A Pressure-Flow Technique for Measuring
Velopharyngeal Orifice Area During
Continuous Speech
DONALD W. WARREN, D.D.s., Ph.D.
ARTHUR B. DuBOIS, M.D.
Philadelphio,
Pen7l81/l'Jonio
The need for roethods which measure adequacy of velopharyngeal closure arises from the requirements for assessingthis function in certain
speech defective individuals, notably those with cleft palates. Many
techniques have been described for studying the velopharyngeal mechanism (U, 7,8,10,13-15), the roost recent of which is Bjork's radiographic
and spectrographic method for relating isthrous area to the acoustical
characteristics
of speech (1)
.
With the development of instruments capable of faithfully measuring
pressure and airflow in the mouth and nase (9), the possibility of calculating velopharyngeal orifice area using hydrokinetic principles (6) became feasible.
Described in this paper is a meÜ1odbased on those principles j certain
preliminary results obtained using normal subjects are also reported.
Studies of surgically and prosthetically treated cleft palate patients will
be reported in subsequentpapers (16,17).
Description of T eehnique
This method is based upon a modification of the Theoretical Hydraulic
Principle and assumesthat the area of an orifice can be determined if the
~~~.E~~_a~~the
rate of airflow tnrougrnr;:Thus,
orificLis
measured simultaneously with
. Area = -:- ~J,e
of Airftow
~~ugh PrP.M1I~
Orifice
Orifice
/;TOrifice
Dift'erentiat
'V ~ \(~ce_Dift'~rentiat~
DeD8itY-~f-Air
Since
--
production
-
of
speech
involves
an
acoustic
disturbance
)
~~ J
suoerim-
PRESSURE-FLOW
53
TECHNIQUE
posed upon respiratory airflow, the aerodynamic phase of phonation can
be used to obtain the nccessaryparameters for application of this theoretical equation (seeAppendix for derivation of equatjon).
Since even in the simplest cases,an equation cannot be developed which
takes jnto account ail of thc details of turbulent, non-unjform and rotational ftow, the derivation was adapted to a fictitious average steady
motjon. The theoretical equation, therefore, neglects several fundamental
considerations and js only an approximatjon to actual ftow conditions in
the oropharynx, i.e., airftow is actually unsteady, non-umform and rotational. The actual area of the orifice will differ from the theoretical ares
becauseof these factors. Bence, to obtai~he actual area from this "rational approximation" a correction facto~, must be introduced.
u.
lU
IC
le
;h
~
Û
e
ia
)f
Experimental Mode.
It is impossible to measure tbis constant k for the velopharyngeal orifice, 80 a simple model was used instead. Furthermore, most orifice equations either assumethat pressure is measured in the narrow ares of the
orifice, or that pressure just past the orifice is equal to pressure in the
orifice. This assumption is valid if the kinetic energy of the gas passing
through the orifice is lost due to turbulence on the nasal Bide of the orjfice.
Again, although unable to verify thjs in the nasopharynx, the model was
used to measure pressure on the nasal Bide of the orifice jn deriving the
correction constant. To the exrent that airftow in the artificial Dosesimulated airflow in the Dose,the above assumption of turbulence would be
valid. The model of the upper respiratory tract was constructed from
plastic tubing and joints. This prototype of the upper speech mechanism
was designed so that the ares of the model velopharyngeal orifice couId
be varied by a series of interchangeable orifices of different size. The
model mouth could ~e opened or closed thereby allowing for simulation
of flow patterns in the pharynx during speech.
M~HOD USEDWITHMODEL.A series of 24 experiments were carried out
on the model. These include 16 jn which the model mouth was closed and
eight in which it was open. This comparison was deemed necessary to
ascertain whether the open mouth, acting as a "sink", affects the equation.
The apparatus used is shown diagrammatically in Fjgure 1. Differential pressure across the model velopharyngeal orifice was measured
with a sensitive Lilly capacitance manometer.1 Connected to the transducer were two catheters2which were placed abovc and below the orifice.
Side holes were drilled in the catheters and the catheter end openings
were plugged with wax so that only the lateral or the "static" pressure
component was measured. Flow through the orifice was measured by a
)
Lilly
capaeitance
ing Co., Philade)phÏ&,
small air di.placement
manometer,
with
Pa. A 8train puge
ÎB alao 8uitable.
. Polyethylene tubing PE 200.
venoUl
preaun
manometer
head. Technitrol
(diBerential
pre8Ure
En~ineertype)
witb
1
ff
.34
Warren, DuB0i8
A~C,-
0
:)
DI',_.",
,.-. T
~ooo
0:E3
0:1::3
0:1=:]
0:E3
v..,-
'VII""" ,.,..r
0.,1...
