摘要
利用TI公司制造的DSP芯片TMS320VC5509和CODEC芯片TLV320AIC23构建一个语音采集系统,为语音信号处理搭建一个有效的平台。具体方法是:采用TMS320VC5509的I2C模块对TLV320AIC23进行配置,并通过TMSVC5509的McBSP实现两者之间的数据通信;利用小波变换在信号突变检测方面的原理在不同尺度下对采集到的语音信号进行小波变换,并引入空域相关函数,利用相邻尺度的相关系数代替传统小波变换方法中的小波系数寻找局部极大值之间的距离。实验结果表明,系统以22kHz的采样频率采集到语音信号;通过计算不同尺度下的相关系数局部极大值之间的距离,能估计出基音周期;与传统的小波变换方法相比,基于空域相关函数的小波变换方法在锐化信号边缘的同时抑制了噪声,对噪声的鲁棒性更强。
A speech acquisition system based on DSP chip TLV320AIC23 and CODEC chip TMS320VC5509 used as an efficient platform of speech signal processing was constructed. The detail procedures were as follows： I^2C module of TMS320VC5509 was used to configure the TLV320AIC23 and data transfer between the two chips was implemented through McBSP on TMS320VC5509. Pitch period of the speech acquired was detected based on the principle of singularity detection of signals with wavelet transform at different scales. An arithmetic based on spatial correlation function which used the correlation coefficients to replace the wavelet transform coefficients to look for the distance between the local maximums was proposed. The experimental results show that speech signal is acquired at sample frequency of 22 kHz and the pitch period is estimated by fmding the distance between the local maximums. Compared with the classic arithmetic using wavelet transform, the arithmetic based on spatial correlation function can enhance sharp edges and suppress noise, and so it has stronger robust to noise.
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2008年第4期805-809,共5页
Journal of Central South University:Science and Technology
基金
中国博士后科研基金资助项目(20060390883)
国家杰出青年科研基金资助项目(60425310)
