Harvard University Library

   

<< Table of Contents

veritasHarvard University Library Notes, For Harvard Library Staff, Number 1339 September 2007

Harvard University Library Notes / September 2007 / No. 1339

Interview: Virginia Danielson

Virginia Danielson

Virginia Danielson

Virginia Danielson has held the Richard F. French Librarianship of HCL's Loeb Music Library since 1999. She is also the curator of the Archive of World Music and oversees Harvard College Library Audio Preservation Services. Prior to her current appointment, she served as Keeper of the Isham Memorial Library. Trained as an ethnomusicologist, she is the author of a book, "The Voice of Egypt": Umm Kulthum, Arabic Song and Egyptian Society in the 20th Century, that won the Alan P. Merriam Prize for best English-language monograph in 1997. She recently co-edited the Garland Encyclopedia of World Music: The Middle East. Currently, she serves as co-director of "Sound Directions: Digital Preservation and Access for a Global Audio Heritage," an NEH-funded cooperative project with Indiana University, and she contributes regularly to scholarly discourse on Middle Eastern music and culture. She occasionally teaches in Harvard's Department of Music. Ginny Danielson was interviewed for Library Notes on August 23.

LN

The Loeb Music Library is a highly complex, specialized library. You're a reference library, a circulating collection, a manuscript repository, and a media center. How is the digital environment affecting your already complex collections and services?

GD

Building a digital library is part of everyone's job. It's not something that you hire somebody else to do. It's an opportunity to engage the expert staff that you already have.

For example, for our digital scores projects, we have two curators plus the reference librarian working on selecting materials. The curators are also involved with metadata creation, because they're the ones who know how to parse works.

Within a work such as an opera that may be three or four hundred pages long, where are users likely to want to go? These are questions that a technologist or a non-music specialist couldn't answer. So the metadata—the division of content—is provided by music specialists and curators who work with the data formats that the Digital Imaging Group [DIG] provides or that Audio Preservation Services [APS] provides in the case of audio work.

LN

So we can't rely on a table of contents?

GD

Scores don't always have tables of contents. And from a young student's perspective, the information that a score offers can be a little opaque. For example, a Roman-numeral 3 at the top of a page would indicate to a specialist the third movement of a piece. Depending on the sophistication of the student, that may or may not be clear. Sometimes there's not even such a thing as a Roman-numeral 3, but simply an indication of a tempo change, in which case, to the person who knows the piece and can read the music and knows where he is in the piece, that indicates the beginning of the third movement. But it might not be apparent to a non-specialist.

Similarly, within an opera where there may be a table of contents, there are always parts that people remember that aren't at the beginnings of acts or scenes. For example, "Parto, parto," from Mozart's Clemenzo di Tito, is a piece that people know and would want to find. But it's not necessarily listed in the table of contents!

LN

Reflecting back: the Loeb Music staff works in a highly integrated way to select and prepare works for digitization.

GD

Yes. Of course we rely heavily on experts in the DIG and in APS and in Preservation to do high-quality digital work, all the more so because our library has been heavily used over the years, and a lot of the items that we're digitizing are fragile. They have to be cleaned and then they have to be treated very carefully. Digitization, then, becomes not just an access tool, but a preservation tool.

LN

Are your scores candidates for scanning in the Harvard-Google Project?

GD

No, though we discussed the possibility at some length. There were already commercial agencies around the world that were digitizing out-of-print scores and making them available for licensing.

We have very specific criteria of our own for digitizing materials, and we choose to put our efforts into works that aren't commonly available, that are scholarly in nature or performative in special ways, that are intended for professional performers or professional scholars. That's rather different from the aim of Google Book Search.

One of the things that make our online opera scores so valuable is that our different editions aren't identical. The presence of a particular piece in one edition or another can be a matter of great importance to someone performing the work, to an editor who wants to publish, or to a student who is looking at the versions and variants of operas. Rossini and Donizetti operas were very often revised with particular singers in mind. Simply defining what the work is can be an adventure if the work did morph as different performing venues and personnel became engaged with them. That's another key factor. It's a reason to digitize, because performers need to see the different versions.

LN

How many scores do you have online?

GD

Right now there are 180. We have LDI funding that should at least double that number, if not more. So getting up in the direction of 500.

LN

Let's consider your audio preservation program. Obviously, working with scores on paper is very different, but there are some similarities that might not be immediately apparent. What are they?