00
00
00
00
'- --c
r::;;:~
~_":' ::J
1
FIGURE 1. Di8lT&lnmatic representation of the apparatu8 ulled for meaeuring
the coefficient k ODthe model "velophuyngeal orifice".
pneumotachograph3connectedto the model Dose.An air cylinder supplied
the ftow necessaryto simulate the aerodynamics of speech. Both pressure
and ftow were recorded on a four channel magnetic tape recorder4 which
contains oscilloscopesfor monitoring each parameter. An oscilloscope recording camera5was utilized to photograph on paper8the data which was
replayed from the tape recorder into a cathode ray oscillograph.7
RESULTS OF ExPERDŒNT8 WITH THE MODEL TO DmERMINE
CoEFFICIENT
k. The results of the experiments witlt tlte model are shown in Table 1.
The coefficient k was computed from the equation
V
k-- A I;t~~~
~
(1)
where Pl is the pressure below the orifice (oral side) and Ps is the pressure above the orifice (nasal side). (See Appendix for definition of oUter
symbols and for derivation of equations 1 and 2.) Although k theoreticali y varies with orifice size, the differences found in the 16 experiments
with the "mouUt" closed were 80 slight that it was decided that the
constant k couId be averaged and treated as a constant. Figure 2 illustrates how little k changesas the area is increased from 2.4.mm2 to 120.4.
mm2.
The "theoretical equation" is thus modified by the correction coefficient
to give the "working equation".
t,
a Pneumotachosraph,
.
Technitrol
Enlineering
Co., Philadelphia,
Pa. (a strain
gRuge
manometer
and ftowmeter is aIso suitable)
Daconl Tape
Recorder,EJectro-Medi-Dyne,Farmincdale,N. 1.
a Gr.. Kymocraph Camera,Gr.. In8trument Co., Quincy, Ma.'m.
. Kodak linocraph papeT;f44, EaatmanKodak Co., Rochester,N. 1.
'Dumont 322-A Cathode Ray Ü8ciliograph, Allan E. Dumont Co., Clifton,
This allowed recordinc two channela at a time; the tape WM ron through twiœ.
N. J.
PRESSURE-FLOW
B
76.1
147
176
166
147
273
297
273
28S
1411
706
735
940
186
128
136
149
1205
1029
793
1381
22
~
18
24
-
TECHNIQUE
.66
.66
.62
.69
.65
.63
.66
.64
.60
.61
.59
MeaD k - .65
v
A.;y,(~
T.. ~~tioo.
W Ch..","km"',
.m"~
.61
",
~"O'oO -"""
Go".
"'w'mo_o'~ooo".m,"",o'~
-~-~A=~"
Th"o."~"oo","m"",
~,::,o::,:;,oo~~~::';'::~~::~
WM ..~o,.."
W:~::O~'i.T=
,::::;';;:;
::,::~.-::::,;;';:::,:o:::;:;",::::,",m:::
;:::~::~~;:~::o:~:
~;
""'o"""~odom
'no" io'ore.' i" -'~"'Ch,
,o,om.,. = Ch,
,'ow,.,bi' ,o~, M' mou"om""" Woomodo
w Ch,0"""
." mm
.7
..
..
..
..;.la
,~
'0
If]
~
55
56
Warren,DuBois
TABLE 2. Calculated aile of model velopharyngeal orifice when using k a! a con
atant value (model mouth cl~).
-
AclNIIlsi: (",III')
1- -
~
. (...:
E~
-
(Mill')
A
13».4
131.5
132.9
131.7
132.0
+12.5
+11.3
+11.6
B
76.1
77.7
77.1
73.0
81.1
+1.6
+1.0
-3.1
+5.0
C
17.1
17.0
16.6
17.1
16.8
-0.1
-0.5
0.0
-0.3
D
2.4
2.3
2.2
2.2
2.3
-0.1
-0.2
-0.2
-0.1
+11.1
TABLE 3. Calculated size of model velopharyngeal orifice wben using k as a con
Itant value (mode! mouth open).
-
~-
AclNalsi" (111111')
C
17.1
D
2.4
c..I.-
.;. <-'>
16.6
18.4
14.7
16.4
1.8
1.6
1.8
1.8
Ener (-->
-0.$
+1.3
-2.4
-1.7
-0.6
-0.8
-0.6
-0.6
It should be noted that when the "mouth" was open the increased effect
of turbulence did not appreciably affect the results (Table 3).
Proeed ore
Ten adult subjects with normal speech patterns and no gro8S dental
deviatioos were selected for this phase of the study. The subjects were
iostructed to use a pitch level which seemedmost natural for them during
the phonation of two test sentences.A recording level meter on the tape
recorder provided limited standardization for loudness level, but strict
contraI was Dot considered necessary since the essential parameters of
;'
PRESSURE-FLOW
kj
i~
~ ,.S ~~-E:3-=--J_.
~~J--l~--,--
1
-~
~""'--
1
-~~~- ,
--
-
,---
-.
