Music Information Retrieval: Query-By-Humming and Source Estimation

October 13, 2006
Bryan Pardo | Northwestern University

A key problem facing us in the 21st century is information retrieval and management – how to retrieve, process, and store the information one seeks from the huge and ever-growing mass of available data, including multimedia. Music, from mp3s to ring tones to digitized scores, is one of the most popular categories of multimedia. Music collections are typically indexed by such features as title, composer, and performer. However, people often wish to perform tasks that require knowledge of their musical content, such as the melody (“What is the name of that song that goes like this ?”) and selection of individual sound sources (“Could you make the flute part a bit louder in this recording?”). To extend the range of ways in which audio recordings can be accessed and manipulated, researchers must create systems that can access and manipulate perceptually relevant structures in the audio signal. This requires advances in areas such as audio source separation (picking out a single instrument), and higher-level structure identification (labeling a sung melody as “Hey Jude”). Systems able to reliably access such features represent a fundamental improvement in our ability to access and manipulate audio data, opening up new applications. In this talk, Bryan Pardo will discuss and demonstrate two interlinked research systems addressing these issue: the VocalSearch system for query-by-humming and the ASE system for automatic separation of sound sources in a stereo mix to isolate individual instruments.

Speaker Details

Bryan Pardo is an assistant professor in the Northwestern University Department of Electrical Engineering and Computer Science with a courtesy appointment in Northwestern University’s School of Music. Bryan’s academic career began at the Ohio State University, where he received both a B. Mus. in Jazz Composition and a M.Sc. in Computer Science. After graduation, he spent several years working as a jazz musician and software developer. As a software developer he worked for the Speech & Hearing Science Department of Ohio State and for the statistical software company SPSS. Bryan then attended the University of Michigan, where he received a M. Mus. in Jazz and Improvisation, followed by a Ph.D. in Computer Science. Over the years, Bryan has also been featured on five albums, taught for two years as an adjunct professor in the Music Department of Madonna University, and worked as a researcher for General Dynamics on machine learning tasks. When he’s not programming, writing or teaching, he performs on saxophone and clarinet throughout the Midwest.