This paper is organized around two main endeavours. On the one hand, we examine currently available Computer Assisted Pronunciation Training (CAPT) systems with a view to establishing whether they meet pedagogically sound requirements. In this respect, we show that many commercial systems tend to prefer technological novelties to the detriment of pedagogical criteria that could benefit the learner more. On the other hand, we more narrowly focus on the crucial issue of computer-generated feedback, which still represents a big challenge for state-of-the-art CAPT technology and discuss its impact on learning. In the final part of the paper, we present the PROO project (Programma voor Onderwijsonderzoek), which is aimed at establishing the effects of erroneous feedback on the acquisition of L2 pronunciation.

Low-proficiency non-native speakers represent a significant challenge for large-vocabulary continuous speech recognition (LVCSR). Acoustic models are confused by a heavy accent

In this paper, we present an acoustic-phonetic approach to automatic pronunciation error detection. Classifiers using

Pronunciation variability is present in both native and foreign words. Since pronunciation variability constitutes a problem for automatic speech recognition (ASR) systems, modeling pronunciation variation for ASR has been the topic of various studies. In most studies, modeling pronunciation variation was attempted within the standard framework used in mainstream ASR systems. Given that some assumptions made within this framework are not in line with the properties of speech signals and the findings in human speech recognition, and that the improvements obtained by modeling pronunciation variation within this framework have generally been small, it might be better to look for a new paradigm in which pronunciation variation can be modeled more accurately. In this paper a novel paradigm for ASR is presented, which has many potential advantages for modeling pronunciation variation.

In this paper we report on a study that was carried out to obtain an inventory of segmental errors in the Dutch of adult learners with different mother tongues (L1s). The errors observed were subsequently examined in detail to select a number of errors that should receive priority in Computer Assisted Pronunciation Training (CAPT) for Dutch as L2.

Prosody—suprasegmental characteristics of speech such as pitch, rhythm, and loudness— is a rich source of information in spoken language and can tell a listener much about the internal state of a speaker. This thesis explores the role of prosody in conveying three very different types of speaker state: paralinguistic state, in particular emotion; pragmatic state, in particular questioning; and the state of spoken language proficiency of non-native English speakers. Paralinguistics. Intonational features describing pitch contour shape were found to dis-criminate emotion in terms of positive and negative affect. A procedure is described for clustering groups of listeners according to perceptual emotion ratings that foster further understanding of the relationship between acoustic-prosodic cues and emotion perception. Pragmatics. Student questions in a corpus of one-on-one tutorial dialogs were found to be signaled primarily by phrase-final rising intonation, an important cue used in conjunc-tion with lexico-pragmatic cues to differentiate the high rate of observed declarative questions from proper declaratives. The automatic classification of question form and

to spontaneous speech too, if it turned out to be feasible for read This paper describes an experiment aimed at determining whether speech. automatic assessment of second language learners ’ fluency in This experiment produced interesting results in two respects, spontaneous speech is feasible and whether it differs from a. fluency assessment by expert raters and b. the relationship automatic fluency assessment in read speech. Spontaneous speech between expert fluency ratings and automatically obtained objective of 60 learners of Dutch was scored for fluency by five raters and fluency measures. With regard to a., the results showed that expert was analyzed by means of a continuous speech recognizer to ratings of fluency in read speech are reliable (Cronbachs ’ varies calculate seven quantitative measures of speech quality known to be between.90 and.96). With respect to b., very high correlations related to perceived fluency. The results show that automatic were found between the expert fluency ratings and the automatic assessment of second language learners ’ fluency in spontaneous measures of fluency: five automatic measures showed correlations speech is feasible, although not all variables suitable for measuring with the fluency scores whose magnitude varied between.77 and fluency in read speech are as effective in spontaneous speech. In.91. The highest correlations were found for rate of speech (between particular, measures that express the rate at which sounds are.86 and.91). Further analyses revealed that two factors are produced without taking pauses into account appear to be important for perceived fluency in read speech: the rate at which unsuitable for measuring fluency in spontaneous speech. speakers articulate the sounds and the number of pauses they make. Furthermore, the correlations between machine scores and human Rate of speech appears to be such a good predictor of perceived ratings are lower for spontaneous speech. Possible explanations are fluency because it incorporates these two aspects. discussed. Since automatic assessment of fluency in read speech turned out to be feasible, we decided to extend this approach to 1.

Prosody, or intonation, is a critically important component of spoken communication. The automatic extraction of prosodic information is necessary for machines to process speech with human levels of proficiency. In this thesis we describe work on the automatic detection and classification of prosodic events – specifically, pitch accents and prosodic phrase boundaries. We present novel techniques, feature representations and state of the art performance in each of these tasks. We also present three proof-of-concept applications – speech summarization, story segmentation and non-native speech assessment – showing that access to hypothesized prosodic event information can be used to improve the performance of downstream spoken language processing tasks. We believe the contributions of this thesis advance the understanding of prosodic events and the use of prosody in spoken language processing towards the goal of human-like processing of speech by machines.

Language Students ’ Pronunciation?

The teaching of a language is, to a large part, still based on written language rather than on the spoken language. Thus, very often, questions of speaking a language plays the role of "cinderella " in language teaching. However, there is not much disagreement that the most natural way to communicate with one other is through spoken communication. In consequence there is a