---=..
~
.
c-
0
57
TECHNIQUE
00 -
:çooo
E::J
1 ..
Po
. ~ 0 0 - 1..J...
0:E:=;300 _1..0
0
~
:- E3
'-
0
r.",-c_.
0
10
"""'
FIGURE 3. Diacrammatic JepreeutatiOD of the apparatue UIed for recordinc
Pre81re and ftow in the upper pharynx during speech.
t
pressure and flow follow each other. The apparatue used ie shown diagrammatically in Figure 3.8 The pressure transducer formerly used in the
model study to record dift'erential pressure across the orifice now measures oropharyngeal pressure with respect to atmospheric pressure. The
pneumotachograph is warmed by an electric current passing through a
loop of nichrome wire. A microphone9 positioned under the subject's chio
ie used for sound pick-up of the test sentences. Pressure, airflow, and the
voiced 8ounds emitted are recorded on the four channel tape recorder, and
the cathode ray oscillograph and oscillograph camera are utilized as previously mentioned for photographing the data.
CALIBRATION.The apparatus was calibrated with each set of experiments. The pneumotachograph was calibrated by airflow into a rotameterlO and the capacitance manometer was calibrated against a water
manometer.
MEASUREMJ:NT
OPOaOPHABYNGEAL
PaESSUBI:.A 8mall partially collapsed
balloon was passed through the left nostril and orifice into the oropharynx
(Figure 4). The baIJoon, made of thin latex, measured approximately 0.8
cm long and 0.4 cm wide. A balloon of these dimensions was found to be
tolerated easily without éausing a gag reflex and was round to give more
reliable readings than one of a smaller sile. Best positioning of the balloon could bc determined by monitoring the pressure record as the subject phonated. If the balloon was correctIf placed at the level of the resting uvula, the contracting soft palate moved upward and backward a~'ay
from it (Figure 5). A cork was then inserted into the left nostril to prevent air leakage and to secure the balloon in position. The heated pneumotachograph was connected to the right nostril by a large snugly-fitting
tube.
TECHNIQUE FOR RELATING OROPHARYNGEAL PRESSURE TO ORIFICE DIF-
7
~
FERENTIALPRESSURE.
ln order to use the area equation, it is necessary to~
. The Lilly manometeraand tubin, have à responaetime of 1/100~c. (9) which is
aPfroximately twice 88fast as the rise time or pre8SUre
or flow durinl speech.
Tumer cry8tal microphone,The Turner Co., Cedar Rapids,Iowa.
"Flowrater Meter, Fischer &; Porter Co., Warminster, Pa.
58
Warren, DuBois
A
C>
\
/\
,<:::>
:1
VrLcork to PluQ Nostril
and Secur. TubinQ
Ta
Flowrn.ter
To Pressure Tronsducer
B
\.,Post.rior
Soft
Pho
,
~
Jeal Wc
Polo te
801100n in Pharynx
Ton9ue
FIGURE 4. Technique for recordinc oropharyngeal preesure and orifice airftow
in DonnaI subjecta. ln A, the flowmeter Î8 connected to the right noetril, and a cork
plugs the left DoRt"1 and ~Ure8 the balloon in position. ln B, the balloon ig placed
in the oropharynx slightly below the uvula.
relate the oropharyngeal pressure obtained during speech to differential
pressure in the velopharyngeal orifice (hereafter referred to as orifice).
Since oropharyngeal pressure is equal to the pressure drap across the
orifice and the nasal pathway, then orifice dift'erential pressure can be
obtained by subtracting the nasal pressure component from oropharyngeaI pressure. To do this the following procedure is performed. A record
is made while the subject, with lips closed, breathes lightly through his
nase.Cineradiographic studies have shawn that during expiration the soft
palate remains in its resting position and the orifire is wide open (18).
PRESSURE-FLOW
B
A
;'
59
TECHNIQUE
--t
Hord Polot.
Soft
Polot.
1
1
Posterior
Phoryngeol.Wali
.l
FIGURE 5. Correct balloon plaœment in oropharynx. ln A, the baUoon je placed
at tip of re8tiDc uvuJa.ln B, the ~ft paJateelevateaaway from the baUoon durinl
speech.
The values obtained for oropharyngeal pressure and nasal airflow are
then plotted on graph paper. Theoretically oropharyngeal pressure measurements during expiration contain both orifice and nasal pressure
components,but in reality, at low nasal Bow rates comparable to those encountered during speech, the orifice component is too small to be recorded.ll Figure 6 illustrates that the pressure-flow graph obtained during
expiration actually Ï8 a measure of nasal resistance at low nasal flow
rates. The wide open orifice pressure component cao be measured only at
flow rates above approximately 200 cc/sec. Therefore, during expiration
with the orifice wide open, oropharyngeal pressure can be considered
equal to the nasal pressure drop.
