NINCH guide home interview table of contents previous interview next interview
On January 25 2001, HATII interviewed Andrea Notman, Registrar for Collections at Harvard University Fine Art Museums, and Peter Siegel, Director for Digital Imaging at Harvard Fine Arts Museums and the Fine Arts Library. The Collections Management System was established to support initiatives within the Museums, such as exhibition programming, internal research interests and educational schemes. The digitization of the source materials not only allows for increased accessibility and cross-fertilization between departments in the Museums, but also enables more fragile objects to be accessed without risk, as well as being an effective method of cataloging new acquisitions.
The Harvard University Fine Art Museums include the Fogg Art Museum (Western art from Middle Ages to present), the Busch-Reisinger Museum (art of German-speaking countries), and the Arthur M. Sackler Museum (ancient, Asian, Islamic, and Later Indian art). Their first efforts with digital imaging started in 1993-4 and they are currently moving into an electronic collections management system which supports digital images. This is an ongoing program.
At the time of the interview the Museums did not have a formal digitization strategy, but were considering developing one. They carried out several discussions about digitizing the different types of collections, rather than a formal collection survey.
Factors that influenced priorities for digitizing were risk management for objects in the collections, exhibition-related activities which guided selection, and cataloging of new acquisitions. They see the digital collections management system as a tool to support day-to-day activities in the Museums, a resource for staff, and a way to overcome physical restrictions of use of the collections. Selection is generally internally-driven by exhibition program, staff research interests, and education activities.
Their experience highlighted the importance of having the collections management system established prior to embarking on digitization. They also observed, that all the layers of documentation are important and should be recorded and preserved. Their experience showed that high-level managerial support is vital.
Some of the obstacles to planning the development of digital deliverables were: assessing the collections and talking to curators; overcoming institutional resistance to change and a high learning curve (e.g. with MESL).
Obstacles to building digital deliverables were: infrastructure problems; the late creation of Harvard Library central Digital Repository; and large data storage needs.
Criteria used in the selection and prioritization of materials for digitization include:
A further criterion, although not in our questionnaire list, was to support collections management.
Intellectual property rights are not a selection criterion, since they might digitize material they do not own the rights to, for purposes of internal documentation.
Although they collaborated with various institutions and organizations, all these were in the periphery of their digitization activities.
Generally they would recommend that you make sure you “clean house” first, before you start collaborations, so you can start from a position of strength.
The purpose of the Digitization program is to build digital collections primarily as a teaching and learning resource, to support research, and assist preservation by reducing wear and tear. Another purpose is to assist in the production of catalogs and publications. They have not produced a formal statement of intent for their digitization activities, which support their basic aim to document the collections.
The program digitized a range of visual material with varied format and nature:
The materials selected for digitization were intended to comprise the entire body of some collections. They considered the question of misrepresenting the collection when only a part is shared with the public. As a general rule the digital deliverables are intended to be re-used, as their main principle is “scan once, use many”.
For representing content they used the Dublin Core DTD, JPEG and XML. They also use the NISO (National Information Standards Organization) Image Preservation metadata (http://www.niso.org/ commitau.html). For describing content, they used the Categories for the Description of Works of Art. For controlling data values they use AAT, Library of Congress subject headings for some photographs, and their own system. There is a lack of documentation about these types of collections, e.g. geographical place names. When organizations develop their own system, they must standardize and consistently document it.
They looked at existing guidelines for digitizing particular document and object types, but did not adopt any. Few were available at the beginning. Although this situation has changed, a higher level of detail is still needed when adopting guidelines within the institution.
They realized that one specification cannot work for all users. The technical specifications vary according to source matter. These are all different from pre-LDI specifications.
The primary intended audience for the digital deliverables is Harvard University students (four-year college and graduate school), as well as Museum users. Audiences wider than the one intended can also share the parts that are publicly available.
