Pronunciation Training Research Articles

Pronunciation Training Facilitates the Learning and Retention of L2 Grammatical Structures

AbstractThe present study investigated the impact of pronunciation training on the learning and retention of second language (L2) grammatical structures. By drawing on the phonological saliency of stress placement in German separable‐ and inseparable‐prefix verbs, the study explored whether enhancing explicit instruction through pronunciation training can boost learners’ accuracy with the target grammatical form. Learners’ command of separable‐ and inseparable‐prefix verbs was tested prior to and after a grammar lesson using a listening task in which participants identified whether target verbs contained a separable or inseparable prefix and a written production task in which participants described pictures using separable‐ and inseparable‐prefix verbs. Results revealed significant and sustained improvement on both the listening and written production task among the grammar + pronunciation group, which suggests that pronunciation training can facilitate the learning and retention of L2 grammatical structures over and above explicit instruction.

Development of a visual app for improving learner's pronunciation with ultrasound and the speech accent archive

Although many language learners desire to improve their pronunciation of a foreign language, there are not many apps to help them do so. Most commercial apps for pronunciation evaluation and training focus on only the acoustic signal. However, few systems give visual movements of native speakers’ lips, tongue, and jaw. In this paper, we describe the ongoing development of and app that is programmed in Swift for iPhone. The app incorporates and links together different kinds of phonetic data for the pronunciation learner—for example, recorded frontal and side videos of a native speaker’s face during pronunciation with an ultrasound movie of the tongue moving in the mouth overlaid. The training text is a paragraph from the Speech Accent Archive. The initial version of this app has two systems. First, listening to English sentences by a native speaker and checking tongue movement with ultrasound. Second, this app has buttons that play from user-chosen words and it also plays in slow motion. The method of the...

Developing ultrasound overlay videos for SENĆOŦEN language learners

This paper reports on an initiative to develop a pronunciation training tool for learners of SENCOŦEN (Salish: Canada). In acquiring new speech sounds, second language learners rely on both acoustic and articulatory information. Ultrasound has been demonstrated to be effective for pronunciation instruction, as it allows learners to visualize tongue movements during speech. However, ultrasound-based instruction can be difficult to implement with large groups or with learners wanting to learn independently. To address these limitations, we developed a technique for creating ultrasound overlay videos which combine ultrasound images of tongue movements with external profile views of a speaker’s head. Ultrasound is particularly useful in SENCOŦEN, as its inventory includes many lingual articulations that are difficult to distinguish without visual cues, such as velar/uvular contrasts, and a complex coronal series (e.g., ť, ťθ, ʧ, ʃ, ƛ, ɫ, l, ľ). SENCOŦEN is critically endangered, but efforts are underway to re...

Teaching Pronunciation with Computer Assisted Pronunciation Instruction in a Technological University

The purpose of this study is to evaluate the effectiveness of computer assisted pronunciation instruction in English pronunciation for students in vocational colleges and universities in Taiwan. The participants were fifty-one first-year undergraduate students from a technological university located in central Taiwan. The participants received an eight-week pronunciation instruction, in which the participants were presented model utterances and corresponding pitch contours of target sentences and then practiced pronunciation of target sentences with computer assisted pronunciation training software. Computerized speaking tests were conducted before, during, and after the training to measure the improvement of pronunciation quality, and a questionnaire was administered at the end of the instruction. The results of repeated measures analysis of variance on the scores of the tests indicate that the pronunciation quality of the participants was significantly improved. The results of this study provide empirical evidence in teaching pronunciation to vocational college students with computer technology.

They prepare you what to say, but do they teach you how to say it?

