencies. Although the Library did consider outsourcing and called for expressions of interest for this new project, a decision for an in-house approach was made. This was to be a learning process, allowing Library staff to gain experience and understanding of the issues in a controlled, secure environment.
The review of the preservation surrogate aspect of the project will indicate whether digitisation is an appropriate method for producing preservation copies. Analogue copies of the more significant collection items will be produced during the trial period and beyond, as required.
Between February and July 2001 Preservation Services advised on the preservation aspects and issues associated with the establishment of this new Digitisation Project. A pilot project to digitise selected Australian sheet music commenced in April and by July the project team was fully staffed and up and running.
The majority of the items being digitised today are paper-based or photographic materials and because digitisation is progressively replacing traditional imaging methods, we thought it was timely to share our experiences of some of the issues with other paper conservators. This paper is in two parts. The first will describe the initial tasks involved in getting the project started from a Preservation Services viewpoint. The second part will talk about the first six months of the project, giving some details of the problems encountered with equipment, software, colour management and preservation issues.
PART 1
ISSUES
The project to digitise pictorial, music, manuscript and map material coincided with a review of the role of the Librarys Photographic Section. This was the beginning of a new era in the Library where digitisation is starting to replace traditional photography and may become the primary means of creating surrogates of collection material for preservation and access. Because of this, a number of issues needed to be addressed prior to the start of the project.
Selection of Material
Collection managers from the special collection areas compiled a list of materials they wished to be digitised. The initial selection included items that had:
- significant research and cultural interest;
- needed minimal preservation, cataloguing and copyright work; and
- were able to be captured easily.
Whereas previous digitising of Pictorial items had been done from transparencies, it was decided that all digitising for this project would be done from original material. This was mainly because the scanned file formats (JPEG derivatives from a proprietary photo CD) produced from transparencies did not meet all preservation and access requirements of the Library.
The first list comprised collections such as Frank Hurley photographic material, Norman Lindsay war poster originals, Geoff Pryor cartoons, Sir Joseph Banks papers, Sydney 2000 Olympic photographs, rare maps and Australian printed sheet music to 1899. Preservation Services staff assessed these items and reported on their condition. Resources needed in terms of condition, treatments required, rehousing and storage materials and provision of staff training, to prepare the items for digitising were also estimated.
Scanning Equipment
Preservation Services was not involved in the choice of equipment for digitisation, but there were issues that, along with collection managers, it was concerned and consulted about. Because the move from analogue to digital in creating preservation surrogates of Library items required that they be done to a suitable standard, a number of issues had to be resolved:
- high fidelity capture
- ability to monitor the quality and effectiveness of the digital objects
- maintenance of the digital copies
- ever-increasing storage capacity
- management of preservation metadata.
Capture and quality control checking are ongoing developments and are discussed in more detail in Part 2. The implementation of DOSS and DOMS is currently underway and is the responsibility of the Librarys Information Technology section (IT). It should be emphasised that the development of a management system is crucial to the effectiveness of any digitisation project and one should be in place before any scanning starts. The Library already possessed several flatbed scanners with the ability to process a wide variety of materials up to A3 size and it was planned to acquire more of the same. To accommodate the larger size items, fragile items and bound items camera set-ups were needed so digital scanning backs were needed. These are attachments which allow existing cameras to make digital images.
The preferred equipment was the Agfa Duoscan T2000XL flatbed scanner and Phase One PowerFX scanning camera backs. The Duoscan has tray inserts for scanning items such as negatives and transparencies of various sizes. This provided the means to digitally image a wide range of items including black and white, as well as colour, negatives, transparencies, glass negatives, photographic prints, printed items, artworks on paper, maps and easel paintings.
Computer hardware and software was also acquired for imaging and colour management. Adobe ® Photoshop 6.0 was used initially to manipulate and control the image capture. After consultant advice and software upgrades other systems have been trialled Part 2 will mention these along with colour management systems. Several storage and management systems were considered with the aim to have a system with a large capacity incorporated into the Librarys computer framework. The new DOMS will be known as the Special Collections Manager (SCM), will replace several current systems such as Pictorial Manager and will provide digital functionality to the Integrated Library Management System (ILMS). The ability to be accessed both in-house and by Internet users was also needed.
