摘要
针对一般的谱相减语音增强算法所带来的“音乐”噪声较为严重的问题 ,提出了一种改进的谱相减语音增强算法——噪声残差消除谱相减法。实验结果表明 ,噪声残差消除谱相减法明显地削弱了一般谱相减法所带来的“音乐”噪声。
To solve the serious problem of the “music” noise brought by the generic spectrum subtraction speech enhancement method, we put forward an improved spectrum subtraction speech enhancement method:Residual Noise Removal Spectrum Subtraction. The experiment results show that this improved method weakens the “music” noise distinctly. Whats more, it notably improves the quality of noisy speech.
出处
《解放军理工大学学报（自然科学版）》
EI
2001年第1期41-44,共4页
Journal of PLA University of Science and Technology(Natural Science Edition)
