Facial Bone Vibration In Resonant Voice Production

Fei C. Chen, Estella P.-M. Ma, and Edwin M.-L. Yiu, Pokfulam, Hong Kong

Summary: Purpose. This study investigated whether resonant voice training would enhance facial bone vibration
during resonant voice production.
Methods. Twelve normal healthy participants undertook four sessions of resonant voice training, each lasted for 30
minutes. Piezoelectric accelerometer was used to measure the vibratory level on the face (nasal bridge and upper lip) and
the perilaryngeal area during the production of nasal consonant /m/ and vowels /a/, /i/, and /u/ before and after the reso-
nant voice training. The extents of vibration of these four sounds among these three sites were compared.
Results. Significant increase in facial bone vibration following resonant voice training was found. The nasal bridge
showed a significantly larger magnitude of increase when compared with that at the upper lip. Different sounds were
also found to facilitate different magnitude of facial bone vibration. Greater magnitude of facial bone vibration was
found with the phonation of /m/, /i/, and /u/ when compared with the phonation of /a/.
Conclusion. Resonant voice training facilitated an increase in facial bone vibration, more so at the nasal bridge area
than around the upper lip. This is hypothesized to contribute to the improved resonant voice production. Sounds that
involve relatively restricted oropharyngeal cavities facilitated a greater extent of facial bone vibration during resonant
voice production.
Key Words: Humming–Vowel–Vibration–Accelerometer–Resonant voice.

INTRODUCTION Positive outcomes of resonant voice therapy have been docu-

Resonant voice is defined physiologically as a voicing pattern
with a laryngeal posture slightly abducted or barely adducted.1
Under this specific type of voice production, the impact stress is
minimal, thus reducing the possibility of vocal fold damage.2–6
Resonant voice production is often taught using auditory-
perceptual judgment supplemented by tactile perception of
facial bone vibration. Bone vibration during resonant voice pro-
duction is generally considered to be prominent at the anterior
alveolar ridge and maxillary bones.6 Perceptually, resonant
voice can be defined as an easy phonation accompanied by a
vibrant sensation around the orofacial area.6,7
Resonant voice training, which has been developed and based
on the concept of resonant voice, has been widely used clini-
cally to treat individuals with hyperfunctional voice disorder
with or without benign vocal lesions, which commonly include
vocal nodules or polyps.8–11 Resonant voice training aims to
help the learners to achieve an optimal voice production using
a relaxed and clear voice with minimal effort. The possibility
of vocal fold injury is, therefore, minimized.12 The training
procedure often requires individuals to feel kinesthetically the
vibration around the nasal bridge area or the skull in general.
The vibration is believed to be the energy transmitted through
the vocal tract to the skull and the skin tissue.13 Resonant voice
training also emphasizes auditory perception of relaxed and
natural voice production.6,10,14,15 The training procedure
generally recommends the use of nasal stimuli for the training
and practice of resonant voice.14,16
Accepted for publication December 19, 2013.
From the Voice Research Laboratory, Division of Speech and Hearing Sciences, The
University of Hong Kong, Pokfulam, Hong Kong.
Address correspondence and reprint requests to Edwin Yiu, Voice Research Laboratory,
Division of Speech and Hearing Sciences, The University of Hong Kong, Pokfulam,
Hong Kong. E-mail: eyiu@hku.hk
Journal of Voice, Vol. -, No. -, pp. 1-7
0892-1997/$36.00
Ó 2014 The Voice Foundation
http://dx.doi.org/10.1016/j.jvoice.2013.12.014
mented rather extensively in the literature. Outcome measures
include improvements in perceptual voice qualities,10,11,17
acoustic perturbation,7,10,11,17 aerodynamic measures,11 elec-
troglottographic measures,1 quality of life changes using self-
assessment indices,6,11,17 and physiological changes.18 It should
be noted that the published studies so far focused on the voice
quality outcome in general following resonant voice training.
However, there has not been any study that specifically investi-
gated the resonant voice training protocol itself. Specifically, lit-
tle is known about the relationship between facial bone vibration
and resonant voice production or whether the tactile sensation of
the facial bone vibration facilitates resonant voice production.
Accelerometer is a vibration sensor using piezoelectric effect.
The sensor converts mechanical energy into electrical energy in
response to the stress applied to it. Miniature accelerometers
have been used in measuring nasalization,19–21 chest wall
vibration,22 and skin vibration on the neck during voicing.13,23