It cannot be denied that research on L2 pronunciation has developed greatly in the last decades, nevertheless the conclusions drawn from such studies are rarely applied in practice, especially in the school curricula. Studies carried out in Poland since early 2000s have proven that pronunciation teaching is still almost absent at schools (apart from the academic level) and that L2 learners are highly critical concerning both their own as well as their teachers’ pronunciation in EFL. Although it has been shown that correct pronunciation is crucial in L2 communication and that this element of L2 competence has been included in the oral part of the school-leaving (<em>Matura) </em>exam (CKE, 2013), the popular claim is that it is still difficult to find any elements of pronunciation training in textbooks designed for this kind of schools. The first aim of this paper is to compare the latest versions of compendia (<em>repetytoria </em>in Polish) usually applied in the last class of this type of school. The books are used to revise all the previously acquired knowledge about an L2 and are supposed to include theory and exercises in all skills and elements of a target language. And as learners graduating from upper-secondary school are about nineteen years old and, moreover, some of them may stop learning L2 intensively after graduation, it is the last chance to work seriously on English phonetics. The study results show that in most cases, although a separate pronunciation module is not included, word-lists in compendia are usually accompanied by audio tracks and, in some cases, even IPA is applied. Another aim of the study is to examine whether upper-secondary L2 teachers pay attention to their learners’ pronunciation and train it. An online questionnaire concerning pronunciation teaching practices was completed by 51 teachers from all over Poland and this did not bring optimistic results.

Large-scale characterization of non-native Mandarin Chinese spoken by speakers of European origin: Analysis on iCALL

In this work, we analyze phonetic and prosodic pronunciation patterns from iCALL, a speech corpus designed to evaluate Mandarin mispronunciations by non-native speakers of European origin and to address the lack of large-scale, non-native corpora with comprehensive annotations for applications in CAPT (computer-assisted pronunciation training). iCALL consists of 90,841 utterances from 305 speakers with a total duration of 142 hours. The speakers are from diverse linguistic backgrounds (spanning Germanic, Romance, and Slavic native languages). The read utterances are phonetically balanced with phonetic, tonal, and fluency annotations. Our findings on iCALL reveal that lexical tone errors are over six times more prevalent than phonetic errors, French speakers are twice as likely to mispronounce Tone 2, 3, 4 when compared to English speakers, native Romance language speakers are more likely to make de-aspiration and aspiration mistakes, and fluency scores correlate inversely with tone and phone error rate.

Beliefs and practices of Brazilian EFL teachers regarding pronunciation

Interest in pronunciation learning and teaching has increased significantly in the past few years. Studies and resources in the area have proliferated, but it is important to know whether they have influenced teachers of English as a foreign language (EFL) and English as a second language (ESL). The purpose of this study was to investigate the beliefs and practices of Brazilian EFL teachers. Convenience and snowball sampling were employed to recruit 60 participants, who completed an online survey on pronunciation teaching and learning. Descriptive statistics was used to analyse trends, while qualitative responses were coded for common topics. The findings suggest that the instructors had generally informed views about pronunciation and positive attitudes toward its teaching. Their teaching practices tended to be traditional: the predominant approach was to deal with word-level features, especially problematic sounds, through repetition as the need arose. Although most of the respondents claimed to be comfortable teaching pronunciation, they reported a wish for more pronunciation training, as have other instructors in prior studies (e.g. Burgess & Spencer, 2000; Foote, Holtby, & Derwing, 2011).

Verifying a holistic multimodal approach to pronunciation training of intermediate Polish learners of English

The paper is an attempt to verify empirically a holistic multimodal approach to pronunciation training developed by Szpyra-Kozlowska (2015). It reports a phonodidactic experiment carried out with two groups of Polish secondary school intermediate learners of English and demonstrates that the tested procedure is both more effective that the traditional listen-and-repeat tasks and more attractive to the participants, as shown in a post-test questionnaire study. Artykul jest probą empirycznej weryfikacji holistycznego multimodalnego podejścia do treningu fonetycznego, zaproponowanego przez Szpyre-Kozlowską (2015). Autorzy relacjonują przeprowadzony przez nich eksperyment fonodydaktyczny z udzialem dwoch grup polskich uczniow na średniozaawansowanym poziomie angielszczyzny i wykazują, ze zastosowana metoda jest zarowno bardziej efektywna niz tradycyjne cwiczenia oparte na imitacji wzorca fonetycznego, jak rowniez bardziej atrakcyjna dla uczniow, co wykazala przeprowadzona wśrod nich ankieta.

