Contact Quotient of Female Singers Singing Four Pitches for Five Vowels in Normal and Pressed Phonations

ARTICLE IN PRESS Contact Quotient of Female Singers Singing Four Pitches for Five Vowels in Normal and Pressed Phonations Kendrich Graemer Ong Tan, Jyväskylä, Finland Summary: Objectives. The present study aimed to investigate the contact quotient (CQ) values of breathy, normal, and pressed phonation types in four different sections of the female singing range. Methods. Electroglottography (EGG) and acoustic signals were recorded from 10 female singing teachers. Five vowels were sung for 1–3 seconds each, in three phonation types—normal, breathy, and pressed, in four pitches representing registration change points in the singing range. CQ values were automatically generated from the EGG signal using VoceVista at 35% threshold level. Sound pressure levels were checked in Praat. Unianova and correlations were performed using an SPSS program. Results. CQ values of female participants in the study yielded ranges of 0.25–0.62 in normal and 0.34–0.73 in pressed. Normal and pressed CQ differed significantly from each other at P < 0.00. Breathy samples were not viable for analysis. A concentration of CQ values from 0.5 to 0.6 of both pressed low and pressed break samples was noted, but CQ values across the pitch range showed no significant trend. Conclusions. Normal and pressed phonation CQ values beyond the speaking pitch varied among the subjects. Pressed phonation CQ values were mostly higher, but the values were only relative to the corresponding normal phonation on the same pitch. Other measurements may be more suitable for measuring vocal fold impact stress in higher frequencies, thus, distinguishing normal from pressed singing. Key Words: Contact quotient–Singing voice–Female–Pressed phonation–Voice range.

INTRODUCTION Phonation types have been physiologically differentiated using contact quotient (CQ) values derived from electroglottographic procedures.1 CQ has been used to evaluate vocal fold adduction in earlier clinical and nonclinical voice studies. CQ is the ratio of vocal fold closure time to the entire vibration cycle,1–3 and its complement is the open quotient (OQ). Electroglottography (EGG) has been shown to be an effective instrument for noninvasive investigation of vocal fold contact and CQ during sustained vowel phonation and singing.4–6 Imaging studies have cross-checked the findings on vocal fold behavior of the less invasive EGG.7–12 There is difficulty, however, in capturing clear glottograms from breathy phonation13 and very high fundamental frequencies.14 Publications on breathy singing measurements were not found. In measuring the CQ values of phonation types classified into breathy, flow, and pressed,15 EGG signals of 30 trained voice users analyzed in VoceVista (VoceVista, Roden, The Netherlands), at 35% threshold level by Kankare et al4 yielded mean CQ values of 0.38 for breathy, 0.47 for normal, and 0.58 for pressed, in habitual speaking pitch. When fundamental frequency is restricted or preplanned, CQ values may deviate from the predicted mean especially as vocal quality changes in the higher ranges.6,16 OQ values—1 minus CQ—have been reported to fall between 0.3 and 0.8 in speech and always greater than 0.5 above the register break frequencies, suggesting a change in laryngeal Accepted for publication February 22, 2017. From the Institute of Education, University of London, Jyvaskyla, Finland. Address correspondence and reprint requests to Kendrich Graemer Ong Tan, Institute of Education, University of London, Taitoniekantie 9 E 813, Jyväskylä 40740, Finland. E-mail: [email protected] Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■ 0892-1997 © 2017 The Voice Foundation. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jvoice.2017.02.014

mechanism.17,18 Female speakers’ sound pressure level (SPL) increased as OQ increased around 70–80 dB, 19 supporting correlation of SPL and OQ in thyroarytenoid (TA) dominant mechanism (M1).18 This may imply that CQ value parameters suggested apply to fundamental frequencies below 392 Hz, before cricothyroid (CT) dominant phonation (M2) takes over. CQ is expected to diminish as the vocal folds stretch for higher frequency vibration, when they lose depth or contact surface.17 Henrich et al18 elaborated that in CT dominant mechanism (M2), OQ inversely correlates with fundamental frequency in both men and women, most especially in countertenors. Other findings suggest that no significant CQ variance is noted in habitual speaking pitch between men and women, but significant differences may be noted in high-pitch sustained phonations,20 which is lower in men than in women.17 The head and whistle register of sopranos may highlight this gender difference.21 However, research investigating phonation types across the entire singing range was not found. Speaking does not require specific fundamental frequencies to be produced, and phonation behavior can be quite different in singing. Vocologists theorize that the physiological limit of the TA muscle causes an involuntary transition between 294 Hz (D4) and 392 Hz (G4).22 As pitch rises, transitioning from TA to CT dominant phonation is said to be effective in avoiding pressed phonation.23 Researchers assert that “[a] raised larynx is often associated with a more hyperfunctional type of phonation. Thus, a narrowing of the pharynx would easily (if not necessarily) lead to a firmer glottal adduction.”24 However, it has been a practice in nonclassical singing to retain TA dominant phonation past the usual pitch limit, in higher volumes and higher larynx positions,25,26 to allow for speech-like qualities.27 Blues, rock, heavy metal, and gospel singers who, despite sounding unhealthy to some listeners, outlived their predicted career and voice limit and “sing better