Svec et al23 investigated the relationship between the intensity
(sound pressure level) of phonation and the level of skin vibration
at the suprasternal notch using an accelerometer. A significant
positive correlation between these two measures was found.23
It has also been demonstrated that the lateral nasal cartilage/
bone is the most sensitive site for the detection of vibration.13,20
This study set out to determine whether resonant voice
training would enhance facial bone vibration during voice pro-
duction in vocally healthy speakers. It was hypothesized that
trained resonant voice production would produce a greater extent
of vibration in the facial bone. A second objective of the study
was to determine whether the use of different vowels and nasal
sound during resonant voice production would produce different
degree of vibration in the facial region (nasal bridge vs lips area).
METHOD
Participants
Twelve vocally healthy individuals (five males and seven females)
with a mean age of 27.25 years (standard deviation ¼ 2.01 years,
2 Journal of Voice, Vol. -, No. -, 2014

FIGURE 1. Placement of vibrodetectors at the nasal bridge, upper

lip, and perilaryngeal site.

range ¼ 24–30 years) participated in this study. All the partici-

pants were native Mandarin speakers with no existing or reported
history of voice/resonance disorders and not receiving medica-
tions at the time of the study that might have any effect on their
voice. All of them passed a screening test for hearing, with hearing
threshold of the better ear better than 30 dB HL at octave fre-
quencies between 250 and 4000 Hz frequency range. None of
the subjects displayed any perceivable nasality in their voice as
determined by the first author (F.C.C.).

