NINCH guide home interview table of contents previous interview next interview
On January 10 2001, HATII interviewed the following members of the Variations project at Indiana University, Bloomington: Jon Dunn, Assistant Director for Technology (Indiana University Digital Library Program); Emma Dederick-Colon, Electronic Music Resources Librarian and Variations Project Coordinator (Cook Music Library); and Michael Fling, Collection Development Librarian (Cook Music Library). The goal of the Variations project is to construct a digital library system to contain the 7000-plus items of audio material that are presently stored in the Cook Music Library. The project also hopes to digitize the objects associated with music, such as scores, texts and videos. The underlying incentive for the project is to create CD sound quality over a broad network, while the project has a subsidiary aim of developing a comprehensive database of scores for distribution.
The Cook Music Library’s Variations project developed from the large amount of assigned listening in courses run by the School of Music. As a result the Library needed to distribute a large amount of music yet maintain control over its playback whilst at the same time improving access to what were frequently limited holdings. Therefore, the Variations project began in 1995 and delivery commenced in 1996 despite hardware and personnel limitations. The focus for the digitization was the assigned listening music, which also suffered from poor playback quality on tapes. The Library’s digitization activities were formalized in a policy statement in 1998/9.
The primary objective of this policy is to serve students’ needs and service special collections on demand digitization, which has the aim of reducing handling. A secondary priority is to make available faculty performances in digital form. The project has been successful in achieving these objectives and brought faculty on board through demonstrations and the involvement of the Music Library Advisory Committee. Two faculty members also served as “guinea pigs” in 1996. Overall obstacles to planning and building the digital deliverables were faculty’s initial resistance to change, technical obstacles such as storage space and CPU time for compression, and copyright and licensing issues.
The primary selection criteria are the materials’ teaching and learning potential and enhanced access, as determined by their designation as assigned listening. The Library will endeavor to place any sound recordings on digital reserve so long as they have a copy. For special collections the material’s historical and cultural value are also priorities and attempts have also been made to select material to build a standard repertoire, but this has been a secondary priority. Labor cost reduction had been mentioned as a motivation only in selecting material. These selection criteria have not changed over time, but the balance of work has changed as reserve material has reached a plateau. The use of student help for digitization was the most significant change over time.
The Cook Music Library has had informal discussions with other libraries about sharing material and has worked with other academic partners in the UK and Japan (through the NSF).
The primary purpose in creating the digital deliverables is as a teaching and learning resource, in response to previous demand. Public access, depending on copyright and the Library holding a copy, is a second priority; for example, public domain scores are accessible. The project has produced various statements of intent, primarily via internal Indiana University publications, mainly emphasizing the easier access the digital reserves provided.
The type of source material digitized includes:
The predominant medium of the materials digitized is Compact Disc. The project aims to be a digital library, not just a reserve, and will digitize the entire piece or performance, whatever corresponds to the bibliographic record and not just the section requested. The project had originally discussed re-purposing the digital deliverables but nothing has been undertaken so far.
The following standards, guidelines or tools are used for representing content:
MARC is used for describing content.
A MARC-based catalog system is used for controlling data values
The project looked for other existing guidelines for digitizing sound materials, in particular for uncompressed WAV files, but they did not exist at the time.
Structural metadata, such as tables of contents in liner notes, are represented in flat text files in a local format. These are then parsed for error checking. The project will probably migrate to XML.
For the project, navigating between ideal and realistic standards is not an issue that arises in relation to WAV files.
The intended targets for the digital deliverables are four-year college and graduate school audiences. As part of a recent NSF grant several faculty members are interested in lifelong, distance and computer mediated learning audiences – for example, to look at a music theory classroom presentation and lesson authoring tool, web based distance music education and the value of a digital music library. The digital deliverables could be used by audiences other than the primary targets; in this case the most likely are researchers and students at Indiana University regional campuses and other universities participating as satellite sites in the NSF project.
