EF[ubgGp
lV G, 쑺 sO, C , ։ Y, J k
47񎩓Au, pp.527, Nov., 2004. [pdf document]

<Abstract>

This paper presents a wavelet denoising. In the proposed method, we use the output signal of wavelet denoising as the input signal again. This process is iterated a few times. From the investigation of segmental SNR and Itakura-Saito distortion measure, we show the speech enhancement using the proposed method reduces both musical noise and speech distortion.