Two assumptions are made. a) During speech,any change in oropharyngeal pressure for a given Bow rate Ï8 caused by a decrease in orifice
size. b) Nasal resistance does DOt change during phonation of the sentence. The latter assumption can be tested by measuring nasal resistance
from the subject's expiration record prior to and immediately following
each sentence.The measuredBow rate for each specific speech element is
then used to determine the nasal pressure component from the pressureflow graph. This nasal component is subtracted from the oropharyngeal
pressuremeasurementto give orifice dift'erential pressure.Any back pressures from having one nostril occluded should be cancelled out with this
procedure.
V ALIDITYOF THEBALLOONTECHNIQUE.
1t was necessary ta determine
whether the balloon or ifs tubing altered the completenessof velopharyn11
Thjs is shawn by direetly me88l1rÏn&the pre8Ure drop &Cros the reatin& orifiœ
in the following way (Figure 7). One catheter is inaerted into the IUbject's left
nostril, ~ured by a cork which plup the noatril and creates a stap&nt column of
air above the orifice. A second catheter, placed in the subject's mouth, is held in
position by clO8iDl the tipa tocether. Both catheters, clOled at their tipe but open
laterally for static pressure measurement. are connected to the dift'erential preSBUre
transduœr. With the heated ftowmeter connected to the riaht nOltril, the IUbject
is asked to breathe out through his Dose. The relUit is a dlft'erential pre.ure and
-1
flow record acro8 the n!8ting, open oriJiœ. When this orifice pre8Ure compoDent ia eubtracted from the previously discU88edpressure-ftow curve at correspondlng
flow rates, it becomes apparent that orifice dift'erential pre.ure is 80 small that it
can be ignored during expiration (Figure 6).
60
Warren, DuBois
1.4
1.3
1.2
-0
N
:I:
É
Co)
'
l'
.8
.
-
Q.. .6
.5
.4
.3
-
.2
Nasal Pressure Drap
Nasal + V.P. Orifice
Pressure Drap
0
50
100
150
200
250 - -300
350-
~cc/sec
FIGURE
preBlre
drap a~
ia actuaJly
6. Pre88ure-flow
&raph
of expiration
to detennine
nasal
and
orifice
compooenta.
At flow rates Jea than
approximateJy
~
CC/1eC the pre8Ure
the reating
open orifice
il tao eman ta be recorded.
The graph,
there(ore,
a measure
of nasal re8Ï8tance
up ta thia point.
geaI closure during speech.Since closure is most complete when phonating
plosive consonantsit was assumedthat if the balloon did have an effect it
".ould be demonstrated most clearly with these sounds. A preliminary
study was undertaken which involved two procedures perfonned by eight
normal adult subjects. Using the test sentenceBessie stayed aU summer
peak nasal ftow rates were compared for the consonant b under the following two experimental conditions: a) the pneumotachograph was connected to the right nostril and balloon inserted into the oropharynx by
way of the plugged left nostril, and b) the pneumotachograph was connected to the right nostril and a cork stopper placed in the left nostril.
ln this procedure, the balloon was eliminated leaving the orifice in a normal empty state.
The peak nasal ftow rates obtained during phonation of the consonant
b were compared statistically using a t test, and no significant difference
PRESSURE-FLOW
TECHNIQUE
61
A
FIGURE 7. Technique for me88l1ring the pre8Ure drop ac~
the resting open
orifice. The frontal view ia shown in A. The _gittal view, in B, showa how one
catheter ia plaœd into the subject'. left nostril, Î8 RCUred by a cork which pJup the
nostril, and creates a stagnant air column abave the orifice. The second cathcter il
placed in the subject'. mouth. Both cathetel'l are connected to a differential pre8Ure
tran8ducer. The aubjeet i. requested to breathe out through bis n~
with his lips
closed; air fIow ia measured by the ft.owmeter connected to the right nostril.
was found betweenthe two procedures. It may be assumedfrom this that
competency of closure is Dot altered by the balloon technique for measuring oropharyngeal pressure. Calibration of the balloon against an open
tipped catheter did reveal, however, a slight discrepancy (3%) in pressure
amplitude measurement,probably due to the slight 8tift'ness of the thin
walled balloon. The balloon was used to keep saliva out of the catheter.
EXPEBIMENTAL
TECHNIQUE.
Each of the 10 subjects was instructed to
perform the following procedures: First, to breathe lightly through his
Dosewith lip8 closed. (This gÎve8 a record for measurement of nasal resistance in the normal subject.) Second, to repeat the sentencesAre you
home papa and My tent is very clean.
The tl'ansitions between various sounds in continuous speech often are
so indi8tinct that boundaries between speech elements are difficult to
identify, if judged by criteria such as .here a consonant ends and a vowel
begins. For this reason, a method based on pressure, airftow, and sound
62
Warren, DuBois
p
J-lJ l\j_LA~_l..1'..~__Ji
s
-l'
1
,
I(MT
IISI'I(I.'ICllJ(AI-
v p~-L-U__ll-JJJ-O.SSec.
