Voice Training and Changing Weight—Are They Reflected in Speaking Fundamental Frequency, Voice Range, and Pitch Breaks of 13-Year-Old Girls? A Longitudinal Study

Voice Training and Changing Weight—Are They Reflected in Speaking Fundamental Frequency, Voice Range, and Pitch Breaks of 13-Year-Old Girls? A Longitudinal Study Elizabeth C. Willis and Dianna T. Kenny, New South Wales, Australia Summary: Objective. Assessment of the voice-change progress of 20 girls (12–13 years) over 1 year by observing changes in speaking fundamental frequency (SFo), voice range, and register pitch breaks in the context of weight, height, voice training, and self-perception. Study Design. One-year longitudinal collective case study. Method. Twenty girls were recorded at the beginning and end of a year; nine girls were recorded another three times. SFo, vocal range, and characteristics were analyzed and interactions between these data assessed against weight and height to indicate pubertal development, and to test the hypothesis that changes in weight, height, SFo, and pitch breaks were related. Effects of training and the girls’ self-perception of their voice use were also assessed. Results. Vocal characteristics changed as the girls passed through different weight ranges. During 47.5–52.4 kg (called band 2) and 52.4–57.5 kg (band 3), there was progressive contraction of vocal range and in some girls a slight rise in SFo between recording times 1 and 5. Both high- and low-pitch breaks were present in 45% of girls’ voices. Girls in band 4 (<57.5 kg) had an increased vocal range, and pitch breaks in vocal-range areas that indicated the development of adult vocal registers. In this study, voice-trained girls were heavier, had higher SFo, used wider speech-range inflection, had a higher vocal range, and greater voice-use confidence; all girls lost confidence in their voice use over the year. Conclusions. In this longitudinal study of twenty 13-year-old girls, voice changes in SFo, vocal range, and pitch-break frequency were synchronous with certain weight ranges. Girls with training registered higher maximum phonational frequency and were more confident in their voice use than girls without training. Key Words: Female adolescent changing voice–Voice range–Vocal registers–Pitch breaks–Longitudinal study. INTRODUCTION Working with adolescent female voices is a daily task for thousands of teachers, choral directors, and voice pedagogues, but many variables affecting these young voices are little understood. Medically based research has focused on normal and pathological voice changes—for example, laryngeal development,1–3 speaking fundamental frequency (SFo) variation,4–6 breathiness,7,8 vocal fatigue,9 time of day variation,10 and menstrual cycle effects.11 Scientific analytical methods have observed voice source changes,12,13 or used techniques such as long-term average spectra to observe voice quality changes.14 Music education and drama studies have examined age-appropriate vocal techniques and repertoire to maximize the efficiency of the speaking and singing voice in performance,15,16 and to investigate voice-affecting psychological aspects.17 However, these previous studies of adolescent girls have been cross sectional, and have focused on contrasting data between pre- and postpubertal girls groups. With the exception of one study by Howard and Welch,18 there have been no recent prospective longitudinal studies that examine aspects of adolescent female voice over the period of peak change. Accepted for publication June 9, 2010. From the Australian Centre for Applied Research in Music Performance, University of Sydney, New South Wales, Australia. Address correspondence and reprint requests to Elizabeth C. Willis, 544 Darling St, Rozelle 2039, New South Wales, Australia. E-mail: [email protected] (E.C.W.), [email protected] (D.T.K.) Journal of Voice, Vol. 25, No. 5, pp. e233-e243 0892-1997/$36.00 Ó 2011 The Voice Foundation doi:10.1016/j.jvoice.2010.06.004

Speaking fundamental frequency Previous research into the adolescent speaking voice established normal speaking ranges for 13-year-old girls, but often used stimuli developed for adults. Duffy5 used ‘‘The Rainbow Passage’’19 to assess the SFo for 24 girls aged 11, 13, and 15 years. Duffy found that the premenarcheal girls had a mean SFo of 260 Hz (median, 264 Hz) [C4] and the postmenarcheal girls recorded a mean SFo of 245 Hz (median, 250 Hz) [B3]. Duffy also reported a phenomenon he called ‘‘frequency breaks’’ in the voices of all his 13-year-old female study participants, which were more prevalent, and of greater extent than in other groups of males and females. Williams et al,8 in their investigation of singing and speaking characteristics of adolescent females aged 11–15 years, used both spontaneous speech (20 seconds on a topic of the subject’s choice) and reading (a passage of 150 syllables at fifth Grade reading level). SFo for pre- and postmenarcheal girls was calculated. Premenarcheal girls in conversation averaged 218.3 Hz [A3] and in reading 225.3 Hz [zA3]. For the postmenarcheal girls, the average SFo was 206.3 Hz [A-flat3] for conversation and 214.8 Hz [zA3] for reading. These results are consistently lower than those in the Duffy study, possibly because of the different tasks used and the wider age range of study participants. Wilson20 compiled a table of normative SFo for children and adolescents, and used a range of studies including Eguchi and Hirsch,21 Duffy,5 and Vuorenkoski et al22 for 13-year-old girls. Averaged results were 246.26 Hz [B3] (range, 239–260 Hz) for less mature, and 241 Hz [zB3] (range, 237–245 Hz) for more mature 13-year-old girls. For SFo analysis of the age group, Wilson recommended reading a selected passage of

