论文部分内容阅读
文章给出了计算机辅助汉语教学系统中语音端点信号的检测和清浊音信号的切分方法:采用短时相对能频积对汉语语音信号的端点进行检测;采用短时相对能频比的方法对语音信号的清浊音进行切分。这两种方法的使用与现有方法相比可以有效地提高汉语语音信号切分的成功率,实验结果表明正确率可达到95%以上。文中通过实验验证了所提出的汉语语音信号切分方法是有效的和可行的。它基本上能够满足计算机辅助汉语教学系统在线切分汉语语音信号的需要,比已有的语音信号切分方法的切分效果有显著提高,为下一步提高语音信号的识别率奠定了基础。
The article gives the method of detecting the endpoint of speech signal and the method of dividing the voiced and unvoiced signal in the computer aided Chinese teaching system. The endpoint of Chinese speech signal is detected by short-term relative energy-frequency product. The method of short-time relative energy-frequency ratio The voiced / unvoiced voice signal is split. Compared with the existing methods, the use of these two methods can effectively improve the success rate of Chinese speech signal segmentation, the experimental results show that the correct rate can reach more than 95%. The experiment proves that the proposed segmentation method of Chinese speech signal is effective and feasible. It basically meets the need of computer-aided Chinese teaching system to segment Chinese speech signals on-line. Compared with existing speech signal segmentation methods, the segmentation effect is significantly improved, which lays the foundation for improving speech recognition rate in the next step.