English spelling among the top priorities in pronunciation teaching: Polglish local versus global(ised) errors in the production and perception of words commonly mispronounced

This paper presents the results of a questionnaire and recording-based study on production and recognition of a sample of 60 items from Sobkowiak’s (1996:294) ‘words commonly mispronounced’ by 143 first-year BA students majoring in English. 30 lexical items in each task represent 27 categories defined by Porzuczek (2015), each referring to one aspect of English phonotactics and/or spelling-phonology relations. Our aim is to provide evidence for the occurrence of local and globalised errors in Polglish speech. This experiment is intended to examine what types of errors, that is, seriously deformed words, whether avoidable, ‘either-or’ or unavoidable ones, as classified in Porzuczek (2015), are the most frequent in production and recognition of words. Our goal is to check what patterns concerning letter-to-sound relations, are not respected in the subjects’ production and recognition of an individual word and what rules should be explicitly discussed and practised in a phonetics course. The results of the study confirm the necessity for explicit instruction on the regularity rather than irregularity of English spelling in order to eradicate globalised and ‘either-or’ pronunciation errors in the speech of students. The avoidable globalised errors which have turned out to be the most numerous in a production task include such areas of English phonotactics as: the letters <-old> and <oll>, ‘mute consonant letters’, ‘isolated errors’ and two categories related to the reduction of unstressed syllables: ‘reduce the vowel in stress-adjacent syllables and in syllables following the stressed one to /ə/ or /ɪ/’ and ‘reduce <-ous>, <-age>, and <-ate> in nouns and adjectives.’ The hope is also expressed that once introducing spelling-to-sound relations becomes a routine procedure in pronunciation training, the strain on part of the students of memorizing a list of true local errors, phonetically challenging pronunciation exceptions, will be reduced to the absolute minimum.

Perspectives upon Integrating Music into Freshman English Pronunciation Training Classroom

This study explores the effects of integrating music into English pronunciation practice. Ninety-five English major students from a technological university located in central Taiwan participated in this project. Six songs were selected and played by means of using YouTube. Students were given a list of vocabulary words selected from the lyrics and learned pertinent pronunciation rules that were augmented with phonetic symbols. They were asked to take quizzes that required memorizing the song’s lyrics. The findings point out that almost all students enjoyed listening to songs and indicated that they had become more familiar with many English songs and the rules of pronunciation by practicing singing and memorizing the lyrics.

A Two-Pass Framework of Mispronunciation Detection and Diagnosis for Computer-Aided Pronunciation Training

This paper presents a two-pass framework with discriminative acoustic modeling for mispronunciation detection and diagnoses (MD&D). The first pass of mispronunciation detection does not require explicit phonetic error pattern modeling. The framework instantiates a set of antiphones and a filler model to augment the original phone model for each canonical phone. This guarantees full coverage of all possible error patterns while maximally exploiting the phonetic information derived from the text prompt. The antiphones can be used to detect substitutions. The filler model can detect insertions, and phone skips are allowed to detect deletions. As such, there is no prior assumption on the possible error patterns that can occur. The second pass of mispronunciation diagnosis expands the detected insertions and substitutions into phone networks, and another recognition pass attempts to reveal the phonetic identities of the detected mispronunciation errors. Discriminative training (DT) is applied respectively to the acoustic models of the mispronunciation detection pass and the mispronunciation diagnosis pass. DT effectively separates the acoustic models of the canonical phones and the antiphones. Overall, with DT in both passes of MD&D, the error rate is reduced by 40.4% relative, compared with the maximum likelihood baseline. After DT, the error rates of the respective passes are also lower than those of a strong single-pass baseline with DT by 1.3% and 5.1% relative which are statistically significant.

Exploring a Phonological Process Approach to Adult Pronunciation Training.

