The Application Of Speech Evaluation System Based On DTW In English Teaching Of Middle School

With the development of globalization and internationalization,the degree of emphasis on English has been raised,more and more people speak English as a second language.Spoken English is a prerequisite for learning English well,in the mode of traditional teaching,teachers pay attention to reading and writing ability only,and neglect the oral English teaching often.Today,with the development of science,more and more software of oral English learning is emerging,however,many of them are lack of real-time feedback.Although some can,but not for the practical demand of English teaching in middle school,teachers can not use it based on the actual needs of teaching.Based on the practical demand of English teaching in middle school,this paper designs and develops the system of English vocabulary pronunciation evaluation.Teachers can make vocabulary standard audio for students according to the progress of teaching.Students can practice according to the standard audio system,and the system can feedback immediately.The research of this paper mainly as follows:Firstly,research and analysis the system of English speech recognition and oral practice in the market,based on this,the author studies English vocabulary pronunciation evaluation as the research goal according to the practical demand of English teachers in middle school.Secondly,studies the basic theory of speech information processing,and select the method of combining the spectral subtraction and the short-term energy with better anti-noise ability and robustness.In the aspect of speech feature,chose the smooth Mel-Frequency Cepstrum Coefficient(SMFCC),the intensity of sound and pitch frequency as the characteristic parameters of the system alignment.Thirdly,based on the analysis of speech pattern matching algorithm according to DTW,put forward an improved algorithm of DTW speech matching.Based on the accuracy of matching,improve the speed of voice matching and ensure the real time of voice evaluation system.Finally,designed and implemented the voice evaluation system,and test the system on the Android platform.The results show that the system can basically evaluate the vocabulary pronunciation of students objectively.The system of speech evaluation completed in this paper,can be applied to the teaching on platform of starC,which increases the interest of students 'oral English learning and improves the efficiency of students' oral learning.Teachers can also understand the pronunciation of each student accurately.