Pressure Col.-50rnmHzO
160
Flow CoLsIOOcc./s.c
4
~
Œ:
4
80
1
40
~
1
0 '"
.
1
1
1
1 f"
""1
1 1
1
1
1
1
1
l,
- 1 II~ -,-,--.
.
./
A
. . J'
1
1 1
1 Il
1 1 11 1 1 11 1
MIYI'
lE.'
111'IYIE~YICIL.AI"
FIGURE 8. Normal oroPharynpal pre8tlre, M>und,orifice airflow and orifice area
record of M~ tent ia veTl/ cl8an. Arrow8 point to where meuurementa were made.
Sound element boundarieaare establiabedaccording to physiological and acoustical
events.
PRESSURE-FLOW
63
TECHNIQUE
p
s
v
_l
III
Al M ..t~
l
~
0.5 Sec.
80
1
1
-N 60r-
PressureCol= 50mm. H1O
Flow Cot.8100cc./sec
E
E
- 40
~
UJ
0:
~ 20;
-.-
0""
area
Ilade.
;tical
1
1
1
1
1
1
.
~
Il Il 1 1
1 1 1 Il
ARIYO!HI 0 IMEI PIAIPIAI
E U
FIGURE 9. A normal oropharyngeal pre88Ure,sound, orifice airftow and orifiœ
area record of Are 1I'OUhome papa.
'es 8
, the
the
ends
ùnal
yof
'fard
midpoint of the voiced sound interval except in the few caseswhere it was
difficult to separate two speech elements. When this occurred the measurements were made at the beginning and end of the voiced sounds.
Since only one measurementwas made for each vowel or consonant, the
value obtained for orifice sile is an instantaneous measure of area at the
reference point used.
IS IS
Resuh.
Typical oropharyngeal pressure,voiced sound,orifice airftow, and orifice
area patterns of the two test sentencesare illustrated in Figures 8 and 9.
Occlusive and continuant consonanœcharacteristically are produced with
high oropharyngeal pressure and little, if any, orifice ftow. Nasal consonants, on the other band, are produced with high orifice ftow and very low
oropharyngeal pressure.
The influence of phonetic content on the area of the velopharyngeal
orifice is evident in the sentenceMy tent Î8 very clean (Figure 10). This
lned
sepDot
this
itted
ants,
the
le at
64
Warren, DuBois
300
- 200
N
.
E
E
<t
.
1
loi 150
.
II:
<t
t
.
.
.
100
50.
o~
..
.
.
0010
0
0
00
t 1
..
r ~
. . ..
.0
"0
.
.
.
0
"00"
00000
..
.
""_0_-
...
..
.
.
.
..
. .
~
M 1 y 1 T 1 E 1 NIT
1 Ils
1 VIE
IRYI CIL
lE AI N
FIGURE 10. Summary of orifiœ area data from 10 normal subjects phonating Mf/
tent U very cleo". Meaaurement8 are inBtantaneous values of orifice I!izc at the reference point of each speech element. Arrowa indicate that actual areas are greater than
the values noted (as explained in text). Note the effect of phonetic content on orifice
size.
effect was verified by statistical comparison of orifice size during production of the vowel e in the word lent and in the ward very; the analysis by a
nonparametric sign test revealed a highly significant difference in orifice
size for this vowel in different phonetic contexœ.
ln certain cases,dcpending upon flow rate, the pressure component due
to nasal resistance was equal to the oropharyngeal pressure recorded.
This meant that the orifice was open too wide for differential pressure to
be measured.When this occurred, an arbitrary value, equal to the lowest
pressure which could be measured at the highest amplifier sensitivity, was
used for the differential pressure value. Since the actual pressure was less
than this amount the calculated area was less than the actual ares. Thus
if the calculated area, using the arbitrary value for differential pressure,
equaled 124 mm2 then it was written as > 124 mm2.
That a nasal consonantalso influencesa closely associatedocclusive consonant is evident from the analysis of the plosive t pair in the ward lent.
Using the t test, a significant difference in orifice size (P < .02) was
found, the orifice being smaller for the initial t than for the terminal t.
These data indicate iliat bath vowels and consonants are influenced by
phonetic content, although the degreeof cffect is much greater for vowels.
PRESSURE-FLOW
65
TECHNIQUE
t
200
t
180
t
t
160
-N
140
ci
W
100
t
.
E 120
E
~
a::
~
t
t
80
60
40
20
0
.
t
.. .
...
A IREI
.~. ,
.~.
..
..
..
.. .
.
Y jOUi
t
&
.
.
..
Hia
IMEt
..
