Speech dereverberation

• Main Author: Chua, Benedict Chi En.
• Other Authors: School of Electrical and Electronic Engineering
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
• Online Access: http://hdl.handle.net/10356/40818
• Institution: Nanyang Technological University

The phenomenon of reverberations has always and will continue to be a potential debating issue in terms of speech processing in an enclosed environment. For environments such as concert theatres and indoor stadiums, reverberations of the original source often tend to increase the “liveness” of the overall sound, as a result, improve the quality of the signal we receive in our ears. However, on flip side, when we discuss about reverberations occurring in a meeting room where a conference call is currently being made, these reverberations usually distort the original signal, and as a result, the listener on the other end of the phone line will experience an “echoy” and degraded speech signal instead of the ideal clean signal. This might lead to misinterpretations during communication through such conference calls, which in turn, might end up causing more serious implications. The interesting aspect of this problem is that we can use various methods to retrieve a signal as similar to the original signal as possible. In the case of this project and report, we will make use of calculated data using one of the simplest and most basic methods of delay-and-sum beamforming to introduce delays to signals received at every microphone in an array to retrieve the desired output. Our main goal of this project is to achieve maximum dereverberation using the delay-and-sum beamforming technique. Through this technique, we can estimate the Direction of Arrival (DOA) of the source signal, and using the appropriate delays calculated at this DOA, attempt to retrieve the source signal. The delay-and-sum beamformer poses several areas which can be potentially improved through further research. As the number of microphones in an array is directly proportional to the accuracy of the algorithm, more microphones in an array is preferable. However, increasing the amount of microphones also requires more computing capability. Improvements can be made in this aspect to improve the efficiency of the algorithm.