Academic journal article Educational Technology & Society

A Multimedia English Learning System Using HMMs to Improve Phonemic Awareness for English Learning

Academic journal article Educational Technology & Society

A Multimedia English Learning System Using HMMs to Improve Phonemic Awareness for English Learning

Article excerpt

Introduction

Phonemic awareness is an important metalinguistic skill which can let students more effectively acquire reading and spelling abilities (Mehta, Foorman, Branum, & Taylor, 2005). While children learn English, an important step is to train them with high phonemic awareness (Leong, Tan, Cheng, & Han, 2005; Goswami & Bryant, 1990). Carreker (2005) stated that phonemic awareness training helps remediate the problems of poor spelling at any age. Learners, who possess high capability of phonemic awareness, have better capabilities in pronunciation-recognition, reading, and spelling (Treiman & Baron, 1983). How to promote learner's phonemic awareness during teaching English has become an essential subject in lecture hour. Also, s/he has more opportunities than others to effectively promote her/his phonemic awareness so as to shorten the learning time of reading and spelling. During teaching pronounces in classes, most teachers in Taiwan often concentrate on teaching learners speech skills. Moreover, they neglect learners' fostering of the recognition capability of phonemic voice. This leads to the fact that the learners cannot have high pronunciation recognition ability. Therefore, the learners are unable to clearly compare their pronunciation differences with correct ones. It raises the problem of inaccurate pronunciation, late speech development, and low letter knowledge (Mann & Foy, 2007).

While learning English in classrooms, teachers often teach students English pronunciation without computer aids. In recent years, due to the growing advancement of information technologies, large amount of multimedia English learning materials have been developed to enhance the learning performance of English pronunciation (Hincks, 2003). The technology of speech analysis has been used for teaching intonational patterns since 1970s (Zinovjeva, 2005). Therefore, speech analysis has been incorporated in much commercial software for English pronunciation. However, the software is still insufficient in offering the feedback to learners for the analysis results of incorrect pronunciations. Thus, the computer assisted language learning (CALL) systems have been successfully developed to improve those limitations such that traditional CALL systems not only perform speech recognizers but also offer the language learning activity and feedback (Wachowicz & Scott, 1999). Moreover, Precoda, Halverson, and Franco (2000) presented a result that the user interface of the CALL system is designed to give pronunciation feedbacks for the learners' pronunciation ability, and a conclusion that the design of the user interface plays an important role to attract learners' attention.

Currently, speech synthesizers and digitally manipulated stimuli have been developed in laboratory studies (Neumeyer, Franco, Digalakis, & Weintraub, 2000). Unfortunately, they are not widely utilized in the design of CALL systems. As a result, linguistic experts, including the Second Language (L2) teachers, exclude to take those software products as tools to teach English pronunciation (Zinovjeva, 2005). The reason is that those systems just provide learners with speech synthesizers. They can not offer learners with learning feedback and high quality of voice. Therefore, our research devises a multimedia English Learning (MEL) system to overcome above two limits, not providing learning feedback and using speech synthesizers.

The Hidden Markov Model (HMM)-based automatic speech recognition (ASR) system has been successfully applied to dictate speech tasks (Nock & Ostendorf, 2003). HMM provides a framework which is broadly used for modeling speech patterns. The hidden Markov model (HMM) is the most commonly employed model for speech recognition. Speech recognition technology based on the HMM using for word spotting and speech recognition, has improved significantly during the past few decades (Liu et al., 2006). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.