An Automatic Vowel Space Generator for Language Learner Pronunciation Acquisition and Correction

An Automatic Vowel Space Generator for Language Learners’ Pronunciation Acquisition and Correction Xinyuan Chao1, Charbel El-Khaissi2, Nicholas Kuo1, Priscilla Kan John1, and Hanna Suominen1, 3, 4 1Research School of Computer Science, The Australian National University, Australia 2College of Arts and Social Sciences, The Australian National University, Australia 3Data61, Commonwealth Scientific and Industrial Research Organisation, Australia 4Department of Future Technologies, University of Turku, Finland {u6456596, charbel.el-khaissi, nicholas.kuo, priscilla.kanjohn, hanna.suominen}@anu.edu.au Abstract Speech visualisations are known to help lan- guage learners to acquire correct pronuncia- tion and promote a better study experience. We present a two-step approach based on two established techniques to display tongue tip movements of an acoustic speech signal on a vowel space plot. First, we use Energy En- tropy Ratio to extract vowels; and then, we ap- ply the Linear Predictive Coding root method to estimate Formant 1 and Formant 2. We invited and collected acoustic data from one Modern Standard Arabic (MSA) lecturer and four MSA students. Our proof of concept was able to reflect differences between the tongue tip movements in a native MSA speaker to those of a MSA language learner at a vocab- ulary level. This paper addresses principle methods for generating features that reflect bio-physiological features of speech and thus, facilitates an approach that can be generally adapted to languages other than MSA. 1 Introduction Second language (L2) learners have difficulties in pronouncing words as well as native speakers (Burgess and Spencer, 2000) which can create in- conveniences in social interactions (Derwing and Munro, 2005). Difficulty in providing pronunci- ation instructions by language teachers add extra challenges on L2 pronunciation training and cor- rections (Breitkreutz et al., 2001). One solution to assist pronunciation acquisition is through the adoption of educational software ap- plications (Levis, 2007). A well-designed language educational software can provide straightforward guidance to correct L2 pronunciation through mul- tiple information sources. One instance of auxil- iary systems is Pronunciation Learning Aid (PLA), which supports language students towards native- like pronunciation in a target language (Fudholi and Suominen, 2018). PLA achieves this via eval- uating students’ produced speech to reflect their pronunciation status. Another instance of auxiliary systems is visual cues, which serves as a friendly and accessible form of feedback to language stu- dents (Yoshida, 2018). Through combining language lecturers’ teaching with auxiliary systems, our aim is to assist students in both a classroom setting and in their individual practices. We present a prototype system that dis- plays visual feedback on tongue movements to as- sist language learners to acquire correct pronuncia- tion in the process of L2 studying. We have adopted a human-centred approach for the development of the system using a design-oriented perspective through applying a methodology that draws from Design Science Research (DSR) (Hevner et al., 2004) and Design Thinking (DT) (Plattner et al., 2009). Unlike machine learning methods, which train deep neural networks to predict articulatory movements (Yu et al., 2018), our proposed system uses vowel space plots based on bio-physiological features to help visualise tongue movements. In …

Content truncated to 500 words