Differences and intersubject variability of loudness discomfort levels measured in sound pressure level and hearing level for TDH-50P and ER-3A earphones.

Abstract

Loudness discomfort levels (LDLs) were measured in dB HL and SPL at discrete frequencies between 500 to 4000 Hz on 31 hearing-impaired ears using TDH-50P and ER-3A earphones. The results revealed no significant differences in the measured sound pressure level (SPL) between the two earphones at all test frequencies. However, with dB HL measurements, statistically significant differences were revealed at 1500 and 4000 Hz between earphone conditions. The results also revealed large intersubject differences in the measured LDL (HL and SPL) for both earphones. The results of this study highlight the difficulty in accurately predicting individual performance from averaged group data.

Full text

Page 1

Washington University School of Medicine
Digital Commons@Becker
Hearing Aid ResearchDivision of Adult Audiology
1997
Differences and intersubject variability of loudness
discomfort levels measured in sound pressure level
and hearing level for TDH-50P and ER-3A
earphones
Michael Valente
Washington University School of Medicine in St. Louis
Lisa G. Potts
Washington University School of Medicine in St. Louis
L. Maureen Valente
Washington University School of Medicine in St. Louis
This Article is brought to you for free and open access by the Division of Adult Audiology at Digital Commons@Becker. It has been accepted for
inclusion in Hearing Aid Research by an authorized administrator of Digital Commons@Becker. For more information, please contact
engeszer@wustl.edu.
Recommended Citation
Valente, Michael; Potts, Lisa G.; and Valente, L. Maureen, "Differences and intersubject variability of loudness discomfort levels
measured in sound pressure level and hearing level for TDH-50P and ER-3A earphones" (1997).Hearing Aid Research. Paper 12.
http://digitalcommons.wustl.edu/audio_hapubs/12

Page 2

• • • • JAm Acad Audiol 5: 390-398 (1994)
Intersubject Variability of Real-Ear Sound
Pressure Level: Conventional and Insert
Earphones
Michael Valente*
Lisa G. Potts*
Maureen Valentet
William Vass*
Joel Goebel*
Abstract
Measures of the sound pressure level (SPL) near the eardrum were determined at discrete
frequencies between 500 and 4000 Hz on 50 ears using TOH-39P and ER-3A earphones with
the attenuator of an audiometer fixed at 90 dB HL. Results revealed significant differences
in the measured SPL between the two earphones at all test frequencies. Results also
revealed large intersubject differences in the SPL measured near the eardrum for both
earphones. The results of this study highlight the large intersubject variability associated with
measuring the SPL at the eardrum and point out the difficulty in accurately predicting
individual performance from averaged group data.
Key Words: Intersubject variability, loudness discomfort level (LOLl, real-ear aided
response (REAR), real-ear insertion response (REIR)
R
use of this technology has been to determine if
the measured real-ear insertion gain (REIG)
"matched" a prescribed REIG. Recently, in­
creased attention has been placed upon using
REM to directly measure the sound pressure
level (SPL) near the eardrum corresponding to
the individual dynamic range between thresh­
old and suprathreshold levels. This dynamic
range, measured in dB SPL near the eardrum,
could then serve as a "target" to determine ifthe
real-ear aided response (REAR) for frequency­
specific or composite speech signals was placed
within the individual dynamic range using ei­
ther single or multiple input levels (Hawkins,
eal-ear measures (REM) have become
increasingly popular over the past sev­
eral years. Up to this point the primary
*Department of Otolaryngology, Washington Univer­
sity School of Medicine, St Louis, Missouri; tDepartment of
Communicative Disorders, St Louis University, SI. Louis,
Missouri; and "Starkey Laboratories, Eden Prairie, Minne­
sota
Reprint requests: Michael Valente, Washington Uni­
versitySchool of Medicine, 517S. EuciidAve., SI. Louis, MO
63100
1987; Seewald et aI, 1987; KaweU et aI, 1988;
Feigin et aI, 1989; Hawkins et aI, 1989; Cox and
Alexander, 1990; Hawkins et aI, 1990; Gagne et
aI, 1991a,b; MacPhersonet aI, 1991;Stelmacho­
wiczandSeewald, 1991;Stuartetal, 1991;Zelisko
et aI, 1992a, b; Valente et aI, 1993a; Skinner et
aI, 1993, 1994).
Instead ofdirect measures ofthe SPL near
the eardrum, several studies have suggested
that the real-ear SPL near the eardrum can be
predicted from audiometric thresholds mea­
sured in hearing level (dB HL) using either
conventional or insertearphones (Etymotic ER­
3A or E-A-R Tone® tubephones) by applying a
set ofaverage transformation values (Leijon et
ai, 1983; Walker et ai, 1984; Libby, 1985; Cox,
1986, 1988; Hawkins et aI, 1987; KaweU et aI,
1988; Skinner, 1988; Bentler and Pavlovic, 1989;
Hawkins et ai, 1990; Gagne et ai, 1991a, b;
Seewald et ai, 1991; Stuart et ai, 1991; Seewald,
1992; Zelisko et ai, 1992a). To illustrate this
point, a probe microphone system (Audioscan,
1992) was recently introduced that contains
software (i.e., SpeechmapTM) using threshold
values (in dB HL) to calculate and display the
predicted loudness discomfort level (LDL) in dB
HL. In addition, the user can also obtain the
•
390

