! ORIGINAL ARTICLE !

Source and filter adjustments affecting the perception of

the vocal qualities twang and yawn
Ingo R. Titze1,2, Christine C. Bergan1, Eric J. Hunter1,2 and Brad Story3
From the 1Department of Speech Pathology and Audiology, The University of Iowa, USA, 2National Center for Voice and
Speech, The Denver Center for the Performing Arts, USA, and 3The University of Arizona, USA

Received 11 February 2003. Accepted 4 September 2003.

Logoped Phoniatr Vocol 2003; 28: 147 !/155

Two vocal qualities, twang and yawn , were synthesized and rated perceptually. The stimuli consisted of synthesized vocal
productions of a sentence-length utterance ‘ya ya ya ya ya,’ which had speech-like intonation. In a continuum transformation
from normal to twang , the area in the pharynx was gradually decreased , along with vocal tract shortening and a decreased
open quotient in the glottal airflow. In a continuum transformation toward yawn , the area in the pharynx was gradually
increased , along with vocal tract lengthening and an increased open quotient. The normal (untransformed) vocal tract area
was pre-determined by earlier studies involving MRI scans of a human subject’s vocal tract. Listeners were asked to rate (on a
scale from 1 !/10) the ‘amount of twang ’ in one listening session and the ‘amount of yawn’ in another listening session. Overall,
the perception of twang increased directly with pharyngeal area narrowing, vocal tract shortening, and decreased open
quotient. The perception of yawn increased with pharyngeal area widening, vocal tract lengthening, and increased open
quotient. Adjustments of one parameter alone yielded less significant perceptual changes than the above combinations, with
open quotient showing the greatest effect in isolation. Listeners demonstrated variable perceptions in both continua with poor
inter-subject, intra-subject, and inter-group reliability.
Key words: perception, timbre, twang , voice, voice quality, yawn .

Ingo R. Titze, PhD, Department of Speech Pathology and Audiology, National Center for Voice and Speech, The University of
Iowa, Iowa City, IA 52242, USA. Tel.: "/1-319-335-6600. Fax: "/1-319-335-6603.

INTRODUCTION AND BACKGROUND
Evaluation of voice disorders and assessment of vocal
performance skill requires the ability to differentiate
vocal qualities. Many vocologists depend upon this
ability on a daily basis, such as teachers of singing,
speech-language pathologists, and otolaryngologists.
Improving the reliability and validity of voice quality
assessment strengthens the value of diagnosis of voice
pathologies and the habilitation of healthy voice
production. The general goal of this study is aimed
toward such an improvement with regard to two
distinct vocal qualities, yawn and twang .
The two qualities are chosen because they appear to
form a dichotomy between a wide vocal tract produc-
tion and a narrow vocal tract production, both of
which may be healthy and efficient. The choice may be
based on dialect and other socio-linguistic factors, but
may also be driven by a need for a wide range in pitch,
timbre and loudness in occupational voice production
(as in singing or public speaking). In a videoendo-
scopic study by Yanagisawa et al . (15), nine profes-
sional singers were asked to find the limits of their
vocal range in six voice qualities: speech, falsetto,
yawn (also called cry or sob), twang, belt, and opera.
Simultaneous activities of the larynx, the pharyngeal
walls, and the soft palate were monitored using a
videoendoscope. Results showed that the larynx rose
in all subjects with the production of higher frequen-
cies, the lateral pharyngeal walls significantly con-
tracted toward the midline in an ‘upside-down V
shape’, with the highest fundamental frequency creat-
ing the narrowest pharynx; in addition, the soft palate
lifted and the velopharyngeal port narrowed consider-
ably with higher frequencies. In a related study,
Yanagisawa et al . (16) investigated three nasal and
two oral vocal qualities. Results demonstrated signifi-
cant interactions of velar and laryngeal functions
during the production of nasal and twang qualities,
suggesting that the source and vocal tract are adjusted
simultaneously to find a best match for a given vocal
production.
Story et al . (9) revealed significant new information
about the three-dimensional nature of the vocal tract,