Instrumentation
Br€uel & Kjær piezoelectric DeltaTron accelerometers (Type
4507-B-002; Br€ uel & Kjær, Naerum, Denmark) were used as
the bone vibration sensors. The dimension of each piezoelectric
accelerometer was 1 3 1 3 1 cm, with a weight of 4.8 g, and a
usable frequency range of 0.3–6000 Hz. The vibration signals
were recorded using the PowerLab system (Power Lab AD In-
struments, Melbourne, Australia) with the Nexus conditioning
amplifier (Type 2693; Br€ uel & Kjær, Naerum, Denmark) and
analyzed with the software program Chart (v 5.4.2; Power
Lab AD Instruments).
Procedures
Placement of piezoelectric accelerometer. One piezo-
electric accelerometer was placed on the right nasal bridge
just above the septal cartilage (Figures 1 and 2). A second piezo-
electric accelerometer was placed just above the upper lip on the
left side of the philtrum. These two measurement sites recorded
the facial bone vibration caused by the resonance of the phona-
tion from the larynx. A third piezoelectric accelerometer was
placed at the laryngeal prominence of the thyroid cartilage (per-
ilaryngeal area) as a control to record vibration from the phona-
tory source (Figure 1). All the accelerometers were taped to the
sites using 3M Micropore medical tape (St Paul, MN). Figure 2
illustrates the taping of an accelerometer to the nasal bridge.
Baseline measurement. A baseline measurement of bone
vibration during voice production was conducted before the reso-
nant voice training. All recordings were carried out in a sound-
treated booth with a background noise of less than 35 dBA.
Participants were seated comfortably in an upright position on
a straight-back chair throughout the recording. The recordings
involved the sustained prolongation of four speech sounds (/m/,
/a/, /i/, and /u/). Each participant was first asked to produce the
nasal consonant /m/ five times; each sustained for 5 seconds us-
FIGURE 2. Placement of accelerometer at the nasal bridge.
ing his/her most comfortable pitch and loudness level. This pro-
cedure was then repeated for each of the three vowels (/a/, /i/, /u/).
Resonant voice training. After the baseline measurement
and before the resonant voice training began, each participant
was first given a 30-minute rest. This was then followed by a
30-minute resonant voice training session. Three more subse-
quent training sessions (ie, a total of four training sessions)
were conducted every other day over a 7-day period. The resonant
voice training protocol used in this study was based on the reso-
nant voice therapy program developed by Verdolini6 and Yiu.17,24
At the beginning of the training, the investigator (first author,
F.C.C.) provided verbal instructions and demonstrations of
resonant voice using the nasal consonant /m/. Participants were
asked to produce a resonant /m/ using a relaxed projection
without creating neck muscle tension. The participants were
instructed to concentrate on the auditory feedback and tactile
sensory information of the vibration around the facial bone.
The training was extended to other materials that included /m/
in combination with /a/, /i/, or /u/. In general, the /m/ lasted for
one second followed immediately by a one-second long /a/, /i/,
or /u/. Thus, subsequently, the combination resulted in the
production of a prolonged /ma/, /mi/, and /mu/, respectively.
Comfortable pitch and loudness levels were emphasized through
out the practice. The instructional steps of the resonant voice
training are listed in the Appendix. At the end of the four training
sessions, the participants needed to be able to produce ‘‘resonant’’
/m/, /ma/, /mi/, and /mu/ as determined by the investigator F.C.C.
using a perceptual criterion which emphasized easy phonations
accompanied by a vibrant sensation.6,7
Fei C. Chen, et al Facial Bone Vibration and Resonant Voice Training
Posttraining measurement. Posttraining measurement
was carried out 30 minutes after the completion of the last (ie,
the fourth) session of resonant voice training on the last day of
the training. The procedures of the measurement were identical
to those of the baseline measurement.
Data analysis
For each recording of vibration, a two-second stable segment in
the middle of each recording was extracted and analyzed with a
frequency range of 1–6000 Hz for the root mean square millivolt-
age (mVRMS). A 2 (Training: pretraining vs posttraining) 3 3
(Site: nasal bridge, upper lip, and perilaryngeal area) two-way
repeated Analysis of Variance (ANOVA) was used on the pooled
data (/m/, /a/, /i/, /u/) to determine if the resonant voice training
and the site of measurement had any effect on the bone vibration
during voicing. Another 2 (Training: pretraining vs posttraining)
3 4 (Sounds: /m/, /a/, /i/, /u/) two-way repeated ANOVAwas car-
ried out to determine which sound facilitated the best vibration
following resonant voice training. Following the recommenda-
tion of Max and Onghena,25 Mauchly test for assumption for