The needs of those with disabilities have not been explicitly recognized in the Variations project, but sites that are part of the DLP program have been checked. The digital deliverables in the Variations project are restricted to the music library building because of copyright. Shortly all of the school of music building will be able to access the material. The profile of users for the digital reserve has been as expected, that is undergraduate music students. For the scores material, the audience has also been the same as anticipated, but there have been additional users from outside Indiana University, particularly eastern Europe and South America.
In-house advice on managing the digitization project was available and the project worked with IBM on technology decisions. The management of the Variations project is entirely integrated into the Library structure, for example the head of the project is the head of the Library and the electronic reserves are part of reserves. There have been some changes in organizational relationships and procedures in that the project is now also part of a larger campus wide Digital Library Program.
Formal project management procedures have taken shape in a recently formed advisory board that wants to involve faculty members. There are no formal managerial quality assurance procedures in place.
A pilot test is being carried out for a project to digitize music scores but was not undertaken for the electronic reserves of the Variations project. In this respect the project staff recognize that Variations is not an example of good practice. The pilot study for the scores project was used for establishing training needs, technical feasibility, workflow analysis, technology forecasting and to assess the benefits to faculty and students. This pilot study is ongoing and it remains to be seen what impact it will have on the design of the project. No formal benchmarking or time and motion studies have been carried out, but the output of each person was examined.
Digitization has been carried out entirely in-house with the majority of equipment bought in (there was some existing audio source equipment). There was not much choice in digitization processes to adopt for audio, and quality decisions were the determining factor. Source equipment ranges from turntables to CD players, and sound cards used for capture are Event Electronics Echo Gina cards. Specialist software used is Sound Forge by Sonic Foundry and Disc to Disk by Microtest.
A full-time director and project co-ordinator in the Music Library work on the project, but each for a small amount of their time. Slightly over one FT equivalent digitizer is employed (up to 4-5 individuals, 10-20 hours a week each) and portions of a system administrator, programmer and technical manager in the Digital Library Program make up the technical staff at approximately 1.25 FTE. An evaluation specialist is available through the Digital Library Program. These staff all have library, music and/or IT backgrounds and have not been re-deployed from other areas. Where necessary the project looked to other people for advice on the technical aspects of digitization.
Training needs are assessed informally and areas of training that have been identified include:
The team members that have received training include:
Training has been organized using the following:
The training has met the needs of the project.
The project is aware of the copyright position of the digital deliverables and only owns the copyright in Indiana University recordings, but even this area is unclear. The copyright status of the digital deliverable is declared in a copyright notice recorded by the digitizer (repeating the copyright notice from the sleeve notes). This policy has been effective because students are unable to copy the streaming format. The material in copyright is digitized under the legal provision for libraries (preservation or fair use). Users are unable to copy any material.
For digital audio material, users can only listen to full length compressed versions.
The streaming format acts as an electronic management system as it is extremely difficult, if not impossible, for users to capture or download during playback. As a result the streaming format has proved an effective means of preventing unauthorized use.
The only conservation work carried out as part of the project is the cleaning of originals, such as LPs. Access to the original materials has only been restricted for items in non-circulating special collections.
The cataloging system in place prior to digitization is the Library electronic catalog and a web based reserves list. The project uses all of this information during the digitization process. The project does not have to locate any core reference material to complete the digital deliverables or alter any material from its original form for digitization. Score material will be rejected for digitization if it has blurred pages, is defaced or the print is poor. Nor will the Library digitize any material that it does not already hold or can easily acquire.
The project has digitized from intermediaries, but the material was also available in its original form. Intermediaries used for audio digitization are:
The original material is cataloged in the Library catalog and the digital deliverables use the same system.
USMARC is used for cataloging the digital deliverables.
Tools used for controlling data values are:
Metadata details are recorded about:
The existing MARC records for the physical item are used as descriptive metadata for the digitized item. Structural and administrative metadata for the digital item are in a separate set of flat text files, created at time of digitization by the digitizer. The catalog for the digital deliverables is available on the internet and the catalog and object are linked through the MARC 856 field.