Guidelines and Training
As noted above digitisation for preservation is being trialled and is likely to become the preferred method of producing virtually all images within the Library. Because of this a reassessment of imaging methods needed to be done.
Flatbed scanners presented the same problems as photocopiers did in terms of the types of item that could safely be placed on them for scanning pastels and thick gouache watercolours, for example, are not suitable. Thus a set of guidelines for staff in the special collection areas was prepared to help them choose the appropriate digitisation method on the flatbed scanner or through the digital camera. These guidelines were designed to minimise the decision-making required by scanner operators and thus speed up the imaging output, without causing any inadvertent damage to the item. Preservation Services developed these guidelines in consultation with collection managers, the Digitisation Project team and Imaging Services staff.
Generally, the guidelines separate materials into two groups those that can be done on the flatbed scanner and those that must be done through the digital camera. The flatbed scanner is appropriate for items up to A3 size (420 x 297mm) that are robust, flat, without friable media, printed with black ink or written with ink. These include photographic material, sheet music, hand and mechanically produced prints and pen/ink drawings.
All large items (A3 and larger), most original coloured art on paper, fragile items, friable media items, bound items, manuscript diaries and letter books, historic process photographs, non-silver photographs, textiles, parchment and vellum documents and easel paintings are to be scanned using the digital camera. Discussion sessions were held with all special collections and Imaging Services staff to familiarize them with the new procedures and to encourage feedback.
THE PILOT PROJECT
The digitising of selected early Australian sheet music commenced in April 2001. This material was in relatively good condition and did not require extensive or complex preservation treatment to get started. Because of the cataloguing requirements needed, curatorial staff from the Music Section did the scanning and several training sessions in materials handling were provided. This training taught staff to do minor treatments such as tear repairs with Filmoplast ® tape, spine stitching removal and flattening of creases.
Procedures for digital imaging in the Library environment were trialled. For access and collection management purposes, each image needs to be given a catalogue record prior to being digitised. Thus a large amount of preparation other than conservation had to be done before the item could even be scanned. Figure 2 shows a schematic of the current workflow.
The pilot progressed well, producing many images, solutions to workflow issues and set the groundwork for the main project.
PART 2
STARTING THE PROJECT
By July 2001, the Digitisation Project was under way. A team had been assembled - seven scanner/cataloguers, two supervisors and a director. It was considered important to have someone with conservation/preservation experience on the team, and one supervisor was selected as the preservation officer. The rest of the team was chosen with an emphasis on cataloguing experience. Preservation Services had worked closely with the pilot team, but was unable to maintain the support needed once the project was under way. Besides undertaking assessments and treatments, the preservation officer needed to be a trouble-shooter to say nothing about acquiring cataloguing skills.
Workflow Issues
As mentioned above seven staff catalogue and scan the images. Most of the work is computer-based and work procedures can get tedious and repetitious, which raises occupational health and safety issues. Staff are therefore given different duties to break up the work schedule. These include assisting with simple conservation treatments and carrying out administrative duties. It is vital that a large area with enough storage room, good ventilation and working space be part of the plan for any digitisation project.
Most pictorial collections selected to be digitised are controlled by a collection level record. For access, file management, and to conform with the description of other Library collection materials, all items require individual catalogue records before scanning. Special Materials Cataloguing staff either catalogue items first or, more commonly, create catalogue templates and guidelines for Digitisation Project staff to use. This has developed into a time consuming task, as most of the guidelines to date have required a number of amendments after use due to the wide variety of materials found in collections being catalogued.
Since the start of the project, there has been a lot of contact between the Digitisation Project team and IT sections of the Library. The move to Microsoft ® Windows 2000 sitting on a UNIX platform has caused some problems which have not yet been fully resolved. This means that the team has to work with the system as best as it can until solutions are found. Problems include; system, server, metadata spreadsheet, along with Special Collections Manager, Pictorial Manager and Pictures Catalogue issues.
