为了减少带噪语音信号的噪声含量,本文根据小波阈值函数降噪法,提出改进型阈值函数和名为newshrink阈值的改进型阈值对带噪语音信号进行降噪处理。在MATLAB平台上设定四组不同信噪比的带噪语音,利用不同的小波基函数、阈值和阈值函数对...为了减少带噪语音信号的噪声含量,本文根据小波阈值函数降噪法,提出改进型阈值函数和名为newshrink阈值的改进型阈值对带噪语音信号进行降噪处理。在MATLAB平台上设定四组不同信噪比的带噪语音,利用不同的小波基函数、阈值和阈值函数对语音信号降噪,并以信噪比和均方误差评价小波降噪效果。实验结果表明,以9层小波分解层数、db6小波基函数、newshrink阈值以及改进型阈值函数为小波变换参数的降噪组合,提高了语音信号信噪比,降低了均方误差,改善了降噪效果。To reduce the noise content in noisy speech signals, this paper proposes an improved wavelet threshold function and a new threshold named “newshrink” for speech signal denoising. Four sets of noisy speech with distinct signal-to-noise ratios are established on MATLAB, and diverse wavelet basis functions, thresholds, and threshold functions are employed to denoise the speech signal. The signal-to-noise ratio (SNR) and the mean square error (MSE) are utilized to assess the effectiveness of wavelet denoising. The experimental results demonstrate that the denoising combination of a 9-level wavelet decomposition level, the db6 wavelet basis function, the newshrink threshold and the improved threshold function as the parameters of wavelet transformation enhances the SNR of the speech signal, reduces the MSE, and enhances the noise reduction effect.展开更多
文摘为了减少带噪语音信号的噪声含量,本文根据小波阈值函数降噪法,提出改进型阈值函数和名为newshrink阈值的改进型阈值对带噪语音信号进行降噪处理。在MATLAB平台上设定四组不同信噪比的带噪语音,利用不同的小波基函数、阈值和阈值函数对语音信号降噪,并以信噪比和均方误差评价小波降噪效果。实验结果表明,以9层小波分解层数、db6小波基函数、newshrink阈值以及改进型阈值函数为小波变换参数的降噪组合,提高了语音信号信噪比,降低了均方误差,改善了降噪效果。To reduce the noise content in noisy speech signals, this paper proposes an improved wavelet threshold function and a new threshold named “newshrink” for speech signal denoising. Four sets of noisy speech with distinct signal-to-noise ratios are established on MATLAB, and diverse wavelet basis functions, thresholds, and threshold functions are employed to denoise the speech signal. The signal-to-noise ratio (SNR) and the mean square error (MSE) are utilized to assess the effectiveness of wavelet denoising. The experimental results demonstrate that the denoising combination of a 9-level wavelet decomposition level, the db6 wavelet basis function, the newshrink threshold and the improved threshold function as the parameters of wavelet transformation enhances the SNR of the speech signal, reduces the MSE, and enhances the noise reduction effect.