They used both in-house and external advice on managing digitization. Digitization activity is a sub-component of the larger organization of the Fine Art Museums. It led to changes and re-evaluation of organizational relationships and procedures. Initially they had the VIA Steering Committee as a formal project management structure, before Peter Siegel was hired to head digitization initiatives.
The QA procedures they use are: ISO image creation and viewing; checking integrity of data; curators vet the materials created; press proofs comply with CMYK at the press houses; use of image metadata; and use of the MD5 checksum algorithm for ensuring the quality of digital files.
They carried out a pilot study digitizing 300 images that assisted them in scheduling, definition of training needs, technical feasibility, user needs, workflow analysis, workflow piloting, and technology forecasting. A small change from the pilot was that conservation now plays a very active role when they start a project. They used Gantt charts, but found that they are ineffective when running five projects simultaneously. The best option is to choose professionals suited for the job in the first place.
Digitization was carried out in-house because of the quality control that they required and because the collection needs to stay in-house. They purchased all the necessary equipment.
|
Technologies Used for Image Digitization |
|
Flatbed scanners (Heidelberg, Topax iX) |
|
Drum scanner (will buy one for collection of X-rays where it can penetrate the density) & for old negatives |
|
Film scanners |
|
Digital cameras |
|
High-end professional cameras (Phase One) |
They established a set of guidelines for data capture procedure and specified color management guidelines. They use grayscales, color charts, SR37 target for black and white 2D calibration, and digital photometers for monitors. They scan raw color proof. One problem is that most commercial targets are designed for synthetic material.
The number of staff who work on the program are as follows:
|
Type of Staff |
Number |
% of time on the project |
|
Director |
1 |
100 |
|
Programmer (extracts metadata automatically) |
1 |
100 |
|
Curator |
9-10 |
5-10 |
|
Digitizer |
1.5 |
100 |
|
Photographer |
3.5 |
100 |
|
Technical Support staff |
5 |
10 |
Most people have a museum background or are industry professionals.
Areas where training needs have been identified are: project management; application of technical standards; preparation and handling of materials for digitization; technical operation of digitizing equipment; post digitization processes (especially color management and preparation for press); and metadata creation. The photographers received appropriate training.
Training has been organized in a variety of ways, in-house using project staff, by attending external courses, with independent study, and by learning on the job. The training did meet the needs of the organization. In addition to this, they attended more industrial-type training, offered by the camera manufacturers.
They are aware of the copyright status of the digital deliverables that the program is creating. Due to the nature of the material (e.g. with artists often owning copyright of analog works), this is a complex area. Although they own the copyright for most of the materials that they digitize, there are some where they do not own the copyright but have a license to use the material. The rights information is available and they will also declare the copyright status of the digitized material in the future. Users of the material can make printouts on paper or download to a PC both thumbnails and lower quality images as long as it is for fair use. Harvard users can also download the high quality images. No electronic management systems are used to control copying.
There is a conservation procedure for original materials. The investigations carried out into the condition of the original materials prior to digitization vary according to material and include conservation survey and stabilization. Risks identified during the preparation or digitization process include heat, especially from lighting. They put UV gels in front of lighting sources and use special equipment. The material is prepared by curatorial and conservation staff before and during digitization. They will only restrict access to originals once the material has been digitized for conservation reasons.
The cataloging and reference systems used are EmbARK and 3D System. The information used from these systems in the digitization process included dimensions, material, title, and artist. They digitized from originals as much as possible, but also used reproductions and intermediaries. They used 35mm slides, photographic prints, 4x5 and 8x10 transparencies.
The originals are cataloged using EmbARK. Eventually the same system will be used for cataloging the digital surrogates when the Library’s Digital Repository Service is fully operational. In the meantime, they record Dublin Core metadata for the digital surrogates using a spreadsheet. They record metadata about the original object, the digital object, the digitization process, technical details, staffing information, and administrative information. The metadata records are created by archivists. The catalog for the digital deliverables is available on the internet.
