Compression Handbook PDF

The compression
Handbook
Third Edition
An overview of the characteristics and
applications of compression amplification
Shilpi Banerjee, Ph.D.

Starkey Hearing Research & Technology
Table of contents
Chapter 1 : Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Essential Terminology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Chapter 2: Basics of Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Characteristics of a Compressor. . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Visualizing Compression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Chapter 3: Applications of Compression. . . . . . . . . . . . . . . . . . . . . . 21

Avoiding Distortion, Discomfort and Damage . . . . . . . . . . . . . . . . 21
Optimizing Use of the Residual Dynamic Range
and Restoring Normal Loudness Perception. . . . . . . . . . . . . . . . . . 22
Maintaining Listening Comfort . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Maximizing Speech Intelligibility. . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Reducing the Adverse Effects of Noise. . . . . . . . . . . . . . . . . . . . . . 26
Chapter 4: Fitting Compression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Factors to Consider. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Prescribing Amplification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Chapter 5: Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Issues Resulting from WDRC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
What is Expansion?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Characterizing Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Measuring Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Chapter 6: Troubleshooting & Fine-tuning . . . . . . . . . . . . . . . . . . . . 42

A Systematic Approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Common Complaints and Solutions. . . . . . . . . . . . . . . . . . . . . . . . 45
Recommended Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2
Chapter 1: introduction
According to the World Health Organization For an individual with normal hearing,
(WHO), hearing loss is one of the leading average conversational speech falls
health concerns around the world (WHO, approximately midway within the dynamic
2005). At the present time, hearing aids range of hearing and coincides with the most
are the most common first step in [re] comfortable loudness level (MCL).
habilitation. On the surface, this seems fairly
straightforward because much is known The most common complaint associated
about the physiology of the auditory system with hearing loss is the inability to hear;
and the psychoacoustics of perception. specifically, the inability to hear soft
However, hearing loss impacts an individual sounds. Figure 1-1B depicts a person with
in numerous ways, making the fitting of a sensorineural hearing loss. Once again
hearing aid a complex process. the white bar is the range of sounds in the
environment, while the blue bar represents
The intact auditory system is capable of the individual’s dynamic range of hearing.
perceiving a wide range of sounds, from the You notice three things immediately. First,
quiet pitter-patter of rain to the loud boom average conversational speech is now barely
of explosives. In Figure 1-1A, the white bar audible to the individual. Second, weak
represents the entire range of sounds, from sounds (for example, gentle rain) are below
extremely weak to extremely intense, that the threshold of hearing and, therefore, too
may occur in an individual’s environment. soft to be heard. And, finally, intense sounds
The weakest sounds that are audible lie at (for example, the boom of an explosion)
the threshold of hearing sensitivity. At the are still perceived as being loud. As a result
opposite end is the loudness discomfort of the threshold increasing and the LDL
level (LDL), representing the most intense remaining the same, the dynamic range of
sounds that can be tolerated without pain. In hearing is considerably reduced compared to
between these two extremes is the dynamic that of a person with normal hearing.
range of hearing (shown by the blue bar).
Intense Too Loud Too Loud

Sounds
Moderate
Sounds
Weak Too Soft Too Soft

Sounds A B
Range of Environmental Sounds Dynamic Range of Hearing
Figure 1-1
The relationship between the range of sounds in the environment and the dynamic range of hearing for
persons with: (A) normal hearing, and (B) sensorineural hearing loss.
Figure 1.1
3
persons with: (A) normal hearing, and (B) sensorineural hearing loss
Human communication is arguably the single This handbook is designed to provide
most important function of the auditory the reader with a working knowledge of
system. Indeed, reduced ability to hear speech compression amplification: what it is, how it
is a major reason for seeking remediation. works, and how it is applied. Also included is
In addition to a reduced dynamic range, the some discussion on the principles of fitting
communication difficulties of a person with compression systems, and troubleshooting
hearing impairment are further complicated problems.
by the dynamic nature of speech itself. As
shown in Figure 1-2, average conversational
speech spans a range of 30dB (Beranek, 1947). Essential Terminology
Note that, in general, vowel sounds (for Before launching into the detailed workings of
example /a/, /u/, and /i/) are low-pitched, a compression circuit, it is important to have
relatively intense, and primarily responsible for some general knowledge of amplification.
making speech audible. On the other hand, Despite the variety available, all hearing aids
consonants (especially unvoiced sounds such have some of the same basic components:
as /th/, /f/ and /s/) are high-pitched, relatively a microphone, an amplifier, a receiver,
weak and carry most of the information that and a battery (Figure 1-3). The microphone
aids in speech understanding. picks up the incoming acoustic signal and
converts it to an electrical signal. The amplifier
Frequency [Hz] then magnifies the electrical signal. Like
Hearing Level in decibels (dB) re: ANSI 1996
125 250 500 1K 2K 4K 8K a loudspeaker, the receiver converts the

0 amplified electrical signal back into an acoustic
signal that is delivered to the ear. Finally, the
20
z v
f
s
th
battery provides the power for the circuit.
p k
40 j md b
n
i a hg
o r
ng
e l
ch
It is essential that the reader understand the
u
60 following terminology that relates to the signal
entering the hearing aid, the amplification, and
80
the sound that is delivered to the ear.
100
Input
120 Input refers to the acoustic signal entering
the hearing aid. Specifically, the American
Figure 1-2
Figure 1.2
Frequency and level of various speech sounds National Standards Institute (ANSI, 2003)
Frequency
at and levellevels.
average conversational of various speech
sounds at average conversational levels defines input level as the sound pressure level
(SPL) at the microphone opening of a hearing
aid. Input level is expressed in dBSPL.
In light of these considerations, the hearing
healthcare professional is faced with squeezing
an elephant into a suitcase when fitting hearing
aids – soft sounds must be made audible
without loud sounds becoming uncomfortable,
and speech should remain at a comfortable
level. Compression amplification is a means
for fitting the world of sound (the elephant)
into the dynamic range of the individual with
hearing impairment (suitcase).
4
Input Amplifier Output
Microphone Receiver
Battery
Figure 1-3
Figure 1.3
Basics components of a hearing aid: microphone, amplifier, receiver and battery.
Basics components of a hearing aid: microphone, amplifier, receiver, and battery
Output Gain
Output refers to the amplified signal that Gain refers to the amount of amplification
is delivered to the ear. The output level is applied to the input signal. Specifically,
expressed in dBSPL. ANSI (2003) defines gain as the difference
between the output SPL in a coupler and
Input/Output Function the input SPL. Gain is expressed in dB. The
An input/output (I/O) function is a graphical mathematical relationship between input, gain
representation of the output of a hearing aid and output is given by the simple formula:
at various input levels. ANSI (2003) defines it as
a single-frequency plot of the coupler SPL on 120
the ordinate (Y‑axis) as a function of the input 110

SPL on the abscissa (X-axis) with equal decibel 100
divisions on each axis; a similar definition is
Output (dBSPL)
90
used by the International Electrotechnical
80
Commission (IEC, 1983a). Figure 1-4 shows
70
a sample I/O function of a hearing aid. In
this example, an input of 50 dBSPL results 60
in an output of 80 dBSPL, while an input of 50

20 30 40 50 60 70 80 90 100
90 dBSPL results in an output of 110 dBSPL. It Input (dBSPL)
can also be seen that the output of the hearing Figure 1-4
Figure 1.4input/output function of a hearing aid.
Sample
aid does not exceed 110 dBSPL. Figure 1-5 Sample input/output function of a hearing aid
(page 6) shows sample I/O functions for

three hearing aids. Notice that they do not Gain = Output – Input
all behave in the same way – inputs of 50 and
90 dBSPL result in different outputs for the Thus, in Figure 1-4, if an input of 50 dBSPL
three hearing aids. results in an output of 80 dBSPL, the gain of
the hearing aid is:
Gain = Output – Input = 80 – 50 = 30dB
5
120
60
110
50
100
40
Output (dBSPL)
Gain (dB)
90
30
80 20
70 10
60 0
20 30 40 50 60 70 80 90 100
50 Input (dBSPL)
20 30 40 50 60 70 80 90 100
Figure 1.6
Input (dBSPL) Sample input/gain function of a hearing aid
Figure 1-6
Sample input/gain function of a hearing aid.
Hearing Aid 1 Hearing Aid 2
Hearing Aid 3
Figure 1.5
Figure 1-5
Sample input/output
Sample functions
input/output for three different
functions for threehearing aids.
60
different hearing aids.
50
Similarly, with an output of 105 dBSPL for an
40
input level of 90 dBSPL, Hearing Aid 3 (in
Gain (dB)
Figure 1-5) has a gain of 15dB. 30
20
Input/Gain Function
10
An input/gain (I/G) function is a graphical
representation of the gain of a hearing aid at 0
20 30 40 50 60 70 80 90 100
various input levels. Figure 1-6 shows a sample Input (dBSPL)
I/G function of a hearing aid. In this example,
the hearing aid provides 30dB of gain for an Hearing Aid 3
input of 50 dBSPL, but only 20dB of gain for Figure 1.7
Figure 1-7
an input of 90 dBSPL. Figure 1-7 shows sample Sample input/gain functions for three
Sample input/gain functions for three different hearing aids.
different hearing aids.
I/G functions for three hearing aids. Notice
that they do not all behave in the same way –
different amounts of gain are applied to inputs Similarly, with gain of 18dB applied to an input
of 50 dBSPL and 90 dBSPL. level of 90 dBSPL, Hearing Aid 3 (in Figure 1-7)
has an output of 108 dBSPL.
Just as gain can be calculated from the input
and output, output is calculated by rearranging An obvious, but frequently overlooked, point
the formula for calculating gain as follows: is that even though the gain of a device
decreases with increasing input level, the
Output = Input + Gain output continues to increase. This occurs
because the decrease in gain is less than the
Thus, in Figure 1-6, if 30dB of gain is applied increase in input level. Figures 1-8A and 1-8B,
to an input of 50 dBSPL, the output of the which show the I/O and I/G functions for a
hearing aid is: hearing aid, illustrate this point. 30dB of gain
applied to an input of 50 dBSPL results in an
Output = Input + Gain = 50 + 30 = 80 dBSPL
output of 80 dBSPL. On the other hand, 20dB
6
of gain applied to an input of 90 dBSPL results the hearing aid are fixed when measuring a
in an output of 110 dBSPL. In this example, frequency response curve. Figure 1-9 depicts
the input level increases by 40dB, while the a sample frequency response curve of a
gain decreases only by 10dB. Thus, although hearing aid for an input of 60 dBSPL. It can
less gain is applied to the 90 dBSPL input than be seen that the output of the hearing aid
to the 50 dBSPL input, the output SPL is still varies across frequencies. Figure 1-10 shows
greater for the 90 dBSPL input. sample frequency response curves for speech
presented at three different input levels. It is
120
not uncommon for the shape of the curve to
110 change as the input level increases.
120
100
110
100
Output (dBSPL)
90
100
90
80
Output (dBSPL)
90
Output (dBSPL)
80
70
80
70
60
70
60
50
60 20 30 40 50 60 70 80 90 100
50 Graph_Blue_FreqResponse_HA_3InputLevels
Input (dBSPL) 250 500 1000 2000 4000 8000
50
Figure 1-8A
Figure
20 30 1.8A
40 50 60 70 80 90 100 Frequency (Hz)
Sample input/output function for a hearing aid.
Input (dBSPL) Figure 1.9
Figure 1-9
60 Sample frequency
Sample responseresponse
frequency curve of a hearing aidatohearing
curve of an input
Figure 1.8A levelaid
of 60 dBSPL.
to an input level of 60 dBSPL.
50
60
40 100
50
Gain (dB)Gain (dB)
30
90
40
Output (dBSPL)
20
80
30
10
70
20
60
1020 30 40 50 60 70 80 90 100
Input (dBSPL) 50
Figure
20 30 1.8B
40 50 60 70 80 90 100 250 500 1000 2000 4000 8000
Sample input/output (A) and input/gain Frequency (Hz)
(B) functions for Input (dBSPL)
a hearing aid
Figure 1.8B Input Level:
Figure 1-8B
Sample
Sample input/output
input/gain (B) function (A)
for aand input/gain
hearing aid. 90 dBSPL 70 dBSPL 50 dBSPL
(B) functions for a hearing aid Figure 1-10
Figure
Sample 1.10 response curves of a hearing aid to speech
frequency
Sample
presented frequency
at three differentresponse
input levels.curves of a hearing aid
to speech presented at three different input levels.
Frequency Response Curve
A frequency response curve is a graphical
representation of the hearing aid output as
a function of frequency. Specifically, the IEC
(1983a) defines it as the SPL developed by a
hearing aid in the ear simulator expressed as
a function of frequency under specified test
conditions. The input level and overall gain of
7
A frequency-gain curve is a graph showing the Peak clipping is one method of controlling or
gain of a hearing aid as a function of frequency limiting the maximum output of a hearing aid.
(Figure 1-11) under specified test conditions. [Alternate methods of output limiting will be
discussed in subsequent chapters.] It is virtually
60 impossible to determine whether or not peak
clipping is occurring merely by examining
50
an I/O or I/G function. The hallmark of peak
40 clipping is that it produces an output signal
Gain (dB)
30 that is distorted, often described as sounding

“scratchy.”
20
10 Distortion
0 Distortion refers to the presence of frequency
250 500 1000 2000 4000 8000
components in the output of a hearing aid
Frequency (Hz)
Figure 1.11
that were not present in the input signal. There
Figure 1-11
Sample frequency-gain curve of a hearing are two types of distortion – harmonic and
Sampleaid
frequency-gain
to an inputcurve
levelofofa 60
hearing aid to an input level
dBSPL.
of 60 dBSPL. intermodulation.
Harmonic distortion is said to occur when