# 2003 Taylor & Francis. ISSN 1401-5439 Logoped Phoniatr Vocol 28

DOI: 10.1080/14015430310018874
148 I. R. Titze et al.
especially in the pharyngeal region. Using magnetic
resonance imaging (MRI), the authors produced vocal
tract area functions for a nearly complete set of
phonemes of the English language for one male
speaker. In a follow-up study (10), similar data were
obtained for a female speaker. Most relevant to the
current study, however, Story et al . (8) investigated the
relationship of vocal tract shape to three voice
qualities: normal , yawn , and twang. Three-dimen-
sional vocal tract shapes and consequent area func-
tions representing the vowels [i, æ, ", u] were obtained
from one additional male and one additional female
speaker, again using MRI. The two new speakers were
trained vocal performers and adept at manipulation of
vocal tract shape to alter voice quality. Each vowel was
performed three times, one for each of the three voice
qualities. Relative to normal speech, the mean area
functions showed that the vocal tract widened and
lengthened for the yawny productions while the vocal
tract narrowed and shortened for the twangy produc-
tions. But the expansions and contractions were not
uniform from glottis to lips, which brings into ques-
tion the relevance of ‘front cavity’ versus ‘back cavity’
adjustments. Resulting acoustic correlates of these
articulatory alterations consisted of the first two
formants (F1 and F2) being close together for all
yawny vowels and farther apart for all twangy vowels.
It is suspected that laryngeal adjustments also play a
major large role in these qualities. In a study by
Bergan, Titze and Story (1), the objective was to relate
the perception of ring voice quality and pressed voice
quality to laryngeal and epilaryngeal parameters. Ring
quality was rated according to the extent of glottal
airflow skewing and the cross sectional area of the
epilaryngeal tube. Pressed quality was rated according
to the extent of the open quotient and the glottal flow
amplitude. Spectrally, both qualities were character-
istic of high frequency energy in the voice, but ring
quality maintained a strong concentration of low
frequency energy, whereas pressed quality did not.
This was in part attributed to the fact that ring
involved a widening of the pharynx (a vocal tract
adjustment that lowers F1), whereas pressed was
strictly glottal in nature.
Past research has demonstrated poor intra- and
inter-subject reliability when rating vocal qualities.
Some of this research has focused on the effect of
experience and professional background on perceptual
ratings of voice quality (3). Not every level of
professionalism (clinical experience, musicianship,
auditory training, etc.) is of equal relevance in the
judgment of vocal quality. For example, in a previous
study dealing with pitch and roughness perception (2),
musicians did outscore non-musicians in reliability,
but in another study the same was not found for the
Logoped Phoniatr Vocol 28
voice quality ratings ring and pressed (1). These latter
qualities apparently had no specific relevance to
musicianship.
With use of the GRBAS scale (grade, roughness,
breathiness, asthenia, strained) for deviant voice
qualities, Dejonckere et al . (4) found that clinical
experience does play a role in inter-rater agreement.
Comparison of internal and external standards in
voice quality judgments through the use of standard
rating scales was also made by Gerratt et al . (5), in
which a set of anchors was presented prior to rating.
Poor listener agreement in mid-range ratings of
breathiness and roughness has been noted (6), but
this is perhaps dependent on the nature of the scale
(e.g., linear versus logarithmic). The reliability of a
visual analog versus an ordinal scale for the perceptual
evaluation of dysphonia in 14 pathological voices was
tested by Wuyts et al . (14). They determined that a
four-point scale is generally sufficient and a visual
analog scale did not increase reliability. Collectively,
these investigations suggest that experience is not to be
discounted, but the type of experience may need to
vary according to the listening paradigm. In the
current study, musical (singing) experience will be
continued as a factor of analysis.
PURPOSE AND RESEARCH QUESTIONS
The purpose of this study was to determine whether or
not specific combinations of source and vocal tract
area adjustments, hypothetically chosen, correlate
with the perception of twang and yawn . In particular,
we focus on ‘back of the throat’ adjustments. Subjects
were presented with the utterance ‘ya ya ya ya ya’,
created with a voice simulation model described in the
procedures below. The utterance had a speech-like
intonation pattern and semantically resembled the
utterance ‘I know, I know, I know’ or ‘Yes, I’ve heard
this before.’ The research questions to be answered
were:
1. How do changes in combined pharyngeal and
epilaryngeal tube area, vocal tract length, and
open quotient of the glottal flow correlate with the
perception of twang and yawn in voice produc-
tion?
2. Does the co-variation of all three of these para-
meters result in a greater perception of twang and
yawn than any parameter alone?
3. How variable are the listener’s abilities to rate
these qualities?
4. Is there a significant difference in inter- and intra-
listener variability between vocal musicians and
non-musicians?
 Twang and yawn qualities 149

