Once More, With Feeling: U.Va. Music Library Helps Develop Digital Format for Music Notation

April 4, 2012 — An international effort spearheaded by the University of Virginia Music Library is paving the way for powerful new digital tools for music scholars.

The Music Encoding Initiative, or MEI, is the name of both a new open-source format for digital music notation and of the organization working to develop it, which includes U.Va. Music Library staff and international collaborators.  

The idea, according to U.Va. music librarian Perry Roland – the project's originator – is to create a standard format for computerized music notation in the same way there are standard formats for online text. Currently, numerous formats exist for digital sheet music, but many are commercial products and none are universally embraced, which makes them difficult for scholars to use, Roland said.

"It's not that we don't currently have any way of doing this, it's that we have too many ways," he said last week, just before presenting a project update at the Scholars' Lab. "Now we're trying to get the cart and the horse in the right order, and get scholars to participate."

Technological advances have made it easy for humanities scholars to search and analyze massive databases of scholarly texts. It takes only seconds to search Shakespeare's entire works or to access the entire written record of the Salem Witch Trials.

However, the lack of a standard format for online musical notation has made it difficult to do similar projects based on notation, said Erin Mayhood, head of the U.Va. Music Library.

"In the music community, there is all of this buzzing," Mayhood said. "Musicologists and music theorists aren't able to do the same kinds of the things that digital humanists in other disciplines are able to do, because of this lack of an ability to put musical notation online in a meaningful way."

A common format for notation would facilitate analysis of databases of musical works and allow scholars to embed textual analysis inside the notation, or compare many pieces of music for common themes and passages. Roland's counterparts at the University of Paderborn in North Rhine-Westphalia, Germany, have already developed software exploring some of these functions. MEI could also allow the inclusion of video or other multimedia elements that are often a part of modern compositions and are impossible to capture in standard notation.

"Scholars, especially those studying old music, often look at different versions of the same piece," Mayhood said. "They might find a piece of it here or and a piece of it there. Some scholars spend a lifetime studying these different versions. So now we could create digital tools that would allow them to analyze the versions and even embed their notes on the differences."

Bonnie Gordon, an associate professor in the music department, said MEI is a potentially wonderful resource for teaching and scholarship.

"I can see it as a fabulous way to work with Jefferson's Monticello Music Collection," Gordon said, referring to a collection of Thomas Jefferson's personal sheet music. "This is a collection that for the most part has not been preserved or digitized and that will be useful for scholars interested in the music of early America. For those of us who work on the distant past, one of our biggest problems is presenting scores to students. Many are not readily available in modern editions and are far too cumbersome for straight-up Xeroxing."

MEI is in its second year of a three-year, $161,175 grant from the National Endowment for the Humanities, and the project's German partners are funded by the German Research Foundation.

Roland, who in addition to his data management background has a graduate degree in music composition from the College of Arts & Sciences, first conceived of using the XML markup language to model musical notation about 10 years ago, and has been steadily working at it since.  

Now, an online community of researchers and scholars is creating a database of musical works in the MEI format and developing new software to take advantage of it. That software ranges from programs that allow composers to create works in MEI to analysis software that allows a researcher to search the entire works of Beethoven for a particular melody, Roland said.

Though the project has grown substantially since its inception, Roland said he believes the work will never really be complete. The Text Encoding Initiative, a similar effort for text-based scholarly work, is about 25 years old and still changing regularly, he said.

"Really what we're shooting for is a community, a group of people who will take this on and manage it," Roland said. "But ideally, it'll never be finished. It'll continue to grow."