1887

oa Litnet Akademies : 'n Joernaal vir die Geesteswetenskappe, Natuurwetenskappe, Regte en Godsdienswetenskappe - Akoestiese analise van die vokale van 'n groep bejaarde Afrikaanse vroue : geesteswetenskappe

 

Abstract

Hierdie artikel doen verslag van 'n ondersoek na die eienskappe van die vokale van Afrikaans soos voortgebring deur 20 bejaarde Afrikaanssprekende vroulike sprekers (gemiddeld 86 jaar oud). Om dit te kon doen, was dit eerstens nodig om al die fonemiese Afrikaanse vokale te versamel, te ontleed en volledig en op 'n insiggewende wyse te beskryf. Hiervoor het ek gekonsentreer op die akoestiese eienskappe van hierdie vokale, maar ek integreer ook die toepaslike artikulatoriese eienskappe van die vokale in die ondersoek. Ek baseer die ontleding en die beskrywing op 'n model wat vir hierdie doel ontwikkel is (Wissing 2012). Ek verwys na 'n interaktiewe vokaalkaart, wat fokus op die artikulatoriese aspek van vokaalproduksie oor die algemeen: http://www.nwu.ac.za/af/content/ctext-vokaalkaart


Wat die akoestiese ontleding en beskrywing betref, gebruik ek die eerste twee vokaalformante, genoem F1 en F2, om 'n oorsigtelike akoestiese vokaalkaart saam te stel. Sodanige kaart is tot groot hulp wanneer verskillende vokaalstelsels, soos die huidige een van die ou groep sprekers, vergelyk word met dié van, sê, jong sprekers. Afgesien van F1 en F2, gee ek ook aandag aan agt ander parameters: vokaalduur, intensiteit (totale intensiteit; intensiteit in BF1 en BF2), grondtoon, harmonisiteit-tot-geraas-verhouding HGV), F3 en die bandwydte van F1, waarvan slegs 'n paar ander (F0, F3 en Duur) naas F1 en F2 soms ook bekyk word - maar dis die uitsondering (vgl. Adank, Van Hout en Smits 2004). In die geval van al die parameters gee ek basiese metings (gemiddeldes, aantal geldige gevalle in die datastel, en standaardafwykings). Ek gebruik die gemiddeldes om ook lyngrafieke van vokale ten opsigte van al tien die parameters te verskaf. Op grond hiervan kan interpretasies en afleidings makliker gemaak word as om bloot die beskikbaarheid van metingswaardes per tabel te gee. Die resultaat van die ondersoek is in baie opsigte nuut en interessant. Dit toon veral dat F1 en F2 'n beperkter waarde het in die klassifikasie van vokale en subgroepe van vokale as wat gewoonlik aangeneem word. In sommige opsigte word die klassifikasie van vokale op grond van die horisontale en vertikale dimensies van artikulasie daarvan sterk ondersteun. Sulke parameters is met name Intensiteit in BF2 (die spektrale area bokant 1000 Hz), HGV en die bandwydte van F1. Die vokaalontledingsmetode en beskrywingsmodel wat in Wissing (2012) voorgestel word, word in hierdie ondersoek ondersteun en blyk baie goed bruikbaar te wees. Dit is ook duidelik uitgewys dat dit onnodig is in die geval van 'n homogene groep sprekers soos hierdie - almal bejaard en vroulik - om van 'n normaliseringsproses gebruik te maak.


This article reports on research on the characteristics of the Afrikaans vowels as produced by 20 elderly Afrikaans-speaking women with an average age of 86. In order to do this, it was necessary first to collect recordings of all the phonemic Afrikaans vowels, to analyse them and then to describe them in a comprehensive and informative way. With this in mind, I focused on the acoustic properties of these vowels, although I also integrated their relevant articulatory properties. I based the analysis and description as well as the evaluation of the results mainly on a descriptive model that had been developed for this purpose (Wissing 2012). Of course, numerous other models and sub-models exist, but this one covers all the acoustic parameters dictated by the project of which this article forms a part, namely .
A secondary, superordinate goal of this project was to compare different varieties with one another. The present research provides an acoustic description of the Afrikaans that was probably spoken approximately 80 years ago. Comparing this variety with other, modern ones in terms of how these varieties are used by present-day speakers of Afrikaans, for example, may be informative.
The method which I used to reach the descriptive goal is mostly in line with those followed in similar research endeavours, such as that of Pols et al. (1973) and Adank et al. (2004). In the present study the speech of 20 elderly female speakers was recorded and then analysed acoustically. The readers had not been selected beforehand - they took part voluntarily. The only condition was that they had to be Afrikaans speakers and should experience no inhibitive problems with regard to hearing or sight. Naturally they also had to be able to read out aloud and fluently. The stimulus material used was in the form of a reading list comprising 12 isolated, monosyllabic words. Each word contained a phoneme vowel of Afrikaans, such as the two members of the minimal pair and . This pair enables one to determine the extent to which unrounding of the latter is present in the speech of this group of speakers, if any.
The vowels in each speaker's recordings were determined, segmented and annotated in the Praat phonetic analysis programme, after which the relevant acoustic information was extracted for further statistical analysis. All data gained by means of analysis was tabulated, on which basis an acoustic vowel chart was constructed. In this chart the vowels were placed according to their F1 and F2 formant values. Line graphs of all the parameters' average values for each vowel were also presented. The parameters involved are the usual first two vowel frequency formants, F1 and F2, and, to a lesser extent also F3, as well as the length of vowels (duration), intensity (total intensity; intensity in BF1 and BF2), fundamental frequency (F0), harmonicity-to-noise ratio (HNR) and the bandwidth of F1 (BdW F1).

Loading

Article metrics loading...

/content/litnet/9/3/EJC129825
2012-12-01
2016-12-07
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error