Regarding the last question, a long-range goal of pharyngeal sections, and 22 oral cavity sections. The
this research is to determine if vocal training and oral cavity areas were not varied, but the areas of the
auditory training can help vocologists make better back of the vocal tract were varied according to the
judgments about voice quality. Given that vocal relation

Aepi
8
: 15n5Nepi
>
p(n $ Nepi $ 1)
>
< ! " #$
Am (n)# Ao (n) 1"(S $1)sin : (Nepi "1)5n5(Nepi "Nphx ) (1)
>
>
: Nphx $ 1 : (Nepi "Nphx "1)5n5(Nepi "Nphx "Noral )
Ao (n)

musicians currently receive some of this training, we where

felt it useful to address the benefit of this training.
Nepi #/number of epilaryngeal sections
Nphx #/number of pharyngeal sections
PROCEDURES Noral #/number of oral cavity sections
Aepi #/cross-sectional area of epilarynx
Twenty listener volunteers were recruited, 10 males S #/scale factor*/can range from 0.5 !/2.0
and 10 females. All of them reported normal hearing. n #/section number
The ages of the volunteers ranged from 20 !/60 years Ao #/original area function
with a mean age of 31.5 years. Amount of musical Am #/modified area function
background or training ranged from none or slight
(ten subjects) to professional vocalists (ten subjects). For the vocal tract length change, the pharynx was
Musicians and non-musicians were also gender increased by 2 sections (from 14 to 16) or decreased by
matched, with five females and five males in each 2 sections (from 14 to 12). Fig. 2 illustrates how the
category. vocal tract area function, including the epilaryngeal
tube, was transformed from the original vowel shape
(dotted) to the twang vowel (solid). The top figure
Stimulus generation
shows the area function during the production of the
Synthetic stimuli were generated with a glottal flow /i/-like part of ‘ya ya ya ya ya’, while the bottom figure
model (11), and a vocal tract area function model as a
filter (13). In this model, 10 glottal source parameters
and 44 vocal tract sections are controllable. The first
six glottal parameters are: peak flow, fundamental
frequency, open quotient, skewing quotient, minimum
glottal flow, and the area of the epilaryngeal tube. The
remaining four glottal parameters are for amplitude
and frequency modulations, which were not used in
this study. Only the open quotient (Qo) of the flow was
varied. In a previous study (1), it was found that co-
variation of Qo and peak flow did not alter the
perception of adduction over Qo alone. Hence, the
co-variation was not repeated here. Skewing quotient
was kept constant at 1.7 and minimum flow was kept
at zero. Fig. 1 shows examples of two different glottal
flow waveforms used, one for an open quotient of 0.3
and one for an open quotient of 0.8.
The vocal tract was first divided into 44 sections of
equal length, the sum of which was 17.5 cm in length
(the average male vocal tract length). These 44 sections Fig. 1. Glottal flow pulses for twang (solid lines) and
were arranged into 8 epilaryngeal tube sections, 14 yawn (dotted lines)