e234 about 180 words and counting as the most practical methods. During recording, the researcher used hand signals to guide the student during a counting task to gradually minimize the pitch range to almost a monotone. This tone was then pitch matched to either a pitch pipe or keyboard. Wilson noted variables affecting SFo pitch include sex, age and physical size, dynamic level used, time of day (although this is controversial),10 the presence of pathology, cultural background, and level of reading ability. Frequency analysis techniques also influenced findings. Previous research used magnetic tape viewers, pitch indicator meters, and frequency matching to keyboards, although this last, still used clinically and pedagogically, requires a degree of musical knowledge. Gackle23–26 described the characteristics and symptoms of female adolescent singing voice in change as having decreased and inconsistent range, breathiness and huskiness of tone, voice breaks and cracking, lowering of the SFo, insecurity of pitch, difficulty with onset of phonation, and noticeable changes in vocal timbre and tone quality. She developed normative SFo recommendations for different stages of adolescent female voice change. Her phases (previously ‘‘stages’’) of change and the average speaking fundamental frequencies associated with them are phase 1 (prepubertal): 260–290 Hz [C4–D4]; phase IIA (pubescencepremenarcheal): 245–275 Hz [B3–C#4]; phase IIB (pubescence-postmenarcheal): 222–275 Hz [A3–C#4], and phase III (young female-postmenarcheal): 210–245 Hz [A3– B3]. Typically, 13-year-old girls would be in phases IIA and IIB—the peak of change. To establish SFo, Gackle recommended using continuous voicing in a ‘‘counting backwards from 20’’ task. At pretest in her first study, 43% of her subjects (all female) were 13 years (range for the whole study was 10.11–15.10 years). Her results showed an average SFo of 220 Hz [A3]. These readings were obtained initially through a correlation of SFo to keyboard pitch, and were later refined using a pitch meter. These findings were consistent with the Williams et al,8 and with Wilson’s composite SFo pitch-range chart.20 Voice range In addressing the general classroom situation, Gackle gave typical singing ranges and tessitura for girls undergoing change. Tessitura is defined by Reid27 as ‘‘the pitch range a singer is capable of producing with the greatest ease regardless of technical limitations.’’ Ranges and tessitura in the Gackle model were phase 1: 233–698 Hz [B-flat3–F5] (tessitura: 294–587 Hz [D4–D5]); phase IIA: 220–784 Hz [A3–G5] (tessitura: [D4–D5]); phase IIB: 220–698 Hz [A3–F5] (tessitura: 247–523 Hz [B3–C5]); and phase III: 220–880 Hz [A3– A5] (tessitura: 220–784 Hz [A3–G5]). Decoster et al,28 in their study of 17 girls who were auditioned members of the Antwerp Cathedral Girls’ Choir (mean age, 12.07 years; age range, 9.9– 16.11 years), validated Gackle’s model of female voice change, but found considerable individuality among the girls in their speaking and singing results. Gackle’s phase III (young adults) accurately described the Antwerp students, with the exception that the girls had a wider singing range (either higher, lower, or both), than given in the Gackle model. The extended range

Journal of Voice, Vol. 25, No. 5, 2011

was attributed both to the girls being selected by audition, and to the girls growing taller, which may have resulted in a larger larynx. Fuchs et al15 studied a sample of 183 children and adolescents (72 boys and 111 girls) with a mean age of 13.5 years (range, 6–19 years). They investigated the effects of training on child and adolescent voices and showed that, when compared with children without any vocal training, the voice range averaged 2.7 semitones higher among those children trained in a choir, and by 5.8 semitones higher among those with individual voice tuition. Changes were detected in the vocal high-range limit, but not the low range because of limitations imposed by anatomical and physiological factors. These high-range changes were attributed to the students learning to use their head/falsetto register. Voice-range extension as a result of increasing age was also discussed, although not considered a factor in the study. Vocal registers and pitch breaks Vocal registers have been defined as ‘‘a group of like sounds or tone qualities whose origin can be traced to a special kind of mechanical (muscular) action,’’27 and as ‘‘perceptually distinct regions of vocal quality as pitch or loudness is changed.’’29 Titze29 and McCoy30 speak of two predominant muscle groups that control registration—the thyroarytenoid predominant muscle group, and the cricothyroid predominant muscle group, which flexibly and dynamically control the shape, length, density, and elasticity of the vocal folds in response to the production of various frequencies. Achieving timbral balance between vocal registers is a major concern for singing teachers as vocal pathology can result if there is not a dynamic relationship between vocal fold length, tension, and mass as pitch changes.31 Frequencies where the voice changes timbre involuntarily are associated with register changes and are called passaggi by singing teachers. These frequencies are particularly prone to pitch breaks because of either inappropriate breath management or muscle loading.29 Sˇvec et al32 discussed the vocal fold bifurcations that result in pitch breaks around register change pitch areas, and found a possible hysteresis between the two overlapping voice-control muscle groups. Thus, pitch breaks can be defined as abrupt involuntary register transitions that affect periodicity.33 Research into child vocal registration by McAllister et al34 found register change in 10-year-old children to be at a mean fundamental frequency of 511 Hz (zB4) (25% higher than in the voices of adults), with a small number of children having a second register change at a mean frequency of 902 Hz (zA#5). This is consistent with Phillips,35 who stated that registration boundaries in the child voice are similar to those of the female voice, with the low register being 196–262 Hz [G3–C4], the middle register 262–523 Hz [C4–C5], and the upper register 523–784 Hz [C5–G5] with a possible extension to 1047 Hz [C6]. However, Keidar et al36 and Titze,33 identified 300– 350 Hz (zE4–F4) as the critical register change frequencies in adults, both men and women, which suggests a register model of lower: 82–330 Hz [E2–E4]/middle: 330–659 Hz [E4–E5]/upper: 659–1318 Hz [E5–E6]. In adolescent female

Elizabeth C. Willis and Dianna T. Kenny

A Study on Voice Training and Changing Weight of 13-Year-Old Girls

voice, Gackle identified 392–494 Hz [G4–B4] as a registerchange break point that develops as part of phase IIA, along with weakness around 262 Hz [C4]. In phase IIB, an additional register transition can appear at 587–740 Hz [D5–F#5] and the voice quality can ‘‘flip’’ between heavy and light timbres. These changes are possibly the acoustical response to changes in tracheal length, and vocal fold structure and bulk that occur at pubescence.29 Pitch breaks indicating that vocal registration may be changing as part of the maturational process have been reported by Duffy.5 He observed three times as many ‘‘frequency breaks’’ among his postmenarcheal 13-year-old female subjects than among his premenarcheal 13-year olds, consistent with Gackle25 who had noted their increased prevalence as a sign of girls being in phase IIA. With this previous research in mind, this prospective longitudinal study of twenty 12–13-year-old girls focused on changes in SFo, vocal range, vocal registers, and pitch breaks in the context of vocal training, changing height and weight, and the girls’ self-perception of their vocal performance. The following hypotheses were formulated. Over the course of the year 1. SFo will lower; 2. Vocal range will depend on the girls’ stage of vocal development, showing contraction before a wider vocal range reappears; 3. Pitch breaks will appear within the vocal range of girls at a particular stage of physical development; 4. Girls’ self-perception of their vocal performance will be affected by their vocal change characteristics.