Digitisation specifications currently in use by the Library are shown in Figure 1 (Tables 1 and 2). Specifications and procedures have been developed to accommodate the needs of particular formats such as for the Librarys large collection of 35mm slides. The small format requires high resolution scanning to accommodate both preservation and access needs. A slide scanner magazine that can automatically scan up to 25 plastic mounted slides at a time was purchased and slides are now scanned at 3000 dots per inch (dpi). This aids in capturing the dynamic range of 35mm film and as a result, more faithful copies have been created. Unfortunately the Library also has large collection of cardboard mounted slides, which can jam when using the new slide scanner magazine. New equipment should be trialled with the range of materials to be digitised before purchase.
Preservation Issues
One of the preservation aims of the Digitisation Project was to provide a digital image that represents the original image and whole object as faithfully as technology allows, that is a preservation surrogate. However, this presents a problem in the case of a small image on a large support. The Pryor collection of 2,600 cartoons and drawings is a good example because, although they were usually on paper larger than A3 size, the actual images were approximately A5 size. High quality digital images have large file sizes and require a large amount of storage space. In this situation, for pragmatic workflow reasons, collection managers in collaboration with the Digitisation Team and Preservation Services agreed that each digital image be cropped to the edge of the original image area. This decision was made partly to assist the workflows as items larger than A3 are usually captured with a digital camera, a process that takes considerably longer. Quality images also usually have a large file size requiring more storage space. Cropping is not a new issue when creating a preservation surrogate, the same decisions have to be made with conventional reformatting techniques of photography and micro-reprography. Now decisions on how to crop individual collections are considered and made as they arrive.
The Library has also considered compressing master digital images, but decided not to go ahead because the simpler the processing applied to master images the greater the likelihood of successful migration when the time comes. Compression is considered to be a kind of manipulation, which would make the digital master a slightly less faithful representation of the original item. It is obvious that a tension exists between capturing the best image for both preservation and access and managing a cost effective digitisation project with prescribed outcomes.
Colour Management
Quality control checking of coloured pictorial material identified that most colours of the digitised images did not match those of the original. A colour consultant was contracted to help develop a better colour management strategy. The first recommendation was to change the viewing environment, so that the monitor is viewed under dim illumination (less than 64 lux), reducing glare and allowing the eye to adapt to it. If the original and the screen image are viewed side by side, the original needs to be viewed under daylight bulbs (4700K) and the monitor should be shielded from the daylight bulb. The monitor should also be situated in an area where all walls, floors, pictures, furniture and signage in the field of vision is grey or black. No major colour distractions are to be placed in front of scanner operators and all glare should be avoided (Frey 2000, Nadish Naoroji private communication 2001).
The consultants next recommendation was to change the existing colour management program by calibrating the scanners and monitors. A Gretag Macbeth® colour target was used to help calibrate the scanner. This colour target is made up of patches of colour photographic paper with different surface finishes and was found to be more suitable for the material being digitised. After colour calibration was complete, the colour profiles were automatically assigned as the images were scanned. The monitors were also profiled to the CIE D65 viewing standard using the X-Rite Monitor Optimizer. Each file or image was converted to the working space (Adobe® RGB 1998). Since then we have created additional colour profiles better suited to specific materials. For instance, the Library has a large collection of modern photographic prints on Fujifilm, AGFA Professional, Kodak and Archival Silver Gelatin photographic paper. After a series of tests, we found the KodaColor target to be most suitable for creating a profile for these more recent emulsions. It is important to recognise that in some cases, colour profiles that suit specific material characteristics have to be developed for best colour capture (Naoroji private communication 2001, CIELAB 2000).
Another problem encountered while digitising was due to the silvering out on a collection of black and white photographic prints. The silvering on the prints manifests as a blue flash when scanned. These flashes are a result of the moving scanning lights hitting and reflecting light from silvered areas, resulting in a blue flash. The solution has been to reposition an image by turning it 90º and rescanning. Another problem is material with metallic paint that could not be reproduced with the flatbed scanner. Flatbed scanners cannot capture the subtle shine or glints of metallic paint and these items are better captured by digital camera.
