Proceedings Fonetik 2009 - Institutionen för lingvistik

More documents

Recommendations

Info

Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm UniversityBreathiness differences in male and female speech. IsH1-H2 an appropriate measure?Adrian P. SimpsonInstitute of German Linguistics, University of Jena, Jena, GermanyAbstractA wellestablished difference between male andfemale voices, at least in an AngloSaxon context,is the greater degree of breathy voice usedby women. The acoustic measure that has mostcommonly been used to validate this differenceare the relative strengths of the first and secondharmonics, H1H2. This paper suggests thatsexspecific differences in harmonic spacingcombined with the high likelihood of nasalitybeing present in the vocalic portions make theuse of H1H2 an unreliable measure in establishingsexspecific differences in breathiness.IntroductionOne aspect of male and female speech that hasattracted a good deal of interest are differencesin voice quality, in particular, breathy voice.Sex-specific differences in breathy voice havebeen examined from different perspectives. Hentonand Bladon (1985) examine behavioural differences,whereas in the model proposed byTitze (1989), differences in vocal fold dimensionspredict a constant dc flow during femalevoicing. In an attempt to improve the quality offemale speech synthesis, Klatt and Klatt (1990)use a variety of methods to analyse the amountof aspiration noise in the male and femalesource.A variety of methods have been proposed tomeasure breathiness:● Relative lowering of fundamental frequency(Pandit, 1957).● Presence of noise in the upper spectrum(Pandit, 1957; Ladefoged and AntoñanzasBarroso, 1985; Klatt and Klatt,1990).● Presence of tracheal poles/zeroes (Klattand Klatt, 1990).● Relationship between the strength of thefirst harmonic H1 and amplitude of thefirst formant A1 (Fischer-Jørgensen,1967; Ladefoged, 1983).● Relationship between the strength of thefirst and second harmonic, H1-H2 (Fischer-Jørgensen,1967; Henton andBladon, 1985; Huffman, 1987; Ladefogedand Antoñanzas Barroso, 1985;Klatt and Klatt, 1990).It is the last of these measures that has mostcommonly been applied to measuring sex-specificvoice quality differences.In this paper I set out to show that relatingthe strength of the first harmonic to other spectralmeasures as a way of comparing breathinessbetween male and female speakers is unreliable.The line of argumentation is as follows. The frequencyof the first nasal formant (F N1 ) can be estimatedto in the region of 200–350 Hz for bothmale and female speakers (Stevens et al., 1985;Maeda, 1993). At a typical male fundamentalfrequency of 120 Hz this will be expressed in anenhancement of the second and third harmonics.By contrast, at a typical female fundamental frequencyof over 200 Hz it may well be the firstharmonic that is more affected by F N1 . Comparisonof H1 and H2 as a measure of breathinesshas to be carried out on opener vowel qualitiesin order to minimise the effect of the first oralresonance, F1. Lowering of the soft palate isknown to increase with the degree of vowelopenness. Although the ratio of the opening intothe oral cavity and that into the nasal port is crucialfor the perception of nasality (Laver, 1980),acoustic correlates of nasality are present whenthe velopharyngeal port is open. It cannot be excluded,then, that any attempt to compare themale and female correlates of breathiness interms of the first harmonic might be confoundedby the sexspecific effects of F N1 on the first twoharmonics, in particular a relative strengtheningof the first female and the second male harmonic.Establishing that female voices are breathierthan male voices using the relative intensities ofthe first two harmonics might then be a selffulfillingprophecy.DataThe data used in this study are drawn from twosources. The first data set was collected as partof a study comparing nasometry and spectrographyin a clinical setting (Benkenstein, 2007).Seven male and fifteen female speakers wererecorded producing word lists, short texts andthe map task using the Kay Elemetrics Nasome-172
Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm Universityter 6200. This method uses two microphonesseparated by an attenuating plate (20 dB separation)that capture the acoustic output of the noseand the mouth. A nasalance measure is calculatedfrom the relative intensity of the two signalsfollowing bandpass filtering of both at ca. 500Hz. The present study is interested only in asmall selection of repeated disyllabic oral andnasal words from this corpus so only the combinedand unfiltered oral and nasal signals willbe used. The second data set used is the publiclyavailable Kiel Corpus of Read Speech (IPDS,1994), which contains data from 50 Germanspeakers (25 female and 25 male) reading collectionsof sentences and short texts. Spectralanalysis of consonant-vowel-consonant sequencesanalogous to those from the first datasetwas carried out to ensure that the signals fromthe first dataset had not been adversely affectedby the relatively complex recording setup involvedwith the nasometer together with the subsequentaddition of the oral and nasal signals.Sex-specific harmonic expressionof nasalityThe rest of this paper will concern itself withdemonstrating sex-specific differences in theharmonic expression of nasality. In particular, Iwill show how F N1 is responsible for a greaterenhancement of the second male and the first(i.e. fundamental) female harmonic. Further, Iwill show that F N1 is expressed in precisely thosecontexts where one would want to measure H1-H2. In order to show how systematic the patternsacross different speakers are, individualDFT spectra from the same eight (four male,four female) speakers will be used. It is importantto emphasise that there is nothing specialabout this subset of speakers – any of the 22speakers could have been used to illustrate thesame patterns.We begin with spectral differences foundwithin nasals, i.e. uncontroversial cases of nasality.Figure 1 contains female and male spectrataken at the centre of the alveolar nasal in theword mahne (“warn”). From the strength of thelower harmonics, F N1 for both the male and femalespeakers is around 200–300 Hz, which iscommensurate with sweep-tone measurementsof nasals taken from Båvegård et al. (1993).Spectrally, this is expressed as a strengtheningprimarily of the first female and the second maleharmonic.It is reasonable to assume that the velum willbe lowered throughout the production of theFigure 1: DFT spectra calculated at midpointof [n] in mahne for four female (top) and fourmale speakers.word mahne. Figure 2 shows spectra taken fromthe same word tokens as shown in Figure 1, thistime calculated at the midpoint of the long openvowel in the first syllable. While there is a gooddeal of interindividual variation in the positionand amplitude of F1 and F2, due to qualitativedifferences as well as to the amount of vowelnasalisation present, there is clear spectral evidenceof F N1 , again to be found in the increasedintensity of the second male and the first female173
Page 1 and 2:
Department of LinguisticsProceeding
Page 3 and 4:
Proceedings, FONETIK 2009, Dept. of
Page 5 and 6:
Page 7 and 8:
Page 9 and 10:
Page 11 and 12:
Page 13 and 14:
Page 15 and 16:
Page 17 and 18:
Page 19 and 20:
Page 21 and 22:
Page 23 and 24:
Page 25 and 26:
Page 27 and 28:
Page 29 and 30:
Page 31 and 32:
Page 33 and 34:
Page 35 and 36:
Page 37 and 38:
Page 39 and 40:
Page 41 and 42:
Proceedings, FOETIK 2009, Dept. of
Page 43 and 44:
Page 45 and 46:
Page 47 and 48:
Page 49 and 50:
Page 51 and 52:
Page 53 and 54:
Page 55 and 56:
Page 57 and 58:
Page 59 and 60:
Page 61 and 62:
Page 63 and 64:
Page 65 and 66:
Page 67 and 68:
Page 69 and 70:
Page 71 and 72:
Page 73 and 74:
Page 75 and 76:
Page 77 and 78:
Page 79 and 80:
Page 81 and 82:
Page 83 and 84:
Page 85 and 86:
Page 87 and 88:
Page 89 and 90:
Page 91 and 92:
Page 93 and 94:
Page 95 and 96:
Page 97 and 98:
Page 99 and 100:
Page 101 and 102:
Page 103 and 104:
Page 105 and 106:
Page 107 and 108:
Page 109 and 110:
Page 111 and 112:
Page 113 and 114:
Page 115 and 116:
Page 117 and 118:
Page 119 and 120:
Page 121 and 122: Proceedings, FONETIK 2009, Dept. of
Page 127 and 128: Proceedings, FOETIK 2009, Dept. of
Page 171: Proceedings, FONETIK 2009, Dept. of
Page 223 and 224:
Page 225 and 226:
Page 227:
Department of LinguisticsPhonetics
show all

Proceedings Fonetik 2009 - Institutionen för lingvistik

Create successful ePaper yourself

Delete template?

Save as template?