混响环境下语音信号优化识别仿真研究

发布时间：2018-02-02 23:54

本文关键词： 混响环境语音增强　出处：《计算机仿真》2017年02期 　论文类型：期刊论文

【摘要】：对混响环境下语音信号进行优化识别,可以提高连续英汉语音的识别度。进行语音信号优化识别时,在混响环境下,应对连续英汉语音增强进行增强处理,但是采用当前的方法进行英汉语音增强时,需要大量的延时单元来进行语音增强,存在增强过程较为繁琐的问题。提出一种基于宽带波束形成的混响环境下语音信号优化识别方法。上述方法先融合于小波滤波器将原始的英汉语音变换到小波域,组建小波域阵列宽带英汉语音信号模型,并对双耳正前方英汉语音进行增强,并抑制其它方位噪声,利用改进的维纳滤波方法将各段波束输出英汉语音数据时间序列进行连接,组建连续的波束输出序列,实现对混响环境下连续英汉语音信号的优化识别。仿真结果证明,所提方法对语音信号优化识别较好,有效地抑制了混响和噪声对连续英汉语音信号的影响。
[Abstract]:The speech signal recognition in reverberation environment can improve the recognition degree of continuous English and Chinese speech. In the reverberation environment, continuous English and Chinese speech enhancement should be enhanced. However, when we use the current method to enhance English and Chinese speech, we need a large number of delay units for speech enhancement. In this paper, a method of speech signal recognition in reverberation environment based on wideband beamforming is proposed. Firstly, the method is fused with wavelet filter to transform the original English and Chinese speech into wavelet domain. . The wideband English-Chinese speech signal model of wavelet array is constructed, and the English and Chinese speech in front of binaural is enhanced, and other azimuth noise is suppressed. The improved Wiener filter is used to link the beam-beam output time series of English and Chinese speech data, and a continuous beam-output sequence is constructed. The simulation results show that the proposed method is better for the optimal recognition of speech signals and can effectively suppress the influence of reverberation and noise on continuous English and Chinese speech signals.
【作者单位】：长江大学文理学院;
【分类号】：TN912.3
【正文快照】： 1_ 目前,随着中国与国际社会在文化、经济、政治等方面交流的日趋频繁,以英语为基础的视频会议系统在各个行业领域中受到了广泛的应用^3]。在实际的视频会议系统中,连续英汉语音信号常常受到混响环境的干扰,致使语音处理系统无法正常的工作,难以获取纯净的原始语音连续信号,

【参考文献】