The production of speech sound classes in adult language learners is affected by (a) interference between the native language and the target language and (b) speaker variables such as time speaking English. In this article, we demonstrate how phonological process analysis, an approach typically used in child speech, can be used to characterize adult target language phonological learning. Sentences produced by 2 adult Japanese English language learners were transcribed and coded for phoneme accuracy and analyzed according to the percent occurrence of phonological processes. The results were interpreted relative to a contrastive analysis between Japanese and English phonetic inventories and developmental norms for monolingual English children. In this pilot study, common consonant processes included vocalization, final consonant devoicing, and cluster reduction. These are processes commonly observed in the speech of children who are typically developing. The process analysis can inform clinical approaches to pronunciation training in adult English language learners. For example, the cycles approach (Hodson & Paden, 1981) may provide more clinical efficacy than an articulatory approach in which phonemes are targeted individually. In addition, a process analysis can enable clinicians to examine the principles of within-class and across-class generalization in adult pronunciation instruction.

A Longitudinal View of Look at the Effectiveness of Elicited Imitation with Computer Assisted Pronunciation Training (CAPT)

<p>It is the aim for most EFL learners to have a clear pronunciation in English; however, pronunciation training has always been a challenge for EFL teachers and learners alike. The advent of Computer Assisted Pronunciation Training (CAPT) seems to be the solution to this problem. This longitudinal study adopted a latent growth curve to describe 30 EFL learner’s (N = 30) development of imitated pronunciation in English as well as their satisfaction regarding CAPT. These 30 participants were randomly assigned to the experimental group (N = 18) and control group (N = 12). There were no significant differences among the participants in their pronunciation of English or attitudes toward CAPT at the beginning of this study. After 16 months of training, the growth curve analysis showed that participants in the CAPT group made significantly greater development in English pronunciation, and they were more satisfied with the CAPT. The large effect size indicated that the grouping was the major effect that led to a significant difference in satisfaction. Moreover, the qualitative data derived from focus-group interviews confirmed the benefits of CAPT in participant’s pronunciation training. These findings suggest that EFL learner’s opportunities to conduct elicited imitation with the help of ASR-CAPT will be helpful for their development of imitated pronunciation in English and increase their satisfaction of it, which is a key aspect in the effectiveness of pronunciation training. </p>

Pronunciation learner autonomy: The potential of Automatic Speech Recognition

Abstract In pronunciation learning, students are often hampered in their attempts to study or practice autonomously by their limited abilities to monitor their speech for errors. Automatic Speech Recognition (ASR) has great potential for providing feedback, allowing students to become more autonomous pronunciation learners. This study examined the effect of ASR use as part of a three-week pronunciation workshop on students' autonomous learning beliefs and behaviors. The study utilized three groups: 1) CONV: conventional face-to-face pronunciation training workshop (n = 15), 2) STRAT: mostly conventional with minimal ASR strategy training (n = 17), and 3) HYBRID: hybrid with half of workshop time using ASR (n = 16). Changes in beliefs and behaviors were tracked using pre-, post-, and delayed post-workshop surveys, along with interviews and weekly learning logs. Results showed that while CONV reported no significant change, groups introduced to ASR, STRAT and HYBRID, significantly increased their beliefs of autonomy from the pre- to post-workshop survey and pointed to the feedback from ASR as enabling them to practice autonomously. However, after the workshop ended, HYBRID reported significantly more time spent on autonomous pronunciation learning and more use of ASR than STRAT and CONV, highlighting the need for a gradualist approach to autonomy through repeated practice with ASR.

発話時の顔面筋の筋電計測

This study proposes the hypothesis that we use different facial muscles in the utterance of English/Japanese words. In order to experimentally prove this hypothesis, we measured and compared the electromyography signals of the facial muscles. The experiments include the comparison of English/Japanese words, the effect of pronunciation training, and the comparison of vowels with similar sound. The experimental results show that the used muscles are different depending on the language, and voice training with teacher voice are effective to activate muscle usage that are not usually used so far. This result indicates possibility to provide foreign language learners about the guide function of foreign language voice training.

