Wavelet Packet 변환에서 적응적 잡음추정과 음질개선

Abstract

In this paper, a noise estimation method and a speech enhancement method are proposed, which can be applied efficiently in wavelet packet transform. At first, the proposed noise estimation method consists of the following two main processes to analyze and estimate efficiently the noise from the noisy speech: 1) a least-squares line is used, which is obtained by applying coefficient magnitudes in node with a uniform wavelet packet transform to a least squares method; 2) a differential forgetting factor and a correlation coefficient per subband are applied, where each subband consists of several nodes with the uniform wavelet packet transform. In particular, this approach has the ability to update noise estimation by using the estimated noise at the previous frame only instead of employing the statistical information of long past frames and explicit nonspeech frames detection consisted of noise signals. The proposed speech enhancement method uses a modified spectral subtraction with noise estimation by a least-squares line method and with and overweighting gain per subband with nonlinear structure, where the overweighting gain is used for suppressing the residue of musical noise and the subband is used for applying the weighted values according to change of signals. The enhanced speech by our method has the following properties: 1) the speech intelligibility can be assured reliably; 2)the musical noise can be reduced efficiently. For the experiment, we used noisy speech with SNR 0 - 15 dB corrupted by various measured such as segmental SNR, log area ratio, weighted spectral slope measure, segmental noise estimation error,and waveform spectrogram analysis of enhanced speech. In various assessments, it was conformed that the proposed noise estimation method and speech enhancement method were better performances than the compared method in various noise-level conditions. Especially, our methods showed reliable results even at low SNR.