Peak Clipping
the output contains frequency components
In general, the output of a hearing aid
that are integer multiples of the input signal
increases as the input level increases. However,
frequency (IEC, 1983a). For example, harmonics
once the output reaches a certain level,
of a 500 Hz input signal would occur at 1000 Hz,
the hearing aid is incapable of producing a
1500 Hz, 2000 Hz, 2500 Hz, and so on. Thus,
louder signal. Maximum output is the highest
harmonic distortion only occurs at frequencies
possible signal that a hearing aid is capable
greater than the input signal frequency.
of delivering, regardless of the input level or
Total harmonic distortion (THD) is the
the gain of the hearing aid. The maximum
summed power of all the harmonic distortion
output of a hearing aid is determined by the
products relative to the power of the original
characteristics of the microphone, amplifier
input signal. THD is typically expressed as a
and receiver. That is, the maximum output of
percentage.
the hearing aid is only as high as the weakest
component in the chain. When the input level
Intermodulation distortion is generated
and gain exceed the maximum output, the
by the interaction of at least two signals of
hearing aid is said to be in saturation.
different frequencies in the input (IEC, 1983a).
It occurs when the frequency components of
As long as the output of the hearing aid remains
a complex input signal combine to generate
below the maximum output, the output signal
additional frequency components in the output
is similar to the input signal, only larger in
signal. For example, if F1 and F2 represent
amplitude (Figure 1-12A). When the sum of
two different frequencies, intermodulation
the input level and gain exceed the maximum
distortion may occur at frequencies
output of the hearing aid, however, the peaks of
corresponding to F2‑F1, 2F1‑F2, 2F2‑F1,
the output signal are clipped at the maximum
3F1‑2F2, and so on. Thus, intermodulation
output (Figure 1-12B). This is referred to as peak
distortion may occur at frequencies above and
clipping. Note that the shape of the output
below the input frequencies.
signal is quite different from that of the input
signal once the peaks are clipped.
8
Both types of distortion may occur Now that the reader is familiar with some
simultaneously in a hearing aid. Distortion of general hearing aid terminology, the next few
any type results in unpleasant sound quality chapters will introduce and describe the topic
and may adversely affect speech intelligibility. of compression.
A Input Ouput
Maximum Output of Device
B Input Ouput
Figure 1-12
Figure
Schematic of1.12
the response of a hearing aid: (A) without peak clipping, and (B) with peak clipping.
Schematic of the response of a hearing aid: (A) without peak clipping, and (B) with peak clipping
9
Chapter 2: Basics of compression
As discussed in Chapter 1, the intact auditory dynamic range of hearing available, also
system has an exquisite ability to hear a known as the residual dynamic range.
wide range of sounds in the environment, That is, the individual is unable to hear weak
from weak to intense (Figure 2-1A); average sounds, average conversational speech
conversational speech falls comfortably is barely audible, and intense sounds are
in the middle of this dynamic range. A heard as loudly as the normal-hearing ear
sensorineural hearing loss reduces the (Figure 2-1B).
Too Loud Too Loud
Too Soft Too Soft
A B
Too Loud Too Loud
Too Soft Too Soft
C D
Figure 2-1
Figure 2.1
persons with: (A) normal hearing, (B) sensorineural hearing loss, (C) sensorineural hearing loss with linear
The relationship
amplification, and (D)between the
sensorineural range
hearing lossof sounds
with in the
compression environment and the dynamic
amplification.
range of hearing for persons with: (A) normal hearing, (B) sensorineural hearing loss,
(C) sensorineural hearing loss with linear amplification, and (D) sensorineural hearing
loss with compression amplification
10
Because the most common complaint associated
with hearing impairment is an inability to hear, 60
it is tempting to alleviate the problem simply 50

by making sounds louder. Figure 2-1C shows
40
the sensorineural hearing loss to which a fixed
Gain (dB)
amount of gain is applied to all sounds – i.e., 30
amplification is linear. This makes weak sounds
20
audible. However, average conversational
speech is now loud, and intense sounds are 10
amplified beyond the upper end of the dynamic
0
range making them uncomfortably, or even 20 30 40 50 60 70 80 90 100
painfully, loud. Ouch! In contrast, in Figure 2-1D, Input (dBSPL)
different amounts of gain are applied to weak, Figure 2.3
Figure 2-3
Input/gain function of a linear
of hearing aidhearing
with 30dBaid
gainwith
and 30 dB gain
moderate and intense sounds – i.e., amplification Input/gain function
a maximum output of 110 dBSPL.
a linear
and a maximum output of 110 dBSPL.
is non‑linear. This squeezes the range of
environmental sounds to fit within the reduced
dynamic range of the person with sensorineural Before we begin discussing compression, it may
hearing loss. Weak sounds are made audible, be helpful to lay the groundwork with a brief
moderate sounds are comfortable, while intense description of linear amplification, the simplest
sounds are loud without being uncomfortable. and most basic amplification system. Figure 2-2
The result is that the hearing aid user perceives shows the I/O function for a linear hearing aid.
the world of sounds in much the same way as a As expected, the output increases as the input
person with normal hearing. to the hearing aid increases. More importantly,
for every 10dB increase in the input level, the
The exact manner in which non-linear output also increases by 10dB. For example,
amplification, or compression, is applied when the input level increases from 50 dBSPL
depends on the goal of the hearing aid fitting as to 60 dBSPL, the output level increases from
well as on the characteristics of the circuit. 80 dBSPL to 90 dBSPL. This 1:1 relationship
continues until the maximum output of the
hearing aid is reached. Notice also, that the
120 gain of the hearing aid is constant at 30dB until
110
Maximum Output the maximum output is reached. This is more
easily seen in Figure 2-3, which shows the I/G
100 function for the same linear hearing aid. The
Output (dBSPL)
90 reduction in gain for input levels greater than

80 dBSPL occurs because the maximum output
80 of the hearing aid cannot exceed 110 dBSPL.
70 Thus, the hallmark of a linear hearing aid is that
it applies a fixed amount of gain regardless of
60 the level of the input signal, until the maximum
50 output of the hearing aid is reached. In contrast,
20 30 40 50 60 70 80 90 100 a non-linear hearing aid applies different
Input (dBSPL) amounts of gain to weak, moderate and intense
Figure 2-2
sounds.
Figure 2.2
Input/output function for a linear hearing aid with 30dB gain
Input/output
and function
a maximum output for a linear hearing aid
of 110 dBSPL.
with 30 dB gain and a maximum output of
110 dBSPL.
11
Characteristics of a Compressor 120
The function and application of compression Linear Compression
110
circuits are defined by their static and dynamic
features. Static features, such as compression 100
threshold and compression ratio, indicate the
Output (dBSPL)
90
behavior of the circuit in response to steady 2dB
input signals (e.g., a running vacuum cleaner or 80
the constant hub-bub of a noisy restaurant). 70 TK

CT
On the other hand, dynamic features, such

60
as attack time and release time, describe
the length of time required for the circuit to 50
20 30 40 50 60 70 80 90 100
respond to a changing input signal (e.g., speech
Input (dBSPL)
in a one‑on‑one conversation).
Figure 2-4
Figure
Sample 2.4 function for a hearing aid showing the
input/output
Compression Threshold compression
Samplethreshold (CT) and function
input/output the threshold
forkneepoint (TK).
a hearing
Compression threshold (CT) is defined as the aid showing the compression threshold (CT)
and the threshold kneepoint (TK)
input SPL which, when applied to the hearing
aid, gives a reduction in the gain of 2 (±0.5)dB 120
TK
with respect to the gain in the linear mode (IEC,
110
1983b). In other words, it is the point on the I/O
function at which the output level is 2dB lower 100 TK
Output (dBSPL)
than it would be if no compression had occurred 90 TK

(i.e., if the processing were linear). Thus, in
Figure 2-4, the CT is 54 dBSPL. Because it looks 80
like the knee of a bent leg, the point at which 70

the slope of the I/O function changes is referred
60
to as the threshold kneepoint (TK). The hearing
aid shown in Figure 2-4 has a TK of 50 dBSPL. 50
20 30 40 50 60 70 80 90 100
Because the CT and TK are within a few dB
Input (dBSPL)
of each other, the two terms are often used
interchangeably, and will be used as such here.
Depending on the purpose, a compression Figure 2-5

Sample input/output functions showing threshold kneepoints (TK)
system may have high or low TKs. A high TK, Figure 2.5
for four different hearing aids.
Sample input/output functions showing
60 dBSPL or greater, is used to limit the output threshold kneepoints (TK) for four different
of a hearing aid so that it does not exceed the hearing aids
Compression Ratio
individual’s loudness discomfort levels and to
Once the input signal is loud enough to activate
maximize listening comfort (for example, Hearing
compression (i.e., when the input level exceeds
Aids 1 and 2, respectively, in Figure 2-5). On
the TK), the compression ratio (CR) determines
the other hand, a low TK, typically set below
how much the signal will be compressed.
60 dBSPL, may be used to improve audibility
Specifically, under steady-state conditions, it
of the softer components of speech and/or
is the ratio of an input SPL difference to the
to restore loudness perception (for example,
corresponding output SPL difference (IEC,
Hearing Aid 3 in Figure 2-5). Finally, like Hearing
1983b). Thus, CR relates a change in the input
Aid 4 in Figure 2-5, a device may have a high and
level (∆Input) to a change in the output level
a low TK to achieve all of these goals.
(∆Output). [The symbol ∆ is pronounced “delta”
12
and stands for a change in the target quantity.] example, Hearing Aid 3 in Figure 2-7). Finally,
CR is calculated using the formula: like Hearing Aid 4 in Figure 2-7, a device may
have a high and a low CR to achieve all of
∆Input
CR = these goals. Although necessary for some
∆Output applications, high CRs are known to adversely
affect the clarity and pleasantness of the
For the hearing aid shown in Figure 2-6, amplified sound (Neuman et al, 1998).
increasing the input from 30 to 50 dBSPL
(∆Input = 20dB) increased the output from
120
60 to 80 dBSPL (∆Output = 20dB). Using the
above formula, the CR of the hearing aid is: 110
100
∆Input 20 =2
:1
CR = = 1:1
Output (dBSPL)
= CR
O
∆Output 20 90
I
80
Similarly, when the input increases from 70 to
70 O
90 dBSPL and the output increases from 90 to
:1
=1
CR
100 dBSPL (Figure 2-6), the CR is 2:1. 60
I
50
CRs are generally expressed in terms of the 20 30 40 50 60 70 80 90 100
number of dB by which the input must change Input (dBSPL)
in order to effect a 1-dB change in the output. Figure 2.6
Figure 2-6
Calculating
Calculating compression
compression ratios
ratios (CRs) from(CRs) from an
an input/output
For example, a CR of 2:1 indicates that a 2-dB input/output
function. function.
∆I = Change I=
in input, ∆O = Change
Change ininoutput.
input,
change in input results in a 1-dB change in O = Change in output.
output. Because the reference condition is

always a 1-dB change in output, the latter part 120
of the ratio may be dropped. Thus, a CR of 2:1 CR=10:1
110
can be simply expressed as 2. R=3
:1
C
1 :1
.5: 2.5
100 R=
1 CR=
C
Output (dBSPL)
Note that, in Figure 2-6, 30dB of gain is

90
applied at and below the TK regardless
of the input level, resulting in linear 80
:1
=1
amplification. Thus, another way to look at CR
=1
:1
70 CR
linear amplification is that it has a CR of 1:1;
compression is applied only above the TK and 60
when the CR is greater than 1:1. 50

20 30 40 50 60 70 80 90 100
Depending on the purpose, a compression Input (dBSPL)

system may have high or low CRs. A high CR, Hearing Aid 1 Hearing Aid 2
5.0 or greater, is used to limit the output of Hearing Aid 3 Hearing Aid 4
a hearing aid so that it does not exceed the Figure 2.7

Figure 2-7
individual’s loudness discomfort levels (for Sample input/output
Sample
for fourratios
different
functions showing
input/output
hearing
compression
functions showing ratios
aids. different hearing aids
(CR)
compression
(CR) for four
example, Hearing Aids 1 and 2 in Figure 2-7).
On the other hand, a low CR, typically set
between 1.0 and 5.0, may be used to improve
audibility of the softer components of speech
and/or to restore loudness perception (for
13
Attack Time and Release Time Attack time (AT) is the time delay that occurs
When the incoming signal changes abruptly between the onset of an input signal loud
in level from below the TK to above it, the enough to activate compression (i.e., input
compressor is unable to change the gain signal exceeds the TK) and the resulting
instantaneously. The dynamic characteristics reduction of gain to its target value. Specifically,
of a compressor refer to the length of time ANSI (2003) defines AT as the time between
required for the compression circuit to respond the abrupt increase in input level from 55 to
to a sudden change in the input. Figure 2-8 is a 90 dBSPL and the point where the output level
schematic representation of how changes in the has stabilized to within 3dB of the steady value
input level produce variations in gain and output for an input of 90 dBSPL. [IEC (1983b) defines AT
over time. as the time interval between the moment when
Input exceeds TK
Input (dBSPL)
TK
A
Input below TK
Time (milliseconds)
Gain (dB)
[Gain is reduced when

B input exceeds TK]
AT RT
Time (milliseconds)
Overshoot
Output (dBSPL)
3dB
C
4dB
Undershoot
AT RT
Time (milliseconds)
Figure 2-8
Figurerepresentation
Schematic 2.8 of the behavior of a hearing aid associated with changes in input. The input level, gain and output
of the hearing aid are shown in panels A, B and C, respectively. The attack time (AT) and release time (RT) are indicated by the
Schematic representation of the behavior of a hearing aid associated with changes in input.
shaded areas. TK = Threshold kneepoint.
The input level, gain and output of the hearing aid are shown in panels A, B and C, respectively.
The attack time (AT) and release time (RT) are indicated by the shaded areas. TK=Threshold kneepoint.
14
the input signal level is increased abruptly by a time that the individual has to listen to sounds
stated number of decibels and the moment when louder than desired. Indeed, ATs as fast as 5 ms
the output SPL from the hearing aid stabilizes are especially desirable when compression is
at the elevated steady-state level within ±2dB. used to limit the maximum output of a hearing
The AT for the normal dynamic range of speech aid. Although RTs may also be a few milliseconds
in computed between input levels of 55 and in duration, the consequences when combined
80 dBSPL, whereas the high-level AT is computed with a fast AT may be undesirable because the
between 60 and 100 dBSPL.] In Figure 2-8, when gain will vary in response to each cycle of the
the input increases to a level above the TK incoming signal resulting in a distorted waveform.
(Figure 2-8A), the gain of the hearing aid does Thus, the RT is generally longer than the AT. An
not change immediately (Figure 2-8B). The result RT of 20 ms is considered fast. The disadvantage
is an overshoot in the output (Figure 2-8C). As of ATs and RTs between 100 ms and 2 s is that
the gain approaches its target, so too does the it causes the compressor to respond to brief
output of the hearing aid – i.e., reaches within sounds, or lack thereof, in the environment. For
3dB of its final value. example, although gain may be reduced in the
presence of speech, it will increase during pauses
Release (or recovery) time (RT) is the time delay in the utterance. This results in a pumping
that occurs between the offset of an input sensation where the level of the background (or
signal sufficiently loud to activate compression ambient) noise increases audibly during pauses
(i.e., input signal falls below the TK) and the and decreases when speech is present. This
resulting increase of gain to its target value. pumping sensation is less problematic for attack
Specifically, ANSI (2003) defines RT as the and release times faster than 100 ms because the
interval between the abrupt drop in input gain changes occur too quickly to be perceived.
level from 90 to 55 dBSPL and the point where A fast AT coupled with a slow RT (greater than or
the output level has stabilized to within 4dB equal to 2 s) will adversely affect the audibility of
of the steady value for an input of 55 dBSPL. speech that follows immediately after the gain
[IEC (1983b) defines RT as the time interval reduction in response to a finger snap or the
between the moment when the input signal click of a pen. Finally, attack and release times of
level is decreased abruptly by a stated number 2 s or slower respond to changes in the overall
of decibels and the moment when the output level of sound in the environment rather than to
SPL from the hearing aid stabilizes at the lower individual events.
steady-state level within ±2dB. The RT for the
normal dynamic range of speech in computed Some compression circuits incorporate
between input levels of 80 and 55 dBSPL; the adaptive or variable release times – that is,
high-level RT is computed between 100 and the RT is adjusted based upon the duration of
60 dBSPL.] In Figure 2-8, the gain of the hearing the triggering signal. Figure 2-9 (page 16) is a
aid does not change immediately (Figure 2-8B) schematic representation of the changes in gain
when the input decreases to a level below the and output associated with changes in the input
threshold kneepoint (Figure 2-8A). The result is level of different durations. Thus, for example, if
an undershoot in the output (Figure 2-8C), until the input is a brief, transient sound such as a door
both the gain and output reach their target – i.e., slam, the RT is fast to have the least possible
to within 4dB of their final values. effect on the audibility of the speech that follows.
On the other hand, an input that is sustained
Depending on the purpose, a compression is indicative of a change in the overall level of
system may have fast or slow attack and release sound in the environment. In this instance, a
times. The faster the AT, the shorter the duration slower RT acts in much the same way as a manual
of the overshoot and the shorter the period of adjustment of the volume control.
15
Input (dBSPL) Input exceeds TK
TK
A
Input below TK
Time (milliseconds)
Gain (dB)
B Gain is reduced when