sphericity was not carried out as this could increase the proba-
bility of errors. As the normality assumption was also violated
(Kolmogorov-Smirnov tests showed P < 0.05), Huynh-Feldt
tests with the degrees of freedom adjustments were used.25
RESULTS
Vibration before and after resonant voice training at
different sites
Figure 3 shows the distribution of vibration pooled across all
sounds (/m/, /a/, /i/, and /u/) before and after resonant voice
training at each recording site, whereas Figure 4 shows the dis-
tribution of bone vibration of each of the sound pooled together
across the three sites (nasal bridge, upper lip, and perilaryngeal
site) at the pretraining and posttraining time points.
Main effect of training. Significant main effect of Training
(F ¼ 50.28, df ¼ 1, P < 0.0001, partial eta squared ¼ 0.17,
observed power ¼ 1.00) was found with the sounds pooled
together at each site.
Main effect of sites. Significant main effect of Sites
(F ¼ 156.04, df ¼ 1.77, P < 0.0001, partial eta
squared ¼ 0.39, observed power ¼ 1.00) was also found.
Planned contrasts showed that the three sites were all significant
different with the perilaryngeal area demonstrating the highest
pooled vibration (mean ¼ 5.65 mV) when compared with the
nasal bridge (mean ¼ 4.49 mV; F ¼ 38.39, df ¼ 1, P < 0.0001,
partial eta squared ¼ 0.14, observed power ¼ 1.00) and the upper
lip (mean ¼ 2.49 mV; F ¼ 471.32, df ¼ 1, P < 0.0001, partial eta
squared ¼ 0.66, observed power ¼ 1.00). In the facial area, the
vibration was significantly higher at the nasal bridge (Figure 4)
than that at the upper lip (F ¼ 95.38, df ¼ 1, P < 0.0001, partial
eta squared ¼ 0.28, observed power ¼ 1.00).
Interaction effect. Significant result was found with the
Site 3 Training interaction effect (F ¼ 61.56, df ¼ 1.68,
P < 0.0001, partial eta squared ¼ 0.21, observed power ¼ 1.00).
Following resonant voice training, the nasal bridge vibration
3
increased significantly more (Figure 4) when compared with the
upper lip (F ¼ 73.26, df ¼ 1, P < 0.0001, partial eta
squared ¼ 0.23, observed power ¼ 1.00) and the perilaryngeal
area (F ¼ 77.46, df ¼ 1, P < 0.0001, partial eta squared ¼ 0.25,
observed power ¼ 1.00). No significant interaction effect was
found between the upper lip and the perilaryngeal area
(F ¼ 2.33, df ¼ 1, P ¼ 0.13, partial eta squared ¼ 0.10, observed
power ¼ 0.33).
Vibration produced by different speech sounds
Figure 3 also shows that the nasal consonant /m/ produced the
largest extent of vibration, whereas /a/ produced the smallest
extent of vibration both before and after resonant voice training.
Main effect of training. Significant main effect of Training
(F ¼ 19.49, df ¼ 1, P < 0.0001, partial eta squared ¼ 0.10,
observed power ¼ 0.99) was found, showing vibration
increased significantly following resonant voice training.
Figure 3 shows clearly that the bone vibration associated with
each of the sounds increased following resonant voice training.
Main effect of sounds. Significant main effect of Sounds
(F ¼ 111.30, df ¼ 2.37, P < 0.0001, partial eta squared ¼ 0.38,
observed power ¼ 1.00) was also found. Planned contrasts
showed that all the sounds were significantly different from
one another (/m/ vs /a/: F ¼ 238.57, df ¼ 1, P < 0.0001, partial
eta squared ¼ 0.57, observed power ¼ 1.00; /m/ vs /i/:
F ¼ 78.30, df ¼ 1, P < 0.0001, partial eta squared ¼ 0.30,
observed power ¼ 1.00; /m/ vs /u/: F ¼ 75.81, df ¼ 1,
P < 0.0001, partial eta squared ¼ 0.30, observed power ¼ 1.00;
/a/ vs /i/: F ¼ 98.19, df ¼ 1, P < 0.0001, partial eta
squared ¼ 0.35, observed power ¼ 1.00; /a/ vs /u/: F ¼ 76.62,
df ¼ 1, P < 0.0001, partial eta squared ¼ 0. 03, observed
power ¼ 1.00; and /i/ vs /u/: F ¼ 5.66, df ¼ 1, P ¼ 0.018, partial
eta squared ¼ 0.03, observed power ¼ 0.66), with /m/ showing
the greatest amount of vibration, followed by /u/, and /i/, with
/a/ demonstrating the smallest extent of vibration.
Interaction effect. Significant Sound 3 Training interaction
effect (F ¼ 13.76, df ¼ 27.23, P < 0.0001, partial eta
squared ¼ 0.07, observed power ¼ 1.00) was also found.
Planned contrasts showed that the /m/ sound demonstrated
the greatest amount of increase in vibration than the /a/
(F ¼ 34.01, df ¼ 1, P < 0.0001, partial eta squared ¼ 0.16,
observed power ¼ 1.00); /i/ (F ¼ 23.06, df ¼ 1, P < 0.0001, par-
tial eta squared ¼ 0.11, observed power ¼ 1.00); and /u/
(F ¼ 21.21, df ¼ 1, P < 0.0001, partial eta squared ¼ 0.11,
observed power ¼ 0.99) following resonant voice training
(Figure 3). The interactions were not significant among the
three vowels (/a/ vs /i/: F ¼ 0.05, df ¼ 1, P ¼ 0.82, partial
eta squared ¼ 0.00, observed power ¼ 0.06; /a/ vs /u/:
F ¼ 0.39, df ¼ 1, P ¼ 0.53, partial eta squared ¼ 0.002,
observed power ¼ 0.10; and /i/ vs /u/: F ¼ 0.82, df ¼ 1,
P ¼ 0.37, partial eta squared ¼ 0.01, observed power ¼ 0.15).
DISCUSSION
The present study was set out to determine whether resonant
voice training would enhance facial bone vibration during
4 Journal of Voice, Vol. -, No. -, 2014