The file formats used for images are JPEG for delivering, and TIFF for capturing, preserving, and delivering. The resolution used is 6Kx7K for capturing, preserving, and delivering images. They used 48 bit-depth for capturing and preserving, and 24 for delivering. Compression is used at delivery stage. The aim of the compression was to improve access and enhance usability. They retain the original scans in uncompressed form and without color matching. Post-scanning processing is carried out on the images, mainly color correction, sharpening of low-resolution images which is done automatically with a script but is visually inspected. The average file sizes created are around 200 MB at capturing and preserving stage, and 2Kx3K for delivering. They use specialized targets and software to measure performance of digital cameras and scanners (e.g. Kodak Q-60 color input target). These same targets are used to generate ICC (International Color Consortium) profiles for color management. They check the dynamic range of their equipment with an MTF target and Delta E (which measures changes in color).
Their experience showed that you should try to make a custom profile for the equipment, document it, and stick to it. For publications, you also need to talk to the publisher about the needs for the publication, talk to their technicians and the academic community to determine what the collection requires. Color management is imperative after capture.
They carry out a total check to ensure the quality of the digital deliverables. Quality control procedures might require a little more time at the beginning, but save a lot of time in the long term.
Users have free open access to both the catalog and the digital deliverables for browsing, but not for manipulating them in any way. More advanced ways of using the material will evolve. Access to some material is restricted to in-house or Harvard University users. They monitor use via automatic data capture.
At the moment no publicity is used to publicize the creation of digital deliverables. In the future, the methods that they will use might include: announcement on website, press release, articles in print media, announcements at conferences and meetings, and announcement at electronic listservs.
They carried out front-end and formative evaluation using focus groups and hands-on testing. They also evaluated informally internally, with other museums and with the publishing industry, which did not lead to major changes, but helped them to confirm a lot of their ideas and raised their comfort level. For example, suggestions from the press industry led them to keep the color bars in the archival image.
The program was funded at $1 million from internal Harvard University funds for the largest part (three-quarters) and about one quarter from private funds, e.g. from the Getty, and the Carpenter Foundation. They are provided sufficiently, but could use more space. They think that using standards saved them money in the long term. There are regular updates for the funding bodies about the progress of the program. They also had to provide a collection survey, a conservation plan, risk assessment, front-end evaluation report, project management plan, cost models, and workflow reports.
All elements of the project will need to be updated. New materials and metadata will need to be digitized fairly regularly; metadata will need regular updating; the user interface will need to be changed; and it will be necessary to change file formats probably within the next two years. They have a preservation strategy that includes file formats, storage media and conditions, and metadata updates. The strategy is based on migration of data. Quality control procedures in the life-cycle management are based on the MD5 checksum algorithm which checks the quality of digital files. Longer-term sustainability is not dependent on self-generating funds. They have not adopted an exit strategy.
As general comments, they mentioned that it is very important to keep appraised of professional industry developments, not just academic ones. Digitization programs require a very professional approach, and must be taken seriously, as they are not for the faint-hearted. It is necessary to talk to colleagues, and establish support groups.
An interesting aspect of this case is that due to the nature of the materials, which is a wide range of artworks, and the nature of activities that digitization might support, such as professional publications, great importance is placed on color management, calibration of equipment, and the adoption of procedures and practices across all steps of the digitization process to ensure faithful color reproduction and high quality images. For this reason, the type of equipment used is at the high end of the market and there are also very large file sizes. In this case the range of uses that digital images will be put to is very wide, going beyond simply viewing on the web towards use as a tool to support all aspects of the Museums operation. Another characteristic was the emphasis on the importance of links with professional publishing and the industrial world. They also mentioned that in several cases they had to design their own custom profiles for the equipment or develop their own system for controlling data values. Their view is that this is acceptable as long as it is standardized, applied consistently and is documented.