The work method used for audio digitization is for one person to digitize and for an automated computer system to save, catalog and link the object.
The WAV file format is used for capturing and preserving audio digital material and MPEG layer 1, level 2 format for delivery. A 44.1 kHz sampling rate and 16-bit stereo bit-rate is used for capturing, preservation and delivery. The delivery compression method used is bit-rate compression at 3.7:1 (from 1.411 Mbps to 384 Kbps) to reduce cost, decrease storage requirements and decrease network bandwidth requirements. The average audio file sizes are 463MB for capture and preservation and 125MB for delivery. No post capture editing is carried out.
Quality control procedures for the digital deliverables are under development. Quality control procedures are in place for the metadata recording and delivery, for example the Reserves Coordinator enters the 856 field and this can be suppressed until the object is ready.
Users have open access to the catalog and the materials are restricted to in-house users. Searching and browsing facilities are the standard keyword and Boolean searches. An encoded sub-set of the metadata can be searched in the same way. The project does not have any special software needs regarding delivery and access. Users are not able to manipulate the digital deliverables apart from the standard playback controls. Users do not have to pay to access the digital deliverables except that all students have to pay a technology fee as part of their tuition fees.
Usage is monitored by automatic data capture and potential users informed of the digital deliverables through articles in print media, conferences, meetings and internal publicity. The most effective of these is internal publicity.
The Variations project has not carried out any front-end, formative or summative evaluations.
The Variations project was largely funded from existing personnel budgets. A new position, Variations Technical Director, was funded 33% by the Cook Music Library, 33% by University Computing Services and 33% by the School of Music. When the Digital Library Program was formed, the function of Variations Technical Director was replaced by parts of three DLP positions: Assistant Director for Technology, Programmer/Analyst, and System Administrator. The University recently received a grant from NSF and NEH through the Digital Library Initiatives Phase 2 (DLI2) program for a new R&D project in digital music libraries, but Variations will still exist as the operational production system. IBM has assisted the project with equipment and other sources have been the Digital Library Program and the School of Music. If the project had had more funding it would have pursued a more vigorous equipment replacement cycle.
Standards have saved the project money in both the short and long-term. In the short term, using standard file formats, MARC bibliographic description, and a web interface allowed the project to assemble a system more easily and more cost-effectively by picking and choosing products from different vendors that can inter-operate with each other. In the longer term, use of standards should assist the project with preservation issues, since standard file formats tend to be supported by software products longer and have published specifications so that migration is less of a problem.
New material and metadata are added daily with periodic changes to metadata and the user interface. Equipment is updated on a three-year cycle. A written preservation or archiving strategy has not yet been developed, although in general, the project has been carried out with long-term access in mind. This includes use of standard file formats. Migration of MPEG files from current tape storage system to a new university-wide mass storage system has recently been completed, with migration of WAV files currently in progress. The Digital Library Program is working with University Information Technology Services to ensure long-term preservation/access of Library objects in this mass storage system (see http://storage.iu.edu/ for more information on this system and UITS’ plan). No quality control procedures have been adopted in the life-cycle management.
The project does not have an exit strategy and should resources not be available sufficient to sustain the digital deliverables, the project staff speculate that they would revert to the previous system of serving reserve material. Loss of the digital deliverables would matter.
The Variations project at the Cook Music Library, Indiana University, is notable as being one of the few large scale sound digitization projects. The most significant feature was the absence of standards or guidelines for sound digitization and quality control and the adoption of de facto standards in file formats, such as WAV, as an alternative. As the Variations project serves only students and faculty of the Indiana Music School it has not faced the copyright problems that would be the most challenging aspect of a similar project that served external users. Indeed, unless an organization owned the copyright in the original materials it seems unlikely that a public service version along the lines of Variations would be possible. The fact that IPR issues are such a potential problem possibly hinders the growth of music digitization and hence the development of standards and guidelines. Nevertheless, the Variations project demonstrates the improvement in access that a networked digital resource can bring and how such technology can be harnessed to improve the Library’s service to its core user group.