input exceeds TK
AT RT AT RT
Time (milliseconds)
Output (dBSPL)
AT RT AT RT
Time (milliseconds)
Figure 2-9
Figure 2.9
Schematic representation of variable release time. The input level, gain and output of the hearing aid are shown in panels A, B, and C, respectively.
Schematic
The attack timerepresentation of variable
(AT) and release time release
(RT) are indicated bytime. Theareas.
the shaded inputTKlevel, gain and
= Threshold output of the hearing aid are shown
kneepoint.
in panels A, B, and C, respectively. The attack time (AT) and release time (RT) are indicated by the shaded areas.
TK=Threshold kneepoint.
Although it is convenient to discuss the static longer than the duration of a typical word
and dynamic characteristics of compression or utterance – the gain does not change
as discrete entities when learning about much between softer and louder phonemes.
them, it is important to understand that they As a result, the effective CR for speech is
interact with each other in systematic ways. much lower than would be expected for
For example, CRs are determined from the steady signals. Figure 2-10 shows the effects
response of the hearing aid to relatively steady of fast versus slow AT and RT on average
signals, such as pure tones or speech-shaped conversational speech. Specifically, note that
noise. For time-varying inputs, such as speech, the range between the upper and lower limits
the effective CR is significantly affected by the of speech is smaller with the fast AT and RT,
AT and RT. When the AT and RT are fast – i.e., indicating more compression of the signal,
shorter than the duration of a phoneme or than with the slow AT and RT.
syllable – the gain changes sufficiently quickly
to amplify softer components more than the Another example of the interaction between
louder components. The result is an effective compression parameters is the observation
CR for speech that is similar to that specified that fast ATs and RTs are more detrimental to
on the basis of steady signals. On the other the perceived sound quality at high CRs than
hand, when the AT and RT are slow – i.e., at low CRs (Woods et al, 1999).
16
Input-Versus Output-Controlled Compression
120
The term automatic gain control (AGC) is
often used to describe compression circuits 110
because the amount of gain applied is
100
automatically determined by the signal level.
Output (dBSPL)
Thus, a level detector is an essential component 90
of any compression circuit. The position of this 80

level detector relative to the volume control
70
influences the operation of the circuit.
60
With input-controlled compression (AGC-I),
50
the level detector is located before the volume 250 500 1K 2K 4K 8K
control (Figure 2-11) and compression acts Frequency (Hz)
on the input to the hearing aid. That is, once Fast Slow
the input exceeds the TK, the compressor
Figure 2-10
is activated and gain is reduced at the pre- Figure 2.10
Effects of fast and slow attack and release times on the response
Effectsaid
of a hearing of tofast and slow
a speech signal.attack and release
amplifier. Therefore, the volume control setting times on the response of a hearing aid to a
has no impact on the compression parameters. speech signal.
As shown in Figure 2-12 (page 18), rotating the

volume control from full-on to -20dB (relative
to full-on) decreases the gain of the hearing
aid above and below the threshold kneepoint.
However, neither the TK nor the CR changes as
a result.
Level Detector
Microphone Pre-Amplifier Volume Control Output Amplifier Receiver
FigureFigure 2-11
2.11
Schematic representation of an input-controlled compression (AGC-1) circuit.
Schematic representation of an input-controlled compression (AGC-I) circuit.
17
It is impossible to determine whether a
120 compression circuit uses AGC‑I or AGC‑O
110 simply by examining a single I/O function or
frequency response curve. The only way to
100
distinguish between the two types of circuits
Output (dBSPL)
n
O
90
Fu
ll is by measuring the response of the hearing
n
tio aid at various settings of the volume control.
VC
ta
Ro
80 n
Although gain changes are seen with both
B
tio
0d
ta
-1
Ro
70 Kneepoints types of circuits, the TK changes only in

B
0d
AGC-O circuits.
-2
60
50 Channels and Bands

20 30 40 50 60 70 80 90 100
Input (dBSPL) Although not strictly a “characteristic” of a
Figure 2.12
Figure 2-12 compressor, a brief discussion of channels and
Sample input/output
Sample functionsfunctions
input/output for an input-controlled
for an input-controlled
compression (AGC-1)(AGC-I)
circuit showing the effectsthe
of volume
bands is in order. The terms are often used
compression circuit showing effects of
control rotation.
volume control rotation. erroneously or interchangeably, but a distinction
is made between the two in this book.
In output-controlled compression (AGC-O),
Frequency bands are independently
the level detector is located after the volume
controlled areas for gain adjustment only.
control (Figure 2-13) and compression acts
Thus, increasing or decreasing the gain in a
on the output of the hearing aid. That is,
frequency band will equally affect the response
the compressor is activated once the output
to weak, moderate and intense sounds at
exceeds the TK. As shown in Figure 2-14,
frequencies within that band. The compression
rotating the volume control from full-on to
parameters are unaffected. In contrast,
-20dB (relative to full-on) decreases the gain of
compression channels allow separate
the hearing aid. More importantly, however, it
adjustments for weak and intense input levels;
results in an increase in the TK.
Level Detector
Microphone Pre-Amplifier Volume Control Output Amplifier Receiver
FigureFigure 2-13
2.13
Schematic representation
Schematic representation of anof an output-controlled
output-controlled compression
compression (AGC-O) circuit. (AGC-O) circuit.
18
the effect on moderate level inputs depends on A hearing aid may have the flexibility to allow
the compression architecture. Thus, in addition adjustments in multiple frequency bands and/
to gain, changes within a compression channel or compression channels. Multiple channels are
may affect the CR and/or the TK. Depending separated by crossover frequencies, which may
on the controls provided by the hearing aid or may not be adjustable. The ability to change
manufacturer, it may be possible to make gain individually in bands and/or channels is
changes across both, frequency bands and useful for two reasons. First, there is a wide
compression channels. variety of audiometric configurations for which
a given hearing aid may be useful. The ability
120 to adjust gain and compression in discrete
110
frequency regions permits customization of
Kneepoints amplification to the individual’s needs. Another
100 advantage of multichannel compression is that
Output (dBSPL)
90 acoustic events in a discrete frequency region

do not affect the response of the hearing aid at
n
O
80
ll
all frequencies. For example, Figure 2-15, shows

Fu
n
tio
VC
ta
the response of a 1-channel and 4-channel

Ro
70
n
tio
B
0d
hearing aid in the presence of a 90 dBSPL tone

ta
-1
Ro
60
B
at 500 Hz. It is clear that the high-frequency

0d
-2
50 response is considerably reduced for the

20 30 40 50 60 70 80 90 100
Input (dBSPL) single-channel hearing aid, but unaffected for
Figure 2.14 the multichannel hearing aid. This multichannel
Figure 2-14
Sample input/output functions for an output-
Sample input/output functions for an output-controlled advantage may also extend to advanced signal
controlled
compression compression
(AGC-O) (AGC-O)
circuit showing circuit
the effect showing
of volume
controlthe effect of volume control rotation.
rotation.
processing features such as noise management.
Although on the surface it appears that a large

number of channels would be advantageous, it
60 is unclear as to whether or not this is the case.
Trine and Van Tasell (2002) have shown that
50
3-4 channels are adequate to fit the majority
40 of audiometric configurations, with or without
Gain (dB)
30 the ability to further fine-tune the frequency

response within each channel (i.e. in frequency
20
bands), provided that the crossover frequencies
10 are adjustable. One argument against the
use of multichannel compression is that it
0
250 500 1000 2000 4000 8000 obliterates the relative intensity relationships
Frequency (Hz) between various speech sounds, which are
1-Channel 4-Channel an important cue for speech understanding
Figure 2.15
Figure 2-15
(Yund and Buckles, 1995). Finally, the sounds
Sample frequency-gain
Sample curves for a curves
frequency-gain 1-channel
forand
a a1-channel
4-channel and noises that occur in our environments are
hearing aid ain 4-channel
and the presencehearing
of a 90dBSPL
aid pure
in thetone at 500 Hz.of
presence
a 90dBSPL pure tone at 500 Hz.
usually not discrete in frequency. Thus, gain
would be affected in fairly broad frequency
regions even for a hearing aid with an infinite
number of channels. This problem is further
exacerbated when there is considerable overlap
across channels.
19
Visualizing Compression (peak clipping). Thus, when he emerges on the
other side of the bridge, he will look different
In summary, consider the following analogy to
(distorted) without his hat.
visualize the basics of compression. Picture a
gondolier guiding his gondola under a bridge.
At high water levels, the gondolier can avoid
The level of the water represents the input level
losing his hat by bending his head forward
and the height of the gondolier the gain of the
(high TK), as shown in Figure 2-16C. Note that
hearing aid. Thus, the output is the sum total of
he must bend his head as far forward as possible
the water level and the height of the gondolier.
(high CR). Also, if the water level rises further, he
The bridge represents the maximum output.
could still lose his hat.
In Figure 2-16A, the water level is low (weak
Another way that the gondolier can avoid losing
sounds) and the gondolier is able to guide the
his hat at high water levels is by bending forward
gondola under the bridge with ease (i.e., no
gently at the waist (low TK) (Figure 2-16D). In
gain reduction is necessary).
addition to being more comfortable (low CR),
this approach also affords him greater flexibility
If the water level rises (moderate or intense
in the event that the water level rises further.
sounds), as in Figure 2-16B, the gondolier will
lose his hat when going under the bridge
A B
C D
Figure 2.16
Figure 2-16
Schematic
Schematicforforvisualizing compression:
visualizing compression: (A) at low(A) atlevels,
water low water levels,
the gondolier theunder
passes gondolier passes
the bridge with his under
hat, (B) atthe
highbridge with
water levels, the gondolier would lose his hat while passing under the bridge. At high water levels, the gondolier could avoid
his hat, (B) at high water levels, the gondolier would lose his hat while passing under
losing his hat by (C) bending his head as far forward as possible, or (D) bending more gently from the waist.
the bridge. At high
water levels, the gondolier could avoid losing his hat by (C) bending his head as far forward as possible,
or (D) bending more gently from the waist.
20
Chapter 3: Applications of Compression
We have already determined that compression provides This type of application is referred to as compression
non-linear amplification. That is, the gain decreases limiting. Figure 3-1 shows a sample I/O function of a
as the input level increases. But, by how much should hearing aid with compression limiting. The following are
the input signal be compressed? At what input level? some desirable characteristics of a circuit designed for
And, how quickly? The answers to these questions this purpose.
depend upon the overall goal of the hearing aid fitting.
Compression may be used to: 120
CR=10:1
110
1. Limit the output of the hearing aid without distortion,
100
2. Minimize loudness discomfort, TK
Output (dBSPL)
90
3. Prevent further damage to the auditory system,
:1
=1
4. Optimize the use of the residual dynamic range, 80
CR
5. Restore normal loudness perception, 70
6. Maintain listening comfort, 60
7. Maximize speech recognition ability, and 50

20 30 40 50 60 70 80 90 100
8. Reduce the adverse effects of noise.
Input (dBSPL)
Figure 3.1
Figure 3-1
Compression circuits are defined by their features – Sample input/output
Sample function forfunction
input/output a hearing for
aid with output aid
a hearing
compression limiting.compression limiting
threshold kneepoint, compression ratio, and attack with output
and release times. At the present time, research offers
no compelling reasons for setting these parameters a
• AGC-O compression is used so that volume control
certain way. Thus, each must be adjusted to achieve a
adjustments do not affect the maximum output of
desired goal. Bear in mind that, to improve usability,
the hearing aid. For example, consider a situation
manufacturers may limit the adjustable parameters.
where the individual increases the volume control
of the hearing aid to listen to a child’s relatively soft
voice and a door is slammed shut. If an AGC-I circuit
Avoiding Distortion,
is used, the output of the device may exceed the
Discomfort and Damage LDL. In an AGC-O circuit, the maximum output could
Distortion, discomfort and damage all have be set just below the LDL and, more importantly,
the same foundation – intense sounds. Intense volume control adjustments would have no impact
sounds force a hearing aid into saturation causing on the maximum output of the hearing aid. Thus, this
distortion. Intense sounds may be amplified beyond application of compression is often called output
the individual’s LDLs causing discomfort. Finally, if compression limiting (OCL).
left unchecked, intense sounds entering a hearing
aid may cause amplification-induced hearing loss. • The TK is set high. Because intense sounds are of
While the latter two problems can be overcome primary concern, TKs of 70 dBSPL or greater are
simply by limiting the maximum output by other typically used. A high CR, greater than 8:1, is used
means, compression is the only way to prevent to prevent the amplified sound from exceeding
distortion. the LDL.
21
• A
fast AT minimizes the overshoot This type of application is referred to as
associated with a rapid increase in input wide dynamic range compression (WDRC).
level. Thus, ATs of 10 ms or less are generally Figure 3-3 shows a sample I/O function of a
used to limit the duration for which the hearing aid with WDRC. The following are
output of the hearing aid exceeds the LDL. some desirable characteristics of a circuit
• The RT is a less critical element because designed for this purpose.
it is associated with decreasing input and
output levels. That said, however, consider 120
the situation where a pot dropping on a tile 110
floor during a conversation forces a hearing
Output (dBSPL)
100
aid into compression limiting, resulting in 90
a severe reduction in gain. If gain releases
80
from compression limiting slowly, the speech
that follows may be inaudible. Thus, an RT of 70
100 ms or shorter is preferred; a circuit with 60
an adaptive release time may also be used. 50

125 250 500 1K 2K 4K 8K
Frequency (Hz)
• Finally, single- or multi-channel compression
Hearing Aid A (with OCL) Hearing Aid B (without OCL)
is suitable for this application.
90dB SPL Input 90dB SPL Input
50dB SPL Input 50dB SPL Input
Figure 3‑2 shows frequency response curves
Figure 3-2
Figure 3.2
for devices with and without OCL (Hearing Sample
Samplefrequency response
frequency responsecurves at input
curves levels
at input of 50
levels andand
of 50
90 dBSPL
90 dBSPL for for
hearing aidsaids
hearing without output
with and compression
without limiting
output compression
Aids A and B, respectively). Both hearing (OCL), Hearing
limiting Aids
(OCL), A andaids
hearing B respectively.
A and B respectively.
aids have identical responses for inputs of
50 dBSPL. However, at 90 dBSPL, the output of
Hearing Aid B is greater than that of Hearing
120
Aid A. This is an important consideration
because the output of Hearing Aid B could 110
exceed the individual’s LDL and, if saturation =2
.3:1
100 CR
occurs, the associated distortion will be high.
Output (dBSPL)
90
Optimizing Use of the 80