METHOD Study design This research used a 1-year longitudinal collective case study of twenty 13-year-old adolescent girls to investigate variations in SFo, vocal range, and register development as a function of weight, height, training, and self-perception of their voice use. Ethics approval was obtained for the study from the Human Research Ethics Committee of the University of Sydney, from the New South Wales Department of Education, Australia and from principals of the four participating secondary high schools. Information concerning the research was distributed via music classes of the relevant schools, and signed consent to participate was obtained from the 20 girls who volunteered, and from their parents. Baseline data was collected at data collection time 1, which was at the end of the girls’ first year at secondary school (year 7). At this time, the girls completed physical, self-report, and vocal assessments. Twelve months later, the same 20 girls agreed to a second assessment (data collection time 5); nine of these same girls made three additional recordings at 2–3-month intervals over the same 12-month period (data collection times 2, 3, and 4). Measures Demographic questionnaire. The girls were asked about their age, vocal training, and general health.

e235

Physical measures. Weight and height measurements were taken during each of the five assessments to investigate whether changes in the independent measures of weight or height correlated to changes in voice profiles. Self-report of speaking and singing voice quality. The subjective responses of the students to their developing speaking voices were assessed on a 10-point visual analog scale (VAS) with the question ‘‘How do you feel your voice functioned generally over the past two weeks?’’ The lowest anchor (0) was ‘‘My speaking voice bothered me a lot’’ to (10) ‘‘My speaking voice did not bother me at all.’’ For singing, the lowest anchor (0) was ‘‘Singing was difficult. I did not like the sound that I made’’ to (10) ‘‘Singing was easy and I liked the sound that I made.’’ In addition, the question was asked ‘‘Have you noticed any changes in your voice over the last 4 months. If yes, describe the changes.’’ SFo assessment tasks. The SFo was assessed by the two tasks of reading, and counting backward from 20. The reading passage, ‘‘Arthur, the Young Rat,’’19 was age appropriate in its storyline format, its length (180 words), and the predominance of monosyllabic syllables. It is used by speech pathologists to assess reading rate and stuttering, and had precedent use in other research of this age group.37,38 Vocal range assessment tasks. This study used glissandi, or sweeping glide vocalizations as recommended by Baken and Orlikoff,39 to assess the maximum phonational frequency range (MPFR), and whether any phonational gaps or other vocal phenomena were present. In this study, phonational gaps were defined as frequency ranges of long-term phonational weakness, which appeared in vocal task data at more than one time of data collection. In contrast, pitch breaks tended to occur during only one of the five data collections. Before each recording session, the six glides were modeled by the researcher, using one breath per glide. The girls then practiced the task of three descending glides over their vocal range (the instruction given was to glide from ‘‘as high as possible to as low as possible’’), followed by three ascending glides (‘‘from as low as possible to as high as possible’’). The girls stood while recording. These vocal tasks were recorded in an acoustically treated studio and used an AKG C477 microphone with windshield (AKG Acoustics, Vienna, Austria) mounted 7 cm from the corner of the subject’s mouth. Recordings were captured using a Behringer Ultragain Pro Preamplifier (Behringer, Willich, Germany) to a Marantz Compact Disc Recorder CDR 640 (Philips Electronics, New York, NY). The recorded data were digitized over the range 0–16 kHz and then analyzed using Phog Interactive Phonetography System (Hitech, Taby, Sweden). Analysis of the fundamental frequency was undertaken using the Soundswell program (Hitech, Taby, Sweden) on an IBM Thinkpad A30. Statistical methods Data were analyzed using SPSS version 16.0. For all analyses, P < 0.05 was considered to be statistically significant. Distributions of data values were examined using summary statistics and box plots. The parameters speak (VAS%), sing (VAS%), minimum frequency (Min_Fo), maximum frequency (Max_Fo), and range were close to being normally distributed

e236

Journal of Voice, Vol. 25, No. 5, 2011

and therefore parametric summary statistics were used. There were only one or two outlying values and no extreme values. Pearson’s correlation coefficient (r) was used to explore relationships between these voice parameters and height and weight. Other voice parameters that were measured related to vocal range and pitch breaks—gap, gap start (Gap_st), gap finish (Gap_fin), low range, and high range. These measurements had 80% zero scores at time 1 and 75% zero scores at time 5, and therefore no statistical analyses were conducted using these variables. Sing (VAS%) was not recorded at time 1. Paired t tests were used to determine the significance of the differences in speak (VAS%), minimum frequency, maximum frequency, and range between time 1 and time 5. Linear mixed models were used to explore the changes in speak (VAS%), sing (VAS%), minimum frequency, maximum frequency, and range over time, to build predictive models using height and weight as covariates and to test the effect on the outcome of whether the girls had undergone voice training. Models were run using the various covariance matrix types (diagonal, compound symmetry, autoregressive, and unstructured), and the model with the lowest 2 log likelihood (2LL) value was considered the best fit. Values of 2LL between consecutive models were subtracted to obtain a chisquare estimate to indicate whether the fit of the model had improved significantly. For all models, an unstructured covariance matrix provided the best fit. Because the sample size was not large, parameters were retained in the model if the chisquare for improved fit was significant or the P value for the parameter was <0.20.

RESULTS Demographic results At the time of the first recording (time 1), the average age of the girls was 13 years (standard deviation [SD] ¼ 3.83; range, 12.6–13.7 years). The girls reported being in good health at each of the five recording periods. Twelve girls had received voice training either through private singing lessons (10 girls) or in a choral setting (two girls). Descriptive statistics and change between times 1 and 5 Table 1 shows the mean values of minimum frequency, maximum frequency, speak (VAS%), height, and weight at times 1 and 5. Minimum frequency did not change significantly over the study period (P ¼ 0.38) but there was a significant

increase in maximum frequency (P ¼ 0.04) and a marginally significant increase in range (P ¼ 0.06). Both height and weight increased significantly (P < 0.0001) with a mean increase of 3.9 cm in height and 5.0 kg in weight. At time 1, minimum frequency was not related to maximum frequency (r ¼ 0.21, P ¼ 0.38) or range (r ¼ 0.34, P ¼ 0.15) but maximum frequency was significantly related to range (r ¼ 0.99, P < 0.0001). There was no significant relationship between minimum frequency, maximum frequency, range or speak (VAS%), height or weight. As for time 1, at time 5, minimum frequency was somewhat but not significantly related to maximum frequency (r ¼ 0.42; P ¼ 0.07), nor significantly correlated with range (r ¼ 0.50; P ¼ 0.03). Maximum frequency and range were highly inter correlated (r ¼ 0.99; P < 0.0001). There was no significant relation between minimum frequency, maximum frequency, range or speak (VAS%), and either height or weight, but sing (VAS%) had a significant positive correlation with height (r ¼ 0.52; P ¼ 0.02) and a trend association with weight (r ¼ 0.41; P ¼ 0.08). Training Analysis of vocally trained and untrained girls at times 1 and 5 showed that the average age for trained girls at time 1 was 12.10 years, and for the untrained girls 12.9 years. In this study, the trained group was heavier on average (52.02 kg, SD ¼ 10.95 at time 1 and 56.87, SD ¼ 11.27 at time 5) than the untrained girls (45.04 kg, SD ¼ 5.06 at time 1, and 49.78 kg, SD ¼ 5.6 at time 5). Linear mixed models were used to investigate changes in voice parameters over times 1 through 5 and to assess interrelationships between voice training, height, and weight as predictors of the outcome. For minimum Fo, inclusion of height and training improved the fit of the model with the 2LL reduced from 588.5 to 573.0 (chi-square ¼ 15.5, df ¼ 2, P < 0.001) but none of the time points were significant predictors of frequency, and voice training was not a statistically significant predictor (P ¼ 0.45). Height was a significant covariate when added to the model with an increase in minimum frequency of 1.46 points for every centimeter increase in height (P ¼ 0.04). When maximum frequency was examined, inclusion of voice training in the model improved the fit significantly with the 2LL reduced from 865.2 to 848.8 (chi-square ¼ 16.4, df ¼ 1, P < 0.001). The models plotted using the estimates of fixed effects are shown in Figure 1A. Values at time 1 were on average 110.6-Hz frequency points below time 5 (P ¼ 0.04). None of the other time points were significantly