Logoped Phoniatr Vocol 28

150 I. R. Titze et al.
Fig. 2. Vocal tract area function transformed from normal (dotted lines) to twang (solid lines). The vowel /i / is on
top and /"/ on the bottom.
shows the area function during the production of the
/a / in the utterance. Note the narrowing in the
pharynx and the overall length reduction. Fig. 3 shows
a similar transformation to ‘yawn’. Note the widened
pharynx area and the elongation of the vocal tract. To
obtain the full dynamic area function simulation over
the entire ‘ya ya ya ya ya’ utterance, a mapping from
formants to modal coefficients was performed accord-
ing to Story and Titze (7).
The shape of the epilarynx tube area was deliber-
ately kept cylindrical in order to control it with a
single parameter (the area Aepi). It is acknowledged
that the epilarynx tube shape can vary substantially
with false fold positioning, but this creates air space
variations less than about 0.5 cm in length (such as the
laryngeal ventrical). The portion of the spectrum
affected by such small cavity lengths (!/5000 Hz)
are not under consideration here. Hence, the tube was
kept cylindrical for simplicity.
It should also be pointed out that the subject from
whom the MRI data were obtained had a relatively
narrow epilarynx tube to begin with (dotted lines near
glottis, Aepi #/0.7 cm2). This made the transformation
to yawn (Aepi #/2.0 cm2) appear larger than the
transformation to twang (Aepi #/0.2 cm2), but in terms
of wave reflection coefficients (difference between
Logoped Phoniatr Vocol 28
areas divided by sums of the areas) the increase and
decrease were approximately equal.
In total, 18 different stimuli were created. These
came from two open quotient values (0.3 and 0.8),
three vocal tract lengths (42, 44, and 46 sections), and
three epilaryngeal area and pharyngeal scale factor
combinations (Aepi #/0.2, S #/0.5; Aepi #/0.7, S #/1.0;
and Aepi #/2.0, S #/2.0).
Acoustic analysis of the stimuli
Discrete Fourier transforms of the /i/-like portion and
the /a / portion of the utterance were calculated. The
window length was 1102 points (selected by two
cursors to identify a somewhat steady portion) and
the sampling frequency was 44 kHz. Results are shown
in Fig. 4. The vowel /i/ is on top and the vowel /a / is on
the bottom. Note that twang shows a ‘whiter’ spec-
trum, with energy spread across the higher frequen-
cies. The fundamental is not the dominant partial for
twang , but there is a greater peak of energy in the
2000!/3000 Hz band. Much of this energy spread
comes from the small open quotient and high-
frequency resonation in the epilarynx tube. For the
yawn quality, energy is concentrated mainly at the
fundamental and near the first or second formant
 Twang and yawn qualities 151

Fig. 3. Vocal tract area function transformed from normal (dotted lines) to yawn (solid lines). The vowel /i/ is on
top and /"/ on the bottom.

Fig. 4. Discrete Fourier Transforms of the /i/-like portion (top) and the /"/ portion (bottom) of the utterance ‘ya
ya ya ya ya’. Solid lines are for twang and dotted lines for yawn .

