The Reform of Pronunciation Teaching in Colleges and Universities by Praat Software From the Perspective of Deep Learning

Introduction

English as an international language plays a pivotal role in today’s international communication process. Clear and understandable speech is the foundation of people’s communication. Whether the pronunciation is correct or not directly affects the expression of meaning. Voice can be divided into vowels and consonants, among which vowels are the core and backbone and are the basis of pronunciation. Due to factors that are difficult to grasp, such as the opening degree and tongue position, it is not easy to master the single tone. Traditional English pronunciation teaching mainly relies on teachers’ demonstration of the pronunciation process and explanation of pronunciation methods. Learners rely on their own auditory perception to listen, imitate, and mechanically practice the teachers’ pronunciation, audio or video, and practice repeatedly after teachers’ error correction. Traditional phonetic teaching lacks objective perceptual evaluation criteria and cannot achieve satisfactory teaching results.

For unit phonetic teaching, current research mainly focuses on the comparison of vowels between English and Chinese, the mechanism of second language acquisition, negative transfer of mother tongue, and interlingual research. Through the comparison of English and Chinese vowels, many scholars have discovered the similarities and differences between the two, and pointed out the essentials of pronunciation. They are helpful and enlightening to Chinese English learners, but the method and correctness of pronunciation are only based on subjective perception, not objective enough. The development of global informatization has provided new means for education. The birth and development of the Internet has provided a wider development space for education. The organic combination of modern information technology and education, that is, the application of artificial intelligence technology in modern education, makes learning an ongoing and lifelong process. Nowadays, the construction of networked courses is on the rise. How to reflect intelligence in the teaching platform is a hot topic of various networked courses. Most of the existing intelligent products are in the stage of experimental and theoretical research, and their degree of intelligence is limited. Therefore, it is of great practical significance to design a network-based artificial intelligence-assisted teaching system. At present, the existing voice analysis software is generally not specially designed for English voice teaching. Teachers need to constantly learn modern educational technology, combine the advantages of existing voice software, establish a humanized visual voice teaching platform, improve their own software operation level and ability, and provide effective guidance and suggestions for students. Although the waveform and spectrum speech analysis software can visually and vividly present speech information, teachers and learners need special knowledge and training to understand and master this kind of visual information, which is difficult. In view of the current problems in English phonetics teaching and the inspiration of multimodal discourse theory, many researchers have been devoted to the research and development of speech visualization software and the research on the practical application effect of speech software. Among them, the most widely used is Praat voice processing software. Praat phonetics software, formerly known as “Praat: doing phonetics by computer,” is a cross-platform multifunctional phonetics professional software, which is mainly used for the analysis, annotation, processing, synthesis, and other experiments of digital speech signals. In addition, it generates various language diagrams and text reports at the same time.