TK
1:1
Residual Dynamic Range and 70
CR
=
Restoring Normal Loudness 60
Perception 50
20 30 40 50 60 70 80 90 100
As indicated previously, hearing impairment Input (dBSPL)
results in a loss of sensitivity for weak Figure 3.3
Figure 3-3
sounds, with little or no loss of sensitivity for Sample input/output
Sample
dynamic
function forfunction
input/output a hearing aid
for with wide
a hearing
aidrange
withcompression.
wide dynamic range compression.
intense sounds. Thus, in order for the range
of environmental sounds to fit within the
residual dynamic range of the individual, more
amplification is required for weak sounds
than for intense sounds. The net result is that
weak sounds are audible, moderate sounds
comfortable, and intense sounds are perceived
as loud without causing discomfort.
22
• A
GC‑I is used because the goal is to achieve
120
a certain degree of audibility and/or loudness
TK CR=10:1
for incoming signals. If average, rather than 110
individual, data are used to derive the amount TK
100 3:1
of amplification necessary, the option may CR=
Output (dBSPL)
exist to increase or decrease the output of the 90
hearing aid by adjusting the volume control. :1
80 =1
CR
• T
he TK is as low as possible in order to make 70 :1
=1
weak sounds audible. Thus, the TK is typically CR
60
set at or below 50 dBSPL. Hearing aids with
TKs as low as 20 dBSPL are said to use full 50
20 30 40 50 60 70 80 90 100
dynamic range compression (FDRC) because
Input (dBSPL)
they aim to compress the entire gamut of
Figure 3-4
environmental sounds in the residual dynamic Sample input/output functions for a hearing aid
Hearing Aid A with wide
Hearing Aid B
dynamic range compression (Hearing Aid A) and a linear hearing
range of the individual. aid with output compression limiting (Hearing Aid B).
Additional gain for
low input levels
• L ow CRs of 4:1 or less can be used because
compression acts over a wider range of inputs. Figure 3.4
Mueller (2002) gives the example that WDRC Sample input/output
Uncomfortably Loud functions for a hearing
aid with wide dynamic range compression
is akin to gentle braking as one approaches Very Loud
(WDRC, Hearing Aid A) and a linear hearing
a stop sign while driving. [By the same token, aid with output compression limiting (OCL,
Loud but Okay
Hearing Aid B).
compression limiting is more like screeching to
Loudness Rating
Comfortable but Loud

halt at the last minute!] Comfortable
• T
he AT and RT may be fast or slow. However, Comfortable but Soft
as indicated earlier, several studies have shown Soft
that fast ATs and RTs significantly degrade Very Soft

20 30 40 50 60 70 80 90 100
sound quality (Neuman et al, 1998; Woods Input (dBHL)
et al, 1999). Thus, the use of ATs slower than Normal hearing (A) Sensorineural hearing loss (B)
about 100 ms and RTs slower than 2 s is Linear with OCL (C) WDRC (D)
advisable. Figure 3-5
Figure
Sample 3.5loudness growth curves for persons with: (A) Normal
Sample
hearing,loudness growth curves
(B) Sensorineural for persons
hearing with: (A) Normalhearing,
loss, (C) Sensorineural hearing
• M
ultichannel compression is used to (B)loss
Sensorineural hearing loss, and
with linear amplification (C) Sensorineuralhearing loss with
output compression limiting
linear amplification and output compressionlimiting (OCL), and
(OCL), and (D) Sensorineural hearing loss with wide dynamic
accommodate audiometric configurations that (D) Sensorineural hearing loss with wide dynamic range compression (WDRC)
range compression (WDRC).
deviate substantially from flat. Large variations
in the degree of hearing loss across the
frequency range also significantly change the
perception of loudness.
As shown in Figure 3-4, the I/O function of a

hearing aid with WDRC (Hearing Aid A) is visibly
different from that of a linear hearing aid with
OCL (Hearing Aid B). In addition to the TK and
CR being lower, Hearing Aid A also provides
more gain than Hearing Aid B at and below
the TK. This additional gain results in improved
audibility of weak sounds.
23
The effects of hearing impairment on loudness for much of the time. There are several ways in
perception and compensation through the use which this problem may be addressed that do
of wide dynamic range compression will be not require the user to constantly adjust the
discussed at some length in the next chapter volume control.
in the context of fitting hearing aids with
compression. For now, consider the loudness First, as discussed in the preceding section,
growth functions shown in Figure 3-5 (page 23). WDRC employs a low TK and CR. This results
A loudness growth function is a graph showing in intense sounds approaching the maximum
the perceived loudness of a sound as a function output of the hearing aid more gradually
of input level. As expected, the loudness than linear amplification coupled with OCL
rating increases with input level. However, (Figure 3-5). The characteristics of a circuit
the shapes of the loudness growth functions designed for this purpose are much the same
are significantly different for persons with as those described in the previous section for
normal hearing and those with sensorineural optimizing use of the residual dynamic range
hearing loss. In addition, note that the “loss and restoring loudness perception. That is,
of loudness” (as measured by the difference multichannel AGC-I compression should be
between the two curves) is greater for sounds used with low TK and CR, and long ATs and RTs.
rated as being soft, with little or no loss for
sounds rated as loud. Linear amplification with The second method of maintaining listening
OCL provides a fixed amount of gain regardless comfort has no official name, but has been
of the level of the incoming signal, making all called mid-level or comfort-controlled
sounds louder by the same amount. Thus, the compression (Byrne and Dillon, 1986).
shape of the loudness growth function remains Figure 3-6 shows a sample I/O function of
the same as that for sensorineural hearing comfort-controlled compression. The following
loss, but is shifted to the left. The result is that are some desirable characteristics of a circuit
weak sounds may still be inaudible, moderate designed for this purpose.
sounds are comfortable, and intense sounds are
too loud. In contrast, when WDRC is used, the • AGC-I compression is used because the
resulting loudness growth curve is very similar goal is to achieve the desired loudness for
to that of a person with normal hearing. In other moderate to intense inputs.
words, weak sounds are now audible, moderate
• The TK is set at approximately 60 dBSPL so
sounds comfortable, and intense sounds are
that compression can operate on moderate
perceived as loud. Note that experienced users
and intense sounds.
of linear amplification may object to the use of
WDRC because they have become accustomed • Low CRs of 4:1 or less can be used because
to intense sounds being too loud and weak compression acts over a wider range of
sounds being inaudible. inputs than for OCL.
• ATs and RTs may be fast or slow, but should

Maintaining Listening Comfort probably be slow to preserve sound quality.
Thus, the use of ATs slower than about
Even when the maximum output of a hearing
100 ms and RTs slower than 2 s is advisable.
aid is appropriately restricted to within the
residual dynamic range, it may still be desirable • Single- or multi-channel compression is
to further reduce the output of the hearing aid suitable for this purpose.
to a level below the maximum output (or LDL)
24
Figure 3-7 compares mid-level compression Maximizing Speech Intelligibility
(Hearing Aid A), WDRC (Hearing Aid B) and
Arguably, the single most important criterion
linear amplification with OCL (Hearing Aid C).
for maximizing speech intelligibility is increased
Note that below the kneepoint, gain for
audibility. That is, speech must be audible before
mid‑level compression is similar to that for linear
one can be expected to understand it.
amplification with OCL; in contrast, gain is greater
for the hearing aid with WDRC. Although it does
Treble-Increase-at-Low-Levels (TILL) was
not increase speech audibility, this alleviates
introduced (Killion, 1991) as a means to improve
the primary disadvantage of WDRC – increased
the intelligibility of speech in single-channel
likelihood of feedback resulting from the provision
devices. As the name suggests, this circuit
of more gain. Above the TK, mid-level compression
provides a high-frequency boost for weak sounds.
takes advantage of reduced gain for intense sounds
When the overall input level is relatively low,
when compared to linear amplification with OCL.
high-frequency consonants are amplified more
As a result, like WDRC, this assists in maintaining
than the low‑frequency vowels. This additional
listening comfort.
high‑frequency gain not only makes the
120 consonants more audible, but also reduces the
upward spread of masking by the more intense
110
:1 vowels. Although high-frequency emphasis
=2
CR
100 is potentially useful at all input levels, it also
increases the likelihood of loudness discomfort
Output (dBSPL)
90
at high input levels (Skinner, 1976). Thus, for
80 TK
:1 intense sounds, the frequency response curve is
=1
70 CR relatively flat. Figure 3-8 shows frequency-gain
60
curves for a TILL circuit at inputs of 50, 70 and
90 dBSPL. Note that, for an input of 50 dBSPL,
50
20 30 40 50 60 70 80 90 100
the high‑frequency emphasis is in addition to
Input (dBSPL)
the gain that would ordinarily be prescribed for
Figure 3-6
Figure 3.6 the hearing loss. Further, the shape of the curve
Sample input/output function for a hearing aid with
Sample input/output function for a hearing aid
mid‑level compression. changes with the input level – i.e., it gets flatter as
with mid-level compression.
the input level increases. Improved customization
120 of the frequency response through the use of
CR=10:1 multiple channels has resulted in the demise of
110
2.5
:1 single‑channel TILL processing.
TK CR=
100
Output (dBSPL)
60
90
50
80 TK
:1
40
=1
CR
Gain (dB)
70 30
:1
=1
CR
60 20
50 10
20 30 40 50 60 70 80 90 100
0
Input (dBSPL) 100 200 500 1K 2K 10K
Frequency (Hz)
Hearing Aid A Hearing Aid B
Input Level: 50 dBSPL 70 dBSPL 90 dBSPL
Figure 3-7 Hearing Aid C
Sample input/output functions for hearing aids with mid-level Figure 3.8
Figure 3-8
Figure(Hearing
compression 3.7 Aid A), wide dynamic range compression Sample
Sample frequency
frequency responsecurves
response curvesshowing
showing treble-increase-
(Hearing Aid B) input/output
Sample and linear amplification
functionswith for
output compression
hearing aids at-low-levels (TILL).
treble‑increase‑at‑low‑levels (TILL).
limiting (Hearing Aid C).
with mid-level compression, wide dynamic range
compression (WDRC) and output compression
limiting (OCL). 25
Today, WDRC is the primary means for levels are shown in Figure 3-9. Speech has a
maximizing speech intelligibility. The following dynamic range of 30dB. As a result, there may
are some desirable characteristics of a circuit be components that are even weaker with linear
designed for this purpose. amplification and OCL than shown in the figure.
• AGC-I compression is used because amount 110

of gain applied depends on the level of the
100
incoming sound.
90
• The TK is as low as possible, at or below
Output (dBSPL)
80
50 dBSPL, in order to make the weaker
components of speech audible. 70
60
• CRs of 4:1 or less can be used because
compression acts over a wider range of 50
inputs. 40
250 500 1K 2K 4K 8K
• ATs and RTs should be faster than the Frequency (Hz)
duration of a typical syllable to provide more Hearing aid A (with WDRC)

50 dBSPL Input 70 dBSPL Input
amplification for the weaker components 90 dBSPL Input
than for the more intense components of Hearing aid B (linear with OCL)
50 dBSPL Input 70 dBSPL Input 90 dBSPL Input
speech.
Region of increased comfort Region of improved audibility
• Multichannel compression is used so that Figure 3-9
Sample frequency
3.9 response curves for a hearing aid with wide
weak consonant sounds can be amplified Figure
dynamic range compression (WDRC) and a linear hearing aid
Sample frequency response curves for a hearing
independently of the more intense vowel with output compression limiting (OCL) to speech inputs of 50,
aid with wide dynamic range compression (WDRC)
70 and 90 dBSPL.
sounds. Increasing the consonant-to-vowel and a linear hearing aid with output compression limiting
(OCL) to speech inputs of 50, 70 and 90 dBSPL.
ratio (CVR) – i.e., the intensity of a consonant
relative to that of a vowel – has been Reducing the Adverse Effects
shown to improve speech understanding of Noise
(Montgomery and Edge, 1988).
A common complaint regarding hearing aids
relates to the effects of noise. Although noise
The advantage of WDRC over linear
management schemes are beyond the scope of
amplification with OCL, from the point of
this handbook, the effects of compression will
view of improving speech audibility, is shown
be discussed. The problem with noise is two-
in Figure 3-9. Consider two hearing aids that
fold. First, amplification of intense sounds, such
are set such that conversational speech is
as a vacuum cleaner, may cause the listener
comfortable for the person with a sensorineural
distress even if the LDL is not exceeded. And,
hearing loss. We have already seen that, for
second, communication is challenging in noisy
intense sounds, the output of the hearing
environments, such as at a party or restaurant.
aid approaches LDL more slowly with WDRC
than with linear amplification and OCL. This
Hearing aids with Automatic Signal Processing
is indicated in the Figure by the region of
(ASP) were introduced in the 1990s to counter
increased comfort. More importantly, weak
the deleterious effect of background noise. The
sounds are made audible by WDRC (indicated
principle is based on two assumptions: (1) the
by the region of improved audibility); the
overall level of sound is relatively high in noisy
same cannot be said of the linear hearing aid
environments, and (2) the hubbub of parties and/
with OCL. Keep in mind also that only average
26
or restaurants is dominated by energy in the (SNR) when the outputs of channels with poor
low frequencies. Thus, ASP hearing aids reduce SNR are reduced relative to the outputs of those
the amount of low-frequency amplification where the SNR is good. It is important to note
for high input levels to alleviate the loudness that the SNR is not improved within any given
of the noise. Low frequency amplification is channel.
restored at low input levels. One can think
of this as Bass-Increase-at-Low-Levels (BILL), Although it is convenient to compartmentalize
which is another common name for ASP the applications of compression, more
circuits. A further advantage of reduced low- than one goal can be achieved in a single
frequency amplification at high input levels is fitting. For example, multichannel WDRC
the decreased likelihood of upward spread of may be used to optimize use of the residual
masking. Figure 3-10 shows the frequency-gain dynamic range, normalize the perception of
curves of BILL circuits for speech presented loudness, maintain listening comfort, maximize
at 50, 70 and 90 dBSPL. Like TILL circuits, BILL the intelligibility of speech, and reduce
processing is seldom used nowadays because of the adverse effects of noise. Figure 3-11
the widespread use of multichannel hearing aids summarizes the compression characteristics
to achieve the same objective. associated with each application.
The compression of sound in multiple 60
channels reduces the adverse effects of noise 50

in several ways that provide an advantage 40
over single‑channel BILL processing. First,
Gain (dB)
30
no assumptions are made regarding the
frequency composition of the noise. Thus, gain 20
is reduced in any frequency region where the 10
input levels are high. Second, gain is reduced 0