FIGURE 3. Pre- and posttraining vibration (mVRMS) at each of the recording sites. Midline of the box is the median. The bottom and top of the box
show the 25th and 75th percentile (the lower and upper quartiles). The top and bottom lines represent the highest and lowest values that are not out-
liers (ie, within 1.5 times the interquartile range). The outliers (between 1.5 and three times the interquartile range) are represented by circles, and
extreme values (more than three times the interquartile range) are represented by asterisks.

resonant voice production in vocally healthy speakers. A sec-
ondary objective was to determine whether the use of different
vowels and nasal sound during resonant voice production
would produce different magnitude of vibration in the facial
region (nasal bridge vs upper lip area). Results showed that
resonant voice training facilitated a significant increase in the
nasal bridge vibration over time. The use of nasal consonant
/m/ produced the greatest magnitude of bone vibration during
resonant voice production.
The effect of resonant voice training on the
magnitude of facial bone vibration
Tactile sensation of facial bone vibration has been used as one of
the training strategies in acquiring resonant voice production.6,14
This involves placing a finger on the nasal bridge to detect the
vibration there to facilitate resonant voice production. The
physiological mechanism is believed to involve the conversion
of acoustic energy to mechanical vibration transmitted via the
Fei C. Chen, et al Facial Bone Vibration and Resonant Voice Training 5

FIGURE 4. Vibration (mVRMS) produced by /m/, /a/, /i/, and /u/ at pre- and posttraining measurement. Midline of the box is the median. The bot-
tom and top of the box show the 25th and 75th percentile (the lower and upper quartiles). The top and bottom lines represent the highest and lowest
values that are not outliers (ie, within 1.5 times the interquartile range). The outliers (between 1.5 and three times the interquartile range) are rep-
resented by circles, and extreme values (more than three times the interquartile range) are represented by asterisks.