ARTICLE IN PRESS 2

Journal of Voice, Vol. ■■, No. ■■, 2017

TABLE 1. Participant Demographics Age

Years Since Training Started

Years Teaching

ACN AKI

42 35

17 19

10 10

Classical Folk

INN ANN ASI ALR

30 31 47 48

10 17 32 29

4 5 15 18

Pop Jazz Pop Jazz Classical Music Theater

OLE TIR HOJ AJI

30 37 31 27

14 18 15 12

7 16 2 6

Pop Jazz Classical Pop R&B Contemporary Commercial Music

Participant

Genre Training

and better throughout the years, regardless of how ‘damaging’ their singing sounds,”28 may prove that narrowed-pharynx singing and firm glottal adduction do not automatically equate to injurious vocal habits. Since vocal injury occurs in both classical singing (expanded pharynx) and belting (narrowed pharynx), limits in healthy and pressed singing may need further clarification. Pressed phonation results from the hyperadduction of the vocal folds and increased subglottal pressure.22,29 Hyperadduction increases the risk of vocal injury due to excessive mechanical stress incurred by the vocal folds with prolonged application.30,31 Pressed phonation has been reported to exhibit long vocal fold closure5 and high-impact stress or force of collision.32,33 EGG studies on pressed phonation measurements in speaking exist,4,34 but there are fewer on pressing in singing. The present study aimed to investigate CQ values and differences among phonation types, especially when pitches are to be “hit” (targeted before phonation) in four different sections of the female singing range. What CQ values result when an experienced singer presses intentionally, when he or she sings in the healthiest way possible, and when he or she sings breathily? What are the quantifiable characteristics of intentionally pressed singing? METHODS Subjects Invitations were sent to singing teachers, and 10 female singing teachers with no known vocal pathology during the time of testing volunteered to record samples. Singing teachers were sought to ensure possession of “the skill to phonate in distinct registers (chest and falsetto) with independent control of the degree of glottal adduction.13” The subjects had a mean age of 36, standard deviation of 7.5, and teaching experience of 2–18 years. They came from both classical and nonclassical backgrounds, and some, with studies in voice science (Table 1). Instruments All recordings took place in a sound-treated studio. Audio signals were captured by a Brüel & Kjaer Mediator 2238 (Brüel & Kjær

Degree and Related Studies B Music, M Vocology MM Music Education Complete Vocal Technique teacher B Music , Vocology units BM Singing Pedagogy, MA Music Education ongoing B Music, Vocology units MA Music Education, Music Theater Estill teacher BM Theater Music and Drama MM Singing Pedagogy, PhD Vocology ongoing B Music Pedagogy B Music

Sound & Vibration Measurement A/S (HQ), Nærum, Denmark) microphone and sound level meter at a distance of 40 cm from the subject’s mouth, and calibrated using 3 kHz audio from a sound generator (Boss TU-120, Roland Corporation U.S., Los Angeles, California). EGG signals were recorded from the two channels of a dual-channel Glottal Enterprises equipment, through electrodes placed on the thyroid cartilage. Signals were transmitted through a Roland Quad-Capture sound card (Roland Corporation U.S., Los Angeles, California). Recordings were done in Sound Forge 7.0 software (Sony Creative Software Inc., Middleton, WI) with a sampling rate of 44.1 kHz and amplitude quantization of 16 bits. Vocal samples Participants were asked to pick four pitches in their vocal range, representing their lowest, their speaking, their break (or upper modal limit), and their high range. This is to account for the comfort range or tessitura unique to each singer,35 and the muscular adjustments from 294 to 392 Hz.14,22 According to Vurma,36 phonatory strategies may depend on the “pitch region’s position along the whole voice range of the singer.” Table 2 shows the

TABLE 2. Participant Pitch Choices/Limits and Their Equivalent in Hertz Participant ACN AKI INN ANN ASI ALR OLE TIR HOJ AJI

PL E3 Eb3 E3 Eb3 B2 C#3 C3 D3 F3 Eb3

PS 165 156 165 156 124 139 131 147 175 156

G3 A3 C4 A3 G3 E3 Bb3 G3 G3 F#3

PB 196 220 262 220 196 165 233 196 196 185

F4 B4 G4 G4 E4 F4 F4 F4 E4 E4

PH 349 494 392 392 330 349 349 349 330 330

C6 A5 D5 Ab5 Bb5 A5 F5 G5 B5 A5

1047 880 587 831 932 880 740 784 988 880

ARTICLE IN PRESS 3

107.9 101.3 91.19 97.55 94.57 93.98 101.6 102.5 105.3 94.22 0.558

0.562 0.426

Abbreviations: CQ, contact quotient; SPL, sound pressure level.