Page 3

predicted SPL measured near the eardrum for
both threshold and LDL.
One shortcoming ofutilizing average trans­
formation values to predict the individual real­
ear SPL is the possibility of large intersubject
variability of the SPL measured near the ear­
drum for either TDH or insert earphones. This
may make it very difficult to predict accurately
the individual thresholds (measured in either
HL or SPL) from algorithms based upon aver­
age group data (Kamm et aI, 1978; Dillon et aI,
1984; Cox, 1985; Hawkins et aI, 1987; Kavell et
aI, 1988; Ross and Seewald, 1988; MacPherson
et aI, 1991).
The present study measured the SPL near
the eardrum for six discrete frequencies be­
tween 500 and 4000 Hz using conventional
(TDH-39P) and insert earphones (ER-3A) with
the attenuator ofthe audiometer fixed at 90 dB
HL. As a result of these measurements, the
magnitude of the intersubject variability was
determined.
METHOD
Subjects
The experimental group included 50 ears
from 25 adult subjects. The test ear of each
subject demonstrated normal middle ear func­
tion (Le., middle ear pressure within± 50 daPa;
static compliance between 0.6 and 1.8 mL)
using a Y226 probe tone from a calibrated GS
1733 middle ear analyzer. Hearing thresholds
were not an important factor for the purposes of
this study and, therefore, are not reported.
Procedures
For each subject, measurements were ob­
tained with TDH-39P (MX41/AR cushion) and
ER-3A (50 ohm) earphones connected to a cali­
bratedMaicoMA39 portable audiometer (ANSI,
1989) with the attenuator fixed at 90 dB HL.
The SPL near the eardrum was measured for
each earphone condition with continuous pure
tones of 500, 1000, 1500, 2000, 3000, and 4000
Hz using a probe tube coupled to a probe micro­
phone from a Frye 6500 real-ear analyzer. Mea­
sures were obtained only once because two
previous studies reported excellent test-retest
reliability for the equipment and procedures
used in this study. One study reported mean
intrasubject test-retest differences ofless than
1 dB for the real-ear unaided response (Valente
et aI, 1990), while the second study reported the
SSPLNaJente et aJ
same results for the real-ear insertion response
(Valente et aI, 1991). In addition, numerous
studies have reported on the test-retest reli­
ability of the ER-3A. For example, Clark and
Roeser (1988) reported that mean intrasubject
test-retest reliability was less than 2 dB for the
ER-3A and that the reliability ofthe ER-3Awas
equivalent to the TDH-50P. Larson et al (1988)
revealed that the standard error of estimate
was 0.9 to 1.5 dB for the ER-3A and TDH-50P.
Wilber et al (1988) reported on the results offive
studies and indicated that the standard devia­
tion for threshold measures was equivalent for
the TDH-39 and ER-3A earphones. Borton et al
(1989) reported that mean test-retest differ­
ences were less than 5 dB for the ER-3A. Fi­
nally, Lindgren (1990) and Frank and Vavrek
(1992) reported that intratester test-retest re­
liability was within 3 dB at 500 and 4000 Hz for
the ER-3A.
The probe tube was marked 30 mm from
the tip, and this mark was placed on the
intratragal notch. Inthe average adult ear, this
would place the tip of the probe tube approxi­
mately 4 mm from the eardrum, which is neces­
sary for accurate measures of SPL (Zemplenyi
et aI, 1985; Gilman and Dirks, 1986; Dirks and
Kincaid, 1987). The probe tube was then taped
into place to prevent movement. Great care was
used to assure that the 30-mm mark remained
in the same position as the diaphragm of the
TDH-39P was placed over the orifice ofthe ear
canal or when the immittance probe cuff from
the ER-3A was placed into the ear canal.
The ER-3A was coupled to the ear canal
using an appropriately sized Grason Stadler
immittance probe cuff. For this study, an
immittance cuffwas placed on a plastic adapter
(ER3-06) connected to the sound outlet tube and
coupled to the ER-3A and then to the ear canal.
Immittance cuffs were used for several reasons.
First, the diameter of the ear canal of several
subjects was either too large or small to suc­
cessfully use the standard foam plug. In a re­
cent article, Frank and Vavrek (1992) reported
that 17 percent oftheir subjects had ear canals
that would not allow the standard foam plug to
be used successfully. On the other hand, the
immittance cuffs used in this study have out­
side diameters varying from 2 to 22 mm. In
addition, the length of each immittance cuff is
16 mm. Insertion ofthe cuff so that the outside
edge was flush with the bowl of the concha
ensured a consistent insertion depth of 16 mm
pastthe openingofthe earcanal for all subjects.
This depth is precisely the 15- to 16-mm inser­
391

