Search

Logopedisten en klinisch linguïsten onderzoeken de taalontwikkeling van jonge kinderen met een vermoedelijke of al vastgestelde taalontwikkelingsstoornis. Ze onderzoeken ook de taalvaardigheid van personen met afasie. Naast gestandaardiseerde tests worden samples spontane taal geanalyseerd. Hiervoor worden uitingen ontlokt aan de patiënten via vaste protocollen. De sessies worden opgenomen, getranscribeerd en vervolgens grammaticaal geanalyseerd. Bij de grammaticale analyse wordt bepaald welke soorten constructies en fouten voorkomen en in welke mate, en dit wordt vergeleken met een norm. Taal- en spraaktechnologie (TST) kan er in prin-cipe aan bijdragen om het proces van transcriptie en grammaticale analyse efficiënter te maken en mogelijk zelfs om de kwaliteit van de assessments te verhogen. In dit artikel richten we ons op de mogelijkheden van TST voor de analyse van kindertaal.

Logopedisten en klinisch linguïsten onderzoeken de taalontwikkeling van jonge kinderen met een vermoedelijke of al vastgestelde taalontwikkelingsstoornis. Ze onderzoeken ook de taalvaardigheid van personen met afasie. Naast gestandaardiseerde tests worden samples spontane taal geanalyseerd. Hiervoor worden uitingen ontlokt aan de patiënten via vaste protocollen. De sessies worden opgenomen, getranscribeerd en vervolgens grammaticaal geanalyseerd. Bij de grammaticale analyse wordt bepaald welke soorten constructies en fouten voorkomen en in welke mate, en dit wordt vergeleken met een norm. Taal- en spraaktechnologie (TST) kan er in prin-cipe aan bijdragen om het proces van transcriptie en grammaticale analyse efficiënter te maken en mogelijk zelfs om de kwaliteit van de assessments te verhogen. In dit artikel richten we ons op de mogelijkheden van TST voor de analyse van kindertaal.

In this paper, we discuss acoustic backing-off as a method to improve automatic speech recognition robustness. Acoustic backing-off aims to achieve the same objective as the marginalization approach of missing feature theory: the detrimental influence of outlier values is effectively removed from the local distance computation in the Viterbi algorithm. The proposed method is based on one of the principles of robust statistical pattern matching: during recognition the local distance function (LDF) is modeled using a mixture of the distribution observed during training and a distribution describing observations not previously seen. In order to assess the effectiveness of the new method, we used artificial distortions of the acoustic vectors in connected digit recognition over telephone lines. We found that acoustic backing-off is capable of restoring recognition performance almost to the level observed for the undisturbed features, even in cases where a conventional LDF completely fails. These results show that recognition robustness can be improved using a marginalization approach, where making the distinction between reliable and corrupted feature values is wired into the recognition process. In addition, the results show that application of acoustic backing-off is not limited to feature representations based on filter bank outputs. Finally, the results indicate that acoustic backing-off is much less effective when local distortions are smeared over all vector elements. Therefore, the acoustic pre-processing steps should be chosen with care, so that the dispersion of distortions over all acoustic vector elements as a result of within-vector feature transformations is minimal.

Item does not contain fulltext 7 p.

Contains fulltext : 76237.pdf (Publisher’s version ) (Open Access) 16 p.

Contains fulltext : 168366.pdf (Publisher’s version ) (Open Access) 13 p.

The ultimate aim of the research reported on here is to develop an automatic testing system for Dutch pronunciation. In the experiment described in this paper automatic scores of telephone speech produced by native and non-native speakers of Dutch are compared with specific, i.e., temporal and segmental, and global pronunciation ratings assigned by three groups of experts: three phoneticians and two groups of three speech therapists. The goals of this experiment are to determinutee (1) whether specific expert ratings of pronunciation quality contribute to our understanding of the relation between human pronunciation scores and machine scores of speech quality, (2) whether different expert groups assign essentially different ratings, and (3) to what extent rater pronunciation scores can be predicted on the basis of automatic scores. The results show that collecting specific ratings along with overall ones leads to a better understanding of the relation between human and automatic pronunciation assessment. Furthermore, after normalization no considerable differences are observed between the ratings by the three expert groups. Finally, it appears that the speech quality scores produced by our speech recognizer can predict expert pronunciation ratings with a high degree of accuracy.

