Abstract:

Phonetics has traditionally been one of the first language levels that learners of a foreign language begin to master in a systematic manner. Regardless of their age and language learning experience, they are usually introduced to the sounds of a new language as soon as possible to ensure that they are able to move on to exploring grammar, lexis and start speaking without spending too much time on the theory that is separated from the practice of communication. The goal of such methodology is to help the students to familiarize themselves with the phonetics of a foreign language and ensure a minimal ability to express their ideas and recognize the sounds.

However, depending on the curriculum and learning goals, the control over pronunciation may diminish over time, students may fail to recognize language interference both at the sound and word level. Without listening practice and thorough guidance, this may lead to various difficulties in perception and comprehension of their speech.

The recent breakthrough in the modern technology facilitated by cloud computing, prominent use of mobile devices and further development of software-as-a-service business model have stimulated the growth of speech synthesis or text-to-speak services, some of which are integrated into dictionaries and machine translation software.

The authors explore the potential and share their experience of using such services as the neural-network-based Readspeaker, Speechify, Google Translate, etc. to teach their students and improve pronunciation of individual sounds in a foreign language, mastering its phonetic system, intonation, logical stress, etc. These tools present a new opportunity to help foreign language students to master pronunciation at the level of sentences rather than individual words, which may lead to a significant improvement of their connected speech skills.

Keywords:

Text-to-speech, speech synthesis, foreign languages, pronunciation, neural network.