Topics

2019.09.13

【Young Faculty:027】Shoichi Koyama, Lecturer: Saruwatari & Koyama lab, Department of Mathematical Engineering and Information Physics

Biography
2009-2014 Researcher, Nippon Telegraph and Telephone Corp.
2014 Ph.D. (Information Science and Technology), The University of Tokyo
2014-2018 Assistant Professor (Research Associate), The University of Tokyo
2016-2018 JSPS Overseas Research Fellow (Paris Diderot University)
2018- Assistant Professor (Lecturer), The University of Tokyo
2018- Researcher, JST PRESTO
2018- H30 UTokyo Excellent Young Researcher

Humans use "sound" as a means of communication and artistic expression. Speech is one of the most primitive tools of communication, and telephones are still widely used for telecommunication in various forms. Sound is also necessary for broadcasting and human-computer interaction. Furthermore, there are various types of artistic expression using sound, such as music.

Our main research topic is acoustic signal processing, which aims to extract necessary information from acoustic signals or synthesize desired sound. By using modeling methods based on physical acoustics and statistical nature, and mathematical techniques such as inverse problems, optimization, and machine learning, we develop new technologies for acoustic signal processing.

One example is sound field reproduction for VR/AR audio systems, which aims to synthesize a desired spatial sound by using multiple loudspeakers. In contrast to stereophonic/surround systems, which is based on human auditory property, sound field reproduction enables to physically reproduce sound space in a large area. In a mathematical sense, such problem is regarded as an inverse problem. So far, we have proposed a wave-domain recording and reproduction method and super-resolution based on sparse representation. Moreover, we have developed a real-time sound field transmission system.

References:
- S. Koyama and L. Daudet, “Sparse representation of a spatial sound field in a reverberant environment,” IEEE J. Sel. Topics Signal Process., 13 (1): 172-184, 2019.
- S. Koyama, N. Murata, and H. Saruwatari, “Sparse sound field decomposition for super-resolution in recording and reproduction,” J. Acoust. Soc. Amer., 143 (6): 3780-3895, 2018.
- S. Koyama, K. Furuya, H. Uematsu, Y. Hiwasaki, and Y. Haneda, “Real-time sound field transmission system by using wave field reconstruction filter and its evaluation,” IEICE Trans. Fundamentals, E97-A (9): 1840-1848, 2014.
- S. Koyama, K. Furuya, Y. Hiwasaki, and Y. Haneda, “Analytical approach to wave field reconstruction filtering in spatio-temporal frequency domain,” IEEE Trans. Acoust. Signal, Lang., Process., 21 (4): 685-696, 2013.

gif

Currently, various systems using audio signal processing, such as smart speakers, have been developed; however, there still remain many issues. I would like to widely contribute to academic and industrial fields by developing new technologies that can lead to new cultures and artistic expressions.

Saruwatari & Koyama lab: http://www.sp.ipc.i.u-tokyo.ac.jp/