【作者】 戴明扬； 徐柏龄；

【Author】 Dai Mingyang Xu Boling (Institute of Acoustics,State Key Laboratory of Modern Acoustics,Nanjing University,210093)

【机构】 南京大学声学研究所近代声学国家重点实验室；

【摘要】 本文对基于全极点模型迭代维纳滤波框架的语音增强方法进行深入研究,指出声门波的失真和畸变是导致该类方法增强语音听感质量不够好的主要原因。本文从声门波的恢复和增强入手,探讨了基于声门波LF模型进行声门波分析和增强的可行性及方法。本文深入讨论了声门波LF模型时间特性参数的提取、声门波LF模型时间特性参数的鲁棒性、以及声门波LF模型时间特性参数一定程度的畸变对语音音质的影响。在上述分析的基础上,本文提出了基于激励声门波码本约束受限的声门波增强方法,模拟实验表明此方法能够有效减小声门波的畸变与失真,提高浊音语音段增强语音听感质量。更多还原

【Abstract】 This paper is intended for improving speech quality enhanced by All-Pole based iterative Wiener filtering.After deep research into the All-Pole based iterative Wiener filtering speech enhancement methods,we obtained the result that low quality of the enhanced speech mainly due to glottal wave distortion,so we probed into the crude way to improve enhanced speech quality by glottal wave restoration.Here we choose LF model as the basis for glottal wave enhancement,discuss the extraction and the robustness of LF model timing parameters.We proposed the All-pole based iterative Wiener filter.and glottal wave codebook constraint speech enhancement method,as shown in the simulation experiments,this method can reduce the glottal wave distortion and improve the enhanced speech quality effectively.更多还原