This article describes how the performance of a Dutch continuous speech recognizer was improved by modeling pronunciation variation. We propose a general procedure for modeling pronunciation variation. In short, it consists of adding pronunciation variants to the lexicon, retraining phone models and using language models to which the pronunciation variants have been added. First, within-word pronunciation variants were generated by applying a set of five optional phonological rules to the words in the baseline lexicon. Next, a limited number of cross-word processes were modeled, using two different methods. In the first approach, cross-word processes were modeled by directly adding the cross-word variants to the lexicon, and in the second approach this was done by using multi-words. Finally, the combination of the within-word method with the two cross-word methods was tested. The word error rate (WER) measured for the baseline system was 12.75%. Compared to the baseline, a small but statistically significant improvement of 0.68% in WER was measured for the within-word method, whereas both cross-word methods in isolation led to small, non-significant improvements. The combination of the within-word method and cross-word method 2 led to the best result: an absolute improvement of 1.12% in WER was found compared to the baseline, which is a relative improvement of 8.8% in WER. Dieser Artikel beschreibt, wie die Leistung eines automatischen Spracherkenners, der niederlandische gesprochene Sprache erkennt, mit Hilfe der Modellierung von Aussprachevarianten verbessert wurde. Fur diese Modellformung wird eine allgemeine Prozedur vorgeschlagen, die - kurz gesagt - darin besteht, dem Lexikon Aussprachevarianten hinzuzufugen, die Phonmodelle erneut einer Lernphase zu unterziehen und Sprachmodelle dabei zu verwenden, in denen die Aussprachevarianten mithineinbezogen wurden. Durch Anwendung einer Gruppe von funf optionalen phonologischen Regeln wurden im Basislexikon zunachst Aussprachevarianten innerhalb von Wortern generiert. Dann wurde mit Hilfe zweier Methoden eine begrenzte Anzahl von Sandhiprozessen (Prozesse auf Wordgrenzen) modelliert. Die erste bestand darin, die Sandhivarianten direkt dem Lexikon hinzuzufugen und bei der zweiten wurden Multiworter gebraucht. Letztendlich wurden die wortinternen Ausprachevarianten mit den zwei Sandhivarianten kombiniert getestet. Die Basisleistung des Spracherkenners, d.h. ohne Anwendung des Modells der Aussprachevariation, betrug 12.75% "word error rate" (WER). Bei Anwendung der wortinternen Aussprachevarianten wurde eine geringe, aber statistisch signifikante Verbesserung von 0.68% WER gemessen. Die Anwendung der zwei Sandhimodelle hingegen ergab einen sehr kleinen, nicht signifikanten Verbesserung. Die Kombination des wortinternen Modells mit dem zweiten Sandhimodell hingegen ergab schlie?slich das beste Ergebnis: eine absolute Verbesserung von 1.12% WER, was einer relativen Verbesserung von 8.8% WER entspricht. Cet article decrit comment les performances d'un reconnaisseur de parole continue (CSR) pour le neerlandais ont ete ameliorees en modelant la variation de prononciation. Nous proposons une procedure generale pour modeler cette variation. En bref, elle consiste a ajouter des variantes de prononciation au lexique et dans le re-apprentissage des modeles de phones en utilisant des modeles de langage auxquels les variantes de prononciation ont ete ajoutees. D'abord, des variantes de prononciation a l'interieur de mot ont ete produites en appliquant un ensemble de cinq regles phonologiques optionnelles aux mots dans le lexique de base. Ensuite, un nombre limite de processus entre-mots ont ete modeles, en utilisant deux methodes differentes. Dans la premiere approche, des processus entre-mots ont ete modeles en ajoutant directement les variantes "entre-mots" au lexique, et dans la deuxieme approche ceci a ete fait en utilisant des "mots-multiples". En conclusion, la combinaison de la methode qui se limite aux processus a l'interieur de mot avec les deux methodes "entre-mots" a ete testee. La performance de base etait un taux d'erreur de 12.75% mots (WER); comparee a cette performance de base, une amelioration petite mais significative de 0.68% dans WER a ete obtenue avec la methode 'a l'interieur de mot', tandis que les deux methodes d'entre-mots en isolation ont mene a des petites ameliorations non significatives. La combinaison de la methode "a l'interieur de mot" avec la methode 2 "entre-mots" a mene au meilleur resultat: une amelioration absolue de 1.12% dans le WER a ete trouvee comparee a la ligne de base, qui est une amelioration relative de 8.8% dans le WER.