TABLE 1. Descriptive Statistics and Change Between Times 1 and 5 Descriptor Minimum Fo (Hz) Maximum Fo (Hz) VAS (Speak %) Height (cm) Weight (kg)

Time 1, Mean (SD)

Time 5, Mean (SD)

Mean Difference (95% CI)

P Value

174.7 (45.5) [F3] 1261.6 (319.0) [zD#6] 78.2 (11.7) 159.5 (7.8) 48.6 (9.4)

183.0 (30.5) [F#3] 1372.0 (321.6) [zF6] 79.4 (11.7) 163.4 (6.9) 53.7 (10.0)

8.3 (11.0, 27.6) 110.6 (6.6, 214.6) 1.2 (7.5, 9.8) 3.9 (3.0, 4.8) 5.0 (4.3, 5.8)

0.38 0.04 0.78 <0.0001 <0.0001

Elizabeth C. Willis and Dianna T. Kenny

A

A Study on Voice Training and Changing Weight of 13-Year-Old Girls

The SFo was higher, and the pitch range used in inflected reading wider for the trained than for the untrained group. SFo for the trained girls averaged 264 Hz [C4] at time 1 (SD ¼ 55.06) and 264 Hz [C4] at time 5 (SD ¼ 56.42), and for the untrained group 252 Hz [zB3] at time 1 (SD ¼ 47.2) and 255 Hz at time 5 [zC4] (SD ¼ 54.08).

2000 Untrained singers Trained singers

1800 1600

Maximum Fo

1400 1200 1000 800 1

2

3

4

5

Time

B

0

-10

Speak (VAS %)

-20

-30 Untrained, weight 40 kg Trained, weight 40 kg Trained, weight 60 kg Untrained, weight 60 kg

-40

0

1

2

3

4

5

6

Time

C

90 Trained singers Untrained singers

80

e237

Sing (VAS %)

70

60

VAS results The overall VAS average for speaking for all the 20 girls at time 1 was 77.75% (range, 61–100%) and for time 5, 75.44% (range, 36.2–100%) indicating minimal loss of confidence for most girls in their speaking voice over the time of the study. Training also affected the way girls felt about their voices, with the trained group feeling more confident about the way they both spoke and sang (speaking 77.4% and singing 69% for the trained group; 71.47% for speaking and 59.16% for singing for the untrained group). For speak (VAS%), inclusion of voice training status and weight improved the fit of the model significantly with the 2LL reduced from 549.8 to 460.2 (chisquare ¼ 89.6, df ¼ 2, P < 0.0001). The models plotted using the estimates of fixed effects for singers of 40 and 60 kg are shown in Figure 1B. Speak (VAS%) for the trained singers was on average 9.78 points above speak (VAS%) of the untrained singers (P ¼ 0.08). There was an inverse but not statistically significant relation between weight and speak (VAS%) with a deficit of 0.42 points in speak (VAS%) for every kilogram of weight (P ¼ 0.16). For sing (VAS%), inclusion of voice training status in a model with an unstructured covariance matrix improved the fit of the model with the 2LL reduced from 415.8 to 407.7 (chisquare ¼ 8.1, df ¼ 2, P < 0.02). There was no further improvement in fit when either height or weight was included as a covariate in the model. The models plotted using the estimates of fixed effects are shown in Figure 1C. Sing (VAS%) for the untrained singers was on average 13.6 points below that of the trained singers (P ¼ 0.16).

50

40

30 2

3

4

5

Time

FIGURE 1. The effect of time and voice training on maximum frequency (A) Estimate of effects of time, voice training, and weight on the girls’ self-perception, speak (VAS%) (B), and sing (VAS%) (C). different from time 5. Maximum frequency for the trained singers was on average 333.4-Hz frequency points above that of the untrained singers (P ¼ 0.005). Height was not a significant covariate when added to the model (P ¼ 0.95). When range was examined, the pattern of changes over time was very similar to the changes seen for maximum frequency. The range for the trained singers was on average 243.0 Hzfrequency points above that of the untrained singers (P ¼ 0.047), but height was not a significant covariate when added to the model (P ¼ 0.27).

Acoustic analyses Speaking fundamental frequency. The average SFo results at times 1 and 5 were task dependent, but stable. For the reading task, the average SFo for 20 girls at time 1 was 259 Hz [C4] (SD ¼ 51.91; range, 234–297.5 Hz [A#3– zD4]), and at time 5 was 260.3 Hz [C4] (SD ¼ 55.49; range, 228.5–289.8 Hz [zA#3–D4]). For the counting task, the results were consistently lower than for reading—at time 1, 235.3 Hz [A#3] (SD ¼ 30.8; range, 196.9–272.4 Hz [G3–zD4]); and at time 5, 236.8 Hz [A#3] (SD ¼ 32.22; range, 191.7–278.2 Hz [G3–C#4]). However, for seven students, SFo for both counting and reading was higher at time 5 than at time 1. The SFo for reading was higher in pitch at time 5 for an additional two students, and SFo for counting was also higher for five students. Comparative results for 20 girls are given in Figure 2. The relationship between training and SFo was investigated, and SFo values for reading and counting at times 1 and 5 are given in Table 2. Although the average SFo for 20 girls was stable in both reading and counting tasks over the year, the vocally trained girls used consistently higher SFo and wider standard

e238

Journal of Voice, Vol. 25, No. 5, 2011 Speaking fundamental frequency & standard deviation for reading & counting - 20 girls at times 1 & 5 400

Frequency (Hertz)

350

300

250

200

1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c 1-r 5-r 1-c 5-c

150 1

2

3

4

5

6

7

8

9

10

Read 1&5

11

12

13

14

15

16

17

18

19

20

Count 1&5

FIGURE 2. The SFo and standard deviation for reading and counting at times 1 and 5, showing rising SFo values over the year for seven of the 20 girls.