maxillary bony structures.7 Resonant voice is described as ‘‘easy
phonation’’ that can produce maximum voice output intensity
with the minimum vocal effort.1–3,6 With resonant voice
production, more energy is expected to transmit through the
skull bones when compared with modal voice production. In
the present study, piezoelectric accelerometer was used as a
noninvasive tool to determine if there was an increase in facial
bone vibration with resonant voice production. The results
showed that resonant voice training facilitated a significant
increase in the nasal bridge vibration but not in the source
(larynx) following resonant voice training. This clearly shows
that the increased vibration is not due to increased energy
emitted by the larynx (source) per se.
One common issue encountered in resonant voice training
is to find the most suitable site on the face that would provide
the best tactile sensation for bone vibration. The results of the
present study also found the magnitude of bone vibration
measured at the nasal bridge area to be significantly higher
than that at the upper lip. One possible explanation is that the
vibration signal from the source (larynx) are attenuated through
the muscular and skin tissues, the thicker the muscle and the skin,
the more attenuation there would be. Therefore, the nasal bridge,
with minimal layer of skin, is theoretically the best part on the
face in transmitting vibration signal. Previous studies have
shown that placing the accelerometer on the nasal bone detected
nasality and frequency response during phonation significantly
better than the other sites on the nose and cheek.13,20 The
present study found the magnitude of bone vibration measured
at the nasal bridge area to be significantly higher than that
measured at the upper lip, which, therefore, supported the use
of the nasal bridge for tactile sensation in resonant voice training.
The magnitude of facial bone vibration produced by
different sounds
The present study also found that different sounds produced
different magnitude of vibration during resonant voice produc-
tion. Nasal humming involving the nasal consonant /m/ is often
used in resonant voice training. It is contended that the humming
of /m/ produces a higher acoustic pressure inside a small cavity
and so a greater extent of bone vibration could be detected dur-
ing humming.7 Humming has been shown to increase the effi-
ciency of voice production by forming a semiocclusion tube
that form a coupling tube between the source (larynx) and the fil-
ter (nasal cavity) during phonation.26 The findings of the present
study supported the use of the sound /m/, that is, humming, in
resonant voice training.14,16 The /m/ produced a significant
amount of bone vibration higher than the nonnasal vowels.
Among those three vowels, the vowel /a/ produced the lowest
magnitude of bone vibration, whereas the vibration produced
by /i/ and /u/ were not significantly different. Phonetically, /u/
and /i/ are both closed vowels and /a/ is an open vowel. Hence
the closed vowels /u/ and /i/ resulted in a semioccluded space,
which increased the intensity and efficiency of the voice pro-
duction.6,26 Thus, the use of humming and closed vowels
6 Journal of Voice, Vol. -, No. -, 2014
would, therefore, facilitate better tactile sensation of vibration
around the facial area during phonation.
Limitations and directions for future research
It should be noted that acoustic or auditory-perceptual charac-
teristics of the resonant voice were not measured in the present
study, and hence, the relationships between facial bone vibra-
tion and acoustic and auditory-perceptual characteristics of
resonance were not the focus of the present study. Whether
bone vibration during voice production reflects the extent of
auditory perception of resonant voice is an important question.
The findings of the present study only showed that following
resonant voice training, the magnitude of facial bone vibration
increased during resonant voice production. Direct relationship
between perceptual resonance and bone vibration could not be
established with the data from the present study. Further study
will be needed to address this issue. Furthermore, the resonant
voice training used the trigger sound /m/, which somehow
might lead to the involvement of the velopharyngeal port in
the production of the vowels, thus increasing the facial bone
vibration because of increased energy through the velopharyng-
eal opening. This would, however, need further investigations
using instrument (eg, flexible nasendoscopy) to document the
physiological changes.
CLINICAL IMPLICATION AND CONCLUSION
The present study provided support for increased facial bone
vibration following resonant voice training. Significant dif-
ferences in the bone vibration produced by different sounds
were also found. The humming of the nasal consonant /m/
and closed vowels /u/ and /i/ produced a significantly higher
magnitude of nasal bridge vibration during resonant voice pro-
duction. It is, therefore, recommended that resonant voice
training should make use of the sounds /m/, /i/, and /u/, as
they would facilitate a better learning of resonant voice produc-
tion using the proprioceptive feedback from the nasal bridge
vibration. Therefore, the traditional method in resonant voice
training, which involves asking individuals to put their fingers
on the nasal bridge to feel the vibration during resonant voice
production, is a justified procedure.
