| 基于基音周期的语音MFCC参数提取 |
| |
| 引用本文： | 陈迪,龚卫国,杨利平.基于基音周期的语音MFCC参数提取[J].计算机应用,2007,27(5):1217-1219. |
| |
| 作者姓名： | 陈迪 龚卫国 杨利平 |
| |
| 作者单位： | 重庆大学,光电技术及系统教育部重点实验室,重庆,400044 |
| |
| 摘 要： | 提出了一种可用于改善说话人识别效果的基于基音周期的可变窗长语音MFCC参数提取方法。基本原理是将原始的语音分解为当前基音周期整数倍长度以内部分及其以外部分，并保留前者舍去后者，以减小训练语音与测试语音的频谱失真。通过文本无关的说话人确认实验，验证了该方法能有效提高说话人确认的识别率，并能提高短时语音的稳定性。
|
| 关 键 词： | 说话人识别 基音周期 MFCC参数 频谱距离 |
| 文章编号： | 1001-9081（2007）05-1217-03 |
| 收稿时间： | 2006-11-17 |
| 修稿时间： | 2006-11-17 |
MFCC feature extraction of speech based on pitch period |
| |
| CHEN Di,GONG Wei-guo,YANG Li-ping.MFCC feature extraction of speech based on pitch period[J].journal of Computer Applications,2007,27(5):1217-1219. |
| |
| Authors: | CHEN Di GONG Wei-guo YANG Li-ping |
| |
| Affiliation: | Key Laboratory for Optoelectronic Technology and System of the Education Ministry of China, Chongqing University, Chongqing 400044, China |
| |
| Abstract: | This paper proposed a MFCC feature extraction method based on pitch period to improve the performance of speaker recognition.The original speech signals were decomposed into two parts: the speech within the length of multiple pitch periods and that out of the pitch period.The former part was reserved to minimize the spectral distortion.The improved MFCC feature extraction method was also used in the verification experiment of the text-independent speaker recognition.The result shows the recognition rate is improved. |
| |
| Keywords: | speaker verification pitch period Mel-Frequency Cepstral Coefficients(MFCC) spectral distance |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机应用》浏览原始摘要信息 |
|
点击此处可从《计算机应用》下载全文 |
