Assistant Professor of Electrical Engineering
Harvard University
"Time-Frequency Representations and Statistical Models for Audio and Auditory Signal Analysis"
Time-frequency representations are ubiquitous in audio signal processing, their use being motivated by both auditory physiology and the mathematics of Fourier analysis. Indeed, information-carrying natural sound signals can often be meaningfully represented as a superposition of translated, modulated versions of a simple window function exhibiting good time-frequency concentration. In combination with statistical models formulated in the space of time-frequency coefficients, such an approach provides a principled way of decomposing sounds into their constituent parts, as well as an effective means of exploiting the local correlation present in the time-frequency structure of natural sound signals. Several resultant applications of interest to practitioners will be addressed in this talk, including examples of speech and audio signal analysis, enhancement, and compression, as well as a discussion of the human peripheral auditory system from the point of view of time-frequency analysis.