A.A
P
1 A
o.;
1
P
1 A
FIGURE Il. Summary of orifice area data from 10 normal subjects phonatinl
Are you home papa. Mea8Urementswere made at the reference points and arrOW8
indicate that actual area value! are creater than the values noted. The area graph
demonBtrates the effect of a oua! CODM)nant on c~ly
connected speech elements.
Note that the interval of preparation for the nasal consonant begins at h.
to
est
'88
~88
US
'as
by
~ls.
As expected, the greatest variation in orifice sile occurred with nasal
consonantsand with the nasalized vowels which precedethese consonants.
Occlusive and continuant consonants which require high oropharyngeal
pressurewere made with the tightest closure and least variation in orifice
8Ïae.
The eft'ect of a nasal consonant upon closely associated speech sounds
is also evident in the sentenceAre yau home papa. ln Figure Il it cao be
observed that the velopharyngeal orifice began to open as far in advance
as the voiceless h in preparation for the nasal m. For example, the mean
value of the m voiced interval was 120 millisec compared to a mean opening interval of 284 millisec prior to initiation of tJle sound. The orifice,
therefore, was open an average of 404 millisec or nearly 3.5 times longer
than the actual m voiced interval.
For comparison, in the sentenceMy tent is very clean, the orifice began
ta open 196 millisec (mean) in advance of the consonant n in clean. Since
the nasal sound lasted only 127 millisec (mean) the orifice was open 2.5
66
Warren,DuBois
times longer tJ1anthe nasal voiced interval. The difference in open interval time between m in the word home in Are you home papa (404 mil lisec) and n in the word clean in My tent iB very clean (323 millisec) cao
best be explained by dift'erencesiD phonetic context. The word clean begins with an occlusive consonant which requires a certain degree of closure for its production. The voiceless h in home, however, Dot requiring
high oropharyngeal pressure, cao be produced with the orifice open. Thus,
the discrepancy occurred becausethe orifice did Dot open until after the
production of c in the word clean in My tent is very clean, whereas it
opened during the production of h in the word home in Are you holne
papa.
DiBeuuion and Coneluaions
TECHNIQUE.
A method has been described for relating changes in velar
structure ta the acoustical characteristics of speech.The technique, based
upon pressure and flow measurements, provides a means for estimating
the functional ares of the velopharyngeal orifice during continuous
speech.Reproducibility of data for the model was good; reprodueibility
for normal subjectB based on repeating the sentences several times was
considered adequate by inspection, i.e., pressure and flow patterns are
similar.
The distinct advantage of this method over other techniques is that all
of the important parameters related to velopharyngeal function (pressure, airflow, orifice area and associated speech characteristics) tan be
asse88edand evaluated tagether. A disadvantage already noted is that
orifice size, when large, cannot be calculated if orifice airftow is small.
Presently, the only other technique available for evaluating orifice size
involves the use of transversal tomography during the production of isolated sustained sounds. Bjork reported in hie excellent monograph (1)
that transversal tomography, cineradiography, and sound spectrography
can be used ta determine orifice cross-sectional arcs during connected
speech. The validity of projection of two-dimensional cineradiographic
data ta orifice ares is dependent upon his suggestionthat the sagittal axis
of the coupling gate between the nasal and oral cavities is related to the
area of the cross-section. Although he considers the relationship linear,
the data he presentssuggesta possible quadratic relation between the two
parameters at small orifice sizes. 1t would be interesting to compare the
functional areas measured with this technique ta the anatomical areas
measured by Bjork's method.
PHONETIC
CONTENT
ANDORIFICESIZE. The data obtained in this study
support the notion that the function of the velopharyngeal sphincter is
primarily related ta the production of consonants and only secondarily
concerned with vo~.els, that is, the area of the orifice during pronunciation of vowels is dependent upon the type of consonant present in the
PRESSURE-FLOW
\11
sbe
at
se
Dl)
lY
~
ic
:is
:le
Ir,
..0
:le
as
Iy
is
Iy
~1e
TECHNIQUE
67
phonetic segment. Bjork (1) previously described nasalization of vowels
which precede nasal consonants; the present results also confinn those
findings.
Nusbaum, Foley and Wells (11) found that individuals may produce
vowels with an open orifice, although this does not necessarily imply a
noticeably nasal t,()ne.Moll (10) suggeststhat velopharyngeal closure on
vowels may vary &ystematically not only as a function of the vowel sound
produced but also as a function of the phonetic context of the vowel. He
reported that low vowels are made with greater velar opening than high
vowels (11). Our related findings disclose that variations in vowel orifice
size stem mainly from the sphincter's foie in adequate consonant production. The degree of influence appears t,() depend upon the type of consonant and whether the vowel is contained within its interval of preparation. The preparation interval for nasal consonants was found t,() be
approximately 2.5 t,()3.5 times longer than the actual sound. However, intervallength, in rom, was modified somewhat by the phonetic content of
the preceding segment,that is, the interval is shortened by the presence
of an occlusive consonant. Non-nasal consonants were aft'ected similarly,
but the degree varied with consonant type. Open consonants, not requiring high oropharyngeal pressure,were influenced to a greater degree than
occlusive or continuant consonants.
