Frequency warped wiener filtering for MEL-LPC based speech recognition

Summary form only given, as follows. This paper presents a frequency warped Wiener filter to enhance Mel-LPC spectra in presence of additive noise. The proposed filter is estimated based on minimization of error signal on the linear frequency scale and then is efficiently implemented in the autocorrelation domain without denoising input speech. The performance of the proposed filter is evaluated by speech recognition experiments under the speech with babble and white noise conditions. The optimum filter order is shown to be comparable to that of Mel-LPC analysis, and thus filtering is computationally inexpensive. As a result, word accuracy is improved by about 20% at most with the proposed Wiener filter.