Logoped Phoniatr Vocol 28

152 I. R. Titze et al.
(2000 Hz for /i / and 700 Hz for /a/), with a rapid decay
and loss of energy above this formant region.
Presentation of stimuli and rating
Subjects were seated in a sound-treated room and sat
approximately 10 feet from a loudspeaker. The volume
of the sound system was set and maintained for all
subjects. A brief introduction was given about the two
vocal qualities, twang and yawn , but no strict or
precise definition was offered. (This was to allow the
subjects to rate more freely the stimuli according to
what they believed to be twang and yawn , without the
influence of the investigator’s personal bias.) Subjects
were allowed to practice with a representative pool of
ten anchors (see below). They were asked to rate the
specific vocal quality on a scale from 1 !/10; ‘1’ would
signify very little (if any) presence of the specific
quality and ‘10’ would signify a large amount of that
specific quality.
Eighteen different stimuli were created and each was
randomly presented 3 times, resulting in a total of 54
presentations for each quality. The repeated presenta-
tion of stimuli allowed for the calculation of intra-
subject reliability. Each quality required about 15 min
to complete, for a total listening time of 30 min. This
relatively short exposure to the signals avoided listener
fatigue and helped control learning effects.
RESULTS
Fig. 5 summarizes the listener’s ratings with extreme
(anchor) conditions (simultaneous variation of open
quotient Qo, vocal tract length Lvt, epilarynx tube area
Ae, and pharyngeal area Ap to maximize the difference
Fig. 5. Perceptual ratings of twang and yawn with
extreme (anchor) combinations of parameters. High
impedance refers to small Qo, narrow and short vocal
tract; low impedance refers to large Qo, wide and long
vocal tract.
Logoped Phoniatr Vocol 28
between the two qualities). These two anchor condi-
tions are labeled as high impedance and low impe-
dance in the graph because they represent everything
wide (low glottal and vocal tract impedance) and
everything narrow (high glottal and vocal tract
impedance). For high impedance, the open quotient
Qo was 0.3, the vocal tract length Lvt was 42 sections,
the epilarynx tube area Aepi was 0.2 cm2, and the
scaling factor S for pharyngeal width was 0.5. Con-
versely, for low impedance, the values were: Qo #/0.8,
Lvt #/46, Aepi #/2.0, and S #/2.0. As a result of the
preliminary anchoring task, the listeners used nearly
the entire range (1 !/10) to distinguish these qualities in
the final test when all stimuli were mixed.
Individual parameter variations and the consequen-
tial perceptions are summarized in Figs. 6!/8. Fig. 6
shows that open quotient alone is a dominant para-
meter. It carries most of the variance of the combined
parameter set. As seen, keeping the vocal tract neutral
and varying only Qo reduced the difference in the
yawn -twang ratings by only about 2 !/3 points in
relation to the anchor ratings in Fig. 5.
Fig. 7 shows the perception of yawn and twang for
vocal tract widening alone. The epilarynx tube area
Aepi and the pharyngeal scaling factor S were in-
creased simultaneously from left to right (e.g., Aepi #/
0.2 and S #/0.5), widening the vocal tract in three
steps. Part (a) is for a small open quotient (0.3) and
part (b) is for a large open quotient (0.8). The reason
both plots are shown is that Qo has such a dominant
effect that it could possibly mask all other effects. The
vocal tract length remained at Lvt #/44 sections in
both cases. Note that vocal tract widening (left to
right) increased the perception of yawn and decreased
the perception of twang for both open quotients.
Fig. 8 shows the perception of yawn and twang with
vocal tract length change alone. Again, part (a) is for a
small open quotient (0.3) and part (b) is for a large
open quotient (0.8). The vocal tract width was kept at
Fig. 6. Perceptual ratings of twang and yawn with
open quotient Qo alone. Other parameters remained
constant as labeled above.