Pressed

0.536 0.446 0.624 0.342 0.588 0.422 0.552 0.54 0.39 0.546 104.7 94.37 82.52 97.16 100.2 83.35 87.26 95.74 97.63 94.61 0.474 0.253 0.376 0.286 0.398

91.62 95.79 87.91 84.58 80.54 89.2 88.9 88.19 88.36 86.59 0.562 0.576 0.53 0.556 0.562 0.514 0.574 0.538 0.51 0.727 82.36 85.02 80.74 82.15 75.11 70.48 74.62 82.41 78.15 73.59 0.456 0.474 0.534 0.328 0.318 0.452 0.496 0.496 0.466 0.62 81.93 86.16 82.79 76.33 79.02 78.21 78.57 81.69 75.02 79.49 0.626 0.58 0.634 0.576 0.602 0.586 0.474 0.504 0.462 0.598 75.75 70.14 72.74 72 70.69 70.31 69.28 75.3 68.92 69.03 0.546 0.436 0.382 0.508 0.354 0.45 0.372 0.476 0.394 0.483 73.06 75.02 77.23 66.65 69.51 71.77 75.06 70.84 70.11 71.04 0.596 0.578 0.54 0.496 0.524 0.544 0.576 0.572 0.494 70.73 69.7 66.18 64.94 66.29 64.9 68.48 65.83 65.36 68.35

SPL

Normal

0.588 0.436 0.454 0.488 0.48 0.378 0.412 0.502 0.33 0.385 ACN AKI INN ANN ASI ALR OLE TIR HOJ AJI

SPL CQ SPL

Pressed

CQ SPL

Normal

CQ SPL

Pressed

CQ SPL

Normal

CQ SPL

Break Speaking Pressed

CQ

Normal

High CQ

RESULTS CQ values of female participants in the study ranged from 0.25 to 0.62 in normal and from 0.34 to 0.73 in pressed. Mean CQ values of all 10 subjects are shown in Table 3 with their corresponding SPL. Breathy samples rarely yielded clear glottograms, so they had to be excluded in the analyses. Intraclass correlation of extracted CQ means yielded 0.548 at a significance of 0.000. Normal and pressed CQ differed significantly from each other at P < 0.001. CQ value tabulation had no significant trend across the pitch range, but graphs showed (Figure 1) concentration of CQ values from 0.5 to 0.6 of both pressed low and pressed break samples. Pressed CQ values in nonspeaking frequencies are mostly higher than normal, but values seem to depend on the corresponding normal value. CQ values were significantly different in each pitch as well at P < 0.001. Normal and pressed SPL also differed at a significance of 0.000. Average difference between normal and pressed was 7.1 dB, but this was less predictable in the highest range of the participants (Figure 2). Maximum mean differences in SPL were found in speaking and break frequencies. SPL in each pitch range differed from each other at P < 0.001. SPL means consistently

TABLE 3. Summary of CQ and SPL Means in Two Phonation Types and Four Sections of the Singing Range

EGG and acoustic signal analyses Recordings were run in VoceVista software at 35% CQ criteria level, as the default of the software and the threshold level most consistent with perceptual evaluations in an earlier study.4 CQ values were assessed per vowel, checking the alignment of markings on the glottograms. EGG phase correction was not performed since participants were all female and VoceVista aligns the EGG and acoustic signals. CQ measurements were tallied in excel, categorized first by pitch, then by vocal qualities. SPLs of each vowel were counterchecked in Praat. Statistical analyses were carried out using SPSS Statistics (IBM SPSS Statistics for Windows, Version 22.0., IBM Corp., Armonk, NY). Intraclass correlation was calculated to determine reliability. Univariate analysis of variance was used to test the differences in the CQ means of the 10 cases. Correlation was computed to determine trends related to phonation type and singing range.