deviation than untrained girls at both times 1 and 5. In Figure 2, the untrained girls are numbers 2, 4, 6, 7, 8, 11, 17, and 20. Mean SFo values for trained and untrained girls are given in Table 2. Vocal range and pitch breaks Vocal range was assessed by analysis of recorded ascending and descending glides. These results were validated through repetition, with data summed across three attempts for descending and three attempts for ascending glides. Analysis of vocal range and pitch break data for all the girls from five data collections showed that pitch breaks occurred in three areas of the vocal range—a low-range pitch break, in the mean average range 132–175 Hz [C3–F3], and a very highrange pitch break in the mean average range 1638–1702 Hz [G#6]. One student developed a medium high pitch break around 943–991 Hz [B-flat5–B5] at time 5, possibly indicating adult registration. Presence of very high and very low pitch breaks pushed the vocal range boundaries beyond the norms expected for this cohort. Over the year, nine of the girls (45%) had either high- or lowrange pitch breaks in their vocal range. At time 1, four girls (20%) had low-range pitch breaks in their glide range (average, 123.75–175 Hz [B2–F3]), and these were present in girls over a wide weight range—35.5–56.4 kg. For the remaining girls, the average lowest glide frequency was 192.7 Hz [G3] (SD ¼ 28.07), and the average maximum glide-note frequency

for all the girls was 1261.6 Hz [D#6] (SD ¼ 319; range, 691– 2078 Hz [F5–B6]). At time 5, five girls had pitch breaks, but the ranges for these varied. Two students had low-range pitch breaks (average, 148–188 Hz [D3–F#3]), two had very highrange pitch breaks (average, 1479–1539 Hz [F#6)], and one girl had a gap at 943–991 Hz (B-flat5–B5). Although trained girls had a higher vocal range, pitch breaks occurred in the voices of both trained (six of 12) and untrained girls (three of eight). High-range pitch breaks were present in the voices of

TABLE 2. Mean SFo for 20 Girls, and for Trained and Untrained Girls at Times 1 and 5 Data Collection

SFo_Read (Hz)

SD (Read)

SFo_Count (Hz)

SD (Count)

All Time 1 Time 5

258.95 260.32

51.91 55.49

235.29 236.81

30.81 32.22

Trained Time 1 Time 5

263.83 263.97

55.06 56.43

236.51 240.70

32.80 35.16

Untrained Time 1 Time 5

251.63 254.84

47.20 54.09

233.48 230.99

27.82 27.82

Elizabeth C. Willis and Dianna T. Kenny

A Study on Voice Training and Changing Weight of 13-Year-Old Girls

both trained and untrained girls (trained: four of 12; untrained: two of eight). For the nine students who participated in the further three data collections (data times 2, 3, and 4), only one girl had a low-range pitch break at time 2 (124–172 Hz [B2–F3]). At time 3, very high-range pitch breaks were present in the voices of three girls (average, 1786–1861 Hz [A6–A#6]), whereas at time 4, one girl had a low-range pitch break (140–152 Hz [C#3–D#3]) and two had high-range pitch breaks (average, 1575–1626 Hz [G6–G#6]). Weight Some aspects of the SFo and voice range results needed clarification, as they were not anticipated. There was a rise rather than fall in SFo pitch for seven of 20 girls between times 1 and 5. In addition, the maximum glide frequency at time 5 was higher at the end of the study, while the minimum glide pitch also rose despite the girls growing an average 3.9 cm over the year. Pitch breaks are related to voice register boundaries—was there a pattern in their presence and placement that indicated development to adult registration? Further investigation included placement of all the data in weight order, and the results can be seen in Figure 3. Table 3 shows increase in weight and height by weight band, and the gradual lowering of SFo. Although the mean SFo remains similar between the three groups (band 1—268.33 Hz [zC4]; band 2—263.5 Hz [C4]; band 3—254.6 [zB3]), there

e239

is a narrower pitch range represented particularly in band 2 (band 1—234–292.7 Hz [A#3–D4]; band 2—242.4–289.8 Hz [zB3–zD4]). The vocal range shows variation at each stage. Band 1 has a wide range of almost two octaves—187.55–1418.67 [zF#3–zF6]. For band 2, the whole range lowered a semitone—172.58–1290.18 Hz [F3–zE6], before band 3 has the narrowest pitch range—191.46–1135.67 Hz [zG3–zC#6]. By band 4, the vocal range is reestablished both lower and higher (168.16–1548.2 Hz [zE3–zG6]). The maximum glide frequency gradually lowers in pitch through bands 1, 2, and 3 before being reestablished higher. The minimum glide frequency gradually lowers between bands 1 and 2 (187.55 Hz [zF#4] and 172.58 Hz [F3], respectively) before rising at band 3 to 191.46 Hz [zG3], and then falling to 168.16 Hz [zE3] for band 4. In band 3, the minimum glide frequency and SFo track each other more closely, averaging only four semitones separation. In comparison, the SFo for band 1 is six semitones above the minimum glide frequency; for band 2, it is seven semitones higher; and for band 4, seven semitones higher. The prevalence of low-range pitch breaks for girls in the band 3 (weight range, 52.6–57.5 kg) may be the result of the close relationship between the minimum glide pitch and the SFo. Weight and self-perception. Analysis of trend VAS% results in relation to weight ranges showed that the girls were most content with their speaking voice in band 1

Mean SFo & glide data ordered by weight 2250 85

2000 1750

75

1250

65

1000 55

750

Weight (kilograms)

Frequency (Hertz)

1500

500 45 250 35

0 1

3

5

7

9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 Data Point

Max Glide Fo

Gap Finish

Gap Start

Min Glide Fo

SFo mean

Weight (kg)

FIGURE 3. The SFo, vocal range, and pitch breaks—all data arranged in weight order, and showing contraction of vocal range particularly for girls in the weight range 52.6–57.5 kg.

e240

96–212 [G2–zG#3] 162.3–174.4 57.5–87.2 Range

227.1–267.8 [zA3–zC4]

891–1634.46 [A5–zG#6] 1548.2 [zG6] Yes, in medium high, high, and low range as wider vocal range returns. 935–2094.01 [A#5–C7] 72–239.83 [D2–zA#3] 168.16 [zE3] 160.3–171.5 168.61 52.6–57.5 62.88 Range Band 4: Mean

235.3–289.3 [A#3–zD4] 248.87 [B3]

Yes, low-range, as vocal range limited. 691.38–1694.4 [F5–zG#6] 1135.67 [zC#6] 154–171.4 165.95 47.7–52.4 54.51 Range Band 3: Mean