The important new point resulting from this study is that a consonant
influences the orifice size of all speechelements (vowels and consonants)
contained within its interval of preparation. The effect is dependent upon
bath the type of consonant and the type of vowel involved. The preparation period for a nasal consonant, being an opening velar movement, increasesthe size of the orifice for preceding sounds.ln contrast, this period
is a closing movement for occlusive or continuant consonants, and orifice
size is decreasedfor the preceding sounds. Because of these many interacting processes,absolute air-tight closure appears to occur only infrequently in continuoUBspeech.
ln a subsequent paper a technique will be described for use with cleft
palate subjects, and a comparison of normal and abnormal pressure, flow,
and orifice area patterns will be made (17).
Summary
A new technique which uses hydrokinetic principles has been described
for studying velopharyngeal function during continuous speech.The principle is based upon a modification of the Theoretical Hydraulic Equation
and assumesthat the area of the velopharyngeal orifice can be determined
if the diflerential pressure across the orifice is measured simultaneously
\\-ith airftow through it. The distinct advantage of this new method is that
parameters related to velopharyngeal closure (orifice size, oropharyngeal
pressure,orifice airftow and acoustic characteristics) can be assessedand
Warren, DuBois
68
evaluated simultaneously. Preliminary data obtained in a study of 10
normal adult subjects support the notion that the velopharyngeal sphincter is primarily concemed with consonant production and only secondarily concemed with vowel production. Variations found in vowel orifice
sile stem mainly from the sphincter's role in consonant production.
Department of Prosthodontica
&hool of Dentistry
University of North CaTolina
Chajel HiU, North CaTolina
ApPEXDIX:
DERlVATION
OF THE HYDROKINETIC
EQUATION.
The
basic
equations used to develop the "theoretical equation" are the equation of
continuity and the energy equation combined in a special form. For developing and discussingthe equation, the following symbols will be used:
A
p
Z
Z
=
area.
~
absolutestatic pressure
.
l
,=
,=
,
ergs / cm l
.;,.;
"
netlcenergy
=
dynes/cm'
'
mterna energy
ki
.
,~,
S = meanaxial speedof ftuid.
V -volume
,
'Y
cm'
",,:."""""
,...
,...:
'..'!:
specific weight of the fluid. . . . .,
ergs/ cml
cm/sec
cm 1
. : , , . . . . . . .. dynes/ cml
D =densityofair(O.OOl)
':
gm/cml
V = volume rate of airflow. . . . . . . .. . .. . . . . . . . .. cml/sec
Consider a steady stream flowing along a channel with rigid, imperviOU8
\valJs,from section Al to a section A, . According to the equation of continuity for ~ady ftow, the mass of ftuid passing any section Al per unit
time is constant and equal to that passing a second section A 2 per unit
time, thus
AISI
= A,s%
(3)
The general energy equation states that as each mass of fluid passes
from At to A%, the incre~ of its total energy, kinetic plus internai, is
equal to the \vork done on it plus the heat added to it. The \vork cloneupon
the ftuid due to the pressurechallge is
(PI VI
-
P2V2)
(4)
Sinœ the weight of air is negligible, the gravitational potential can be neglected. Thu8 the general energy equation becomes
(Zt%+ Z;,)
(Ztl + Z;I) = (PIVt - P,V%)
(5)
-
lt is 888umedin equation (5) that there is no heat transfer (adiabatic
flow). Now ordinary subsonic aerodynamics is an exarnple of incompressible ftow 80
li
(6)
PRESSURE-FLOW
(Zu
+
Z'2)
-
+ Z'I) = (Pl
(Zkl
-
TECHNIQUE
69
P2)V
(7)
Since practically no work of comp~ion or expansion is done and no
heat is transmitted to or from the walls. the internaI encrgy Z, cannot
change from either of these causes.However. some heat is generated by
',-ail resistance and in the continuaI damping out of eddy currents by viscosity. If Z~ representsthe quantity of thjs dissipated enel'gy. then
Z'2 - Zil = Z~
(8)
and
Z.2
- Z.I =
(PI
Sinœ flow is actually non-unifonn,
represented by
Zt
- P2) V -
Z~
(9)
the average kinetic energy may be
s' +
= 2g
.
