基于语音感知的英语口语发音自动校准系统

Automatic Calibration System of Spoken English Pronunciation Based on Speech Perception

为提升英语口语发音质量,设计基于语音感知的英语口语发音自动校准系统。通过语音感知传感器与英语口语发音处理器设计系统硬件,采集英语口语发音数字信号,并以此为基础,对信号进行预加重、分帧加窗处理,获取信号特征MFCC系数,将MFCC系数作为训练数据集,构建发音检错模型,以检测到的错误发音为依据,对英语口语发音自动校准。实验结果表明系统英语口语发音检错正确率与校准发音质量评分更高,具备更加优质的性能。

In order to improve the quality of spoken English pronunciation,an automatic calibration system for spoken English pronunciation based on voice perception is designed.It designs the system hardware through the voice sensor and the spoken English pronunciation processor;collects the digital signal of spoken English pronunciation,and based on this,perform pre-emphasis,frame and window processing of the signal is taken to obtain the signal feature-the MFCC coefficient,and the MFCC coefficient as a training data set,a pronunciation error detection model is constructed to automatically calibrate the spoken English pronunciation based on the detected false pronunciation.Experimental results show that the designed system has a higher correct rate of spoken English pronunciation and a higher score for the quality of calibrated pronunciation,which fully proves that the designed system has better performance.