The laboratory of signal processing IC prototyping was established in August 2002. This lab is located at Room 409C of Ge Zhi Building with a space of approximately 45 M2. The lab supervisor is Dr. Hu, who graduated from the Department of Electrical Engineering at National Cheng Kung University, Taiwan in 1985 and received his MS. and Ph. D. degrees from University of Florida, respectively in 1990 and 1993, both in Electrical Engineering. He was with the Department of Electronic Engineering at National I-Lan University since he returned to his country in 1993. His research was initially emphasized on the speech synthesis, coding, and enhancement, but attempts were made to deal with image processing and multimedia communications. Recent efforts were also directed to practical issues such as integrated circuit as well as embedded system design.
This lab aims at the research and development of techniques pertaining to signal processing. In general, there are two kinds of commonly encountered signals, namely, speech and image, which are of great interest to human beings. After sampling and digitization, a signal can be processed through a computer of which the numerical crushing power is capable of carrying out data analysis in an efficient way. Regardless of which signal, the research fields can be further categorized into feature extraction, noise suppression, coding, recognition, enhancement, etc. Once an investigable subject is settled, the researcher may systematically probe into the possibility of technical upgrading and/or performance improvement. Since the relevant techniques are becoming mature in the past decade, we believe it is the time to reap benefits from theoretic exploration. Hence, in order to push ahead the technical integration and application, a great deal of efforts in our lab were devoted to IC design and embedded system.
The research topics can be normally classified into two groups. Theory-oriented topics focus on performance improvement and algorithmic refinement, while the implementation-oriented topics place more emphasis on practical utility and hardware realization. Generally speaking, the simulation and verification for the developed theory were often undertaken in the working environment of Matlab. We go through a series of processes, which consist of the modeling setup and derivation, performance analysis of procedural flow, and assessment of experimental data. Academic publication is anticipated if the results are satisfactory. In case the developed technique has its market value, we can further transfer the proven algorithm into a program based on high-level languages. Useful application software can be put to use by incorporating the developed program with a friendly user interface.
Due to the rapid advancement of semiconductor technology, the 3C products have exhibited a new trend in style in recent years. The pursuit in light weight and flimsy size relies on the support of IC design. To fulfill the need of versatile purposes, the development of an embedded system to cope with miscellaneous functions in accordance with customers’ demanded. Whenever the realization is taken into account, we proceed with the hardware implementation from the viewpoint of system integration by combining the concepts of IC prototyping and embedded system design. In our Lab, the most frequently used tool is the FPGA development kit distributed by Altera Incorporation. The processing flow for either the video or audio signals ought to be validated prior to rushing into hardware design. Depending on its complexity, the underlying algorithm can be transformed into a fixed-point version using C++ or directly converted to the hardware description language. The actual logical circuitry is synthesized subsequent to the functional and timing verification. On the other hand, the implementation of an embedded system calls for the placement of a processor and other peripheral devices in addition to the associated software design and operation system management. Such a link evidently points out the importance of hardware/software codesign, which is exactly the biggest challenge to Lab members. The following are the developing research topics in our Lab. If you are interested in one of them, we welcome you to join our research team.
- Algorithm and chip design for low-bit-rate speech coding
- Development of embedded system and IC design for Mandarin Text-to-speech synthesizer
- Voice conversion via parametric text-to-speech synthesizer
- Techniques for noise reduction and bandwidth expansion of narrowband speech
- Image enhancement and noise suppression for low-cost video devices
- Architectures of fast algorithms in multimedia signal processing