Fig. 7. Perceptual ratings of twang and yawn with
vocal tract area changes alone: (a) small open quotient
(Qo #/0.3) and (b) large open quotient (Qo #/0.8).
Ratios on the abscissa indicate epilarynx tube area
Aepi and scale factor S of the pharynx.
the nominal condition (Aepi #/0.7 cm2 and S #/1.0).
Note that vocal tract length change (from 42 sections
to 46 sections, a 10% change) had the least effect on
the perception of yawn and twang. The difference was
at most 1 rating point (out of 10).
STATISTICAL ANALYSIS
A mixed model ANOVA was performed for subject,
parameter type, and parameter magnitude. The pur-
pose was to determine the relative impact each
controlled parameter had on the slope of a regression
line, and therefore on the resulting rating. Tables 1 !/4
summarize the results. In all tables, R represents the
perceptual rating (either yawn or twang).
For the twang continuum, with parameters inde-
pendently regressed (Table 1), the results may be
interpreted as follows: Based on the slope of the
regression equation, if the open quotient Qo were to be
increased by 1 unit (from 0.3 to 0.8 is 1/2 unit), the
twang rating would decrease by 8.83 points, almost the
entire scale. If the epilarynx tube area alone were to be
increased by 1 cm2 (a condition not presented to the
listeners), the rating would decrease by 1.08 points, or
about 10% of the scale. If the pharyngeal area alone
Twang and yawn qualities 153
Fig. 8. Perceptual ratings of twang and yawn with
vocal tract length changes alone: (a) small open
quotient (Qo #/0.3) and (b) large open quotient
(Qo #/0.8).
were to be increased by 1 cm2 (also not presented), the
rating would decrease by 1.32 points. Finally, if the
vocal tract length were to be increased by 1 section,
the rating would decrease by 0.21 points (although this
last measure did not reach significance).
With the parameters simultaneously regressed (Ta-
ble 2), the slope between the epilarynx tube area and
the twang rating became positive. In other words, as
the epilaryngeal tube area was increased, so did the
rating of twang . This seems contradictory. However,
the slope between the pharyngeal area and twang
rating became more negative, suggesting that there
was a statistical interaction between these two vocal
tract areas. The interaction was an obvious one. The
parameters were forced to co-vary by formula. A gain
in one slope was offset by a loss in the other to satisfy
the co-variance. Previous results have shown that
pharyngeal area and epilarynx tube area move in
opposite directions for the perception of vocal ‘ring’
(1). Thus, twang would have been confused with ring if
the two parameters had been allowed to drift in
opposite directions. For this reason, we varied Aepi
and S together in Eq. (1).
For the yawn continuum (Tables 3 and 4), the
results may be interpreted as follows: For single
Logoped Phoniatr Vocol 28
154 I. R. Titze et al.

Table 1. Summary statistics for the ‘twang’ continuum (one parameter independently regressed)

Type of manipulation Regression equation t-value F-value

Open quotient R #/10.44 $/8.33(Qo) $/25.71 (p B/0.0001) 661.05 (p B/0.0001)

Epilaryngeal area R #/6.62 $/1.08(Aepi) $/5.71 (p B/0.0001) 32.55 (p B/0.0001)
Pharyngeal area R #/7.12 $/1.32(Aphx) $/5.76 (p B/0.0001) 33.21 (p B/0.0001)
Vocal tract length R #/14.77 $/0.21(Lvt) $/2.29 (p B/0.0228) 5.24 (p B/0.0028)

Table 2. Summary statistics for the ‘twang’ continuum (all parameters simultaneously regressed)

Type of manipulation Regression equation t-value F-value

Open quotient R #/22.31 $/8.83(Qo) $/33.36 (p B/0.0001) 1112.66 (p B/0.0001)

Epilaryngeal area "/2.08(Aepi) 1.46 (p B/0.1454) 2.13 (p B/0.1454)
Pharyngeal area $/3.85(Aphx) $/2.22 (p B/0.0275) 4.91 (p B/0.0275)
Vocal tract length $/0.21(Lvt) $/5.15 (p B/0.0001) 26.48 (p B/0.0001)

Table 3. Summary statistic for the ‘yawn’ continuum (one parameter independently regressed)

Type of manipulation Regression equation t-value F-value

Open quotient R #/0.21"/8.92(Qo) 27.18 (p B/0.0001) 738.91 (p B/0.0001)

Epilaryngeal area R #/4.16"/0.98(Aepi) 5.22 (p B/0.0001) 27.23 (p B/0.0001)
Pharyngeal area R #/3.69"/1.22(Aphx) 5.33 (p B/0.0001) 28.42 (p B/0.0001)
Vocal tract length R #/ $/6.85"/0.27(Lvt) 3.03 (p B/0.0027) 9.17 (p B/0.0027)