The focus in automatic speech recognition (ASR) research has gradually shifted from isolated words to conversational speech. Consequently, the amount of pronunciation variation present in the speech under study has gradually increased. Pronunciation variation will deteriorate the performance of an ASR system if it is not well accounted for. This is probably the main reason why research on modeling pronunciation variation for ASR has increased lately. In this contribution, we provide an overview of the publications on this topic, paying particular attention to the papers in this special issue and the papers presented at 'the Rolduc workshop'.11Whenever we mention 'the Rolduc workshop' in the text we refer to the ESCA Tutorial and Research Workshop "Modeling pronunciation variation for ASR" that was held in Rolduc from 4 to 6 May 1998. This special issue of Speech Communication contains a selection of papers presented at that workshop. First, the most important characteristics that distinguish the various studies on pronunciation variation modeling are discussed. Subsequently, the issues of evaluation and comparison are addressed. Particular attention is paid to some of the most important factors that make it difficult to compare the different methods in an objective way. Finally, some conclusions are drawn as to the importance of objective evaluation and the way in which it could be carried out. Die Forschungsrichtung der automatischen Spracherkennung (ASR) hat sich nach und nach vom Erkennen isolierter Worter in Richtung Erkennung frei gesprochener Sprache entwickelt. Das hat zur Folge, da?s die Aussprachevariation, so wie sie in der freien Rede zutage tritt, bei der Spracherkennung ein intervenierender Faktor geworden ist. Die Leistung eines ASR-Systems wird namlich erheblich beeintrachtigt, wenn man diesen Faktor nicht berucksichtigt. Dies ist vermutlich der Hauptgrund dafur, warum die systematische Berucksichtigung der Aussprachevariation bei der ASR in letzter Zeit stark zugenommen hat. Dieser Artikel stellt einen Uberblick der Literatur zu diesem Thema dar, wobei den Beitragen in diesem 'special issue' sowie denen des 'Rolduc workshop' besondere Aufmerksamkeit geschenkt wird. Zunachst werden die wichtigsten Unterschiede der zahlreichen Arbeiten zur Modellbildung der Aussprachevariation diskutiert. Dann folgt eine Besprechung der Beurteilung und des Vergleichs verschiedener Methoden, die der Modellbildung zugrunde liegen. Dabei wird den wichtigsten Faktoren, die einen objektiven Vergleich der Methoden erschweren, besondere Aufmerksamkeit geschenkt. Letztendlich schlie?sen sich einige Schlu?sfolgerungen im Hinblick auf die Relevanz objektiver Beurteilung und deren mogliche Realisierung an. Le centre d'interet dans la recherche de la reconnaissance automatique de la parole (ASR), parti des mots isoles, s'est engage vers le discours conversationnel. Par consequence, la quantite de variation de prononciation presente dans le discours dont nous rapportons les resultats a graduellement augmente. La variation de prononciation deteriorera la performance d'un systeme ASR si l'on n'en rend pas compte. C'est probablement la raison principale pourquoi la recherche dans le domaine de la modelisation de la variation de prononciation pour ASR a augmente recemment. Dans cette contribution on fournit une vue d'ensemble des publications sur ce sujet, et en particulier on refere aux articles de cette edition speciale et aux contributions presentees dans les sessions qui ont eu lieu a 'Rolduc'. D'abord, les caracteristiques les plus importantes qui distinguent les diverses etudes sur modelisation de variation de prononciation sont discutees. Puis les questions d'evaluation et de comparaison sont adressees. Une attention particuliere est pretee a certains des facteurs les plus importants qui rendent difficile de comparer les differentes methodes d'une maniere objective. Enfin quelques conclusions sont tirees quant a l'importance de l'evaluation objective et de la facon dans laquelle elle pourrait etre effectuee.

The automatic parametrization of the first derivative of glottal flow is studied. Representatives of the two types of methods used most often for parametrization were tested and compared. The chosen representatives are all based on the Liljencrants–Fant model. As numerous tests were needed for a detailed comparison of the methods, a novel evaluation procedure is used which consists of the following stages: (1) use the Liljencrants–Fant model to generate synthetic flow pulses; (2) estimate voice source parameters for these synthetic flow pulses; and (3) calculate the errors by comparing the estimated values with the input values of the parameters. This evaluation procedure revealed that in order to reduce the average error in the estimated voice source parameters, the estimation methods should be able to estimate noninteger values of these parameters. The proposed evaluation method was also used to study the influence of low-pass filtering on the estimated voice source parameters. It turned out that low-pass filtering causes an error in all estimated voice source parameters. On average, the smallest errors were found for a parametrization method in which a voice source model is fitted to the flow derivative, and in which the voice source model is low-pass filtered with the same filter as the flow derivative.