100 200 500 1K 2K 10K
only in frequency regions where a great deal Frequency (Hz)
of noise is present; gain and audibility in the
Input Level: 50 dBSPL 70 dBSPL 90 dBSPL
remaining channels are unaffected. Third, when
Figure 3.10
Figure 3-10
the spectra of the signal and noise are different, Sample
Sample frequency
frequency response
response curvesshowing
curves showing bass-increase-
at-low-levels (BILL).
multichannel compression may provide a slight bass‑increase‑at‑low‑levels (BILL).
improvement in the overall signal-to-noise ratio
Application Input-or-Output Threshold Compression Attack & Number of

Control Kneepoint Ratio Release Times Channels
Avoid Distortion, Single

Discomfort and Damage Output High High Fast
or Multiple
Optimize Use of Residual
Dynamic Range & Restore Faste or Slow
Input Low Low Multiple
Loudness Perception (slow preferred)
Maintain Listening Input WDRC: Low WDRC: Low Fast or Slow Single
Comfort Mid-Level: High Mid-Level: High (slow preferred) or Multiple
Maximize Speech
Input Low Low Fast Multiple
Intelligibility
Figure 3-11
Figure
Summary of3.11
compression applications and associated characteristics.
Summary of compression applications and associated characteristics.
27
Chapter 4: Fitting Compression
Factors to Consider aids – recall that WDRC provides more gain at low
input levels than linear amplification. Hearing aids
Simply knowing that an individual has a hearing
utilizing WDRC are now available that provide the
impairment is not sufficient for selecting and
additional gain needed for severe and profound
fitting a hearing aid. At a minimum, a case history
hearing losses.
and audiometric evaluation are required. These
will provide the professional with insight into
Type of Hearing Loss
audiological and non-audiological factors that
A hearing loss always results in a loss of sensitivity
must be taken into account when recommending
for soft sounds. However, the impact on the
the use of amplification.
perception of sounds at suprathreshold levels
differs dramatically depending on the type of
Non-audiological considerations include perceived
hearing loss. Further, each presents unique
hearing difficulty, cosmetic concerns, cognitive
challenges to the hearing aid fitting.
capabilities, and physical limitations. In addition
to assisting the professional in selecting a style of
A conductive loss, due to pathology in the outer
hearing aid and suitable features, knowledge of
and/or middle ear, results in an attenuation
the individual’s needs and lifestyle can be used to
of the incoming sound. While the degree of
set appropriate expectations. Although important,
attenuation may vary with frequency, it applies
discussion of these factors is outside the scope of
equally to weak, moderate and intense sounds
this handbook.
(Figure 4-1A). The effect is similar to wearing
earplugs – the hearing threshold, MCL, and LDL
Degree of Hearing Loss
are all elevated by the amount of conductive loss
It has been reported that the amount of
(Walker, 1997). As a first approximation, simply
perceived benefit from hearing aids is directly
amplifying all sound levels by the same amount –
correlated with the degree of hearing loss
i.e., linear amplification – is a reasonable solution
(Gobalek et al, 1988) – that is, the more severe
(Figure 4-1B). However, compression may also be
the hearing loss, the greater the benefit from
used to avoid the consequences of operating the
amplification. However, this should not be
hearing aid at or close to its maximum output for
construed as a contraindication for fitting
much of the time.
hearing aids to mild hearing impairments.
In contrast, a sensorineural hearing loss results in
Persons with mild to moderately severe hearing
a greater loss of sensitivity for weak sounds than
loss benefit tremendously from the improved
for moderate or intense sounds. In fact, LDLs may
audibility and increased listening comfort offered
not even increase beyond the normal range for
by WDRC, preferably used in combination with
hearing losses of 50dBHL or less (Kamm et al,
OCL. Traditionally, linear amplification, with or
1978). Thus, some form of compression is highly
without OCL, has been used for individuals with
desirable for those with sensorineural hearing loss.
severe to profound hearing losses. Contrary to
Unless otherwise specified, the discussion in this
the popular belief that this population cannot
book relates to issues surrounding sensorineural
take advantage of increased audibility for soft
hearing loss, with or without a conductive
sounds, the primary reason for not using WDRC
component.
has been the lack of sufficiently powerful hearing
28
Too Loud Too Loud
Too Soft Too Soft
A B

Figure 4-1
Figure 4.1
The relationship between the range of sounds in the environment and the dynamic range of hearing for persons with:
(A) conductive hearing loss, and (B) conductive hearing loss with linear amplification.
The relationship between the range of sounds in the environment and the
dynamic
Audiometric range of hearing for persons with:
Configuration (A) conductive
Loudness Growth hearing loss,
Due and (B) conductive
to variations hearing
in etiology, loss
as well as thewith linearThe
amplification
type and degree of hearing loss, audiometric
anatomy and physiology of the cochlea, the degree configuration and residual dynamic range all
of hearing loss often changes across frequency. contribute to variations in the growth of loudness
The typical hearing impairment manifests itself as with increasing input levels. In general, more
a greater loss of sensitivity at high frequencies than amplification is required for weak sounds than
at low frequencies. Atypical configurations, such as for intense sounds; the amount of amplification
precipitous, reverse slope, and “cookie bite,” are required for moderate sounds lies in between
also encountered. these extremes. The goal is for weak sounds to
be perceived as soft, moderate sounds to be
The amount of amplification necessary to perceived as comfortable, and intense sounds
compensate for a hearing impairment depends to be perceived as loud by the hearing aid user.
on the degree of hearing loss. Multichannel Varying amounts of gain depending on the input
compression provides the professional with level can only be achieved through the use of
a way to finely tailor the frequency and gain compression.
characteristics to an individual’s audiometric
configuration across a broad range of frequencies. Although the loudness growth function can be
Trine and Van Tasell (2002) have shown that 3-4 predicted to some extent, the use of individual
channels, with adjustable cross-over frequencies, loudness data may enhance customization of the
are generally sufficient for this purpose. hearing aid fitting. [As described in the previous
chapter, a loudness growth function is a graph
Residual Dynamic Range depicting the perceived loudness of sound as
The upper end of the range of hearing is limited by a function of input level.] These data can be
the LDL, which may or may not be affected by the obtained via testing under headphones, in a
hearing loss (Kamm et al, 1978). Thus, the residual sound field, or directly through the hearing aid (a
dynamic range of a person with impaired hearing feature available in many high-end digital hearing
is considerably smaller than that for a person with aids). While sound in principle, the benefits of the
normal hearing. How the range of environmental time spent obtaining loudness growth data for
sounds is squeezed into this residual dynamic individuals are debatable.
range depends on the goal of the hearing aid
fitting. Regardless of the goal, compression must
be applied to the incoming sound.
29
Unilateral Versus Bilateral Amplification audibility with WDRC! The problem is related
There are several documented advantages to the perception of moderate-to-intense
of fitting hearing aids bilaterally – increased sounds. Recall that, with linear amplification,
speech understanding in background noise intense sounds are amplified more than
and/or reverberation (Nabalek and Pickett, necessary, even when average conversational
1974), better sound quality (Balfour and speech is comfortable. Although undesirable,
Hawkins, 1992), improved localization ability long time users of linear hearing aids have
(Byrne et al, 1992), avoidance of auditory become accustomed to overamplification of
deprivation (Silman et al, 1984), and binaural moderate‑to-intense sounds. Viable options
summation of loudness. Only the last of these include mid‑level compression or easing into
benefits directly impacts the hearing aid fitting. WDRC over a period of weeks or months.
Binaural summation refers to the phenomenon

whereby a sound is louder if heard in two ears Prescribing Amplification
than in one. This loudness increase is to the Given the multitude of hearing losses and the
order of 3dB at threshold (Dermody and Byrne, complexity of hearing aids, the only practical
1975), and approximately 6dB for moderate way to begin the hearing aid fitting process
(Christen, 1980) and intense sounds (Scharf is to prescribe amplification based on some
and Fishken, 1970). Interestingly, binaural audiological characteristic of the person with
loudness summation has little or no effect impaired hearing. Prescriptive approaches use
on the LDL (Hawkins et al, 1987). This implies a formula to generate an amplification target,
that, in a bilateral fitting, less gain (and more which serves as a reasonable starting point.
compression) is required in each hearing aid
(compared to a unilateral fitting) to restore The earliest prescriptions were developed to
loudness perception. Further, this can be fit the available linear technology. Thus, linear
achieved without causing loudness discomfort. prescriptive formulas – like NAL-RP, POGO,
Berger and Libby 1/3 gain – provide a single
Previous Hearing Aid Use target for moderate sounds, because the
WDRC can be used to restore loudness most common and important sound to which
perception for new users of amplification. we listen is average conversational speech.
However, the onset of hearing loss is gradual and However, these do not take into consideration
its impact insidious. It has been reported that the the level-dependent gain characteristics of
typical individual lives with hearing loss for about modern hearing aids. This has led to the
10 years before seeking help (Kochkin, 1991). development of non-linear prescriptions aimed
Such a person has generally forgotten what it was at restoring loudness perception and improving
like to hear “normally,” resulting in a distorted the intelligibility of speech.
recollection of the perception of loudness. Thus,
less than the prescribed amount of gain and/or Depending on the underlying philosophy,
more compression may be used initially, with the non-linear prescriptive approaches can be
possibility of moving closer to the prescription as broadly divided into two categories: loudness
the individual acclimates to the amplified sound. normalization and loudness equalization. As
the term suggests, loudness normalization
Experienced users of linear amplification attempts to restore normal loudness perception
present a different challenge. WDRC is often within each frequency band. Presumably, this
rejected with the complaints like “these hearing also results in normal loudness perception of
aids are too soft,” and “I hear better with my the overall sound – i.e., once all the frequency
own hearing aids.” This is in spite of improved bands are combined. Loudness equalization,
30
on the other hand, aims to maximize speech Second, corrections are applied to the target to
intelligibility by making all frequency bands accommodate conductive and mixed hearing
of speech contribute equally to the loudness losses. And, third, it takes into account the
of speech. Although not normal within each limited ability of the impaired ear to extract useful
frequency band, normal loudness perception information from an audible speech signal (Ching
is restored for the overall sound. The four most et al, 1998). This is especially true for severe-to-
commonly-used, widely-accepted and well- profound hearing losses.
researched non-linear prescriptive approaches
will be discussed – NAL-NL1,DSL [i/o],FIG6 and The prescription is dependent upon the threshold
IHAFF/VIOLA. Variations in the core principles at a particular frequency, the 3-frequency
result in different prescriptions from each of average threshold, the slope of the audiogram
these approaches. Figure 4-2 compares the between 500 and 2000 Hz, and the long-term
I/O functions prescribed by each of these average speech spectrum (LTASS) (Byrne et
approaches for a sensorineural hearing loss of al, 1994). For a speech input of 65 dBSPL, the
60dBHL. The direction and magnitude of these target generated by NAL-NL1 is the same as that
differences are further related to the degree of prescribed by NAL-RP. Although traditionally
hearing loss and audiometric configuration. shown as real‑ear insertion gain (REIG) (Figure 4-3,
page 32), targets can be displayed in various
120 forms, including insertion gain, aided gain or I/O
function measured in a real ear, 2‑cc coupler or ear
110
simulator. A customized prescription is obtained
Real-Ear Output (dBSPL)
100 by measuring the individual real-ear unaided gain

90
(REUG) and real ear-to-coupler difference (RECD).
The number of channels in the hearing aid can
80 be specified. Also, appropriate corrections are
70 applied for bilateral fittings. Finally, the fitting
software may provide recommendations regarding
60
the dimensions of the vent and the type of tubing
used in the earmold.
20 30 40 50 60 70 80 90 100
Input (dBSPL)
Figure 4-4 (page 32) shows the real-ear aided
Linear (NAL-RP) NAL-NL1 DSL [i/o] response (REAR) prescribed by NAL-NL1 for a flat
Fig6 IHAFF/VIOLA hearing loss of 60dBHL. Amplification targets for
Figure 4.2
Figure 4-2 input levels of 50, 65 and 80 dBSPL are displayed.
Prescribed input/output function at 2000 Hz for a sensorineural
Prescribed input/output function at 2000 Hz for a
hearing loss of 60dBHL. NAL‑RP = National Acoustic
sensorineural
Laboratories, hearing loss of 60dB.
Revised, Profound.
NAL-RP = National Acoustic Laboratories, DSL [i/o]
Revised, Profound. Desired Sensation Level [input/output]
NAL-NL1 (Cornelisse et al, 1995) is based on the premise
NAL-NL1 (National Acoustic Laboratories, non- that the desired sensation level for a variety of
linear, version 1; Dillon, 1999) specifies insertion sounds can be predicted from threshold and
gain at 1/3‑octave frequencies between 125 LDL. Based on loudness normalization, the
and 8000 Hz. The NAL-NL1 formula has three goal of this approach is to ensure audibility
unique features. First, the goal is to maximize for weak sounds, provide comfort for average
speech intelligibility while maintaining the overall conversational speech, and prevent discomfort
loudness equivalent to that for a person with for intense sounds. There are actually two DSL
normal hearing. Thus, loudness equalization is [i/o] procedures. In DSL [i/o] linear, the I/O
the underlying philosophy for this approach. function is a straight line over a wide range of
31
inputs, i.e., the CR is fixed. Note that, despite 120
the use of the term “linear,” compression is 110
UL hi
still being applied over a range of input levels. 100 UL n

DSL [i/o] curvilinear, on the other hand, has a 90
DR hi
curved I/O function where the CR is constantly 80
changing. Because DSL [i/o] linear is more
Output (dBSPL)
70
commonly used and better developed than 60 TH hi
DSL [i/o] curvilinear, only the former will be 50
discussed here. Although developed and 40
DR n
widely used for the fitting of children, this 30
approach can be successfully applied to adult 20
hearing aid fittings. 10 TH n
60 Hearing Impaired UCL (UL hi )

Hearing Impaired Threshold (TH hi )
50
Normal UCL (UL n )
40 Normal Threshold (TH n )
REIG (dB)
30 Hearing Impaired Dynamic Range (DR hi )
20 Normal Dynamic Range (DR n )
10 Figure 4-5
Comparison of the dynamic ranges of persons with normal and
impaired hearing.
0
250 500 1K 2K 4K 8K Figure 4.5
Frequency (Hz) Comparison of the dynamic ranges of persons with normal and
impaired hearing.
DSL [i/o] attempts to map the normal dynamic
Input Levels: 80 dBSPL 65 dBSPL 50 dBSPL
range into the residual dynamic range of the
Figure 4-3
Figure 4.3
Real-ear Insertion Gain (REIG) prescribed by NAL-NL1 for a flat,
individual with impaired hearing. As shown in
Real-ear Insertion hearing
Gain (REIG)
loss.prescribed by
60dBHL sensorineural
NAL-NL1 for a flat, 60dB sensorineural hearing loss.
Figure 4-5, the dynamic range is bounded by
the threshold of audibility – THn and THhi for
120
normal and hearing-impaired, respectively –
and the upper limit of comfort – ULn and ULhi.
110
An I/O function can be derived by connecting
100 these points (Figure 4-6). Note that the I/O
REAR (dBSPL)
function has three distinct segments: a region