Table 4. Summary statistics for ‘yawn’ continuum (all parameters simultaneously regressed)

Type of manipulation Regression equation t-value F-value

Open quotient R #/ $/15.3"/8.91(Qo) $/36.44 (p B/0.0001) 1327.88 (p B/0.0001)

Epilaryngeal area $/4.75(Aepi) $/3.60 (p B/0.0004) 12.95 (p B/0.0004)
Pharyngeal area "/6.99(Aphx) 4.35 (p B/0.0001) 18.93 (p B/0.0001)
Vocal tract length "/0.27(Lvt) 7.26 (p B/0.0001) 52.78 (p B/0.0001)

parameter regression (Table 3), if the open quotient Qo
were to increase by 1 unit (its entire theoretical range),
the yawn rating would increase by 8.92 points (also
nearly the entire range). If the epilarynx tube area
alone were to increase by 1 cm2, the rating would
increase by 1.22 points. If the vocal tract length were
to increase by 1 section (out of 44), the rating would
increase by 0.27 points (although this last measure,
once again, did not reach significance). Thus, we have
a direct relationship between the parameters and the
perception of yawn . With simultaneous regression
(Table 4), the relationship between the epilarynx tube
area and the resulting rating of yawn once again
became a negative one; as the epilarynx area in-
creased, the perceptual rating of yawn decreased.
Again, this was a reflection of the forced co-variance
between epilarynx area and pharynx area in our
design.
To determine if musicians are less variable than
non-musicians in their judgments of yawn and twang,
the within-subject variance was assumed to be the
same for all 18 stimuli of the test. A pooled estimate of
the within-subject variance was then made. This
pooled variance (within musician type) was asympto-
tically normal and a two-sample t-test was performed
between musicians and non-musicians. For the twang
continuum, the musicians were slightly more variable
than the non-musicians, with a t-value of $/1.78 (p B/
0.0210) and an F-value of 1.78 (p B/0.0210). In the
yawn continuum, however, the musicians showed

