94 research outputs found
Developing Speech Resources from Parliamentary Data for South African English
No full textAbstractThe official languages of South Africa can still be classified as under-resourced with respect to the speech resources that are required for technology development. Harvesting speech data from existing sources is one means to create additional resources. The aim of the study reported on in this paper was to improve the harvesting and transcription accuracy of a corpus derived from parliamentary data. This aim was achieved by improving on the text normalisation process and pronunciation modelling as well as by iteratively training more accurate in-domain acoustic models. In this manner, more data could be harvested with higher confidence than using baseline pronunciation dictionaries and out-of-domain speech data
The limitations of data perturbation for ASR of learner data in under-resourced languages
No full textTopic modelling to support English text selection for translation into South Africa\u27s other official languages
No full textAppropriate training data is a prerequisite for the development of natural language processing (NLP) techniques. Vast amounts of language data are typically required to develop NLP tools that perform at state-of-the-art level. Such abundant resources are currently only available in a few languages. The remaining languages have to find alternative ways to become ``NLP-enabled\u27\u27. The aim of the study reported on here is to make more language data available to support NLP development in the official languages of South Africa. In this paper we present the idea of generating text data by means of translation. We also propose the use of topic modelling to identify text in a highly resourced source language that will yield meaningful translations in under-resourced target languages. More specifically, the paper describes how topic modelling was used to identify English Wikipedia articles that should be suitable for translation into South Africa\u27s 10 other official languages
The perception and identification of accent in spoken Black South African English
No full textCan mother tongue speakers of the Nguni and Sotho languages determine each other's first language (L1) based on their English accent? Contrasting claims have been made in this regard. While Black South African English (BSAE) can be distinguished clearly from Standard South African English (SSAE) on the levels of both production and perception, insufficient evidence exists that such a distinction can be made between Nguni-English and Sotho-English. This study investigates the question of perceivable differences in BSAE accents by means of two perceptual experiments. The first aim of the experiments is to ascertain whether participants from either the Nguni or Sotho language group can determine whether a particular speaker has an SSAE or aBSAE accent. The second aim is to determine whether L1 Nguni and Sotho listeners can identify a speaker's L1 group by listening to English words and sentences pronounced by Nguni and Sotho L1 speakers. Lastly, we investigate whether there is any correlation between listeners' judgement of speakers' accent and their ability to determine a speaker's L1. The results of both perceptual experiments contradict the notion that different mother tongues influence BSAE to such a degree that the speaker's L1 is easily perceived. However, some correlation was found between perception of accentedness and the correct identification of L1.Southern African Linguistics and Applied Language Studies 2007, 25(1): 91–10
Objective measures to improve the selection of training speakers in HMM-based child speech synthesis
No full text
A language application for Health Science students : a study on user experience
No full textDissertation (MA)--University of Pretoria, 2016.South Africa is home to 11 official languages and speakers of these languages communicate with one another on a daily basis. Such multilingual communication occurs throughout the country, especially at hospitals and clinics. Every so often, someone needs to visit a healthcare facility and then it is difficult for the patient to find a health professional that speaks a language he/she understands. Some universities in South Africa, including the University of Pretoria, address this matter by teaching students an additional language to enable them to communicate with their patients.
This study aimed to assist the University of Pretoria in this endeavour by providing three custom-designed, mobile-assisted Sepedi language learning applications to students from the Faculty of Health Sciences enrolled for the Sepedi language module. The students used the applications as supplementary tools for their studies over nine weeks and then completed a questionnaire on user experience. The questionnaire was used to determine whether the students perceived the mobile applications to be useful supplementary tools to their studies and whether they had a clear preference for a specific application.
The results of this user experience study report a positive response to the applications, including strong preferences made by the students who participated in the study.African LanguagesMAUnrestricte
The Usefulness of Imperfect Speech Data for ASR Development in Low-Resource Languages
No full textWhen the National Centre for Human Language Technology (NCHLT) Speech corpus was released, it created various opportunities for speech technology development in the 11 official, but critically under-resourced, languages of South Africa. Since then, the substantial improvements in acoustic modeling that deep architectures achieved for well-resourced languages ushered in a new data requirement: their development requires hundreds of hours of speech. A suitable strategy for the enlargement of speech resources for the South African languages is therefore required. The first possibility was to look for data that has already been collected but has not been included in an existing corpus. Additional data was collected during the NCHLT project that was not included in the official corpus: it only contains a curated, but limited subset of the data. In this paper, we first analyze the additional resources that could be harvested from the auxiliary NCHLT data. We also measure the effect of this data on acoustic modeling. The analysis incorporates recent factorized time-delay neural networks (TDNN-F). These models significantly reduce phone error rates for all languages. In addition, data augmentation and cross-corpus validation experiments for a number of the datasets illustrate the utility of the auxiliary NCHLT data
Gauging the accuracy of automatic speech data harvesting in five under-resourced languages
No full textRecent research on deep-learning architectures has resulted in substantial improvements in automatic speech recognition accuracy. The leaps of progress made in well-resourced languages can be attributed to the fact that these architectures are able to effectively represent spoken language in all its diversity and complexity. However, developing advanced models of a language without appropriate corpora of speech and text data remains a challenge. For many under-resourced languages, including those spoken in South Africa, such resources simply do not exist. The aim of the work reported on in this paper is to address this situation by investigating the possibility to create diverse speech resources from unannotated broadcast data. The paper describes how existing speech and text resources were used to develop a semi-automatic data harvesting procedure for two genres of broadcast data, namely news bulletins and radio dramas. It was found that adapting acoustic models with less than 10 hours of manually annotated data from the same domain significantly reduced transcription error rates for speaking styles and acoustic conditions that are not represented in any of the existing speech corpora. Results also indicated that much more automatically transcribed adaptation data is required to achieve similar results
Investigating the feasibility of harvesting broadcast speech data to develop resources for South African languages
No full textSufficient target language data remains an important factor in the development of automatic speech recognition (ASR) systems. For instance, the substantial improvement in acoustic modelling that deep architectures have recently achieved for well-resourced languages requires vast amounts of speech data. Moreover, the acoustic models in state-of-the-art ASR systems that generalise well across different domains are usually trained on various corpora, not just one or two. Diverse corpora containing hundreds of hours of speech data are not available for resource limited languages. In this paper, we investigate the feasibility of creating additional speech resources for the official languages of South Africa by employing a semi-automatic data harvesting procedure. Factorised time-delay neural network models were used to generate phone-level transcriptions of speech data harvested from different domains
- …