CQ

pitches chosen by the singers. Prolonged vowel voicing was chosen for this study, as in most EGG analyses.4 Each recording was done in a single closed door session, with further instructions delivered only through a microphone. Although certain vowels and frequency changes could raise or depress the larynx, electrodes were not adjusted further after attachment. On the above pitches, subjects were asked to sing five vowels—i, e, a, o, u—for 1–3 seconds each, in each phonation type—normal, breathy, and pressed. Participants started with one pitch and one phonation type, singing all five vowels in succession. Then, they proceeded to the next pitch in the same phonation type or to another phonation type on the same pitch. EGG and acoustic signal strength were monitored through the Sound Forge window. In the event that the singer’s voice cracked or the signal exceeded the set maximum volume, particularly in the high ranges, the entire set of five was repeated after recalibration. Glottograms were not checked prior to ending the sessions.

SPL

CQ Values of Breathy, Normal, and Pressed Phonation Types

Low

Kendrich Graemer Ong Tan

ARTICLE IN PRESS 4

Journal of Voice, Vol. ■■, No. ■■, 2017

FIGURE 1. Graph of normal versus pressed contact quotient values in four pitches.

FIGURE 2. Sound pressure levels in decibels per pitch range in normal and pressed phonations. increased as pitch rose, in both phonation types, across all 10 subjects. SPL and CQ correlations yielded coefficients of 0.617 in low pressed, 0.576 in speaking pressed, and below 0.36 in other sections. However, differences in vocal tract adjustments and formants, which usually occur above the register break, were not explored in this article. DISCUSSION This study attempted to investigate the measurements of three phonation types across the singing range of 10 female singing teachers. EGG signals and audio signals were simultaneously recorded as subjects sang five vowels in three phonation types, in four pitches representing their low, high, speaking, and register break ranges. All audio samples yielded stable SPLs, but the CQ values were

not as easy to obtain. Breathy samples were collected, but they did not generate clear glottograms. As previously concluded by Herbst and Ternstrom,13 EGG measurement seems unsuitable for breathy phonation. Hence, breathy phonation was not included in the analyses. Moreover, artifacts and waveform changes caused by vertical movement of the larynx above the speaking pitch13 were not noted during the single session recordings. In cases where extraction of an accurate CQ value was not possible for some vowels (Table 4), means were calculated from the available CQ values. Three CQ means were labeled “missing” in SPSS. One participant generated distorted glottograms in all vowels of her low pressed sample, possibly due to electrode placement issues as the larynx moved vertically, so the values of all vowels were considered inaccurate and not suitable for analysis. Two participants’