90
of linear gain below the TKlow, a region of
80 compression between TKlow and TKhigh, and
70 a region of output limiting above TKhigh. At a
minimum, threshold data are required to obtain
60
target information. Patient-specific information
50 such as RECD, real ear-to-dial difference
250 500 1K 2K 4K 8K
Frequency (Hz) (REDD), real‑ear unaided response (REUR),
and LDL can be entered to customize the
Input Levels: 80 dBSPL 65 dBSPL 50 dBSPL prescription. The style of hearing aid and type
Figure 4-4
Figure 4.4
Real-ear aided response (REAR) prescribed by NAL-NL1 for a flat
of compression can also be entered to provide
Real-ear aided response
hearing (REAR)
loss. prescribed by
NAL-NL1 for a flat 60dB sensorineural hearing loss
a more accurate target.
32
120
Fig6
110
UL
hi Fig6, named for its original appearance in an
100 TK high article by Killion and Fikret-Pasa (1993), is a
90 loudness normalization procedure based on
80 average data. Killion and Fikret-Pasa identified
OUTPUT LIMITING
70 three types of sensorineural hearing loss based
Output (dBSPL)
TH
hi COMPRESSION
60 on loudness perception. Type I hearing losses
50 demonstrate a threshold sensitivity loss of about
TK low
40
40dB with normal perception of intense sounds
LINEAR
30
(Figure 4-8A, page 34). For a Type II loss, the
20
TH n UL n loss of threshold sensitivity of about 60dB is
0 10 20 30 40 50 60 70 80 90 100 110 120 accompanied by a small decrease in sensitivity
Input (dBSPL) to intense sounds (Figure 4-8B, page 34).
Figure 4.6
Figure 4-6
Input/outputfunction
Input/output function prescribed
prescribedby DSL
by [i/o]
DSLfor[i/o]
a sensorineural hearing loss
for a sensorineural Finally, the loss of sensitivity for both soft and
of 60 dB. THn=normal threshold of hearing sensitivity, ULn=normal upper limit
hearing loss of 60dBHL. THn=normal threshold of hearing
of comfort, THhi=hearing-impaired threshold of sensitivity, ULhi=hearing-
sensitivity,
intense sounds is substantial for a Type III loss
impaired ULn=normal upper limit of comfort, THhi=hearing-
upper limit of comfort.
impaired threshold of sensitivity, ULhi=hearing-impaired upper (Figure 4-8C, page 34). The distance between the
limit of comfort.
loudness growth curves for persons with normal
and impaired hearing provides an indication of
Figure 4-7 shows the REAR prescribed by the gain required.
DSL [i/o] for a flat hearing loss of 60dBHL.
Amplification targets for input levels of 50, 65 IHAFF/VIOLA
and 80 dBSPL are displayed. In the mid 1990s, a group of engineers, clinicians
and researchers formed the Independent
120 Hearing Aid Fitting Forum (IHAFF) to address
110
the issue of non-linear amplification. Believing
that there was little direction for the dispensers
100 in terms of fitting non-linear amplification, they
REAR (dBSPL)
90 developed a fitting protocol named for the

group. The IHAFF protocol is comprised of three
80
components: a loudness scaling procedure
70 (Contour Test; Cox et al, 1997), a fitting algorithm
60 to specify hearing aid characteristics (Visual
Input/Output Locator Algorithm, VIOLA), and
50
250 500 1K 2K 4K 8K a measure of the fitting outcome (Abbreviated
Frequency (Hz) Profile of Hearing Aid Benefit, APHAB; Cox and
Input Levels: 80 dBSPL 65 dBSPL 50 dBSPL Alexander, 1995). The goal of the VIOLA fitting
Figure 4-7 algorithm is to normalize loudness perception
Figure
Real-ear aided4.7
response (REAR) prescribed by DSL [i/o] for a flat,
and will be the focus of discussion for this
Real-ear hearing loss.
aided response (REAR) prescribed by DSL [i/o]
for a flat, 60dB sensorineural hearing loss. handbook.
33
VIOLA generates a target 2-cc coupler output
for weak, moderate and intense sounds based
on the predicted loudness growth function for
warble tones (Figure 4-11). The line connecting
these target values represents the desired I/O
function. According to Valente and Van Vliet
(1997), a TK of 40-45 dBSPL should be used.
It is recommended that target I/O functions
be obtained for at least one low and one high
frequency (e.g., 500 and 3000 Hz, respectively).
Type I Type II Type III

UL UL UL
L L L
C C C
S S S
20 40 60 80 100 20 40 60 80 100 20 40 60 80 100
Input (dBSPL)
Normal hearing Sensorineural hearing loss Gain required to restore normal loudness perception
Figure 4-8
Figure 4.8
Loudness growth curves for three types of sensorineural hearing loss (Killion and Fikret-Pasa, 1993).
S = Soft, C = Comfortable, L = Loud, UL = Uncomfortably Loud
Loudness growth curves for three types of sensorineural hearing loss
(Killion and Fikret-Pasa, 1993). S = Soft, C = Comfortable, L = Loud,
UL = Uncomfortably Loud.
60 120
50 110
40
REIG (dB)
100
REAR (dBSPL)
30
90
20
10 80
70
0 10 20 30 40 50 60 70 80 90 100
Audiometric Threshold (dBHL) 60
Intense Sounds Moderate Sounds 50

250 500 1K 2K 4K 8K
Weak Sounds Elevation in LDL
Frequency (Hz)
Figure 4-9
Figure 4.9
Real-ear insertion
Real-ear gain (REIG)
insertion prescribed
gain (REIG) by FIG6
prescribed for various
by FIG6 for various Input Levels: 80 dBSPL 65 dBSPL 50 dBSPL
degrees of hearing
degrees loss. loss.
of hearing Predicted elevation
Predicted in loudness
elevation in loudness
discomfort levellevel
(LDL)(LDL)
is also shown. Figure 4-10
Figure 4.10
discomfort is also shown.
Real-earReal-ear
aided response (REAR) prescribed
aided response by FIG6by
(REAR) prescribed for a flat
60dB sensorineural
FIG6 for a flathearing loss.
60dB sensorineural hearing loss.
34
120
110
110
100
2cc Coupler Output (dBSPL)
90 100
REAR (dBSPL)
80 90
70 80
60
70
50
40 50 60 70 80 90 60
Input (dBSPL at Hearing Aid Microphone)
50
Target Output 250 500 1K 2K 4K 8K
Intense speech Intense warble tones Loud Frequency (Hz)
Moderate speech Moderate warble tones Average
Weak speech Weak warble tones Soft Input Levels: 80 dBSPL 65 dBSPL 50 dBSPL
Figure 4.11
Figure 4-11 Figure 4.12
Figure 4-12
Targetinput/output
Target input/output function
function prescribed
prescribedby
byIHAFF/Viola at 3000 Hz. Real-ear
Real-ear aidedaided response
response (REAR) (REAR) prescribed
prescribed by by
IHAFF/VIOLA at 3000 Hz. IHAFF/VIOLA for for
IHAFF/Viola a flat 60dBHL
a flat 60dBsensorineural hearing
sensorineural loss.
hearing loss
The shaded regions in Figure 4-11 represent 120

target levels for weak, moderate and intense
110
speech. These are higher than the target levels
for warble tones because speech is a broadband 100
REAR (dB SPL)
signal (i.e., power summation). The targets 90

can be further customized by measuring the
80
individual loudness growth function via the
Contour Test. 70
60
Figure 4-12 shows the REAR prescribed by 250 500 1K 2K 4K 8K
Frequency (Hz)
IHAFF/VIOLA for a flat hearing loss of 60 dBHL.
NAL - NL1 DSL [i/o]
Amplification targets for input levels of 50, 65
Fig 6 IHAFF / VIOLA
and 80 dBSPL are displayed.
Figure 4-13Figure 4.13
Comparison of the Real-ear aided response (REARs)
Comparison of the Real-Ear aided response
As is obvious from Figures 4-2, 4-4, 4-7, 4-10 prescribed for a sensorineural hearing loss of 40dBHL and
(REARS) prescribed for a sensorineural hearing
an input level of 65 dBSPL.
and 4-12, the various prescriptive approaches loss of 40dB HL and an input level of 65 dB SPL.
differ in the amount of gain they recommend

for weak, moderate and intense sounds, at low
and high frequencies. Figure 4-13 compares
the real-ear output prescribed by the various
approaches for an input of 65 dBSPL. To some
extent, discrepancies are to be expected due
to the differences in the underlying philosophy.
This discrepancy poses a dilemma from the
point of view of choosing the “right” approach.
However, it may be a moot point when one
considers the prescriptive method a good
starting point rather than a final destination.
35
Chapter 5: Expansion
Issues Resulting from WDRC environmental sounds. Some individuals are

especially sensitive to these sounds and, as such,
As we have seen throughout this handbook,
are eager for an immediate solution. Moreover,
WDRC provides additional gain for weak inputs, as
counseling will not alleviate the problem when it is
compared to linear amplification. The advantage
caused by audible circuit noise. Gain reduction has
of this approach is increased speech intelligibility,
traditionally been the second line of attack in this
especially for soft speech, and normalized
situation. Specifically, the gain for weak sounds is
loudness perception. The bad news is that it also
decreased in an attempt to reduce the audibility of
amplifies other soft sounds, including the drone
environmental and/or circuit noise. The drawback
of a refrigerator, the creaking of floor boards and
of this solution is that it defeats the purpose of
the hum of an air conditioner. How can this be a
using WDRC in the first place – weak sounds are
disadvantage when a person with normal hearing is
now no more audible than they would be with
able to hear these everyday sounds?
linear amplification or, worse, with no amplification
at all. Further, the audibility of important speech
Presbycusis, the most common cause of hearing
sounds is adversely affected. Thus, alleviating one
loss in adults, typically has a gradual onset. As
problem has given rise to a second, and arguably
a result, by the time the hearing loss becomes
more serious, issue.
noticeable to the individual, many of the weaker
sounds in the environment may not have been
A more viable solution to the predicament, and
heard for some time. WDRC enables amplification
one that is commonly available in modern digital
of weak sounds to within the audible range. This
hearing aids, is the use of expansion. Expansion
can be overwhelming to an auditory system that
is designed to make a hearing aid sound silent
has been deprived for a long period of time. On a
in quiet environments. One can think of it as a
different note, individuals who have good residual
noise reduction strategy for quiet environments.
hearing (i.e., thresholds that are within or close
The use of expansion can lead to greater listener
to the normal range) at some frequencies may be
satisfaction by reducing the intensity of weak
able to hear noise that is inherent to the hearing
environmental sounds and circuit noise, without
aid, such as circuit noise that is amplified by the
sacrificing speech audibility and intelligibility.
hearing aid. Either of these scenarios may result in
an unpleasant experience and, in turn, hearing aids
that spend most of their time in a drawer. To avoid What is Expansion?
this rejection, the problem must be addressed.
Consider the following example. Figure 5-1A
The first approach to tackling the “problem” of compares the residual dynamic range of a person
increased audibility for weak sounds is counseling. with sensorineural hearing loss to the range of
Potential hearing aid users often do not realize sounds in the environment. As expected, weak
that, in addition to hearing the soft voice of a sounds are inaudible, while intense sounds
grandchild, they will also hear other sounds in are perceived as loud. Figure 5-1B illustrates
the environment that they may not have heard in the effect of compression amplification – weak
a long time. It is important that they understand sounds are amplified to a greater extent than
that these sounds are also heard by persons moderate and intense sounds. This allows the
with normal hearing and that, over time, they range of environmental sounds to fit within the
will become accustomed to the abundance of residual dynamic range of the individual.
36
While this improves the intelligibility of speech, identical. In other words, compression is applied to
it also makes unwanted sounds audible, such moderate and intense sounds, while expansion is
as the quiet drone of a refrigerator. Figure 5-1C used only for weak sounds.
shows the effect of expansion in combination with
compression. The band representing weak sounds Figure 5-2 shows an I/O function for a linear
is wider than those for moderate or intense sounds. hearing aid with OCL. A fixed amount of gain is
Further, the weakest of the weak sounds are below applied to all input levels, as is characteristic of
the dynamic range, and only the more intense linear amplification. Specifically, for every 10dB
components of the weak sounds are audible to the change in input level, the output also changes
individual. This is the principle behind expansion by 10dB. When expansion is applied at low input
– the softer sounds are amplified less than louder levels (i.e., at or below the TK of 50 dBSPL), the
sounds. In hearing aids, expansion is only applied output changes by 20dB for a 10dB change in input
to the weakest sounds in the environment. A level (orange curve in Figure 5-2). Note that the I/O
comparison of Figures 5-1B and 5-1C shows that function above the TK is unaffected by whether or
the bands representing moderate and intense are not expansion is applied.
Too Loud Too Loud Too Loud
Too Soft Too Soft Too Soft
A B C
Figure 5.1
Figure 5-1
The relationship between the range of sounds in the environment and the dynamic range of hearing for persons with sensorineural hearing loss:
The relationship between the range of sounds in the environment and the dynamic range of hearing for persons with
(A) without amplification, (B) with compression amplification, and (C) with compression amplification and expansion.
sensorineural hearing loss: (A) without amplification, (B) with compression amplification, and (C) with compression
amplification and expansion.
Like compression, expansion delivers different
120
amounts of gain depending on the input level.
110 However, in expansion, the gain increases as
100 the input level increases. This is in contrast to
Output (dBSPL)
90
TK
compression where the gain decreases as the
80
input level increases. The change in gain as a
function of input level is most easily illustrated
70
by an I/G function. Figure 5-3 (page 38) shows
60
the I/G function of a linear hearing aid, and the
50
20 30 40 50 60 70 80 90 100
effect of expansion. In a linear hearing aid with
Input (dBSPL) no expansion, the gain remains constant below
Linear with OCL Expansion
kneepoint. In a hearing aid with expansion,
however, little or no gain is applied to input levels
Figure 5.2
Figure 5-2
Sample input/output functions showing linear
of 20 dBSPL or lower. As the input level increases,
Sample input/output functions showing linear amplification and
amplification and output compression limiting
output compression limiting (OCL), with and without expansion so does the gain, until it reaches a maximum of
(OCL), with and without expansion below the
below the threshold kneepoint (TK).
threshold kneepoint (TK). 30dB at the TK (i.e., 50 dBSPL). Note once again
that the I/G function above the TK is unaffected by
whether or not expansion is applied.
37
60 120
50 110
Gain (dBSPL)
TK 100
40
TK
Output (dBSPL)
30 90
20 80
1
10 70
.5:
=0
XR
60
20 30 40 50 60 70 80 90 100
50
Input (dBSPL) 20 30 40 50 60 70 80 90 100
Linear with OCL Expansion Input (dBSPL)
Figure 5.3 Figure 5.4

