Proceedings Fonetik 2009 - Institutionen för lingvistik

Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm UniversityChildren’s vocal behaviour in a pre-school environmentand resulting vocal functionMechtild Tronnier 1 and Anita McAllister 21 Department of Culture and Communication, University of Linköping2 Department of Clinical and Experimental Medicine, University of LinköpingAbstractThis study aims to shed some light onto the relationshipbetween the degree of hoarseness inchildren’s voices observed at different timesduring a day in pre-school and different aspectsof their speech behaviour. Behaviouralaspects include speech activity, phonation time,F0 variation, speech intensity and the relationshipbetween speech intensity and backgroundnoise intensity. The results show that childrenbehave differently and that the same type of behaviourhas a varied effect on the differentchildren. It can be seen from two children withotherwise very similar speech behaviour, thatthe fact that one of them produces speech at ahigher intensity level also brings about an increaseof hoarseness by the end of the day inpre-school. The speech behaviour of the childwith highest degree of hoarseness on the otherhand cannot be observed to be putting an extremeload on the vocal system.IntroductionSpeaking with a loud voice in noisy environmentsin order to making oneself heard demandssome vocal effort and has been shown toharm the voice in the long run.In several studies on vocal demands for differentprofessions it has been shown that preschoolteachers are rather highly affected andthat voice problems are common (Fritzell,1996; Sala, Airo, Olkinuora, Simberg, Ström,Laine, Pentti, & Suonpää 2002; Södersten,Granqvist, Hammarberg & Szabo, 2002). Thisproblem is to a large extent based on the needof the members of this professional group tomake themselves heard over the surroundingnoise, mainly produced by the children present.It is reasonable to assume that children’svoices are equally affected by background noiseas adult voices. As children most of the timecontribute to the noise in a pre-school settingthemselves – rather then other environmentalfactors as traffic etc. – they are exposed to thenoise source even more potently as they arecloser to the noise source. Another factor pointingin the same direction is their shorter bodylength compared to pre-school teachers.In an earlier study by McAllister et al.,(2008, in press), the perceptual evaluation ofpre-school children showed that the girls’voices revealed higher values on breathiness,hyperfunction and roughness by the end of theday, which for the boys was only the case forhyperfunction.In the present study the interest is directedto speech behaviour of children in relation tothe background noise and the affect on the vocalfunction. Diverse acoustic measurementswere carried out for this purpose.The investigation of speech activity is chosento show the individuals’ liveliness in thepre-school context and includes voiced andvoiceless speech segments, even non speechvoiced segments such as laughter, throat clearing,crying etc. In addition, measurements onphonation time were chosen, reflecting the vocalload. It should be noted that in some parts ofspeech intended voicing could fail due to theirregularity of the vocal fold vibrations – thehoarseness of the speaker’s voice. Therefore,both measurements, speech activity and phonationtime were considered important. In a studyon phonation time for different professions,Masuda et al. (1993) showed that the proportionfor pre-school teachers corresponded to20% during working time, which is considereda high level compared to e.g. nurses with a correspondinglevel of 5.3% (Ohlsson, 1988). Havingthese findings in mind, the degree of children’sspeech activity and phonation time andthe consequences for perceived voice quality isan interesting issue.Other factors for the analysis of vocal loadconsist of F0 including F0-variation and speechintensity including intensity variation. A vocaltrauma can be based on using high fundamentalfrequency and high vocal loudness (and hyperfunction,which this study is not focusing on inparticular), as Södersten et al. point out.One further aspect that may increase therisk for voice problems is the need of a speakerto be heard over background noise. Therefore,120