Kendrich Graemer Ong Tan

TABLE 4. Missing Values in CQ Table Normal

Pressed

INN

ANN

ASI

ALR

OLE

TIR

HOJ

AJI

ACN

AKI

INN

ANN

ASI

ALR

OLE

TIR

HOJ

0.34 0.39 0.33 0.48

0.61 0.58 0.58 0.61 0.60

0.64 0.56 0.54 0.56 0.59

0.53 0.52 0.53 0.55 0.57

0.49 0.52 0.51 0.49 0.47

0.54 0.58 0.54 0.50 0.46

0.57 0.53 0.52 0.57 0.53

0.52 0.47 0.58 0.59

0.56 0.56 0.58 0.57 0.59

0.60 0.57 0.49 0.45 0.36

0.60 0.64 0.64 0.62 0.64 0.59

0.58 0.60 0.57 0.58 0.58 0.57

0.54 0.65 0.64 0.63 0.64 0.61

0.50 0.57 0.57 0.58 0.59 0.57

0.52 0.65 0.56 0.58 0.60 0.62

0.54 0.57 0.58 0.58 0.59 0.61

0.54 0.45 0.55 0.44 0.49 0.44

0.57 0.48 0.56 0.53 0.49 0.46

0.49 0.50 0.52 0.47 0.43 0.39

0.63 0.60 0.50 0.51 0.59 0.61

0.58 0.57 0.57 0.56 0.60 0.58

0.63 0.48 0.53 0.60 0.50 0.54

0.58 0.54 0.57 0.55 0.58 0.54

0.60 0.55 0.57 0.56 0.60 0.53

0.59 0.47 0.56 0.50 0.56 0.48

0.47 0.53 0.61 0.58 0.59 0.56

0.50 0.53 0.53 0.56 0.54 0.53

0.46

0.53 0.52

0.56 0.33 0.34 0.34 0.33 0.37

0.56 0.60 0.58 0.60 0.59 0.57

0.51 0.44 0.37 0.52 0.41 0.37

0.57 0.56 0.58 0.54 0.54 0.54

0.54 0.57 0.52 0.55 0.54 0.52

0.51

0.41

0.73 0.59 0.55 0.53 0.56 0.50

0.34

0.59

0.42

0.55

0.54

0.39

0.55

Pitch

Vowel

Low

i e a o u

0.57 0.59 0.61 0.58 0.59

0.43 0.42 0.42 0.46 0.45

0.43 0.46 0.46 0.47 0.45

0.48 0.50 0.51 0.47 0.48

0.52 0.45 0.47 0.49 0.47

0.39 0.35 0.41 0.40 0.34

0.37 0.43 0.42 0.42 0.42

0.49 0.51 0.53 0.51 0.47

0.33 0.36 0.34 0.31 0.31

Mean i e a o u

0.59 0.55 0.56 0.57 0.54 0.51

0.44 0.47 0.46 0.36 0.44 0.45

0.45 0.40 0.37 0.36 0.42 0.36

0.49 0.52 0.55 0.49 0.50 0.48

0.48 0.37 0.34 0.36 0.32 0.38

0.38 0.51 0.37 0.39 0.46 0.52

0.41 0.36 0.43 0.38 0.37 0.32

0.50 0.46 0.47 0.47 0.51 0.47

0.33 0.36 0.43 0.42 0.42 0.34

0.39

Mean i e a o u

0.55 0.47 0.44 0.45 0.46 0.46

0.44 0.46 0.47 0.47 0.53 0.44

0.38 0.50 0.61 0.60 0.49 0.47

0.51 0.35 0.32 0.31 0.31 0.35

0.35 0.33 0.31 0.31 0.32 0.32

0.45 0.47 0.39 0.56 0.47 0.37

0.37 0.47 0.51 0.52 0.51 0.47

0.48 0.49 0.49 0.50 0.51 0.49

0.39 0.45 0.48 0.51 0.45 0.44

0.48

Mean i e a o u

0.46 0.49 0.45 0.43 0.50 0.50

0.47 0.24 0.27 0.25

0.53 0.32 0.34 0.39 0.39 0.44

0.33 0.26 0.31 0.29 0.25 0.32

0.32 0.39 0.38 0.37 0.40 0.45

0.45

0.50 0.61 0.54 0.53 0.57 0.56

0.50 0.40 0.40 0.43 0.45 0.45

0.47

0.62 0.56 0.55 0.57 0.57 0.54

0.56 0.53 0.53 0.48 0.57 0.57

0.58 0.36 0.35 0.36 0.39

0.57 0.66 0.64

Mean

0.47

0.25

0.38

0.29

0.40

0.56

0.43

0.56

0.54

0.37

0.60

Low Speaking

Speaking Break

Break High

High

0.48 0.50 0.47

0.66 0.54 0.66

0.58 0.49 0.53 0.44

0.38 0.38

AJI

0.53 0.64 0.64 0.58 0.60 0.76 0.72 0.70

Abbreviation: CQ, contact quotient.

ARTICLE IN PRESS

AKI

CQ Values of Breathy, Normal, and Pressed Phonation Types

ACN

5

ARTICLE IN PRESS 6 normal phonation in the high range (A5 and B5) displayed the same signal images as breathy phonation. EGG signals may “weaken or disappear” at the very high parts of the singing range.14 They could have been executed in flageolet register which occurs around D#5-D621; however, characteristics of flageolet or M3 remain unverified by research.17,37 Intraclass coefficient excluding the missing values indicated fair reliability. CQ mean values of phonation on a predetermined speaking pitch at 35% CQ critical level in the present study were 0.44 (SD 0.064) for normal and 0.564 (SD 0.062) for pressed, consistent with the previous study on female speech without pitch restrictions—0.47 (SD 0.07) for normal and 0.58 (SD 0.09) for pressed at approximately 169–260 Hz.4 CQ values in this study capped off at 0.727, fairly consistent with previous studies with a maximum CQ value of approximately 0.7,16,38,39 although earlier studies have even recorded CQ values of up to 0.7713 and 0.86.40 CQ values above the register break (frequencies) are expected to be below 0.5,18 lower than CQ values in low frequencies,6,41 but this did not apply to all subjects and to all break and high frequencies in normal phonation (Table 3). Statistical analyses on pressed CQ values did not lead to specific trends, despite steady increases in SPL as pitch rose, but graphical representation showed a notable concentration of low and break values between 0.5 and 0.6 in Figure 1. Being situated in the lower and upper limits of TA dominant phonation, this concentration could be related to motion restriction due to excessive vocal fold tension in extreme ranges described by Titze.22 Nevertheless, the convergence at the 0.5–0.6 range of pressed phonation at low and register break ranges may be relevant only to the participants of this study, since CQ values greater than 0.7 have been reported previously.13,40 Loud chest singing (M1) only generated CQ values above 0.6 in an earlier study,42 and belting, or chest voice singing above the register break, yielded a mean CQ of 0.59 in another study.40 The CQ values of intentionally pressed phonations in the present study ranged from 0.39 to 0.73, with a mean of 0.54. Congruence of intentionally pressed values in this study to intentionally healthy belt values in earlier studies could imply that beyond a certain frequency range (speaking), CQ magnitude is no longer reliable in determining pressedness and impact stress. While this could also mean that the subjects’ adduction was not forceful enough to be pressed phonation, due to concerns of trained singers for vocal health mentioned by Sandgren,43 the two participants with missing values in the normal high samples had clear glottograms in pressed phonation, indicating firmer vocal fold contact than their perceived resonant voice. Pressed phonation has been characterized by high CQ values, but it is not as straightforward in high-frequency CT dominant phonation. Even classical singers with more training exhibited higher CQ values than average resonant voice estimates in singing higher notes in light register,18,36 possibly to sing longer phrases in one breath or sing loudly.44 Moreover, (classical) high voices were noted to be prone to vocal nodules.45 Membranous medialization and cartilaginous adduction of the vocal folds46 may be more active in the high range when pressing. “When the frequency increases, tension and rigidity are not equally distributed in the different layers”17 of the vocal fold. Furthermore, as Herbst et al47