Figure 5-3 Figure 5-4Identifying the threshold kneepoint for
Sample input/gain
Sample input/gain functionslinear
functions showing showing linear and
amplification Identifying the threshold
expansion (TK) kneepoint (TK) forthe
and calculating expansion and
expansion
amplification
output compression and output
limiting (OCL), compression limiting
with and without expansion calculating the (XR)
expansion ratio (XR) from anfunction.
input/output function.
ratio from an input/output
(OCL),
below the with kneepoint
threshold and without expansion below the
(TK).
threshold kneepoint (TK).
Characterizing Expansion A low TKexp (input levels of 50 dBSPL or lower)

Expansion is characterized in much the same permits the use of more gain to ensure maximum
way as compression, via a threshold kneepoint, speech audibility. In this scenario, the hearing aid
expansion ratio, and attack and release times. may be in expansion for only a small portion of
There are, however, a few key differences which time and only in very, very quiet environments. As
will be highlighted here. [Note that, at the a consequence, the problem of a “noisy hearing
time of this writing, there are no standards for aid” may not be alleviated at all. On the other
defining the characteristics of expansion.] hand, a high TKexp (input levels greater that 40
dBSPL) will result in a quiet hearing aid, but at the
Expansion Threshold expense of speech audibility.
Expansion threshold (XT) is the input level
below which expansion operates. For the Expansion Ratio
hearing aid shown in Figure 5-4, the XT is 50 Like its counterpart in compression, the
dBSPL. Because it looks like a bent knee in an expansion ratio (XR) is an indicator of the
I/O function, the XT is also referred to as a TK. extent to which the input signal is expanded.
Specifically, XR is calculated using the formula:
The potential for confusion arises when ∆Input
expansion is coupled with compression, both
XR =
∆Output
of which are characterized by a TK. The TKs for
both expansion and compression frequently For the hearing aid shown in Figure 5-4,
occur at the same input level. Thus, as shown increasing the input from 40 to 50 dBSPL
in Figure 5-5A, expansion operates below the (∆Input = 10dB) increased the output from
TK, while compression takes over above it. A 60 to 80 dBSPL (∆Output = 20dB). Using the
less commonly implemented design is one above formula, the XR of the hearing aid is:
where the TK for compression (TKcomp) is higher
than the TK for expansion (TKexp) (Figure 5-5B). ∆Input = 10 0.5
XR = =
Note that amplification in the portion of the I/O ∆Output 20
function in between TKexp and TKcomp is linear
(i.e., CR=1:1).
38
120 120
A B
110 110 1
.2:
.4 :1 =2
TK
R=2 CR
100 C 100
Output (dBSPL)
Output (dBSPL)
90 90
:1
=1
TK comp
CR
80 80
:1
0.4
70 70 TK exp
XR=
:1
0.4
XR=
60 60
50 50
20 30 40 50 60 70 80 90 100 20 30 40 50 60 70 80 90 100
Input (dBSPL) Input (dBSPL)
Figure 5-5
Figure
Sample input/output 5.5 for hearing aids with: (A) the same threshold kneepoint (TK) for compression and expansion,
functions
and (B) a threshold kneepoint
Sample for compression
input/output (TKcompfor
functions ) higher than the
hearing threshold
aids with: kneepoint for expansion
(A) the same (TKexpkneepoint
threshold ). for compression
and expansion, and (B) a threshold kneepoint for compression (TK comp) higher than the threshold
kneepoint for expansion (TK exp ).
CRs are always greater than 1 when compression

is applied. Further, the larger the number, the 120
greater is the degree of compression applied. 110

On the other hand, XRs between 0 and 1 100
indicate the application of expansion. Figure 5-6
TK
Output (dBSPL)
90
shows I/O functions for three hearing aids with
the same TK but different XRs. Hearing Aid A, 80
with an XR of 1, is linear below the TK. The 70

I/O function for Hearing Aid C (XR=0.4) has
60 :1
:1
1
a steeper slope than that for Hearing Aid B =1

.7:
0.4
XR
=0
XR=
XR
(XR=0.7). A steeper expansion slope indicates a 50

20 30 40 50 60 70 80 90 100
greater degree expansion. Thus, in general, the Input (dBSPL)
smaller the XR (i.e., closer to 0), the steeper is Figure 5-6
Sample input/output functions showing three hearingAid
Hearing aids
B with
the slope of the I/O function and the greater is Hearing Aid A
the same threshold kneepoint (TKexp) and different expansion
ratios (XR). Hearing Aid C
the degree of expansion.
Figure 5.6
The higher the XR (i.e., closer to 1), the more Sample input/output functions showing three
hearing aids with the same threshold kneepoint
likely it is that weak sounds will be amplified (TK exp ) and different expansion ratios (XR).
to within the residual dynamic range. In other

words, the likelihood that a person will be able
to hear the weakest of sounds is greatest at
high XRs. On the other hand, low XRs are known
to adversely affect the intelligibility of speech
at low input levels (Plyler et al, 2005a). Either
problem may be exacerbated by the exact
location of the TK. Specifically, applying a lot of
expansion (low XR) at high TKs may drastically
reduce speech audibility, while applying only a
little expansion (high XR) at low TKs may result
in the hearing aid sounding noisy.
39
Attack and Release Times Informal Listening
Because expansion only occurs below the TK, With expansion turned off, listen for quiet
attack and release are defined differently for sounds (e.g., whirr of a computer, hum of the air
expansion than for compression. Attack into conditioner, etc.). Turn on expansion and listen
expansion occurs when the signal level drops for those same sounds. The hearing aid should
below the TK, while release from expansion sound much quieter with expansion turned on.
occurs when the signal level exceeds the TK. The difference between the two conditions will
be more prominent at high TKs and low XRs.
A fast AT limits the amplification of ambient
sounds in the environment, and implies that 60
the hearing aid may go into expansion during
50
pauses in speech. Slow RTs are known to
adversely affect speech intelligibility (Plyler et 40
Gain (dB)
al, 2005b). This occurs because gain is
30
considerably reduced when the hearing aid is
in expansion, and is slow to recover when the 20
input level increases; the small amount of gain 10

may render speech inaudible. The effects of
0
both of these problems are magnified if the 250 500 1K 2K 4K 8K
TK is high and the XR is low. Thus, in theory, it Frequency (Hz)
appears that a combination of a slow AT and Expansion Off

a fast RT offers the best balance, making weak Expansion On (4 sec wait)
Expansion On (2 sec wait)
environmental sounds inaudible with little effect Expansion On (no wait)
on the audibility of speech. Figure 5-7
Gain measured in a 2-cc coupler with expansion off and on,
using a 70 dBSPL broadband signal. The “wait” time indicates
the duration for which the signal was on before a measurement
Measuring Expansion was made.
Figure 5.7
Gain measured in a 2-cc coupler with expansion off
Expansion operates on low-level inputs. and on, using a 70 dBSPL broadband signal. The
Ambient noise levels in most environments are Coupler“wait
Measurement
time” indicates the duration for which the signal
was on before a measurement was made.
often too high (i.e., above the TK) for expansion The effect of expansion can be measured
to engage. As a result, it can be difficult to hear using a standard 2‑cc coupler, a test box and
and/or measure the effects of expansion. a broadband input signal such as white, pink,
speech-shaped or composite noise. Figure 5-7
While a sound treated room or anechoic shows the gain of a hearing aid for a composite
chamber are not necessary, an important noise with expansion turned off and an overall
prerequisite for listening to and/or measuring input level of about 70 dBSPL (or above the
expansion is that the environment be quiet. TKexp of the hearing aid). To complete the
Objections to the hum of the refrigerator or measurement, turn expansion on and turn the
circuit noise occur primarily when the hearing stimulus in the test box off for longer than the
aid user is in a quiet environment, such as AT to allow expansion to engage. Once this has
reading a book in the living room. Thus, a quiet occurred, turn the stimulus on again and make
office, away from the sounds of copy machines another measurement immediately (“Expansion
and people talking, will suffice for evaluating On (no wait)” in Figure 5-7). You should find
expansion. that the gain of the hearing aid is lower with
expansion on than with expansion off. The
longer the interval of time between turning on
the stimulus and making a measurement, the
40
greater the gain of the hearing aid, as shown by
“Expansion on (2 sec wait)” and “Expansion on
(4 sec wait)” in Figure 5-7. The duration of time
required for the gain to return to the expansion
off condition is an indication of the expansion
RT of the hearing aid.
While expansion is a useful tool for the

audiologist, it should be used judiciously –
experienced hearing aid users and those with
severe hearing loss may welcome the added
audibility provided by WDRC.
41
Chapter 6: Troubleshooting & Fine‑Tuning
The prescriptive approaches described previously problems that the clinician may encounter and one
are based on average and/or predicted data. As approach to dealing with them. A more detailed
such, they serve as a good starting point. Persons list can be found in the section on “Common
with impaired hearing come with a variety of Complaints and Solutions.”]
experiences with hearing. It is a rare individual who
does not require some further adjustment after the Most users of hearing aids are unable to describe
initial fitting of a hearing aid. This chapter describes their complaints in hearing aid terms – gain,
techniques for troubleshooting the individual’s frequency, compression, etc. In a survey of
complaints and fine-tuning the hearing aids. The approximately 300 audiologists, Jenstad et al (2003)
ultimate goal should always be to maximize speech reported over 40 frequently-used descriptors – “too
intelligibility, with minimal loudness discomfort. loud,” “booming,” “can’t hear well in noise,” “in a
barrel,” and so on. Thus, a single problem may be
Before attempting to fine-tune a hearing aid, described in multiple ways! The only way to get to
consider this: we tend to gravitate toward the root of the problem is by asking appropriate
conditions with which we are most familiar, and questions. For example, if the hearing aids
hearing aid users are no exception. For example, an reportedly “sound funny,” find out if this occurs with
experienced user of linear amplification generally all sounds or only with the individual’s own voice.
prefers it over WDRC, even though the latter may The key to translating patient complaints is a good
provide better speech understanding and listening working knowledge of the acoustic characteristics
comfort. Another example of preference based of sounds. Figure 6-1 shows examples of sounds
on experience is that of a new hearing aid user that are commonly associated with complaints.
preferring amplification that essentially matches Thus, if the patient’s own voice “sounds funny,”
the response of the unaided ear. Thus, at the initial the problem may be simplified to a need for
fitting, the clinician should keep in mind how these adjusting the response of the hearing aid to intense,
experiences may affect the outcome of the fitting. low‑frequency sounds.
A Systematic Approach FREQUENCY
There are two primary challenges to fine‑tuning LOW HIGH
a hearing aid fitting – recognizing the source and

possible solutions to the problems being described, Humming Computer Rustling of Leaves
and understanding the compression architecture of
LOW
Purring Kitten Ticking Clock

the specific hearing aid. Vowel Sounds Consonant Sounds
INTENSITY
Recognizing the Problem and Solution

A systematic three-step approach is recommended
Slamming Doors Running Water
for recognizing the problem and identifying
HIGH
Traffic Noise Crackling Paper

potential solutions: (1) Simplify the individual’s Own Voice Clanking Dishes
complaints, (2) determine the direction of change
needed, and (3) identify the control(s) to effect the
desired change. [Note that the examples cited in Figure 6.1
Figure 6-1
Acoustic characteristics
Acoustic of some
characteristics ofeveryday sounds. sounds.
some everyday
this section are only meant to represent the types of
42
120 120
A B
110 110
100 G 100 G
Output (dBSPL)
Output (dBSPL)
90 90
80 80
70 G’ 70 G’
60 60
50 50
20 30 40 50 60 70 80 90 100 20 30 40 50 60 70 80 90 100
120
Figure 6-2 C
ctions for Sample input/output functions for three approaches to decreasing 110
usting the gain the gain for weak sounds by 10 dB. The original and adjusted gains
are indicated by G and G’ respectively. The blue shaded area shows 100
creasing the TK the effect of the adjustment in the desired region; the gray shaded Output (dBSPL) G
increasing the area demonstrates secondary effects of the change.

90
and
The original 80
ndicated by G
e blue shaded 70 G’
the adjustment
e gray shaded 60
ndary
50
20 30 40 50 60 70 80 90 100
Input (dBSPL)
Determining the direction of change needed may be to decrease the gain for intense,
is the next step in addressing the problem. low‑frequency sounds.
The good news is that, once the problem has
been clarified, audiologists generally agree With the root of the problem and a possible
on the course of action. Frequently, this is the solution recognized, identifying the control(s)
easiest part of the process because the direction to effect the necessary change should be
of change is immediately apparent from the obvious. For example, to alleviate the problem
complaint. For example, the expressions “too of the patient’s own voice sounding funny due
loud,” “uncomfortable,” “boomy,” “harsh,” to a plugged sensation, the gain for intense,
etc. indicate too much amplification, at least low-frequency sounds may be decreased by
in some frequency regions and/or at some adjusting a control in the low-frequency channel
intensities. However, some complaints, such as that manipulates gain for loud sounds. However,
that of a plugged sensation, are more difficult the appropriate control is sometimes either not
to interpret. For example, it may be that, when available or cannot be adjusted directly. In such
the ears are plugged (with hearing aids), sounds situations alternative solutions must be sought
in the environment are harder to hear but one’s (for example, increasing the CR to decrease
own voice sounds louder than usual. Thus, the gain for intense, low-frequency sounds),
if sound of the patient’s own voice “sounds which requires a firm grasp of the mechanics of
funny” due to a plugged sensation, one solution compression.
43
Understanding the Compression Architecture intense sounds (90 dBSPL). In Figure 6-2B, the
One potential complication to fine-tuning gain for weak sounds is reduced by increasing
hearing aids based on patient complaints is that the TK from 50 to 60 dBSPL. However, this
alleviating one problem may inadvertently give action also decreases the gain for moderate and
rise to another. intense sounds. Finally, in Figure 6‑2C, the gain
for weak sounds is reduced simply by decreasing
For example, consider the complaint of soft the amount of gain without affecting the TK.
speech being too loud, which requires a reduction This impacts not only the gain for weak sounds,
in the gain for weak sounds. Figure 6-2 (page 43 but also that for moderate and intense sounds.
) shows three different ways in which gain for Thus, it appears that, regardless of the exact
inputs of 50 dBSPL may be decreased by 10dB. mechanism of gain change, secondary effects of
In Figure 6-2A, the TK is increased from 50 to 70 the adjustment are virtually unavoidable. Whether
dBSPL. This results in a 10‑dB decrease in gain for or not this poses a problem depends on the
inputs below 50 dBSPL (shown by the blue shaded accompanying complaints. If moderate and/or
region), indicating that the desired outcome of intense sounds are also reported as being loud,
reducing gain for weak sounds is achieved. The then this secondary effect is desirable. On the
gain for moderate sounds (50-70 dBSPL) is also other hand, if the adjustment results in moderate
inadvertently reduced (shown by the gray shaded and/or intense sounds becoming too soft, then a
region), but there is no impact on the gain for compromise must be made.
120 120
A B
110 110 TK’
TK’
100 TK 100 TK
Output (dBSPL)
Output (dBSPL)
90 90
80 80
70 70
60 60
50 50
20 30 40 50 60 70 80 90 100 20 30 40 50 60 70 80 90 100
120
Figure 6-3 C
Sample input/output functions for three approaches to adjusting 110
s for three the threshold kneepoint. The original and adjusted kneepoints are
hreshold indicated by TK and TK’ respectively. The blue shaded area shows
100 TK
the effect of the adjustment in the desired region; the gray shaded
adjusted
Output (dBSPL)
area demonstrates secondary effects of the change.

