Discipline: Computer Sciences & Information Management
Subcategory: STEM Research
David Heise - Lincoln University
Computational Research on Music & Audio (CRoMA) was launched at Lincoln University in 2015 with the support of the National Science Foundation through an HBCU-UP award. The purpose of this project is to establish a research program to study aspects and applications of computational audio signal processing. This effort has an interdisciplinary focus, drawing from disciplines such as computer science, engineering, mathematics, psychology, and music. Further, the project aims to specifically include undergraduate students in the research activities. In year two, the project has: a) directly supported five undergraduate students and one graduate student as research assistants, b) supported summer research for the PI at the Center for Interdisciplinary Research on Music, Media & Technology (CIRMMT, housed at McGill University), c) fostered interdisciplinary collaborations between researchers in the region and beyond, and d) enabled presentation of work by students and the PI at regional and international conferences. A particular area of research that has advanced within this project is the development of a method to incorporate attention into computational auditory scene analysis (CASA) using ‘focal templates’. Attention is the ability of a listener to filter-out sounds not conforming to an expected pattern, and this ability comes naturally to humans in everyday life (consider the ‘cocktail party’ problem, or a busy conference poster session). Incorporating attention into CASA is a complex problem, and the technique of focal templates has been successfully applied to the detection of bees buzzing within environmental audio. Results to-date are encouraging and suggest that the method may have promise in more challenging contexts. A foundation for CRoMA has been established through this project, and the PI is looking for opportunities to sustain this effort through forged collaborations and other sources of support.
Funder Acknowledgement(s): National Science Foundation, Award #1410586
Faculty Advisor: None Listed,