research

Research

4C Industry (Consumer Electronics, Digital Mobile, Computer Multimedia, Automotive Electronics)-Digital Home-Electroacoustic Technology

Current research directions

Group	Topics
NN-array speech enhancement and separation	parameter estimation: SPP, PSD, SCM, ATF/RTF/RIR, source counting
	localization
	enhancement: denoising, dereverb, AEC/AFC
	separation: LSTSC
	sensor fusion
Audio Telepresence	model matching: TIKR, LCMV, GSC, NN-based model matching
	scalable enhancement vs. ambience
	binaural rendering
	global sound field synthesis (SFS) using a mic-NN-spk system: sound bar
	applications: goggles, drones, sound bar
Machine condition monitoring and diagnosis	DL: semi/self-supervised learning, reassembling
	array board (FPGA)
	acoustic camera (SOFI): kernel interpolation, NN-inverse problem, non-intrusive modal testing
	generative AI: diffusion model
Loudspeaker array	manifold learning: source and sensor interpolation via kernel and NN
	ATF estimation
	algorithms: TIKR, LCMV, kernel interpolation, DNN
	applications: audio focusing and zone control, ANC
	loudspeaker parameter ID, protection, compensation

Past research directions

Active noise control (ANC)	Research and development of active electronic silencer
	Development of active 3D sound canceling headset and cell phone
	Communication echo cancellation and noise suppression
	Development of active vibration isolation platforms
	Active rotor vibration control
	Active vibration control of piezoelectric cantilever beam
Noise and vibration diagnosis and control (NVDC)	Intelligent rotating machinery real-time sequential analysis system
	Measurement and prevention of noise and vibration
	Array microphone source recognition system (Nearfield Acoustic Beamformer, NABF)
	Vehicle noise and vibration diagnosis and control (NVH)
Acoustics	Boundary element method (BEM) sound field analysis
	BVSR Sound Field Analysis
	Analysis of acoustic inverse problems (Inverse problem)
	Sound field and structural interaction analysis
	Sound field omniphase technology (NAH/STSF)

Research results

1. Development of electroacoustic transducer and panel speaker

In the past few years, I have invested a lot of effort in the research of the Planar Speaker. The plane speaker is a new electro-acoustic technology that subverts the design concept of traditional speakers. Its diaphragm is different from the cone-shaped diaphragm of traditional speakers, the plane speaker adopts a flat flexible diaphragm, which can be used for the design of the speaker, and it can be used for the design of the speaker. utilization of electromagnetic fields, piezoelectric material or other means of excitation of this flat diaphragm produces sound field radiation. Planar horn with small size, not occupying space, wide sound range, omnidirectionality, and insensitivity to spatial boundary conditions, etc. It is a technology with great market potential and has attracted extensive attention from the industry. To date, individuals have developed theoretical analyses of system integration, GA/Taguchi optimization design, test procedures and performance evaluation. Application development and productization have been carried out, such as the development of moving coil and piezo actuators, large flat speaker matrix audio-visual system, small mobile communications, computer multimedia, medium-sized PDAs with flat speakers and other devices with productization potential. In addition, the individual also conducts microphone arrays, research and development of various advanced electroacoustic transducers such as the artificial electronic ear. Several papers have been published in J. Acoust. Soc. Am. (JASA) and other reputed journals.


Matrix large flat speaker AV system	Computer multimedia flat speaker array		Ultra-thin flat speakers for mobile phones and MP3s

DTV 3D audio		HRTF measurement system	Audio DSP engine

2. Audio signal processing

In recent years, multimedia systems have evolved to integrate video, audio, and computer animation, providing a more virtualized environment.（Virtual Reality, VR),and interactive and real-time information. These systems have a wide range of applications in many fields, such as headphones, desktop multimedia, home theater, VRTV, stereo, telepresence, distance learning, high-definition television, car audio, flight/driving simulators, hearing aids, and so on. My past research has also covered many topics related to the core technology of audio digital signal processing, such as array signal processing, adaptive filtering, multi-rate signal processing, sensorless subwoofer electronic compensation, and virtual sound field reproduction technology. 目前在JASA與ASME/J. Vibration (JVA)Several academic papers have been published in JASA and ASME/J. Vibration (JVA). Several patents are pending.


Virtual 5.1-channel headphones	HRTF numerical calculation grid

4C applications	3D Audio Processor for Handheld Devices

Computer multimedia personalization3D Audio/Video System	HRTF 3DSound field positioning

Click for On-Line Demo

3. Active noise and vibration control (ANVC) technology and system prototype development

Active noise and vibration control (ANC) often involves an approximation of the so-called Inverse Dynamics, which is very similar to the principle of Inverse Filtering used in audio signal processing such as Crosstalk Cancellation. I have put a lot of effort into the research and development of active noise and vibration control technology in recent years, and have obtained preliminary results, which are summarized as follows.

