Time-frequency plane Wiener filtering for robust processing of speech signals
Adrian Ang, Ee-Luang Ang, A. Benjamin Premkumar, A. S. Madhukumar · 2002
This paper uses discrete wavelet transforms together with Wiener filtering to enhance noisy speech for further processing. The speech signal is mixed with various levels of white and coloured noises and decomposed into several levels using wavelet analysis. The output signal at each level is processed separately using iterative Wiener filtering. An efficient speech coding system is developed for the enhanced speech bands.