Page 4

Journal ofthe American Academy ofAudiologyNolume 5, Number 6, November 1994
tion depth recommended by the manufacturer
for a "deep" insertion. Finally, Borton et al
(1989) reported no significant differences in
threshold when ER-3A earphones were con­
nected to eitherfoam plugs or immittance cuffs.
To measure the SPL near the eardrum, the
reference microphone was "disenabled," and
the measured SPL was read directly from the
video monitor when activating the "Calibrate
Probe" software of the Frye 6500. Finally, the
probe microphone was calibrated daily using
the procedures suggested by the manufacturer
and all treatment levels ofearphone (TDH-39P
and ER-3A) and frequency (500, 1000, 1500,
2000, 3000, and 4000 Hz) were counterbal­
anced.
RESULTS AND DISCUSSION
Intersubject Variability
The range of intersubject variability in the
SPL measured near the eardrum for the TDH­
39P ranged from 9 dB at 1000 Hz to 36 dB at
4000 Hz (row 3; Table 1 and Fig. 1). In compari­
son, the range ofintersubject variability for the
ER-3A ranged from 12 dB at 1000 Hz to 29 dB
at 2000 Hz (row 6; Table 1 and Fig. 2). Even if
the highest and lowest data points from Figures
1 and 2 were removed, the intersubject variabil­
ity would still remain rather large. By remov­
ingthese extremes, therange ofthe intersubject
variability for the TDH-39P was reduced to 20,
8, 14, 12, 21, and 23 dB at 500 to 4000 Hz,
respectively. For the ER-3A, the range was
reduced to 18,8,16,17,15, and 17 dB at 500 to
4000 Hz, respectively.
Table 1 Mean, Standard Deviation (SO), ?
and Range of Measured Real-Ear SPL for the ?
TOH-39P and ER-3A Earphones at ?
Six Test Frequencies· ?
Frequency (Hz)
Earphone
500
1000 1500 2000
3000
4000
TDH-39P
Mean
SD
Range
99.3
4.9
20.0
99.0
2.4
9.0
98.7 103.1
3.5
16.0
101.1
5.6
30.0
95.2
7.3
36.0
4.6
23.0
ER-3A
Mean
SD
Range
88.9
5.8
23.0
92.9
2.9
12.0
96.3
43
21.0
99.5
5.6
29.0
92.6
4.0
20.0
88.7
5.6
25.0
Mean
Difference
104
61
2.4
3.6 8.5 6.5
N = 50 ears.
*Also provided is the mean difference in the measured
SPL between earphones.
Several possibilities may account for the
large intersubject variability ofthe SPL meas­
ured near the eardrum for the two earphones.
First, the subjects included in this study had
eardrum compliance that was within the nor­
mal range of 0.6 to 1.8 mL. Preves and Orton
(1978) reported that small differences in ear­
drum compliance, even for those that are within
.the normal range (0.31 to 1.20 cc), can result in
as much as a 6.5-dB difference between inser­
tion and functional gain. These authors did not
report if this variable was frequency depend­
ent. Dirks and Kincaid (1987) report on the SPL
measured for 3000 Hz at the eardrum and at
Figure 1:
Frontal Bone (FB) vs. Right Mastoid (RM) Placement
Mean Interaural SPL Difference:
30
--e-- REAC-LEAC: RM Difference
--e-- REAC-LEAC: FB Difference
20
1 0
o
- 1 0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Frequency (kHz)
Figure 1 Individual SPL mea­
sured near the eardrum at 500 to
4000 Hz for the TDH-39P ear­
phone (N = 50). The "0" repre­
sents the right ear and "X" repre­
sents the left ear. Also included
are ±1and 2 standard deviations
(SD).
392

