David Temperley, Ph.D., is Associate Professor of Music Theory at the Eastman School of Music and an internationally renowned scholar of computational musicology. Temperley came to the Eastman School in 2000 after completing his Ph.D. at Columbia University, where he studied with Fred Lerdahl, and a post-doctoral fellowship at Ohio State, where he worked with David Huron. Temperley’s research has spanned a range of issues including music cognition, harmony in rock, rhythm in traditional African music, and music-language parallels. His first book, The Cognition of Basic Musical Structures (MIT, 2001), won the Society for Music Theory’s Emerging Scholar Award; his second book, Music and Probability (MIT, 2007), explores computational music research from a probabilistic perspective. Temperley’s Melisma system, a computational toolkit for music analysis, is widely used. He also has a strong secondary interest in computational linguistics; he created the Link Grammar Parser with Daniel Sleator and John Lafferty. This semester, Temperley is teaching a new course called “Computational Models of Music.”
———-
PI: What specific technology or technologies have helped you most in your primary research in music cognition?
The core of my research is developing computational models of music cognition. I try to create computer programs that can perform tasks that humans also perform, such as tracking the beat in a piece of music, generating expectations for the next note in a melody, or generating a satisfying expressive performance of a piece. The idea is that this will shed light on how humans perform those processes. Obviously, computer technology plays a central role in work of that kind. I also use computers for statistical analysis of music—for example, counting up the frequencies of different kinds of musical patterns in pieces. This is important for my computational models, since they are mostly statistical and probabilistic in nature and must be trained on data from actual music.
PI: You also do experimental work; what kind of role does IT play in that?
A big component of my work is running experiments on human subjects to test my models. Recently, I did an experiment studying the emotional connotations of melodies. Technology plays a role in my experimental work in many ways. I often run experiments on the Internet now. I write computer programs to generate a different web page for each participant in the experiment, so that, for example, each participant hears the stimuli in a different random order. Participants can enter their responses on the computer as well. And, of course, I use statistical software to analyze the data.
PI: Are there kinds of technology that are unique to computational music research?
Yes, various kinds. There is software for creating music notation and converting notation into computer-readable data. And there are tools for editing audio data as well. I use these tools a lot—in generating stimuli for experiments or training data for my models.
PI: Does your work have practical applications in the area of information technology?
There are certainly indirect applications. Many companies now are looking around for ways of intelligently processing music. For example, companies want to be able to take music from the Internet and classify it as to its style, emotional content, and so on. To do those tasks, one has to be able to extract basic information from the music like its key and meter—the kind of problem that I work on.
PI: Have there been any unique IT challenges you’ve had to overcome?
Just keeping abreast of all the new developments and technologies is quite a challenge!
PI: Are there any disadvantages to the explosion of information technology, from your point of view?
It’s certainly allowed everyone to be much more productive. For example, doing background research on a topic—to find what else has been written about it—is vastly easier than it was a couple of decades ago. The fact that nearly all journals are online now, and can be searched electronically, makes a tremendous difference. I have to say, though, I miss library research sometimes: There’s something satisfying about going to a library shelf and finding a book that you can hold in your hands. And I still prefer reading hard copies to reading on a computer screen. But I’m no Luddite—modern technology has given us many wonderful things.