242.4–289.8 [zB3–zD4] 103–255.81 [G#2–zC4] 254.6 [zB3] 191.46 [zG3]

Yes, more low-range pitch breaks as vocal range starts to contract. 124–267.8 [B2–zC4] 172.58 [F3] 144–159.8 163.25 35.5–47.5 50.37 Range Band 2: Mean

234–292.7 [A#3–D4] 263.5 [C4]

935–2066.43 [A#5–zC7] 1290.18 [zE6]

Yes, both high and low-range pitch breaks. 1418.67 [zF6] 187.55 [zF#3] 154.00 42.75 Band 1: Mean

268.33 [zC4]

Max Glide Fo (Hz) Min Glide Fo (Hz) SFo Mean (Hz) Height (cm) Weight Range Weight (kg)

TABLE 3. Mean Values for Weight, Height, SFo, Vocal Range, and Pitch Breaks for Four Groups Arranged by Weight

Vocal Range and Pitch Breaks

Journal of Voice, Vol. 25, No. 5, 2011

(35.5–47.5 kg), registering 79.3% (VAS) for speech, and 65.05 (VAS%) for singing; and least satisfied with their speaking and singing voice in band 3 (52.6–57.5 kg), registering 66.92% (VAS) for speech, and only 54.3 (VAS%) for singing. Girls in band 2 (42.7–52.4 kg) were most happy with their singing voice, registering 74.12 (VAS%) for singing and 77.07 (VAS%) for their speaking voice. For band 4, girls (<57.5 kg), speaking VAS% was 72.45 and singing (VAS%) 61.28. Individual subjects over five data collection points Figure 4 shows the longitudinal case data for four girls who have been selected to display the range of individual variability within the group data. Subject 15 (Figure 4A) weighed the least of the girls in the study. She was most confident in speaking at time 3 (87.4% on the VAS), but was least confident about her speaking voice at time 4 (52.6%). She was confident in singing at time 2 (80.3%), but this measure had dropped to 41.6% (VAS) at time 5. Subject 12 (Figure 4B) demonstrates the pitch-break instability and lower pitch range that occurred around 47 kg after rapid weight change. She was most confident in her speaking voice and singing voice at time 2 (speaking, 85.1%; singing, 83.2%), and least confident at time 3 (speaking, 71.8%; singing, 59.8%). Subject 7 (Figure 4C) was the girl with least training and least confidence in her voice use, feeling most confident in her speaking voice at time 1 (VAS speak: 62%) and least confident at time 3 (30.3%). In singing, she was most confident at time 5 (36.5%), and least confident at time 2 (21.9%). Subject 14 (Figure 4D) was one of the heaviest girls whose pitch-break data seems to indicate that she had reached adult registration by time 5. She was one of the two girls to lose weight during the study, both girls being among the heaviest in the study. She was the most highly trained of the girls, and was very confident in both her speaking and singing, with maximum VAS measure for speaking at time 3 (99.2%), and for singing at time 2 (96.3%). She was least confident at time 1 for speaking (68%), and time 5 for singing (78.1%). DISCUSSION Speaking fundamental frequency The stability of mean SFo results between times 1 and 5 in this study was not anticipated, as all previous studies had indicated that there would be a fall in SFo pitch. This was the trend between girls in different weight ranges, but not between SFo values from the beginning to the end of the study. The average SFo for reading at time 1 was 259.7 Hz [C4] (SD ¼ 51.91; range, 234–297.5 Hz [zA#3–D4]) and at time 5, 260.3 Hz [C4] (SD ¼ 55.49; range, 228.5–289.8 Hz [zA#3–D4). These results agree with Duffy5 for reading for less mature students 260 Hz, and 245 Hz for more mature 13-year olds and in the upper range of acceptable limits given by Wilson.20 The results for counting were also stable, but consistently lower than for reading—at time 1, 235.3 Hz [A#3] (SD ¼ 30.8; range, 196.9– 272.4 Hz [G3–C#4]); and at time 5 236.8 Hz [A#3] (SD ¼ 32.22; range, 191.7–278.2 Hz [G3–C#4]). This is consistent with Vuorenkoski et al.22 Variation between adult reading and counting SFo results was noted by Schultz-Coulon40

Elizabeth C. Willis and Dianna T. Kenny

A Study on Voice Training and Changing Weight of 13-Year-Old Girls

e241

FIGURE 4. Longitudinal data for age, weight, height, SFo, vocal range, and pitch breaks for subject 15 (A), subject 12 (B), subject 7 (C), and subject 14 (D). who showed that reading in adults produced the highest SFo, counting the next lowest, and spontaneous speech the lowest SFo. Results from this study indicate this is also true for reading and counting among 13-year-old girls, and suggests that a wider SFo range is acceptable for this age group. Only the Williams et al8 study used spontaneous speech and a reading task, and their results were lower—218.3 Hz [A3] for premenarcheal and 206.3 Hz [G#3] for postmenarcheal girls. Their study also involved a wider age range—11–15-year olds, and the results from all girls were averaged—a technique that seems to produce a lower mean SFo than when 13-year olds are studied in isolation, as it also occurred in the first Gackle study23,25 where only 43% of cohort were 13 years. In summary, the task requested to assess SFo strongly influences the result, with the difference in this 1-year longitudinal study of 13-year olds varying the mean SFo by two semitones (C4–A#3). Reading consistently produced a result in the high end of the acceptable range as given by Duffy5 and Wilson,20 whereas counting produced a result within the range given by Eguchi and Hirsch,21 and Vuorenkoski.22

Vocal ranges and pitch breaks Gackle’s framework for vocal change in female adolescent voices was developed for application to the general population, rather than among highly trained singers.24,25 As such, it is

applicable to this longitudinal study, which observed the vocal changes of an average class of 13-year-old girls over 1 year. In this study, the vocal range development was as Gackle described—gradual contraction, then reestablishment of full vocal range, but in this small longitudinal study, this occurred in certain weight ranges. Range contraction was most evident among girls in the weight range 52.6–57.5 kg. This was observable despite the technique of glide analysis used to assess MPFR range, compared with Gackle who assessed singing range and tessitura by asking the girls to sing an ascending, then descending scale to [ah]. We anticipated that these different techniques would give different results. Correlations between height and minimum glide pitch were also considered, as Groom41 had observed that these measures were linked in boys’ voices, but this was not evident with the girls in this study. At time 1, the average height of the girls was 159.5 cm and the minimum glide frequency was 174.7 Hz [F3]. At time 5, the average height was 163.4 cm, and the average minimum frequency was 183 Hz [F#3]. Figure 3 indicates that higher pitch at time 5 may be due to vocal instability of girls’ voices as they move through change, particularly over the weight range 48–58 kg. If the lightest girls in the 35.5–47.5 kg range are compared with girls heavier than 57.5 kg, then the average minimum glide frequency drops from 187.55 Hz [F#4] to 168.16 Hz [zE3].

