Field Evaluation of a Single-Word Pronunciation Training System

Many learning tasks require substantial skills training. Ideally, the student might benefit the most from having a human expert – a teacher or trainer – at hand throughout, but human expertise remains a scarce resource. The second-best solution could be to do skills training with a computer-based self-training system. This vision of the computer as tutor currently motivates increasing efforts worldwide, in all manner of fields, including that of computer-assisted language learning, or CALL. But, as pointed out by Hincks [2003], along with the growth of the CALL area comes a growing need for empirical evidence that CALL systems have a beneficial effect. This point is reiterated by Chapelle [2002] who defines the goal for Computer Assisted Second Language Research as the gathering of evidence for the effect of CALL and instructional design. This paper presents results of a field test of our pronunciation training system which enables immigrants and others to self-train their pronunciation skills of single Danish words. 1. The Pronunciation Trainer The prototype of the Danish Pronunciation Trainer (DPT) combines a graphical user interface and an Automatic Speech Recogniser. The main window is shown in Figure 1.1. Figure 1.1. Main window of the Pronunciation Trainer. Our aim has been to create an interface which is as simple as possible and intuitively usable by the learner. The window displays the Danish training word, sometimes a simplified ‘phonetic’ transcription using regular orthography, and a translation into English. Ordinary users are not proficient in standard phonetic representation, but we have found that they are helped by simplified phonetic transcription when training to pronounce Danish words whose pronunciation is irregular. This help is often required, Danish being phonetically highly irregular. What the simplified phonetic transcription does is help the learner avoid endlessly following wrong pronunciation intuitions without getting the pronunciation right. The English translation, we have found, is appreciated because (i) most Danish trainees speak English already and (ii) appreciate knowing the meaning of the word they train to pronounce. Given (i), the interface language of the main window is English. Having clicked the ‘Next word’ button and (i) read the written word, its translation and (possibly) its simplified phonetic transcription, the learner may choose to pronounce the word by clicking the ‘Speak’ button. Alternatively, the learner may (ii) click the ‘Play audio’ button to listen to a native Danish speaker pronouncing the word, or (iii) both hear and see the word being pronounced by a native Danish speaker on video by clicking the ‘Play video’ button. The training sound files used in (ii) and (iii) are identical. Having pronounced the word, the learner receives feedback on pronunciation quality. A scoring system was devised for providing immediate feedback to the learner, ranging from 0 to 2, 2 being the highest and 0 the lowest. The learner is given both the numerical score and a graphical indication of success or failure, as illustrated in Figure 1.2. Figure 1.2. Achieving the maximum pronunciation score. The scoring system works as follows. Typically, the speech recogniser returns an ordered multi-word n-best list of words in its vocabulary which best match the acoustic signal produced by the learner, with the best acoustic match on top. If the word to be pronounced is on top of the n-best list, the learner receives the top score and only the top scoring word is presented in the feedback to the learner as shown in Figure 1.2. If the word to be pronounced is not on top of the n-best list but is somewhere else in the list, the learner receives the score 1 and the entire n-best list is shown in the feedback. The feedback informs the learner with which other words in