Appendix
Instructions of resonant voice training
1. The participant will sit comfortably and relaxed during
the training process.
2. The participant will produce ‘‘uh-um’’ or sound /m/ softly
as in acknowledging someone’s question in a relaxed
manner. The investigator will provide a demonstration.
3. The participant will sing /m/ up and down a musical scale
to find their suitable pitch that would result in the
maximum resonance/vibration. The participant is encour-
aged to relax their neck (laryngeal) muscles during the
voice production. The selected pitch will be used as the
note for reference in further training. The investigator
will provide a demonstration.
4. The participant will feel the facial vibration by putting a
finger at the nasal bridge during voice production. The
participant will also be asked to feel a possible tingling
sensation around the lips due to the kinesthetic feedback.
5. The investigator will comment on the voice quality of
humming sound produced by the participant based on
the parameters of gentle onset, resonance, loudness, and
pitch level.
6. The participant will hum at a comfortable pitch and then
sustained a vowel /a/ at the end of the /m/. Smooth tran-
sition from the sound /m/ to /a/ is emphasized. The partic-
ipant will be reminded that the voice should be produced
in a relaxed manner. The investigator will provide a
demonstration.
7. The investigator will comment on the quality of the sound
/ma/ produced by the participant.
8. Repeat steps 6 and 7 for the sound /m/ . /i/ and /m/ . /u/.
REFERENCES
1. Verdolini K, Druker DG, Palmer PM, Samawi H. Laryngeal adduction in
resonant voice. J Voice. 1998;12:315–327.
2. Verdolini K. Resonant voice therapy. In: Verdolini K, ed. Voice and Speech’s
Guide to Vocology. Iowa City, IA: National Center for Voice and Speech;
1998:34–35.
3. Verdolini K, Titze IR. The application of laboratory formulas to clinical
voice management. Am J Speech Lang Pathol. 1995;4:62–69.
4. Verdolini K, Hess MM, Titze IR, Bierhals W, Gross M. Investigation of
vocal fold impact stress in human subjects. J Voice. 1999;13:184–202.
5. Jiang JJ, Titze IR. Measurement of vocal fold intraglottal pressure and
impact stress. J Voice. 1994;8:132–144.
6. Verdolini K. Case study: resonant voice therapy. In: Voice Therapy: Clini-
cal Studies. 2nd ed. San Diego, CA: Singular Publishing; 2000:46–61.
7. Titze IR. Acoustic interpretation of resonant voice. J Voice. 2001;15:519–528.
8. Chen SH, Huang J, Chang W. The efficacy of resonance method to hyper-
functional dysphonia from physiological, acoustic and aerodynamic as-
pects: the preliminary study. Asia Pac J Speech Lang Hear. 2003;8:200–203.
9. Roy N, Weinrich B, Gray SD, Tanner K, Stemple JC, Sapienza CM. Three
treatments for teachers with voice disorders: a randomized clinical trial.
J Speech Lang Hear Res. 2003;46:670–688.
10. Verdolini-Marston K, Burke MK, Lessac A, Glaze L, Caldwell E. Prelim-
inary study of two methods of treatment for laryngeal nodules. J Voice.
1995;9:74–85.
11. Chen SH, Hsiao TY, Hsiao LC, Chung YM, Chiang SC. Outcome of resonant
voice therapy for female teachers with voice disorders: perceptual, physio-
logical, acoustic, aerodynamic, and functional measurements. J Voice. 2007;
21:415–425.
12. Stemple JC, Glaze LE, Klaben BG. Clinical Voice Pathology: Theory and
Management. 3rd ed. San Diego, CA: Singular Publishing Group; 2000.
13. Munger J, Thomson S. Frequency response of the skin on the head and neck
during production of selected speech sounds. J Acoust Soc Am. 2008;124:
4001–4012.
14. Yiu EM-L. Resonant voice: Hong Kong humming. In: Behrman A,
Haskell J, eds. Exercise for Voice Therapy. San Diego, CA: Plural Publi-
shing; 2008.
15. Behrman A, Haskell J. Resonant Voice. Exercises for Voice Therapy. San
Diego, CA: Plural Publishing; 2008.
16. Boone DR, McFarlane SC, Von Berg SL. The Voice and Voice Therapy. 7th
ed. Boston, MA: Allyn & Bacon; 2005.
17. Yiu EM-L, Ho EY-Y. Short-term effect of humming on vocal quality. Asia
Pac J Speech Lang Hear. 2002;7:123–137.
18. Ogawa M, Hosokawa K, Yoshida M, Yoshii T, Shiromoto O, Inohara H.
Immediate effectiveness of humming on the supraglottic compression in
Fei C. Chen, et al Facial Bone Vibration and Resonant Voice Training
subjects with muscle tension dysphonia. Folia Phoniatr Logop. 2013;65:
123–128.
19. Horii Y. An accelerometric measure as a physical correlate of perceived
hypernasality in speech. J Speech Lang Hear Res. 1983;26:476.
20. Lippmann RP. Detecting nasalization using a low-cost miniature acceler-
ometer. J Speech Hear Res. 1981;24:314–317.
21. Stevens K, Kalikow D, Willemain T. A miniature accelerometer for detect-
ing glottal waveforms and nasalization. J Speech Hear Res. 1975;18:594.
22. Sundberg J. Chest wall vibrations in singers. J Speech Hear Res. 1983;26:
329–340.
7


23. Svec JG, Titze IR, Popolo PS. Estimation of sound pressure levels of voiced
speech from skinvibration of the neck. J Acoust Soc Am. 2005;117:1386–1394.
24. Yiu E, M- L. Hong Kong humming. In: Behrman A, Haskell J, eds. Exercises
for Voice Therapy. 2nd ed. San Diego, CA: Plural Publishing; 2013:70–72.
25. Max L, Onghena P. Some issues in the statistical analysis of completely ran-
domized and repeated measures designs for speech, language, and hearing
research. J Speech Lang Hear Res. 1999;42:261–270.
26. Titze IR. Voice training and therapy with a semi-occluded vocal tract:
rationale and scientific underpinnings. J Speech Lang Hear Res. 2006;49:
448–459.