e242 Pitch breaks were of interest as indicators of register development. They occurred in the vocal glides of 45% girls over the course of the study, and were found in three areas of the vocal range—the very low range (average, 131.84–174.94 Hz [C3– F3]; range, 100–208 Hz [G#2–G#3]), and the very high range (1637.993–1701.947 Hz [zG6–G#6]; range, 1471–1954 Hz [F#6–B6]). Both very high and very low pitch breaks were present in girls in weight bands 1, 2, and 4, but only low pitch breaks were present in group 3 because of range contraction at this stage of vocal change. When present, they pushed the minimum and maximum glide boundaries beyond mean averages, and in this study, were synchronous with times of rapid weight change, to changes in SFo, and particularly when high pitch breaks were present, lowered the VAS value of how the girls felt about their vocal performance. For subject 14, a low pitch break (z100–140 Hz [G2–C#3]) at time 1, followed by high pitch breaks at times 3 and 4 (z1650– 1742 Hz [G#6–A6]), then the establishment of a pitch break at 943–991 Hz [zA5] showed a strong octave relationship with the other pitch breaks. Nair42 gives A#5–B5 as the lift point from head voice into the flageolet range, which would indicate the development of adult vocal registers in subject 14. Willis and Kenny43 observed a similar pattern in boys, and this study indicates it may also occur in girls, but in extremes of range that do not affect function. This vocal instability was often greatest after rapid changes in weight. Gackle25 observed pitch breaks in girls’ voices, with those girls potentially alto in voice type. Low- and high-range pitch breaks occurred in both trained and untrained voices, so that the more intense voice use caused over a wider vocal range involved in voice training did not seem to be a factor in pitch-break development in this study. Pitch breaks possibly reflect changes due to growth of the vocal tract, and length and mass changes of the vocal folds because of the development of the mature ligamentous layering structure.2,3,34,44 In this study, the peak of pitch-break activity was identifiable among the girls in the weight range 56–62 kg. Fuchs et al15 found a link between training, and the highest note of the vocal range, measuring on average 2.7 semitones higher for those chorally trained compared with untrained singers, and 5.8 semitones above for those students who had been individually trained. This study found a difference not only in the maximum frequency of the singing range, but also in SFo depending on the stage of change. Values in this study for the mean maximum frequency for trained singers at time 1 was 1404.64 Hz [zF6] and about the same at time 5– 1436.32 Hz [zF6]. For untrained singers, these values were 1047.04 Hz [C6] (also five semitones difference) at time 1, but 1276.1 Hz at time 5 [D#6] (only two semitones difference). This study indicated that the voice range of girls contracts when they weigh between 52.6 and 57.5 kg. This indicates that these results vary depending on the stage of the subjects’ vocal change, especially if they are moving from band 3 weight range (52.6–57.5 kg) to band 4 (<57.5 kg). Because of this, the effects of changing voice also needed to be considered when assessing the effects of vocal training in young female voices.

Journal of Voice, Vol. 25, No. 5, 2011

Weight and height This study investigated whether there was synchonicity between changes in weight and height, and changes in vocal development in the female adolescent voice. Wilson20 had noted that size affected SFo, but previous research had indicated that weight and height affected different vocal parameters. As a result, these measures were assessed independently rather than using body mass index. With height, Decoster et al28 noted the wider vocal range of her research cohort, which she attributed to their greater height, and hence longer vocal tracts. Groom41 noted that greater height synchronized with lower minimum voice range, and this was confirmed in this study, where girls in the 144– 159.8 cm range had a mean minimum glide pitch of 187.55 Hz [zF#3], whereas those in the 162.3–174.4 cm range had a minimum glide pitch of 168.16 Hz [zE3]. This study also noted variations to vocal range synchronous to certain weight ranges. Pitch breaks also occurred around times of rapid weight change (both gain and loss), rather than around height gain.

Self-perception of voice Voice training improved girls’ self-confidence in voice use. When asked about their speaking voice, trained singers averaged 9.78 (VAS%) points above those for untrained singers, displaying a greater satisfaction and confidence in their vocal performance. All students lost confidence in their speaking voices at data point 2, but trained singers less so than untrained singers. For singing, the results for untrained singers averaged 13.6 (VAS%) points below trained singers in their level of confidence, and taller, heavier girls were happier with both their singing and speaking voices. The range of self-confidence was particularly obvious when comparing subjects with no voice training, who tended to be lacking in confidence in their voice use, and subject 14, who had been a member of a choir from the age of 5 years, had received individual singing training from 7 years of age, and was very confident in her voice use.

Conclusions Limitations. Various features and limitations of this study are acknowledged and need to be taken into consideration. Differences in data collection techniques, for example, can produce different results. Gackle’s very practical use of singing vocal ranges and tessituras served a different function to the use of glide technique to elicit MPFR in this study. We had a relatively small sample of girls in this study, which was designed to observe acoustic and perceptual vocal changes in girls from an average class of 13-year olds over 1 year. Collection of data on the menarcheal status of the girls was considered beyond the scope of this study due to its educational setting in high schools, as opposed to a medical setting. The value of this study lies in its longitudinal nature, the opportunity to observe the variability of vocal change among a given population, and participants’ responses to their voice change. The results from this longitudinal study, presenting both group trends and case studies, highlight the extent of the individual variability that occurs in the context of group trends, and shows