1/1-
(10 )
in which 1/12
is the energy 1088due to non-unifom1 flo\\"",
FinaUy, in nearly aIl casesturbulent eddying ftow occurs and this component bas a kinetic energy of its own, in addition to that of the axial motion already present, Thus, if the averageamount of this ne", kinetic energy
be denoted by Z,2(Z, for turbulence), the complete expression is
Zt
S2 +
= 20
1/12
+ Z,2
(11)
12
Now replacing V with 1/., in equation (9) and rearranging,
13
x
=
Z4
+
(~22
-
~12)
-
+ (Z122 Zn')
(14)
The term x representsthe net eft'ect of rt'sistanœ, non-uniformity and
turbulenœ. If, between Al and A2, we neglect these factors represented
by x, a simplified approximate equation cao be written.
Thug
Stl
2g
~11 =
(PI
-
1
PI) -
"Y
15)
11
~
70
Warren, DuBois
50 that
Bt'- sa'- 2g(y)
16)
If we JetV be the volume rate of ftow then
(V/A) = S
(17)
acco~ng to the conti nuit y equation and
AIS.
=
- Ji
AtBl
(18)
If we DOWassumeAI to be very small compared to AI 50 that AI«
then SI » S) , we can neglect S)I compared to SIlo Bence
'1!1- -=-~)
Â1 t
(19)
l'
and
s = ..y2g
or rearranging
A~-
~
v
-
-
since
'Y-Dg
then
-
AI
v
~.~.
~2(~
(23)
Refereneea
1. BlOR",
L.,
Velopharyngeal
function
in connected
speech.
Acta
Radinl.
Suppl.
202,
1981.
2. BUNCKB,H. J., J.., Manometricevaluationof palatal function in cleft palate
patients,Ploatic reCOPIaIT.
81Irg.,13,148-158,1969.
3. CALNAN,
1953.
4. CHA88,
rec~r.
J. S..
R.
A.,
8urg.,
Movementa
An
objective
of
the
.,ft
evaJuation
palate.
of
Brit.
J.
palatopharynseal
piaatic
8urg.,
competence.
6,
286-296,
PIa,tic
16, 23-.19,1960.
5. CooPER,H. K., and HOFMANN,
F. A., The application of cineftuorography with
imap intensification in the field of plutic surgery, dentiBtry and speech.PIa.tic
recon,tr. 8urg.. 16, 135-137,1955.
6. GoaUJf, R.. and GO8LJN,S. G.. Hydraulic fonnula for calculation of the area of
the atenotic mitral valve, other cardiac valves, and central circulatory ahunts.
Amer. Heart., 41,1-29,1951.
7. HAGDTT, R. F., and H~,
F. S., Velopharyngeal cloeure: an index of
speech. PlGItic rec~r.8urg.,I",
290-298,1954.
l
PRESSURE-FLOW
TECHNIQ1.TE
8. HOF~N, F. A., Cineradiographic data reduction of 50ft palate movement. ln S.
Pruzansky (Ed.), p. 337. Congentibl Anomalies of the Face and Associated Strnctures. Springfield, Illinois: Charles C Thomas, 1961.
9. LILLY, J. C., Methods in Medical ReBearch. J. H. Comroe, Jr. (Ed.), Vol. 2, p
113-121.Chicago: Year Book Publ., 1950.
10. MOLL, K. L., CineBuorographic te<'hniques in speech research. J. speech heal;llg
Res., S, 227-2U, 1960.
Il.
MOLL, K. L., Velophaf)'ngeal
1962.
closure
on vowels.
J. speech hearing
Res., 5, 30-37,
12. NUSBAUM,E. A., FOLEY,L., and \\'ELLS, CHARLOTTE,
Experimental studies of the
firmness of the velar-pharyngeal occ.lusion during the production of Englil;lt
vowels. Speech M onogr., t, 71-80, 1935.
13. N"LÉN, B. O., Cleft palate and speech. Acta. Radiol. Suppl. 203,1961.
14. RANDALL,P., O'HARA, A. E., and BA~E8, F. P., A simplified x-ray technique for
the study of soft palate function in patients with poor speech. Plastic recon3tr.
Surg., "1, 345-356, 1958.
15. SUBTELNY,J., Roentgenography applied to the study of speech. ln S. Pruzltnl;k~.
(Ed.), p. 309. Congenital Anomalies of the Face and Associnted Str!lctllr".~.
Springfield, Illinois: Charles C Thomas, 1961.
16. \\'ARREN,D. W., Roie of the ."elophnryngeal Orifice in Speech: .\Qrmnl Pres.~"rf
and FloU' Relationships and the Effects upon theu I)f Cleft Palate. l"npublisht'd
PhD. diS8ertation, Univ. Penn~'lvania, 1963.
17. WARREN, D. W., Role of the velophar)'ngeal
and cleft palate subjects. (in pfPparation)
orifice
in speech:
Studies
in norulal
18. WARREN,D. W., and HOFMANN,F. A., A cineradiographic study of velophal,.ngeal
closure.Plastic reconstr.Surg.,t8, 656-669,1961.