10/8/97

CONTACT: David F. Salisbury, News Service (650) 725-1944;
e-mail: david.salisbury@stanford.edu

New computer music center comes to Stanford

It all started with a passion for Bach.

After studying physics at Harvard and pursing advanced degrees in electrical engineering and operations research at Stanford, Walter Hewlett decided to immerse himself in the magical music of Johann Sebastian.

So the son of the co-founder of Hewlett-Packard signed up for Stanford's Doctor of Musical Arts program and began studying the great composer and performing his works on the organ. By 1980, when he got his third advanced degree, however, Hewlett realized that he was far from alone in his aspirations.

"If you go to the library, you can see that for 200 years people have been studying Bach, and each generation has written several books about Bach counterpoint or the structure of Bach's music. So I realized I'm just one in a long line of people who've wanted to do this. But then I thought I'm in a different position because I can add something, or do something that these others couldn't," Hewlett recalled.

That extra something was to bring the power of the computer to bear on the study of Bach's music, and music in general. It was something that he had the background, education and training to do.

So Hewlett embarked on a project to "encode" all of Bach's music in a form that computers could use. About that time he heard that Eleanor Selfridge-Field ­ the musicologist who had authored one of his favorite books, Venetian Instrumental Music from Gabrieli to Vivaldi ­ was between positions. So he hired her to assist him in his effort. In particular, he asked her to find out what other similar research was taking place around the country.

After a few years, the two realized that they had embarked on a long-term effort, one that would benefit from an institutional structure and identity. So they decided to set up a foundation to achieve this goal. Because they didn't want to limit their efforts to music, they decided to name it the Center for Computer-Assisted Research in the Humanities (CCARH). It was established in 1984 in Menlo Park.

Since then, CCARH has established an international reputation in the area of applying computers to musicology, the scholarly study of music, including historical research, musical theory and the physical nature of sound. The center's efforts are complementary to those of Stanford's Center for Computer Research in Music and Acoustics (CCRMA), which has pioneered new methods for using computers to create music. "We do all the things with computers and music that don't make noise," Selfridge-Field explained.

Last May CCARH moved to campus. "After working for 10 years in Menlo Park, I became concerned about the possible lack of continuation of our efforts," Hewlett said. He had seen a number of excellent research projects simply disappear when the principals either retired or lost interest. He decided that the best way to ensure that the center's research has lasting value is to attract and train graduate students. So he negotiated an agreement with the university and the music department to house CCARH for a five-year trial period. It now occupies a suite of rooms in Braun Music Center and one of the trailers behind CCRMA. At the end of five years the center and the university will review the situation and decide whether it makes sense to continue.

"Because of the nature of their work, they were already working closely with people at CCRMA and those relationships have been facilitated by their presence on campus," said Steve Hinton, associate professor and chair of the music department. "There is also hope that what they are doing will have important ramifications for musicologists, composers and other humanities scholars on campus."

Hewlett and Selfridge-Field started on this effort before the birth of the personal computer. He began the encoding project using a Hewlett-Packard minicomputer running a customized operating system that his friend, David Woodley Packard, had developed in order to store and print classic Greek text. At that time, computers could not even print music. It was not until the dot-matrix printer was introduced that Hewlett was able to get the computer to print musical notation. Even then he had to reprogram the printer's character set.

"When we started, the field was totally devoid of infrastructure," Hewlett said. As a result, much of his efforts have been devoted to creating the tools required to store and reproduce musical notation, and he has been forced to put aside his early goal of using the computer to analyze the structure of Bach's music. "We've opened up all the fundamental questions, but it will probably be someone in the next generation who will use the expert musical databases that we have developed to begin answering them," he said.

Still, researchers at the center have made a number of contributions in the field. Hewlett developed a two-step process for encoding musical notation. In the first step, the sequence of notes is fed into the computer. The second step involves filling in the specific notations that composers add, such as pizzicato or staccato.

MIDI, the computer system used to control music synthesizers, cannot capture all the detail required to perform classical music accurately, Selfridge-Field said. She recently completed Beyond MIDI: The Handbook of Musical Codes, which details more than 40 methods of encoding musical information on computer. The book is published by MIT Press and was released this summer.

To aid communications within this specialized field, the center publishes an annual directory of computer-assisted research in musicology. Originally, it consisted of a simple listing of projects that resulted from Selfridge-Field's initial project for Hewlett. Over the years the authors have given the publication a formal journal format and now call it Computing in Musicology: An International Directory of Applications.

By now, the CCARH staff has encoded hundreds of works by Bach, Beethoven, Corelli, Handel, Haydn, Telemann, Mozart, Vivaldi and other similar composers using Hewlett's process. They are using this musical database in a number of different ways.

One application is producing printed scores and individual parts for performers. For a number of older works, individual parts do not exist, so orchestra music directors must create them by cutting up photocopies of the master score and then pasting together the parts for each performer. Sometimes, music publishers rent individual parts for thousands of dollars. The center intends to publish individual parts for a number of classical works on their website, where they will be freely available to all.

CCARH materials have been used to create several recordings of major works of the 18th century. One is a compact disc of Handel's opera Ariodante directed by Nicholas McGegan at the Göttingen Handel Festival in Germany. Another is the first "program-it-yourself" CD to appear commercially that lets listeners reconstruct any of nine historical versions of Handel's Messiah from the variations performed by the Philharmonia Baroque and recorded on the disc.

The direction the center takes in the future depends a lot on where the field goes, Hewlett said. "Right now, musicology is not very computer oriented. But I think there will be a tremendous shift when today's computer-literate students move into the profession."

Hewlett and Selfridge-Field hope to contribute to this shift by providing the computer tools and data that these scholars will need. "There's a lot to be said for simple staying power," Hewlett said. "As Woody Allen once said, 'Ninety percent of success is showing up.' "

-30-

By David F. Salisbury