| 一种基于频谱特征的语音信号识别方法及硬件实现 |
| |
| 引用本文： | 蔡建,方勇,音春,徐洪伟. 一种基于频谱特征的语音信号识别方法及硬件实现[J]. 电声技术, 2003, 0(7): 11-14 |
| |
| 作者姓名： | 蔡建 方勇 音春 徐洪伟 |
| |
| 作者单位： | 四川大学,电子信息学院,四川,成都,610064 |
| |
| 摘 要： | 在客观实验中运用短时分析技术，着重研究语音信号的频谱特征，并采用硬件方式实现对语音信号的识别，得到如下结论：基于频谱特征的模型对纯语音信号能保留语音自然度95％以上，对纯噪声信号能压缩噪声95％以上。
|
| 关 键 词： | 频谱特征 语音信号识别 短时分析 傅里叶变换 压缩率 |
| 修稿时间： | 2003-01-17 |
A method for voice signal identification based on spectral characteristics and its hardware implementation |
| |
| CAI Jian,FANG Yong,et al.. A method for voice signal identification based on spectral characteristics and its hardware implementation[J]. Audio Engineering, 2003, 0(7): 11-14 |
| |
| Authors: | CAI Jian FANG Yong et al. |
| |
| Abstract: | By using short time Fourier transform in the experiment,the study of spectrum characteris-tics is focused and voice signal is identified u sing hardware.The model based on spectrum can keep more than95%of naturalnes s for pure voice signal and compress more than95%of noise for pure noise sign al. |
| |
| Keywords: | spectrum Fourier transform short time analysis compression ratio |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
