A number of aspects of speech training involve assessing the quality of the student's speech. It is of interest to determine whether such speech quality analysis can be done automatically. This thesis provides a preliminary answer to that question by proposing and then evaluating a set of quality measures for comparing the quality of two segments of speech. Speech quality is taken to be the lack of defects in the articulatory and prosodic components of speech. It is a non-quantitative definition from speech pathology that can meet the needs of speech training. Speech defects common among deaf children and students of English as a second language are reviewed, and classified according to this scheme. The speech quality measures are based on a linear prediction model of speech, and adapt several techniques from the field of speech recognition. Evaluations using speech with known quality defects show that the articulatory measures are effective in detecting most of the common errors of articulation, with the exception of ones between nasal sounds. The prosodic quality measures of loudness and timing give very useful indications of syllable stress and voicing errors. The timing measure is derived from the optimal time-warping curve between the two utterances, and provides an accurate means of tracking speed variations in speech. Differences between speakers tend to mask articulatory quality errors, but have little effect on the prosodic quality measures. An articulatory distance measure is proposed that partly counters these interspeaker differences. Work remains to be done in a number of key areas, but the results of this preliminary investigation suggest that automatic speech quality analysis by computer is practical and may one day become a versatile tool for speech training. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Unknown

Thesis (MPhil (General Linguistics))—University of Stellenbosch, 2009. / ENGLISH ABSTRACT: The term "schizophrenia" refers to a psychiatric condition which affects an individual's thought and speech (Eaton and Chen 2006). The verbal expression of schizophrenics can therefore be used as a tool for insight into the nature of schizophrenia as well as the cognitive processes of schizophrenics (Wróbel 1990:1). This thesis reports on a comparative evaluation of three pragmatic assessment tools, namely the Pragmatic Protocol (Prutting and Kirchner 1987), the Profile of Communicative Appropriateness (Penn 1985), and the Framework for Assessing (Children's) Conversational Skills (Rumble 1988), in order to establish which of these tools is most suitable for assessing the first language (L1) and second language (L2) pragmatic abilities of late bilingual schizophrenics. Four late bilingual schizophrenic patients participated in this study. A thirty minute informal interview was conducted with each of the participants in both their L1 and L2 and the speech samples were transcribed and then analysed by means of each of the pragmatic assessment tools. A careful examination of the results yielded by the three assessment tools, showed, firstly, that when presenting the results of a pragmatic assessment of schizophrenic speech, it is crucial that both quantitative and qualitative information be included: if the latter is excluded, a significant amount of information is hidden from the clinicians and/or linguists doing the assessment, as well as the people to whom they report their findings. Secondly, with respect to the characteristics of schizophrenic speech, the three instruments used in this study show that whereas most of the aspects of schizophrenics' linguistic abilities seem intact, their pragmatic skills are definitely impaired. Thirdly, regarding differential symptomatology in bilingual schizophrenics, this study concludes that none of the three assessment tools contributes to a better understanding of this phenomenon, and that, in fact, it is highly unlikely that any pragmatic assessment tool would be able to capture this phenomenon. Finally, it is recommended that clinicians assess bilingual patients in both languages, whenever possible, in order to determine the full range of symptoms experienced by the patient, to gain a better indication of the severity of the illness and to track the progress of the illness. / AFRIKAANS OPSOMMING: Die term "skisofrenie" verwys na 'n psigiatriese toestand wat 'n individu se denkprosesse en spraak beïnvloed (Eaton en Chen 2006). Die verbale uitinge van skisofrene kan dus gebruik word om insig oor die aard van skisofrenie, sowel as die kognitiewe prosesse van skisofrene, te verkry (Wróbel 1990:1). Hierdie tesis lewer verslag oor 'n vergelykende evaluering van drie pragmatiese assesseringsinstrumente, naamlik die "Pragmatic Protocol" (Prutting en Kirchner 1987), die "Profile of Communicative Appropriateness" (Penn 1985), en die "Framework for Assessing (Children's) Conversational Skills" (Rumble 1988), om sodoende vas te stel watter een van hierdie drie die mees gepaste instrument is vir die assessering van tweetalige skisofrene se pragmatiese vaardighede in hul eerstetaal (T1) en tweedetaal (T2), spesifiek in gevalle waar die T2 later (d.w.s. nie binne die eerste sewe lewensjare nie) verwerf is. Vier sulke tweetalige skisofrene het deelgeneem aan die studie. Daar is met elkeen van die deelnemers 'n informele onderhoud gevoer vir 30 minute in hul T1, gevolg deur 30 minute in hul T2. Die onderhoude is getranskribeer en daarna geanaliseer deur middel van elk van die drie assesseringsinstrumente. 'n Noukeurige ondersoek en vergelyking van die resultate van die drie instrumente het eerstens getoon dat dit belangrik is om die resultate van 'n pragmatiese analise van skisofreniese spraak op beide 'n kwantitatiewe en kwalitatiewe wyse aan te bied: wanneer kwalitatiewe inligting weggelaat word, bly 'n betekenisvolle hoeveelheid van die informasie verborge vir die klinici en/of taalwetenskaplikes wat die assessering doen, asook die mense aan wie hulle hulle bevindinge rapporteer. Tweedens, met betrekking tot die eienskappe van skisofreniese spraak, wys die drie instrumente wat in hierdie studie gebruik is dat alhoewel meeste aspekte van skisofrene se taalvaardighede ongeskonde is, hulle pragmatiese vaardighede ooglopend aangetas is. Derdens, rakende differensiële simptomatologie in tweetalige skisofrene kom hierdie studie tot die gevolgtrekking dat geen van die drie instrumente bydra tot 'n beter begrip van hierdie verskynsel nie, en dat dit selfs hoogs onwaarskynlik is dat enige pragmatiese assesseringsinstrument hierdie verskynsel sou kon vaslê. Uiteindelik word daar aanbeveel dat klinici, wanneer dit ookal moontlik is, tweetalige pasiënte in beide tale behoort te assesseer om sodoende die volledige reeks van simptome wat 'n pasiënt ervaar vas te stel, om 'n beter aanduiding te bekom oor die erns van die siekte, en om die progressie van die siekte te volg.