Journal of Voice, Vol. ■■, No. ■■, 2017

suggested, EGG may not be sensitive to posterior glottal adduction in the falsetto register. Accordingly, CQ may not be a suitable measure of pressedness in the high singing range. Since the midpoint of the membranous vocal fold, where impact stress tends to be highest,48 cannot be isolated in EGG signals, imaging could be more reliable in evaluating hyperfunctional singing,49 especially in higher frequencies. Henrich et al18 stated that OQ correlates with SPL in M1 and fundamental frequency correlates to OQ in M2. Vurma36 agreed that fundamental frequency may be a more significant determinant of CQ in high parts of the singing range. Similarly, two subjects (AJI and ANN) in the present study had almost equal mean SPLs for normal and pressed high, and one soprano (ASI) had a lower SPL for pressed than normal high note despite a higher CQ. Moderate correlations between CQ and SPL were only noted in the low and speaking ranges of the subjects. These reports could mean that even the intense sound quality commonly observed in auditory evaluation of pressed phonation does not characterize highimpact stress phonation in high-frequency singing, particularly of sopranos. If ever, evaluation criteria used in auditory perception may also not be applicable to high-frequency pressed phonation. “Although frequently used, much of the literature suggests there is little reliability in voice perceptions,” but reliability is enhanced by experience.50 However, further research is needed to verify this possibility and to rule out effects of vocal tract adjustments.

CONCLUSIONS Normal and pressed phonation CQ values beyond the speaking pitch varied among the subjects (0.25–0.73). Both normal and pressed phonations did not exhibit distinct trends within each category. Pressed phonation CQ values were mostly higher than the normal counterpart, but the values were relative to normal phonation on the same pitch. This could support the idea that finding the resonant voice, which is neither pressed nor breathy,5 suitable for specific frequencies throughout a person’s singing range36 may be necessary to ensure safety from vocal fatigue. The magnitude of CQ that sounds pressed and feels pressed still warrants further investigation. Subjective accounts of singers on the production of each vocal quality may provide information on sensations and strategies used for each phonation type. A perceptual study on the audio signals in this study involving voice experts may also validate the pressedness of intentionally pressed singing and suggest new criteria for identifying pressed high singing. Finally, speed of vocal fold closure, which influences loudness and resonance, may be added to the estimated values from the EGG waveforms. The sample size, however, is too small to generalize findings. Breathy phonation is not always detected in EGG procedures, since EGG instruments respond essentially to vocal fold contact. Moreover, in this study, the EGG’s limitations became more apparent in the higher ranges, especially as larynges moved more vertically and vocal folds lost depth in CT dominant phonation (M2) and possibly flageolet (M3). Other measurements may be more suitable than CQ in gauging vocal fold impact stress in higher frequencies, thus, distinguishing between normal and pressed, since intentionally pressed singing resulted to values