GD

The medium changes, but the problems often don't. You need a specialist for your descriptive metadata and your division of content. For esoteric material, audio engineers do not necessarily know when tuning stops and performance begins.

LN

How are the divisions determined?

GD

You create a file that a curator can listen to and mark the divisions using the time code that the digital file has and says, okay, at this minute and these seconds, put an index marker.

Fragility is often a huge issue. Because of the fragility of magnetic media, you have to assume that you'll only be able to play the item once. And so you have to have your machines cleaned and everything set up so that if that is the last play you get, it's a good one. Which means that you can't be stopping and starting unnecessarily and fooling around with re-plays at that point. You've got to be able to run the tape as best you can.

LN

Your audio preservation program is known to be on the cutting edge. What are the current priorities?

GD

One is the "Sound Directions" project conducted by Harvard and Indiana University. The project looks at existing guidelines for audio preservation and takes them "out for a spin." We drew up a set of 40 best practices, and at Harvard, we built a suite of 40 pieces of software—that will be available free of charge, as open-source software—to streamline and mechanize various parts of the process, particularly the metadata collection. The best practices will be published online in September of 2007, and the software will be available at that time too for anyone to use.

On a more local level, APS works on faculty and student requests for classroom use and research all the time. They also work on the Woodberry Poetry Room collection and the Milman Parry Collection. At the moment, they're working on a project in Iranian oral history, which documents life in Iran in the 1970s, before the ouster of the Shah.

LN

What's the relationship of APS to the Harvard Film Archive?

GD

We have done work for the Film Archive. But film preservation and audio preservation are two different things, because film scholars and technologists don't necessarily want to lose their original media. The objective there is to preserve the film as film, and to rehabilitate it. Whereas the new paradigm for audio preservation is, "Don't worry about the carrier. Go for the content."

Video is the great, uncharted terrain. I'm actually working with a specialist in New York to get some best practices drafted that will last us until the DRS can take video.

LN

Music publishing has changed a great deal in recent years. How does that affect Loeb Music?

GD

Our biggest headache with music publishing involves new music and contemporary composers. These are not high-profit areas for publishers, and the frustration hit universities first at the faculty level, where composers were having trouble getting compositions published at all.

We began to find that publishing companies were making scores available for rental only—not for purchases. Which is fine if you're a symphony orchestra. It's not so fine if you're a library. Composers themselves were unable to get copies of their own work. Given that people come to particular universities to study with particular composers, and that those composers want to teach their own pieces, this can be a real problem. It's prompted quite a number of composers to create their own web sites and to market compositions on their own.

Since we intend to collect comprehensively in this area, it means that, instead of going to a handful of vendors, we have to go to dozens of little web sites and make sure they're up and running—and then try to make sure that everything that a composer has written is actually available on them. It's not really expensive work in terms of paying for the object, but it can be terribly expensive work in terms of labor.

There's another side to the publishing coin. Many of the current composers don't work in the domain of printed music, but rather work in a domain of electronics. We now must be able to accept media that in some cases we have little or no familiarity with. We are collecting in much more performative media than we ever have before.

LN

Many of our colleagues report a higher level—or a changing level—of engagement with students and faculty than in the past. Is this the case at Loeb Music?

GD

Students arriving at the University now—as opposed to ten years ago—have more questions about library resources and particularly the online resources. The questions aren't so much about how you do research, but instead, "What is it that this online tool includes that that one doesn't?" Or "How do I get at music reviews in newspapers as opposed to music reviews in journals?" Or, in the domain of classical culture, "How do I get at music reviews in Parisian journals from the 19th century?" And, of course, "How do I access music in the domain of popular culture?"

We also engage with faculty and students in important ways that you wouldn't strictly call instruction. For example, the "Bernstein in Boston" course involved Sarah Adams and Liza Vick from the inception. The products of the course included an exhibit curated by Sarah with the students and a web site that includes online resources that Liza Vick manages. I don't know that anyone would be comfortable saying that Liza and Sarah "instructed" the students, but it was a very, very productive engagement with the teaching.

There's a lot of engagement that goes on. Many of our staff members sit in on courses. We keep track of course lists for every semester, so that when books come in that we know are relevant to a course, we notify the faculty member immediately. The fact is that we are two feet from the department chairman's doorstep. When our faculty or students need something, they come right in and ask.

 

<< Previous Article

 

Links:

Current Issue

Contact Library Notes

 

Return to the top.