Page 5

120'1r----------------------------------,
ER-3A Earphones
1<0
:00
80
70
- < ~ 1 ] ~ l x - ~ 1
x -1 -">
I
"­
~
l< <
~ ... , .
~ ~ . "i
o
o
~
o
~
SSPLNalente et al
Figure 2 Individual 8PL
measured near the eardrum at
500 to 4000 Hz for the ER-3A
earphone (N=50). The "0" rep­
resents the right ear and "X"
represents the left ear, Also in­
cluded are ± 1 and 2 standard
deviations (80),
O-LI-----j-____-j-__-j-_-j-__-j-_-j-____......l
500
1000
2000
Frequency (Hz)
varying probe positions in the ear canal for
eardrums with average, low-normal, and high­
normal impedance. They report that eardrums
having high-normal impedance will result in
SPLs that are lower than measured at the
eardrum with average impedance. For ear­
drums with low-normal impedance, they report
that the measured SPL will be higher than
measured in an eardrum with average imped­
ance. For both conditions, the difference in­
creases as the distance from the probe to the
eardrum increases.
Second, the procedure used in this study for
probe placement assured that the distance from
the orifice ofthe ear canal to the tip ofthe probe
tube was equal across subjects. However, the
distance from the end ofthe immittance cuff of
the ER-3A and diaphragm of the TDH-39P to
theeardrum probablyvaried quite widely across
the 50 ears due to intersubject differences in
actual canal length. Gilman and Dirks (1986)
and Chan and Geisler (1990) report that the
measured SPL at probe positions as far as 12
mm from the eardrum may be as much as 4 dB
less at higher frequencies relative to probe
positions closer to the eardrum. The likely pres­
ence of intersubject variability of canal length
led Bruell et al (1976) to call for developing
transferfunctions basedupon individual equiva­
lent volumes of the residual ear canal and
eardrum in order to more accurately predict the
SPL at the eardrum. As notedbyBentler(1989),
"wide intersubject variability ofresonance am­
plitude may be related, in part, to the small,
although significant, differences in probe-to­
eardrum distance differences among subjects"
(p.286).
4000
A third likely cause of the resulting
intersubject variability may be related to slit
leak, which affects the reliability of measures
at 500 Hz and below (Bruell et al 1976; Borton
et aI, 1989).
The results of this study seem to question
the validity and clinical accuracy as advocated
by some to predict the individual real-ear SPL
for threshold and suprathreshold measures
from averaged group data. This is clearly illus­
trated in Figures 3 and 4.
As mentioned earlier, a probe microphone
system (Audioscan) recently introduced a new
software package called Speechmap®. This soft­
ware calculates the predicted LDL (dB HL) (left
side of Figs. 3 and 4) as well as the predicted
threshold and LDL in dB SPL measured near
the eardrum (right side of Figs. 3 and 4) from
audiometric threshold entered in dB HL.
The dashed upper line in the left side of
Figure 3 is the threshold (dB HL) measured for
one subject using the TDH-39P earphone. The
lower solid line is the predicted LDL (dB HL).
The lower dashed line represents themeasured
LDL (dB HL). As can be seen, the agreement
between measured and predicted LDL is quite
good. On the right side of Figure 3 is the pre­
dicted threshold (lower solid curve) and LDL
(upper solid curve) measured in dB SPL near
the eardrum. The lower dashed line represents
the measured SPL for threshold, while the up­
per dashed line represents the measured SPL
for LDL. Again, the agreement between meas­
ured and predicted SPL for threshold and LDL
is quite remarkable.
Figure 4 reports the same measures for a
second subject. The upper dashed line on the
393