Corey I. Cheng's Home Page

Contact Information

Dolby Laboratories Research Division 100 Potrero Ave. San Francisco, CA 94103
e-mail: cnc@dolby.com, coreyicheng@gmail.com
URL: http://music.dartmouth.edu/~corey

Biosketch

The serious biosketch
Corey I. Cheng (b. 1972) studied physics at Harvard University (B.A. 1994), electro-acoustic music at Dartmouth College (M.A. 1996), and electrical engineering at the University of Michigan (M.S.E. 1998, Ph.D. 2001). Corey’s music has appeared at International Computer Music Conferences(ICMC's) in Hong Kong and Greece. In addition to composition, Corey’s engineering research interests include perceptual audio coding, spatial audio and Head-Related Transfer Functions (HRTF’s), applications of wavelets to electro-acoustic music, and econosonometrics – the study of sound level, trade volume, and price indices in "open cry-out" commodities trading pits. Corey is currently a Staff Engineer in the Research and Development department at Dolby Laboratories, San Francisco.

The not-so-serious biosketch
Corey has finally made it out of grad school. While in grad school, his student office was located directly underneath his advisor's desk, where his butt was intentionally placed within convenient kicking distance. As a student, he made scads of money, got plenty of sleep, was well-fed, and wasn't bitter at all. In obtaining degrees from Harvard and Michigan, Corey plans to follow in the footsteps of Ted Kaczinski, as his Michigan friends include the Bra Lady, the Queen of Despair, a priest, and a pimp. Corey is a proud alumnus of the University of Michigan Ballroom Dance Team, and is currently learning the Cha-Cha.

Research

Head-related Transfer Functions (HRTF's)

Description
Previous psychophysical studies have concluded that knowledge of the acoustic filtering properties of the pinna, head, and torso are important in human localization of sounds in space. The combined effects of such filtering are summarized by a single set of spatially dependent filters known as Head-Related Transfer Functions (HRTF’s). HRTF’s are can be empirically measured, and are commonly approximated as FIR filters. The current research focuses on alternate, spatial domain representations of HRTF's and the HRTF signal processing and perceptual implications of such representations.

Online Ph.D. Thesis: Visualization, Measurement, and Interpolation of Head-Related Transfer Functions (HRTF's) with Applications in Electro-Acoustic Music
README (text, ~2kB)
Complete Thesis (PDF, ~5.1 MB)
Abstract (PDF, ~8kB)
Frontmatter (PDF, ~81kB)
Chapter 1 (PDF, ~178 kB)
Chapter 2 (PDF, ~1376 kB)
Chapter 3 (PDF, ~774 kB)
Chapter 4 (PDF, 2.2 MB)
Chapter 5 (PDF, 407 kB)
Chapter 6 (PDF, 13 kB)
References (PDF, 65 kB)
3-D Sound examples from thesis

Papers on HRTF's
Cheng, Corey I. and Wakefield, Gregory H. “Spatial Frequency Response Surfaces: An Alternative Visualization Tool for Head-Related Transfer Functions (HRTF’s).” International Conference on Acoustics, Speech, and Signal Processing (ICASSP99), Phoenix, Arizona: 1999. (PDF format ~1926kB)

Cheng, Corey I. and Wakefield, Gregory H. “Spatial Frequency Response Surfaces (SFRS’s): An Alternative Visualization and Interpolation Technique for Head-Related Transfer Functions (HRTF’s).” AES (Audio Engineering Society) 16th International conference on spatial sound reproduction, Rovaniemi, Finland: 1999. (PDF format ~2461 kB)

Cheng, Corey I. and Wakefield, Gregory H. "Introduction to Head-Related Transfer Functions (HRTF's): Representations of HRTF's in Time, Frequency, and Space (invited paper)." Proceedings of the 107th Audio Engineering Society (AES) 107th Convention, New York: 1999.

The response of a subject's right and left ears at ~3027-3125 Hz. This style of plot is a Spatial Frequency Response Surface (SFRS), and shows how much energy the right and left ear canals receive as a function of spatial location (azimuth, elevation).

Econosonometrics

Description
Econosonometrics is the study of sound level, trade volume, and price indices in "open cry-out" commodities trading pits. Although many traditional "open cry-out" trading pits have been replaced by computerized transaction handling, "open cry-out" pits provide traders with acoustic information not available to electronic traders. The current research focuses on how acoustic information recorded at the Chicago Board of Trade corresponds to important financial quantities such as trade volume and price indices.

Articles on Econosonometrics
Burns, Greg. "Exchange noise can be boon to traders." The Chicago Tribune, Business Section, 4/11/99, front page: Chicago, IL: 1999. (JPG image ~535 kB)

Burns, Greg. "CBOT's time-recording a 'problem': Researchers." The Chicago Tribune, Business Section, 4/13/99, page 3: Chicago, IL: 1999. (JPG image ~614 kB)

Total energy recorded from a microphone placed over the 30-year government bond "open-cryout" futures trading pit at the Chicago Board of Trade (CBOT), plotted as a function of time. This record shows relative overall acoustic activity in the pit on May 25, 26, 27, 28, and May 31, 1998. Note: on May 28 (Friday), the pits closed early; on May 31 (Monday, Memorial Day), the pits were closed.

Wavelets

	Description Wavelet analysis is one of many generalized time-frequency analysis methods which give a signal's frequency content at a certain point in time. A strength of wavelet analysis is its ability to capture transient information, as the time support for different frequency bands varies inversely with center frequency of each band. Wavelets have been used in applications such as denoising and signal compression. The current research applies wavelet signal procesing to audio for denoising, equalization, and excitation purposes. On-line Master's thesis: Wavelet Signal Processing of Digital Audio with Applications in Electro-acoustic Music. HTML version PDF version (requires Adobe Acrobat Reader. Best viewed with Netscape v4.0 and higher. All sections of the thesis are ~82kB or less, except for Appendix L, which is 398kB) Papers on wavelets Cheng, Corey I. “High Frequency compensation of low sample-rate audio files: A Wavelet-based spectral excitation algorithm.” Proceedings of the 1997 International Computer Music Conference (ICMC ‘97), Thessaloniki, Greece: 1997. (PDF format ~26 kB)
Wavelet analysis of a signal is an analysis of the signal using a non-uniform division of the time-frequency plane (time is on the horizontal axis; frequency is on the vertical axis). Each rectangle in the figure corresponds to a single wavelet coefficient with a different time and frequency support range.

This simple score is a musical representation of the non-uniform division of the time-frequency plane shown above. Each note corresponds to one rectangle in the time-frequency plane shown above, and each instrument's dynamic corresponds to the degree of shading, or associated energy, of each rectangle.