This study presents a new frequency domain parameter, Parabolic Spectral Parameter (PSP), for the quantification of the glottal volume velocity waveform. PSP is based on fitting a parabolic function to the low-frequency part of a pitch-synchronously computed spectrum of the estimated glottal flow. PSP gives a single numerical value that describes how the spectral decay of an obtained glottal flow behaves with respect to a theoretical limit corresponding to maximal spectral decay. By analyzing speech signals of different phonation types the performance of the new parameter is compared to three commonly used time-based parameters and to one previously developed frequency domain method.

Five perception experiments were conducted that investigated how the perceived prominence of F0 maxima in accented syllables in Dutch is affected by the variation of F0 minima that is supposed to relate to variation in global pitch range. The purpose of the first two experiments was to test the predictions of a model in which the reference line is directly given by an interpolation between observable F0 minima. The results showed that the model was inadequate, and confirmed earlier research suggesting that the reference line is calculated in a less direct way. The next three experiments investigated the role of the F0 of the unaccented portions of speech at the beginning ("onset") and at the end ("offset") of the contour, and show that only the (low) onset is used to calibrate the reference line. The results also suggest that longer onsets affect the abstract reference more than do shorter onsets.

Item does not contain fulltext 16 p.

Contains fulltext : 76261.pdf (Publisher’s version ) (Open Access)

Item does not contain fulltext

In recent years it has become common practice among speech researchers to report transcription agreement coefficients, when describing findings based on phonetic transcription. Although such coefficients are intended to give an indication of the degree of transcription accuracy, in reality it is not clear to what extent high agreement coefficients do indeed guarantee great transcription accuracy. In this paper it is argued that the most commonly used agreement coefficient, percentage agreement, has three major disadvantages when applied to phonetic transcription: (a) it is based on the assumption that agreement between transcription symbols is all-or-none, (b) it is strongly influenced by chance agreement, and (c) it gives no account of the criteria used for transcription alignment. After a detailed discussion of these drawbacks, an alternative approach to calculating transcription (dis)agreement is proposed. In this approach experimentally derived feature matrices are used as input to a program that alig...

Contains fulltext : 74981.pdf (Publisher’s version ) (Open Access)

Item does not contain fulltext

Contains fulltext : 76257.pdf (Publisher’s version ) (Open Access)

A rating instrument is described that can be used to assess the results of stuttering treatments. The instrument is designed for use with naive listeners. It yields a comprehensive and detailed description of the speech quality in terms of articulation, phonation, pitch, and loudness; in addition, it includes a naturalness scale. Analysis of ratings obtained with the instrument show that naturalness is a multidimensional characteristic. Moreover, the speech characteristics that determine the naturalness ratings appear to be different pretreatment, posttreatment, and at follow-up treatment. The psychometric characteristics of the instrument are analyzed in detail. It is concluded that mixing of samples of stutterers and nonstutterers in one rating experiment may artificially inflate the reliability of the ratings. Also, ratings on equal-appearing interval scales cannot be interpreted in an absolute sense. Solutions for this methodological problem are suggested.

The fact that the oral flow of both males and females normallycontains an appreciable do component during the “closed glottis” interval of vowel sounds produced at normal loudness levels indicates that glottal leakage is a very common phenomenon. In this paper the acoustic consequences of glottal leakage are studied by means of a computer simulation. The effects of two different types of leaks were studied, i.e., of (a) a linked leak: an opening (at least partly) situated in the membranous glottis and caused by abduction, and (b) a parallel chink: a leak that can be viewed as an opening which is essentially separated from (parallel to) the time-varying part of the glottis. The results of our simulations show that a moderate leak may give rise to appreciable source-tract interaction which becomes most apparent for a parallel chink. In the time domain it manifests itself as a ripple in the glottal flow waveform just after closure. In the frequency domain, the spectrum of the flow through a glottis with a leak (both linked and parallel) is characterized by zeros at the formant frequencies. The major difference in spectral effects of a linked leak and a parallel chink is the spectral slope. For a parallel chink it is of the same order of magnitude as in the no leakage case or even slightly flatter. In the case of a linked leak, the spectral slope falls off much more rapidly. These findings suggest that the amount of do flow alone is not a very good measure to use for voice efficiency measures.

The perceptual cross-over point between a Dutch ‘‘flat-hat ’’ contour with an early final fall and one with a late final fall is sensitive to the segmental structure of the syllable accented by the fall. The findings are interpreted in terms of segmental influences on the alignment of accented targets (H*): voiced segments in the syllable onset and coda pull it leftwards and rightwards, respectively, while voiceless segments in the onset also exert a leftward pull. The P-centre as defined in the work of Pompino-Marschall does not explain the variation observed. It is concluded that phonetic implementation rules align the intonational targets with reference to the beginning of the syllable, but additionally allow the voiced segments in the onset and rhyme to centre the location to a certain extent within the voiced stretch.