Workflow
The workflow chart (Figure 2), describes the processing of pictorial items to be digitised. Items are firstly chosen by collection managers in collaboration with the Digitisation Team and Preservation Services staff. Items are then assessed as to their suitability for scanning or photographing with a digital camera. After assessment and any necessary treatments, collection items are catalogued onto the National Bibliographic Database (NBD) via the Kinetica service. Treatments may vary from minor to complex and currently take up 25% of the preservation officers time. Finally the catalogued items are ready to be digitised by scanner or digital camera. Additional and time consuming routine digital camera work being undertaken by Imaging Services has also affected Digitisation Project workflows and completion.
Once items are scanned they are automatically placed into the DOSS. The last two stages of the Digitisation Process flow chart (Figure 2), indicate when images are uploaded into the Pictures Catalogue. This part of the process is currently being upgraded. The new Special Collection Manager database will record processing, management and technical information (metadata) about digital collection items. The database is designed to support the upload of images of pictures, maps, manuscripts and sheet music to the DOSS server. Benefits include the elimination of digitisation spreadsheets and extended filenames, faster uploads, uploads controlled by the digitiser, ability to check and review processes online and improved management of master and derivative images and technical metadata.
CONCLUSION
With the implementation of the Digitisation Project, the Library has further committed itself to the pursuit of digital strategies in creating wider internet access to its collections. The decision to create digital surrogates for both preservation and access is a courageous one which the Library will have to review regularly, time will tell its success. The importance of the involvement and contribution of preservation staff in planning and setting-up digitisation projects cannot be emphasised strongly enough. This extends from the provision of training sessions to the intricacies of image resolution and metadata preservation staff have to be involved along with information technology, imaging services, cataloguing and collection curatorial staff.
References
Colour Manager Manual, The CIE Lab colour space, 2000 Available:http://www.linocolor.com/colorman/
Elisabeth Novitski, Internal handout,2001
Franziska Frey, Measuring Quality of Digital Masters, Online. Available:http://www.rlg.org/visguides4.html.
Linda Groom, Internal minute, 2000
Author Biographies
Chesley Engram, National Library of Australia, Manager, Treatment Programs, Parkes, ACT, 2600. Tel. +61 2 6262 1566. Fax +61 2 6257 1703. Email cengram@nla.gov.au.
While completing the BAppSc (Conservation) course at University of Canberra, Chesley worked at the Australian War Memorial Paintings Conservation section. On graduation, specialising in paintings and paper, he commenced as conservator at the National Library of Australia and from 1990 - 1998 he was its Exhibitions Conservator. He is currently manager of the treatments program in the Collections Preservation section of Preservation Services. His interests are photographic, paper and paintings conservation and recently, matters digital.
Lisa Jeong-Reuss, National Library of Australia, Preservation Officer, Parkes, ACT, 2600, Tel. +61 2 6262 1406 Fax. +61 2 6273 5763. Email ljeong@nla.gov.au
Lisa graduated from the University of Canberra in 1998 with a Bachelor of Applied Science (Conservation). Since graduating, she has worked as a paper conservator at the Australian War Memorial, National Archives of Australia and Australian Maritime Museum. Lisa has been employed at the National Library of Australia for the past 3 years in Paper Conservation. She is currently the Preservation Officer of the Digitisation Project ensuring attention is given to preservation aspects of the project work.
Figure 1: Standards for creating digital images
Standards for creating digital images
Table 1: Colour material
| Material | Pixel depth | Resolution |
|---|---|---|
Colour, reflective,
including:
|
RGB 24 bits per pixel |
300dpi; 600 dpi for items with fine detail or which are very small (smaller than A5) |
| Colour transparencies > 35mm | RGB 24 bits per pixel |
2000 dpi |
| Colour negatives and transparencies 35mm | RGB 24 bits per pixe |
3000 dpi |
| Colour negatives larger than 35mm | RGB 24 bits per pixel |
2000 dpi |
Table 2: Black and white material
| Material | Pixel depth | Resolution |
|---|---|---|
Black and white
reflective material, including:
|
Grey scale 8 bit |
300 dpi; 600 dpi for items with fine detail or which are very small (smaller than A5) |
| Black and white negatives 35mm | Grey scale 8 bit |
3000 dpi |
| Black and white negatives larger than 35mm | Grey scale 8 bit | 2000 dpi |
(Groom, 2001)