A. Extension of the algorithm：In the past, the control methods used by scholars in this field have been limited to LMS and other methods, but I have tried to apply other control methods, such as LQG, H2, H∞, L1, etc., to active control.

B. System Implementation：Individuals implement active control systems with analog circuits and digital signal processors to complete the integration of hardware and software.

C. Realized application examples：I have applied active control technology to such applications as air-conditioning/automotive electronic silencers, intelligent structures, active headphones, subwoofers, automotive cabin noise control, rotor active control systems, and active vibration isolation platforms. Compared to the traditional passive noise control method, active control has many advantages such as good low-frequency effect, small size, lightweight, no back pressure, programmable flexibility, etc., which is a very promising new control technology. I have published about 30 papers in this field in international journals such as JASA, ASME/JVA, J. Sound Vib.

Applications of ANC Technology Click

4. Intelligent Online Monitoring and Diagnostic System for Mechanical Systems

Over the past three years, I have also been involved in mechanical monitoring and diagnostics, such as vibration noise monitoring and troubleshooting of turbo compressors, fan automobile engines, intelligent order analysis of rotating machinery combined with adaptive Kalman Filter and Fuzzy Logic, as well as software and hardware systems for on-line real-time monitoring.

5. Other relevant research findings

In addition to the major research results mentioned above, my past research in acoustics and signal processing has covered many other techniques related to this project, such as sound field, structural intermodulation calculations, and sound field radiation modal analysis, which are the basic research on planar loudspeakers; an array microphone source identification system, which utilizes array signal processing techniques to calculate the orientation of the source; the Boundary Element Method (BEM) resonance and radiation analysis, which is an important numerical method for complex sound field calculations; and the Spatial Transformation technique, which is an important numerical method for the calculation of sound fields; and the Spatial Transformation technique (SPT), which is a method for calculating sound field radiations. Boundary Element Method (BEM) resonance and radiation analysis is an important numerical method for the calculation of complex sound fields; Spatial Transformation of Sound Field (STSF) is an important method for the calculation of sound fields. Spatial Transformation of Sound Field (STSF) is a useful tool for sound field visualization. These research results have been published in several academic journals.

A. Development of Sound Field Image Processing Technology：I have developed a technique for analyzing noise sources with complex geometries, which is different from the general FFT planar holographic method. Based on the boundary element method, the sound pressure or particle velocity data of the holographic plane is measured near the noise source, and the near-field or surface sound field can be back-calculated inwardly, or the radiation form of the far-field can be calculated outwardly. This method can effectively analyze the sound field radiation of irregularly shaped noise sources, but for higher frequency sound fields, it is necessary to use more dense boundary elements, and using the proposed inverse filtering method, the inverse reconstruction can sometimes be up to about 10 wavelengths. This method can be applied to noise source determination, non-contact structural modal analysis, structural energy flow analysis, far-field radiation shape calculation, etc., and has great potential for industrial applications (subsequently, B&K Instruments in Denmark developed a new product based on this concept, called the STSF system).

B. Application of Boundary Element Method to Noise and Vibration：I have developed a complex numerical method. The finite element method (FEM) and the boundary element method (BEM) are used to analyze the structural part and the sound field part of such problems, respectively. The two methods are combined to form an equation of motion with pressure and displacement field variables. The natural frequencies and vibration shapes are obtained by the singular value decomposition method. It is found that when there is a strong interlocking phenomenon, the natural frequency is significantly shifted due to the radiated load of the sound field, but the vibration shape and the nodal line are not significantly changed. The results of this study will provide a numerical analysis and design tool for the structural dynamic stability of the general sound field and structural interaction systems.

C. Numerical and Experimental Analysis of Sound Field and Structure Interaction Phenomena：A numerical and experimental method for predicting low-frequency surface vibrations due to internal sound pressure oscillations in structures has been developed. A modified acoustic field/structural modal testing technique is utilized to establish the response pattern between internal sound pressure and surface vibration, considering the structure and acoustic field as a coupled system and the loading effect of the acoustic field. The research contributes to the fact that it is not easy to find analytical or numerical solutions for structural/acoustic field systems with complex geometries in practical applications, such as automobile engine cylinder combustion chambers, missile pressurized engine combustion chambers, pressurized storage tanks, and nuclear reactors, etc. The method can be used to estimate structural surface vibrations from pressure measurements taken by pressure sensors, and it provides a useful experimental prediction technique.

« Return »