ARTICLE IN PRESS Kendrich Graemer Ong Tan

CQ Values of Breathy, Normal, and Pressed Phonation Types

close to healthy belt CQ in earlier studies of 0.5–0.6. Further research is also needed among men and in inverse-filtering audio signals, especially for pressed phonation. Acknowledgments The author would like to especially thank Professor Anne-Maria Laukkanen for her guidance throughout this project. He would also like to express his gratitude to Professor Graham Welch, Senior Laboratory Technician Jussi Helin, Pekka Rahkonen, Dremon Salas, Alfredo and Paz Ong Tan, and the singing teachers who generously shared their time and voices to the study. REFERENCES 1. Baken RJ, Orlikoff RF. Clinical Measurement of Speech and Voice. San Diego, CA: Singular Thomson Learning; 2000. 2. Rothenberg M, Mahshie JJ. Monitoring vocal fold abduction through vocal fold contact area. J Speech Lang Hear Res. 1988;31:338–351. 3. Davies P, Lindsey GA, Fuller H, et al. Variation of glottal open and closed phases for speakers of English. Proc Inst Acoust. 1986;8:539–546. 4. Kankare E, Laukkanen A-M, Ilomäki I, et al. Electroglottographic contact quotient in different phonation types using different amplitude threshold levels. Logop Phoniatr Vocol. 2012;37:127–132. doi:10.3109/14015439 .2012.664656. 5. Verdolini K, Druker DG, Palmer PM, et al. Laryngeal adduction in resonant voice. J Voice. 1998;12:315–327. doi:10.1016/S0892-1997(98)80021-0. 6. Morris RJ, Okerlund DA, Craven EA. First Passaggio transition gestures in classically trained female singers. J Voice. 2016;30:1–9. doi:10.1016/ j.jvoice.2015.05.002. 7. Miller DG, Schutte HK, Doing J. Soft phonation in the male singing voice: a preliminary study. J Voice. 2001;15:483–491. doi:10.1016/S0892 -1997(01)00048-0. 8. Henrich N, d’Alessandro C, Doval B, et al. On the use of the derivative of electroglottographic signals for characterization of nonpathological phonation. J Acoust Soc Am. 2004;115:1321–1332. doi:10.1121/1.1646401. 9. Schutte HK, Miller DG. Measurement of closed quotient in a female singing voice by electroglottography and videokymography. In: 5th International Conference Advances in Quantitative Laryngology, Groningen, The Netherlands, April 27–28. 2001. 10. Herbst CT, Lohscheller J, Švec JG, et al. Glottal opening and closing events investigated by electroglottography and super-high-speed video recordings. J Exp Biol. 2014;217:955–963. doi:10.1242/jeb.093203. 11. Pedersen F. Electroglottography compared with synchronized stroboscopy in normal persons. Folia Phoniatr Logop. 1977;29:191–199. doi:10.1159/ 000264088. 12. Kitzing P, Sonesson B. A photoglottographical study of the female vocal folds during phonation. Folia Phoniatr Logop. 1974;26:138–149. 13. Herbst CT, Ternstrom S. A comparison of different methods to measure the EGG contact quotient. Logop Phoniatr Vocol. 2006;31:126–138. doi:10.1080/ 14015430500376580. 14. Daffern H, Howard DM. Voice source comparison between modern singers of early music and opera. Logoped Phoniatr Vocol. 2010;35:68–73. doi:10.3109/14015439.2010.482861. 15. Titze IR, Verdolini Abbott K. Vocology: The Science and Practice of Voice Habilitation. Salt Lake City, UT: National Center for Voice and Speech; 2012. 16. Salomao GL, Sundberg J. What do male singers mean by modal and falsetto register? An investigation of the glottal voice source. Logop Phoniatr Vocol. 2009;34:73–83. doi:10.1080/14015430902879918. 17. Roubeau B, Henrich N, Castellengo M. Laryngeal vibratory mechanisms: the notion of vocal register revisited. J Voice. 2009;23:425–438. doi:10.1016/ j.jvoice.2007.10.014. 18. Henrich N, D’Alessandro C, Doval B, et al. Glottal open quotient in singing: measurements and correlation with laryngeal mechanisms, vocal intensity, and fundamental frequency. J Acoust Soc Am. 2005;117:1417. doi:10.1121/ 1.1850031.

