Proceedings Fonetik 2009 - Institutionen för lingvistik

Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm UniversityThe parameters speech activity and phonationtime solely do therefore not give a clear explanationto vocal fatigue. However, child B,who suffers from vocal fatigue by the end of theday presents us with an amount of phonationtime comparable to what has been found forpre-school teachers (Masuda et al. 1993).Hoarseness vs. fundamental frequencyChild B presents us with particularly high meanF0 - much higher then F0 under controlled condition- and a broad F0-range in the morning,followed by an increase of hoarseness later inthe day. However, Child C with fairly high F0-increase over the day and a high F0-range is notaffected by a change of hoarseness for theworse. Child A with fairly stable mean F0 andF0-range over the day presents us with higherdegree of hoarseness after the morning recordings.Child D with comparably stable meanF0 and F0-range on the other hand improvedthe voice condition over the day.The use of high F0, a high F0-range solelydoes not seem to account for voice deterioration.Hoarseness vs. speech intensityThe child (B) producing speech at highestloudness level is the one that suffers most fromincreased voice problems later in the day.Strenuous speech production with high intensitytherefore seems to be an important parameterto take into consideration when accountingfor voice problems in children.Hoarseness vs. speech intensity andbackground noiseThe children react in a different way to the levelof background noise. Being exposed to a highlevel of background noise, one of the activechildren (B) seems to be triggered for a loudvoice use, whereas one other speech activechild (C) does not behave in the same way, butproduces a much softer voice. The child reactingwith a stronger voice (B) also responds withincreased hoarseness later in the day.As has been presented in the results, thechildren never produce speech at a loudness of9.1dB above background noise, the level thathad been found by Södersten et al. to occur forpre-school teachers. However, normalizationwith regard to microphone distance for thechildren’s speech might be a question to beconsidered, since no comparable normalisationhas been carried out for the recordings of thebackground noise.General discussionIt can be found that child B refers to a typicalchild in a risk zone who suffers from voiceproblems by the end of the day due to hazardousvoice use: being a lively child in a verynoisy environment leads to making use of aloud, strong voice with a relatively high fundamentalfrequency and a high fundamental frequencyrange, resulting in vocal fatigue at theend of the day, reflected by an increase ofhoarseness. The results for this child agree withSödersten et al. that producing high fundamentalfrequency at high vocal loudness can lead tovocal trauma.Child A on the other hand does not showany particularly unusual voice use. However thedegree of hoarseness was already very high inthe first recording made in the morning and increasesfurther in during the day. The high degreeof hoarseness could have had influence onthe calculation of different acoustic measurements,e.g. phonation time is fairly low, becausethe algorithm is not able to pick periodicalparts of speech. Even the low measure ofintensity might have been affected, since voicedsounds show stronger intensity, which might belacking due to the child’s high degree ofhoarseness. It should however be noted thatchild A does not present us with high numberof measurements on speech activity, so that alow degree of phonation time is unlikely to bebased on period picking problems. This childmight have a predisposition for a hoarse voice,or an already obtained voice problem.Child C seems to be typical for a lively childwith high speech activity (Table 1) and a broadF0-range (Table 2). On the other hand, thischild shows lowest speech intensity (Table 3)which seems to be a good prerequisite to preventvoice problems. This child presents uswith a somewhat higher degree of hoarsenessthen child B, but there is no change for worseover the day. When taking a look at how childC uses her voice, one can find out that she ishumming a lot by herself.Child D is most lively in the afternoon,where the degree of hoarseness is lowest (Table1). Obviously, this child is seems to need towarm up the voice during the day, which resultsin highest activity later in the day combinedwith best voice condition.124