Item does not contain fulltext 14 p.

Item does not contain fulltext

Item does not contain fulltext 3 p.

Item does not contain fulltext

Item does not contain fulltext J. Svartvik Directions in Corpus linguistics. Proceedings of the Nobel Symposium 82, Stockholm, 4-8 August 1991 : ,1996

Item does not contain fulltext 111 p.

The paper describes the development of a fully automatic multimodal information system for the consumer market. The system will be able to provide information on a large number of topics via a single telephone number. The eventual system will integrate interactive voice response, speech recognition, speaker verification, direct dial in, calling line identification, facsimile and electronic mail. The present version is limited to DTMF input and voice and facsimile output. The architecture of the system described in the paper allows successive addition of other technologies. >

The ultimate aim of the research reported on in this paper is to develop a system for automatic assessment of foreign speakers' pronunciation of Dutch. The aim of the experiment was to determine whether pronunciation ratings assigned by human experts could be predicted on the basis of scores calculated by an automatic speech recognizer. To this end 20 native and 60 non-native speakers of Dutch read ten phonetically rich sentences over the telephone. The automatic speech recognizer was trained with read speech of 4019 Dutch subjects with varying regional accents. The results show that the human scores can be accurately predicted, even in the case of telephone speech. Analysis of the various types of human ratings and automatic measures provides more insight into the relationship between human and machine scores and indicates how the automatic measures can be further improved to achieve even greater predictive power.

Contains fulltext : 75051.pdf (author's version ) (Open Access) 4 p.

Item does not contain fulltext 3 p.

Item does not contain fulltext 4 p.

Item does not contain fulltext

Contains fulltext : 75050.pdf (author's version ) (Open Access) 4 p.

Item does not contain fulltext

Contains fulltext : mmubn000001_350223920.pdf (Publisher’s version ) (Closed access) X, 227 p.

Item does not contain fulltext

So far most attempts to automate Directory Assistance services focused on private listings, because it is not known precisely how callers will refer to a business listings. The research described in this paper, carried out in the SMADA project, tries to fill this gap. The aim of the research is to model the expressions people use when referring to a business listing by means of rules, in order to automatically create a vocabulary, which can be part of an automated DA service. In this paper a rule-base procedure is proposed, which derives rules from the expressions people use. These rules are then used to automatically create expressions from directory listings. Two categories of businesses, viz. hospitals and the hotel and catering industry, are used to explain this procedure. Results for these two categories are used to discuss the problem of the over- and undergeneration of expressions.

A novel approach to articulatory-acoustic feature extraction has been developed for enhancing the accuracy of classification associated with place and manner of articulation information. This "elitist" approach is tested on a corpus of spontaneous Dutch using two different systems, one trained on a subset of the same corpus, the other trained on a corpus from a different language (American English). The feature dimensions, voicing and manner of articulation transfer relatively well between the two languages. However, place information transfers less well. Manner-specific training can be used to improve classification of articulatory place information.

Item does not contain fulltext 3 p.

Item does not contain fulltext 4 p.

A novel framework for automatic articulatory-acoustic feature extraction has been developed for enhancing the accuracy of place- and manner-of-articulation classification in spoken language. The "elitist" approach focuses on frames for which neural network (MLP) classifiers are highly confident, and discards the rest. Using this method, it is possible to achieve a frame-level accuracy of 93% for manner information on a corpus of American English sentences passed through a telephone network (NTIMIT). Place information is extracted for each manner class independently, resulting in an appreciable gain in place-feature classification relative to performance for a manner-independent system. The elitist framework provides a potential means of automatically annotating a corpus at the phonetic level without recourse to a word-level transcript and could thus be of utility for developing training materials for automatic speech recognition and speech synthesis applications, as well as aid the empirical study of spoken language.

Item does not contain fulltext

This paper describes a prototype of a multimodal railway information system that was built by extending an existing speech-only system. The purpose of the extensions is to alleviate a number of shortcomings of speech-only interfaces.

Contains fulltext : 76203.pdf (author's version ) (Open Access) 9 p.

Item does not contain fulltext 20 p.

Contains fulltext : 76196.pdf (author's version ) (Open Access) 4 p.

Item does not contain fulltext

Contains fulltext : 76459.pdf (Unknown ) (Open Access) 6 p.