7

19. Sapienza CM, Stathopoulos ET, Dromey C. Approximations of open quotient and speed quotient from glottal airflow and EGG waveforms: effects of measurement criteria and sound pressure level. J Voice. 1998;12:31–43. doi:10.1016/S0892-1997(98)80073-8. 20. Awan SN, Awan JA. The effect of gender on measures of electroglottographic contact quotient. J Voice. 2013;27:433–440. doi:10.1016/j.jvoice.2013.03.007. 21. Garnier M, Henrich N, Crevier-Buchman L, et al. Glottal behavior in the high soprano range and the transition to the whistle register. J Acoust Soc Am. 2012;131:951–962. doi:10.1121/1.3664008. 22. Titze IR. Principles of Voice Production. Iowa City, IA: National Center for Voice and Speech; 2000. 23. Thurman L, Welch GF, eds. Bodymind & Voice: Foundations of Voice Education, Vol. 1. Revised Ed. Iowa City, IA: National Center for Voice & Speech; 2000. 24. Laukkanen A-M, Björkner E, Sundberg J. Throaty voice quality: subglottal pressure, voice source, and formant characteristics. J Voice. 2006;20:25–37. doi:10.1016/j.jvoice.2004.11.008. 25. Kayes G. Singing and the Actor. London: A. & C. Black; 2004. 26. Weekly EM, Lovetri JL. Follow-up contemporary commercial music (CCM) survey: who’s teaching what in nonclassical music. J Voice. 2009;23:367– 375. doi:10.1016/j.jvoice.2007.10.012. 27. Potter J. Vocal Authority: Singing Style and Ideology. New York, NY: Cambridge University Press; 1998. 28. Sadolin C. Complete Vocal Technique. Third ed. Copenhagen: CVI Publications; 2012. 29. Sundberg J. The Science of the Singing Voice. DeKalb, IL: Northern Illinois University Press; 1987. 30. Titze IR. Acoustic interpretation of resonant voice. J Voice. 2001;15:519–528. doi:10.1016/S0892-1997(01)00052-2. 31. Solomon NP. Vocal fatigue and its relation to vocal hyperfunction. Int J Speech Lang Pathol. 2008;10:254–266. 32. Verdolini K, Chan R, Titze IR, et al. Correspondence of electroglottographic closed quotient to vocal fold impact stress in excised canine larynges. J Voice. 1998;12:415–423. 33. Verdolini K, Hess MM, Titze IR, et al. Investigation of vocal fold impact stress in human subjects. J Voice. 1999;13:184–202. 34. Kankare E, Laukkanen A-M. Quasi-output-cost-ratio, perceived voice quality, and subjective evaluation in female kindergarten teachers. Logop Phoniatr Vocol. 2012;37:62–68. 35. Kayes G. How does genre shape the vocal behaviour of female singers? 2014. 36. Vurma A. Phonatory strategies of male vocalists in singing diatonic scales with various dynamic shapings. J Voice. 2016;31:254.e17–254.e29. doi:10.1016/j.jvoice.2016.06.018. 37. Miller DG, Schutte HK. Physical definition of the “flageolet register”. J Voice. 1993;7:206–212. 38. Kangasmäki L. Complete vocal technique -moodit oikeaoppisesti ja epätaloudellisesti tuotettuina. 2012. 39. Thalén M, Sundberg J. Describing different styles of singing: a comparison of a female singer’s voice source in “Classical”, “Pop”, “Jazz” and “Blues”. Logoped Phoniatr Vocol. 2001;26:82–93. doi:10.1080/140154301753207458. 40. McCoy S. Voice pedagogy: a classical pedagogue explores belting. J Sing. 2007;63:545–549. 41. Miller DG, Švec JG, Schutte HK. Measurement of characteristic leap interval between chest and falsetto registers. J Voice. 2002;16:8–19. 42. van Lankeren H. Voice characteristics of amateur female tenors are comparable with those of male tenors. Logoped Phoniatr Vocol. 2010;35:51– 58. doi:10.3109/14015431003667660. 43. Sandgren M. Becoming and being an opera singer: health, personality, and skills. 2005. doi:10.1037/e530422013-001. 44. Sundberg J. Where does the sound come from? In: Potter J, ed. The Cambridge Companion to Singing. Cambridge, UK; New York, NY: Cambridge University Press; 2000. 45. Brodnitz FS, Froeschels E. Treatment of nodules of vocal cords by chewing method. AMA Arch Otolaryngol. 1954;59:560–565. 46. Herbst CT, Qiu Q, Schutte HK, et al. Membranous and cartilaginous vocal fold adduction in singing. J Acoust Soc Am. 2011;129:2253. doi:10.1121/ 1.3552874.

ARTICLE IN PRESS 8 47. Herbst CT, Howard D, Schlömicher-Thier J. Using electroglottographic real-time feedback to control posterior glottal adduction during phonation. J Voice. 2010;24:72–85. doi:10.1016/j.jvoice.2008.06.003. 48. Jiang JJ, Titze IR. Measurement of vocal fold intraglottal pressure and impact stress. J Voice. 1994;8:132–144.

Journal of Voice, Vol. ■■, No. ■■, 2017 49. Freeman E, Woo P, Saxman JH, et al. A comparison of sung and spoken phonation onset gestures using high-speed digital imaging. J Voice. 2012;26:226–238. 50. Sofranko JL, Prosek RA. The effect of experience on classification of voice quality. J Voice. 2012;26:299–303.