Logoped Phoniatr Vocol 28


somewhat less variability than the non-musicians, with
a t-value of $/2.18 (pB/ 0.0523) and an F-value of
5.46 (p B/0.0523). In general, the musicians showed
only a slight advantage over the non-musicians for this
pooled analysis. When comparing standard deviations
between groups on the same stimuli, the non-musi-
cians showed greater variability than the musicians for
only 60% of the stimuli.
DISCUSSION AND CONCLUSIONS
It was hypothesized that certain pre-selected source
and filter adjustments contribute to the perception of
the vocal qualities twang and yawn . The source
parameter open quotient had the greatest effect on
these qualities, being small for twang and large for
yawn . The combination of pharynx area and epilarynx
tube area had the next greatest effect, being narrow for
twang and wide for yawn . The additional parameter
vocal tract length (and, in particular, pharyngeal
length) had the third greatest effect, being short for
twang and long for yawn . The results do not exclude
the possibility that other parameters may contribute to
the perceptions of these qualities, but given that the
selected parameters capture many of the physiologic
changes observed endoscopically, some confidence in
the results is warranted.
These results can be further summarized in terms of
acoustic impedances of the source and filter. Twang is
a high impedance configuration (a smaller open
quotient of the glottis and a narrower entry to the
vocal tract) and yawn is a low impedance configura-
tion (a larger open quotient of the glottis and a wider
entry to the vocal tract). Both configurations may be
efficient for maximum power transfer from the glottis
to the lips (12), and the choice for using twang or yawn
in speech therapy may not be one of vocal efficiency or
vocal health, but rather one of finding the greatest
range of pitch, loudness and timbre in a given voice.
There may also be a cultural or artistic preference, or a
choice based on regional dialect. A superficial argu-
ment in preference of yawn (from a therapeutic point
of view) might be that larger open quotients and an
overall low-impedance configuration reduce vocal fold
collision, but the low impedance configuration would
probably also call for larger amplitudes of vibration,
thereby negating the argument for less collision.
A weakness of the study is that listeners were not
asked to rate the naturalness of the synthesis. Informal
listening by the investigators did not suggest major
perceptual distractions, but a few may have been there
for highly discriminate listeners. Formal tests of
naturalness would have been desirable, but the choice
was made to limit the scope of the study.
Twang and yawn qualities 155
ACKNOWLEDGEMENT
This work was supported by the National Institutes of
Health, grant number 5 1R01 DC04224-03.
REFERENCES
1. Bergan C, Titze IR, Story B. Perception of two vocal
qualities in a synthesized vocal utterance: ring and
pressed voice. J Voice (in press).
2. Bergan C, Titze I. Perception of pitch and roughness in
vocal signals with subharmonics. J Voice 2001; 15: 165 !/
75.
3. DeBodt MS, Wyts FL, Van deHeyning PH, Croix C.
Test-retest study of the GRBAS scale: influence of
experience and professional background on perceptual
rating of voice quality. J Voice 1997; 11: 74 !/80.
4. Dejonckere PH, Remacle M, Fresnel-Elbaz E, Woisard
V, Crevier L, Millet B. Reliability and clinical relevance
of perceptual evaluation of pathological voices. In:
McCafferty G, Coman W, Carrol R, editors. XVI.
World Congress of Otorhinolaryngology Head and
Neck Surgery, Sydney 1997. Proceedings, Vol 2. Bo-
logna: Monduzzi Editore; 1997. p. 1699 !/703.
5. Gerratt BR, Kreiman J, Antonanzas-Barroso N, Berke
GS. Comparing internal and external standards in voice
quality judgments. J Speech Hear Res 1993; 36: 14 !/20.
6. Kreiman J, Gerratt BR. Validity of rating scale measures
of voice quality. J Acoust Soc Am 1998; 104: 1598 !/608.
7. Story B, Titze I. A preliminary study of speech
transformation using empirically defined articulatory
modes. J Acoust Soc Am 1999; 105(2), pt 2, 2pSCa5,
Berlin, Germany, March 1999.
8. Story BH, Titze IR, Hoffman EA. The relationship of
vocal tract shape to three voice qualities. J Acoust Soc
Am 2001; 109: 1651 !/67.
9. Story BH, Titze IR, Hoffman EA. Vocal tract area
functions from magnetic resonance imaging. J Acoust
Soc Am 1996; 100: 537 !/54.
10. Story BH, Titze IR, Hoffman EA. Vocal tract area
functions of an adult female speaker based on volu-
metric imaging. J Acoust Soc Am 1998; 104: 481 !/7.
11. Titze IR, Mapes S, Story B. Acoustics of the tenor high
voice. J Acoust Soc Am 1994; 95: 1133 !/42.
12. Titze IR. Regulating glottal airflow in phonation:
application of the maximum power transfer theorem. J
Acoust Soc Am 2002; 111: 367 !/76.
13. Titze R, Story B. Acoustic interactions of the voice
source with the lower vocal tract. J Acoust Soc Am
1997; 101: 2234 !/43.
14. Wuyts FL, DeBodt MS, Van deHeyning PH. Is the
reliability of a visual analog scale higher than an ordinal
scale? An experiment with the GRBAS scale for the
perceptual evaluation of dysphonia. J Voice 1999; 13:
508 !/17.
15. Yanagisawa E, Estill J, Kmucha T, Leder SB. The
contribution of aryepiglottic constriction of ‘‘ringing’’
voice quality */A videolaryngoscopic study with acous-
tic analysis. J Voice 1989; 3: 342 !/50.
16. Yanagisawa E, Mambrino L, Estill J, Talkin D. Supra-
glottic contributions to pitch raising: videoendoscopic
study with spectral analysis. Ann Otol Rhinol Laryngol
1991; 100: 19 !/31.
Logoped Phoniatr Vocol 28