Pronunciation training of Japanese prosody: Progress in a development of CALL system

Pronunciation training of Japanese prosody: Progress in a development of CALL system

English Pronunciation Training and Custom-Made Speaking Site for Training and Testing

English Pronunciation Training and Custom-Made Speaking Site for Training and Testing

Dynamic time warping in phoneme modeling for fast pronunciation error detection

The presented paper describes a novel approach to the detection of pronunciation errors. It makes use of the modeling of well-pronounced and mispronounced phonemes by means of the Dynamic Time Warping (DTW) algorithm. Four approaches that make use of the DTW phoneme modeling were developed to detect pronunciation errors: Variations of the Word Structure (VoWS), Normalized Phoneme Distances Thresholding (NPDT), Furthest Segment Search (FSS) and Normalized Furthest Segment Search (NFSS). The performance evaluation of each module was carried out using a speech database of correctly and incorrectly pronounced words in the Polish language, with up to 10 patterns of every trained word from a set of 12 words having different phonetic structures. The performance of DTW modeling was compared to Hidden Markov Models (HMM) that were used for the same four approaches (VoWS, NPDT, FSS, NFSS). The average error rate (AER) was the lowest for DTW with NPDT (AER=0.287) and scored better than HMM with FSS (AER=0.473), which was the best result for HMM. The DTW modeling was faster than HMM for all four approaches. This technique can be used for computer-assisted pronunciation training systems that can work with a relatively small training speech corpus (less than 20 patterns per word) to support speech therapy at home.

Speaker-Adaptive Acoustic-Articulatory Inversion Using Cascaded Gaussian Mixture Regression

This paper addresses the adaptation of an acoustic-articulatory model of a reference speaker to the voice of another speaker, using a limited amount of audio-only data. In the context of pronunciation training, a virtual talking head displaying the internal speech articulators (e.g., the tongue) could be automatically animated by means of such a model using only the speaker’s voice. In this study, the articulatory-acoustic relationship of the reference speaker is modeled by a gaussian mixture model (GMM). To address the speaker adaptation problem, we propose a new framework called cascaded Gaussian mixture regression (C-GMR), and derive two implementations. The first one, referred to as Split-C-GMR, is a straightforward chaining of two distinct GMRs: one mapping the acoustic features of the source speaker into the acoustic space of the reference speaker, and the other estimating the articulatory trajectories with the reference model. In the second implementation, referred to as Integrated-C-GMR, the two mapping steps are tied together in a single probabilistic model. For this latter model, we present the full derivation of the exact EM training algorithm, that explicitly exploits the missing data methodology of machine learning. Other adaptation schemes based on maximum-a posteriori (MAP), maximum likelihood linear regression (MLLR) and direct cross-speaker acoustic-to-articulatory GMR are also investigated. Experiments conducted on two speakers for different amount of adaptation data show the interest of the proposed C-GMR techniques.

Visual Feedback of Tongue Movement for Novel Speech Sound Learning.

Pronunciation training studies have yielded important information concerning the processing of audiovisual (AV) information. Second language (L2) learners show increased reliance on bottom-up, multimodal input for speech perception (compared to monolingual individuals). However, little is known about the role of viewing one's own speech articulation processes during speech training. The current study investigated whether real-time, visual feedback for tongue movement can improve a speaker's learning of non-native speech sounds. An interactive 3D tongue visualization system based on electromagnetic articulography (EMA) was used in a speech training experiment. Native speakers of American English produced a novel speech sound (/ɖ/; a voiced, coronal, palatal stop) before, during, and after trials in which they viewed their own speech movements using the 3D model. Talkers' productions were evaluated using kinematic (tongue-tip spatial positioning) and acoustic (burst spectra) measures. The results indicated a rapid gain in accuracy associated with visual feedback training. The findings are discussed with respect to neural models for multimodal speech processing.