Speech enhancement (SE) plays an important role in modern-day communication of human speakers and automatic speech recognition (ASR). In this talk, an overview of SE problems, including denoising, dereverberation, and source separation problems, is given. A plethora of SE techniques have been suggested in literature on the basis of spatial cues, spectro-temporal sparsity, statistical independence, etc. State-of-the-art SE approaches are compared with the newly emerging learning-based approaches in the face of adverse acoustic conditions. These two classes of SE approaches are evaluated in terms of various objective and subjective metrics. The results have revealed that each approach has its own pros and cons for various SE applications. Leveraging the complementary power of two approaches to maximize the synergy of both in terms of performance and robustness will be an ongoing research.

Mingsian R. Bai was born in 1959 in Taipei, Taiwan. He received a bachelor’s degree in Power Mechanical Engineering from National Tsing Hua University in 1981. He also received a master degree in Business Management from National Chen-Chi University in 1984. He left Taiwan in 1984 to enter graduate school of Iowa State University and later received a MS degree from Mechanical Engineering in 1985 and a Ph. D. from Engineering Mechanics and Aerospace Engineering in 1989.
In 1989, he joined the Department of Mechanical Engineering of National Chiao-Tung University in Taiwan as an associate professor and became a professor in 1996. He moved to National Tsing Hua University in 2010 and served as a distinguished professor in Power Mechanical Engineering
and the director of Telecom-Electroacoustics-Audio (TEA) laboratory.
He was also a visiting scholar to Center of Vibration and Acoustics, Penn State University, University of Adelaide, Australia, and Institute of Sound and Vibration Research (ISVR), UK in 1997, 2000, 2002, respectively. He has rather diverse research interests in acoustics, spanning acoustic array systems, audio signal processing, electroacoustic transducers, vibroacoustic diagnostics, active noise and vibration control, and so forth.
He currently serves as an active consultant and a project leader in these areas in industry. He has published over 231 papers, 5 books (including Engineering Acoustics and Acoustic Array Systems published by IEEE/Wiley), and 36 granted patents. He has received numerous academic awards including two Outstanding Research Awards from National Science Council in 2006 and 2010. Professor Bai is a senior member of the Institute of Electrical and Electronics Engineers (IEEE), American Society of Mechanical Engineers (ASME), the Audio Engineering Society (AES), the Acoustical Society of America (ASA), the Acoustical Society of Taiwan, and the Vibration and Noise Control Engineering Society in Taiwan. He is an associate editor under transduction and engineering acoustics and a technical committee member in signal processing (TCSP) of the Journal of Acoustical Society of America (JASA) since July 2012 and a guest editor of Asian Journal of Control in 2012. He was elected as a Fellow of Acoustical Society of America (ASA) in 2014 and a Fellow of American Society of Mechanical Engineers (ASME) in 2021.