Elizabeth C. Willis and Dianna T. Kenny

A Study on Voice Training and Changing Weight of 13-Year-Old Girls

that girls in all phases of voice change will be present in a group of 13-year olds. It is hoped that this study will help both teachers and clinicians to guide these young girls through this time of maximum vocal instability. REFERENCES 1. Kahane JC. A morphological study of the human prepubertal and pubertal larynx. Am J Anat. 1978;151:11–19. 2. Kahane JC. Growth of the human prepubertal and pubertal larynx. J Speech Hear Res. 1982;25:446–455. 3. Hirano M, Kurita S, Nakashima T. Growth, development and aging in human vocal folds. In: Bless DM, Abbs JH, eds. Vocal Fold Physiology: Contemporary Research and Clinical Issues. San Diego, CA: College Hill Press; 1983:22–43. 4. Michel JF, Hollien H, Moore P. Speaking fundamental frequency characteristics of 15, 16 and 17 year-old girls. Lang Speech. 1965;9:46–51. 5. Duffy RJ. Fundamental frequency characteristics of adolescent females. Lang Speech. 1970;13:14–24. 6. Whiteside SP, Hodgson C, Tapster C. Vocal characteristics in preadolescent and adolescent children: a longitudinal study. Logoped Phoniatr Vocol. 2002;27:12–20. 7. Vennard W. Singing: The Mechanism and the Technic. New York: Carl Fischer, Inc; 1967. 63. 8. Williams B, Larson GW, Price DW. An investigation of selected female singing- and speaking-voice characteristics through comparison of a group of premenarcheal girls to a group of post-menarcheal girls. J Singing. 1996;52: 33–40. 9. Welham NV, Maclagan MA. Vocal fatigue in young trained singers across a solo performance: a preliminary study. Logoped Phoniatr Vocol. 2004;29: 3–12. 10. van Mersbergen MR, Verdolini K, Titze IR. Time-of-day effects on voice range profile performance in young, vocally untrained adult females. J Voice. 1999;13:518–528. 11. Meurer EM, Garcez V, Eye Corleta H, Capp E. Menstrual cycle influences on voice and speech in adolescent females. J Voice. 2009;23:109–113. 12. Barlow CA, Howard DM. Voice source changes of child and adolescent subjects undergoing singing training—a preliminary study. Logoped Phoniatr Vocol. 2002;27:66–73. 13. Barlow C, Howard DM. Electrolaryngographically derived voice source changes of child and adolescent singers. Logoped Phoniatr Vocol. 2005; 30:147–157. 14. Welch GF, Howard DM. Gendered voice in the Cathedral choir. Psychol Music. 2002;30:102–120. 15. Fuchs M, Meuret S, Geister D, Pfohl W, Thiel S, Dietz A, Gelbrich G. Empirical criteria for establishing a classification of singing activity in children and adolescents. J Voice. 2008;22:649–657. 16. Fuchs M, Meuret S, Thiel S, Ta¨schner R, Dietz A, Gelbrich G. Influence of singing activity, age and sex on voice performance parameters, on subject’s perception and use of their voice in childhood and adolescence. J Voice. 2009;23:182–189. 17. Monks S. Adolescent singers and perceptions of vocal identity. Br J Music Educ. 2003;20:243–256. 18. Howard DM, Welch GF. Female chorister development: a longitudinal study at Wells, UK. Bulletin of the Council for Research in Music Education. 2002;153:4. 19. Darley F, Spriesterbach DC. Diagnostic Methods in Speech Pathology. New York: Harper & Row, Publishers, Inc; 1978. 20. Wilson DK. Voice Problems of Children. 3rd ed. Baltimore, MD: Williams and Wilkins; 1987. 21. Eguchi S, Hirsch IJ. Development of speech sounds in children. Acta Otolaryngol 1969;(suppl 257):1–51.

e243

22. Vuorenkoski V, Lenko HL, Tjernlund P, Vuorenkoski L, Perheentupa J. Fundamental voice frequency during normal and abnormal growth, and after androgen treatment. Arch Dis Child. 1978;53:201–209. 23. Gackle L. The effect of selected vocal techniques for breath management, resonation and vowel unification on tone production in the junior high school female voice (dissertation for PhD). Florida, USA: University of Miami; 1987. 24. Gackle L. The adolescent female voice: characteristics of change and stages of development. Choral J. 1991;31:17–25. 25. Gackle L. Understanding voice transformation in female adolescents. In: Thurman L, Welch G, eds. Bodymind and Voice-Foundations of Voice Education. Iowa City, IA: The VoiceCare Network; National Center for Voice and Speech; Fairview Voice Center; Centre for Advanced Studies in Music Education; 2000:739–744; 814–820. 26. Gackle L. Finding Ophelia’s voice: the female voice during adolescence. Choral J. 2006;47:28–37. 27. Reid C. A Dictionary of Vocal Terminology. An Analysis. New York: Joseph Patelson Music House Ltd; 1983. 372. 28. Decoster W, Guesquiere S, van Steenberge S. Great talent, excellent voices—no problem for pubertal girls? Logoped Phoniatr Vocol. 2008; 33:104–112. 29. Titze IR. Principles of Voice Production. Englewood Cliffs, NJ: Prentice Hall Inc; 1994. 30. McCoy S. Your Voice—An Inside View. Princeton, NJ: Inside View Press; 2006. 31. McKinney JC. The Diagnosis and Correction of Vocal Faults. Nashville, TN: Broadman Press; 1983. 32. Sˇvec J, Schutte HK, Miller DG. On pitch jumps between chest and falsetto registers in voice: data from living and excised human larynges. J Acoust Soc Am. 1999;106(3 Pt1):1523–1531. 33. Titze IR. A framework for the study of vocal registers. J Voice. 1988;2: 183–194. 34. McAllister A, Sederholm E, Sundberg J. Perceptual and acoustic analysis of vocal registers in ten-year-old children. In: McAllister A, ed. Acoustic, Perceptual and Physiological Studies of Ten-Year-Old Children’s Voices. Stockholm: Kongl Carolinska Medico Chirurgiska Institutet; 1997:15. Study III. 35. Phillips K. Teaching Kids to Sing. New York: Schirmer Books; 1992. 36. Keidar A, Hurtig RR, Titze IR. The perceptual nature of vocal register change. J Voice. 1987;1:223–233. 37. Harries M, Hawkins S, Hacking J, Hughes I. Changes in the male voice at puberty: vocal fold length and its relationship to fundamental frequency of the voice. J Laryngol Otol. 1998;112:451–454. 38. Williams J, Welch G, Howard D. An exploratory baseline study of boy chorister vocal behaviour and development in an intensive professional context. Logoped Phoniatr Vocol. 2005;30:158–162. 39. Baken RJ, Orlikoff RF. Clinical Measurement of Speech and Voice. San Diego, CA: Singular Thomas Learning; 2000. 40. Schultz-Coulon HJ. Bestimmung and Beurteilung der individuellen mittleren Sprechstimmlage. Folia Phoniatr. 1975;27:375–386. 41. Groom M. A descriptive analysis of adolescent male voices during the summer time period. In: Rufola EM, ed. Proceedings: Research Symposium on the Male Adolescent Voice. Buffalo, NY: State University of New York at Buffalo Press; 1984:80–85. 42. Nair G. Voice Tradition and Technology: a State-of-the-art Studio. San Diego, CA: Singular Publishing Group Inc; 1999. 43. Willis EC, Kenny DT. Relationship between weight, speaking fundamental frequency, and the appearance of phonational gaps in the adolescent male changing voice. J Voice. 2008;22:451–471. 44. Stathopoulos ET. A review of the development of the child voice: an anatomical and functional perspective. In: White P, ed. Child Voice. Stockholm, Sweden: KTH Voice Research Centre; 2000:1–12.