K and 90
aded area
ment in 80 TK’
haded
70
y effects
60
50
20 30 40 50 60 70 80 90 100
Input (dBSPL)
44
A word of caution against assumptions about Common Complaints
the function of various compression controls. For
and Solutions
example, consider the commonly held notion that
the gain for weak sounds can be decreased by Just as there are multiple ways to describe
increasing the TK. The I/O functions in Figure 6-3 a problem, there are often several solutions
demonstrate three ways of increasing the TK by to a given problem (Jenstad et al, 2003). The
10dB. In Figure 6-3A, this action does indeed following is a list of some common complaints
decrease the gain for weak sounds. In contrast, not and recommended solutions. Each problem is
only does increasing the TK not decrease the gain first categorized according to the general nature
for weak sounds in Figure 6-3B, it actually increases of the complaint, followed by a more specific
the gain for moderate and intense sounds! In description of the complaint, possible causes,
addition to not alleviating the primary complaint, recommended solutions, and other factors
this adjustment may give rise to a secondary and options that should be considered. Note
complaint of moderate and intense sounds that this is not intended to be an exhaustive
being too loud. Finally, as shown in Figure 6-3C, catalog of problems and solutions. Rather,
increasing the TK may decrease gain for all sounds. it should be treated as a general guideline.
Further, as indicated in the preceding section,
These examples highlight the need for the precise mechanism for addressing some of
comprehending the compression architecture of these changes will depend on the compression
the particular hearing aid being adjusted. Stated architecture of the particular hearing aid.
differently, the clinician must understand the
impact that fine-tuning will have in the intended
region as well as any secondary effects. This is
especially useful for predicting and minimizing
secondary complaints.
Complaint “My Voice Sounds Boomy”

“Sounds like I’m speaking with my head in a barrel (or tunnel),”
Own Voice Quality
Alternate “my voice echoes,” “my voice sounds hollow,”

Descriptions “my voice sounds like I have a cold,” or “my ears feel plugged.”
Possible Cause Excessive low‑frequency gain at high input levels.
Recommended Reduce gain for intense, low-frequency sounds.

Solution(s)
• O
cclusion due to the presence of the hearing aid in the ear. To test, turn off
the hearing aid and have the person vocalize /i/; the problem is occlusion if the
Other “boomy” sensation persists. Address the problem acoustically by increasing the
Considerations size of the vent.
• Individual may have forgotten the sound of his/her own voice due to
long‑standing hearing loss.
45
Own Voice Quality Complaint “My Voice Sounds Muffled”
Alternate “Hearing aid cuts out when I speak,” “hearing aid is weak.”
Descriptions
Possible Cause Insufficient low-frequency gain for high input levels.
• Increase gain for intense, low-frequency sounds.

Recommended
Solution(s) • Increase maximum output at low frequencies.
Other • Individual may be accustomed to linear amplification.

Considerations • M
ay be associated with complaints about other intense sounds (e.g., noisy
restaurants, music, etc.) sounding muffled.
Complaint “My Voice Sounds Distorted”

Own Voice Quality
Alternate “Hearing aid crackles when I speak,” “my voice sounds unnatural.”
Descriptions
Possible Cause Hearing aid is in saturation.
Recommended • Change to compression limiting.

Solution(s) • Increase maximum output.
• Decrease gain for intense, low-frequency sounds.
• P
roblems may be associated with complaints about other intense sounds
Other (e.g., noisy restaurants, music, etc.) sounding distorted or unnatural.
Considerations • Individual may have forgotten the sound of his/her own voice due to
long‑standing hearing loss.
• Evaluate distortion of hearing aid per ANSI specifications (ANSI, 2003).
46
speech intelligibility Complaint “I Hear Better Without My Hearing Aids”
Alternate “Difficulty understanding speech in quiet,” “speech is unclear,”

Descriptions “hearing aid makes no difference,” “difficulty understanding TV (or radio).”
Possible Cause Insufficient gain.
• Increase overall gain.

Recommended • Increase gain for high-frequency sounds.
Solution(s)
• D
ecrease overall gain (if person reports reducing the volume control to
improve listening comfort).
• Individual may have poor speech recognition ability unrelated to audibility

Other and/or unrealistic expectations regarding impact of hearing aid on speech
understanding.
Considerations
• H
earing aid may not be perceived as being loud enough due to previous
experience with linear amplification.
Complaint “I Can’t Hear Well at Restaurants”

speech intelligibility
Alternate “Don’t hear well in noise,” “difficulty understanding speech in background noise.”
Descriptions
• Increase overall gain.

Recommended
• Increase gain for high-frequency sounds.
Solution(s)
• Decrease gain for low-frequency sounds.
• Individual may have poor speech recognition ability unrelated to audibility and/or
unrealistic expectations regarding impact of hearing aid on speech understanding.
Other
• Hearing aid may not be perceived as being loud enough due to previous experience
Considerations
with linear amplification.
• Recommend use of technology to improve the SNR (e.g., directional microphones).
47
speech intelligibility Complaint “I Hear People at a Distance Better Than Those at My Table”
Alternate “Hear better at a distance.”

Descriptions
Possible Cause Insufficient gain for high input levels.
Recommended • Increase gain for intense sounds.

Solution(s) • Increase maximum output.
• M
ay be associated with complaints about other intense sounds
Other (e.g., own voice, music, etc.) sounding muffled.
Considerations • If sounds at a distance are reported as being too loud, consider decreasing gain
for weak sounds. May be associated with complaints about other weak sounds
being too loud.
Complaint “Hearing Aids are Too Loud”
Alternate “Background noise is too loud,” “voices are too loud,” “hearing aid is boomy,”
Descriptions “sirens are too loud,” “refrigerator hum is too loud.”
loudness
Possible Cause Excessive gain.
• Decrease overall gain if all sounds are too loud.

Recommended • D
ecrease gain for intense sounds if only intense sounds are too loud.
Solution(s) ecrease gain for weak sounds if only weak sounds are too loud.
• D
• Use expansion if only weak sounds are too loud.
Other New hearing aid users may have become accustomed to a quieter world due to
Considerations long-standing hearing loss.
48
Complaint “Loud Sounds are Uncomfortable”
Alternate “Sounds are painful,” “sound of clattering dishes is too loud,”

Descriptions “sound of running water is uncomfortable.”
loudness
Possible Cause Excessive output for loud and/or transient inputs.
• Decrease the maximum output.

Recommended • Decrease gain for intense sounds.
Solution(s)
• Decrease gain for intense, high-frequency sounds.
Other New hearing aid users may have become accustomed to a quieter world due to
Considerations long-standing hearing loss.
Complaint “Hearing Aids are Too Soft”
Alternate “Not loud enough,” “sounds are not clear.”

Descriptions
loudness
• Increase overall gain if all sounds are too soft.

Recommended • Increase gain for soft sounds if weak sounds are too soft.
Solution(s) • Increase gain for loud sounds if intense sounds are too soft.
• Increase maximum output if intense sounds or all sounds are too soft.
• M
ay be associated with complaints regarding difficulty understanding speech
Other (especially in quiet), or other intense sounds (e.g., own voice, music, etc.)
Considerations sounding muffled.
• H
earing aid may not be perceived as being loud enough due to previous
experience with linear amplification.
49
Complaint “Hearing Aids are Noisy”
Alternate “Refrigerator hum is too loud,” “noisy when I’m reading in my living room.”
Descriptions
sound quality
Possible Cause Excessive gain for low input levels.
Recommended • Use expansion.

Solution(s) • Decrease gain for soft sounds.
Other Evaluate circuit noise of hearing aid per ANSI specifications (ANSI, 2003).
Considerations
Complaint “Hearing Aids have Static”
Alternate “Sounds are distorted,” “crackling sound in noisy environments.”

Descriptions
sound quality
Possible Cause Hearing aid is in saturation.
• Change to compression limiting.

Recommended
• Increase maximum output.
Solution(s)
• Decrease gain for intense sounds.
• P
roblem may be associated with complaints about intense sounds (e.g., noisy
Other restaurants, music, own voice, etc.) sounding distorted or unnatural.
Considerations
• E
valuate distortion of hearing aid per ANSI specifications (ANSI, 2003).
50
sound quality Complaint “Hearing Aids Cut in and Out”
Alternate “Pumping sound,” “background noise fades in and out,” “level of sound in hearing
Descriptions aids fluctuates.”
Possible Cause Short time constants.
Recommended • Increase attack and/or release time.

Solution(s) • Decrease compression ration (i.e., make the hearing aids more linear).
Other Verify that the problem is not related to intermittency.

Considerations
Complaint “Hearing Aids Shut Down with Loud Sounds”

sound quality
Alternate None.
Descriptions
Possible Cause OCL with slow release time.
Recommended • Increase maximum output.

Solution(s) • Use faster release time.
Other Verify that the problem is not related to intermittency or battery drainage.
Considerations
Complaint “Sounds are Too Sharp”

sound quality
Alternate “Sounds harsh,” “sounds tinny,” “too much treble.”

Descriptions
Possible Cause Excessive high-frequency gain (relative to low-frequency gain).
Recommended • Decrease high-frequency gain.

Solution(s) • Increase low-frequency gain.
Other Individual’s auditory perception may be distorted due to long-standing

Considerations high‑frequency hearing loss.
51
Complaint “Sounds are Too Dull”
Alternate “Sounds boomy,” “sounds hollow,” “too much bass.”

Descriptions
sound quality
Possible Cause Excessive low-frequency gain (relative to high-frequency gain).
Recommended • Decrease low-frequency gain.

Solution(s) • Increase high-frequency gain.
Other • M
ay be associated with occlusion due to the presence of the hearing aid
Considerations in the ear.
• Individual’s auditory perception may be distorted due to long-standing
hearing loss.
Complaint “Hearing Aids Whistle”
Alternate “Squeal.”
Descriptions
feedback
Possible Cause Excessive gain.
• Decrease gain for weak, high-frequency sounds.

Recommended
• Decrease gain for weak sounds.
Solution(s)
• Decrease overall gain.
• A
ddress excessive acoustic leakage due to ill-fitting hearing aid or a vent that is
Other larger than necessary.
Considerations • U
se electronic feedback solutions such as notch filters and feedback cancellers.
52
Complaint “Chirping Noise”
Alternate “Whooping sound,” “beeping sound.”

Descriptions
feedback
Possible Cause Interaction between acoustic leakage and compression (Figure 6-4).
• Decrease gain for weak, high-frequency sounds.

Recommended • Decrease gain for weak sounds.
Solution(s)
• Decrease overall gain.
• A
ddress excessive acoustic leakage due to ill-fitting hearing aid or a vent that is
Other larger than necessary.
Considerations • Use electronic feedback solutions such as notch filters and feedback cancellers.
Acoustic leakage
(feedback occurs)
Hearing aid gain Input level to hearing aid

increases microphone increases
Input level to hearing aid Hearing aid gain

microphone decreases decreases
Feedback is eliminated
Figure 6-4
Mechanism behind “chirping” feedback. Figure 6.4
Mechanism behind “chirping” feedback.
53
Recommended Reading
Dillon, H. (2000). Hearing Aids. Boomerang Press: Turramurra, Australia.
Mueller, H. G., Hawkins, D. B., & Northern, J. L. (1992).

Probe Microphone Measurements: Hearing Aid Selection and Assessment.
Singular Publishing Group, Inc.: San Diego, CA.
Sandlin, R. E. (2000). Textbook of Hearing Aid Amplification: Technical and

Clinical Considerations. Singular Publishing Group, Inc.: San Diego, CA.
Valente, M. (2002). Hearing Aids: Standards, Options and Limitations.

Thieme Medical Publishers, Inc.: New York, NY.
Valente, M. (2002). Strategies for Selecting and Verifying Hearing Aid Fittings.

Thieme Medical Publishers, Inc.: New York, NY.
Villchur, E. (2000). Acoustics for Audiologists.

Singular Publishing Group, Inc.: San Diego, CA.
Vonlanthen, A. (2000). Hearing Instrument Technology for the Hearing

Healthcare Professional. Singular Publishing Group, Inc.: San Diego, CA.
54
References
American National Standards Institute (2003). American National Standard: Specification of

Hearing Aid Characteristics (ANSI S3.22-2003). New York: Acoustical Society of America.
Balfour, P. B., & Hawkins, D. B. (1992). A comparison of sound quality judgments for
monaural and binaural hearing aid processed stimuli. Ear & Hearing, 13(5), 331-339.
Beranek, L. L. (1947). The design of speech communication systems.

IRE Proceedings, 35, 880-890.
Byrne, D., & Dillon, H. (1986). The National Acoustics Laboratories’ (NAL) new procedure for
selecting the gain and frequency response of a hearing aid. Ear and Hearing, 7(4), 257-265.
Byrne, D., Dillon, H., Khanh, T., Arlinger, S., Wilbraham, K., Cox, R., et al. (1994).
An international comparison of long-term average speech spectra.
Journal of the Acoustical Society of America, 96(4), 2108-2120.
Byrne, D., Noble, W., & LePage, B. (1992). Effects of long-term bilateral and
unilateral fitting of different hearing aid types on the ability to locate sounds.
Journal of the American Academy of Audiology, 3(6), 369-382.
Ching, T. Y.-C., Dillon, H., & Byrne, D. (1998). Speech recognition of hearing‑impaired
listeners: Predictions from audibility and the limited role of high-frequency amplification.
Christen, R. (1980). Binaural summation at the most comfortable loudness level (MCL).
Australian Journal of Audiology, 2(2), 92-98.
Cornelisse, L. E., Seewald, R. C., & Jamieson, D. G. (1995). The input/output formula:
A theoretical approach to the fitting of personal amplification devices.
Cox, R. M., & Alexander, G. C. (1995). The abbreviated profile of hearing aid benefit.
Ear and Hearing, 16(2), 176-186.
Cox, R. M., Alexander, G. C., Taylor, I. M., & Gray, G. A. (1997).

The contour test of loudness perception. Ear and Hearing, 18(5), 388-400.
Dermody, P., & Byrne, D. (1975). Loudness summation with binaural hearing aids.
Scandinavian Audiology, 2(1), 23-28.
55
References (continued)
Dillon, H. (1999). NAL-NL1: A new prescriptive procedure for fitting non-linear hearing aids.
Hearing Journal, 52(4), 10, 12, 14, 16.
Golabek, W., Nowakowska, M., Siwiec, H., & Stephens, S. (1988). Self-reported benefits of hearing aids by
the hearing-impaired. British Journal of Audiology, 22(3), 183-186.
Hawkins, D. B., Prosek, R. A., Walden, B. E., & Montgomery, A. A. (1987). Binaural loudness summation in
the hearing impaired. Journal of Speech and Hearing Research, 30(1), 37-43.
International Electrotechnical Commission (1983a). Hearing Aids. Part 0: Measurement of Electroacoustic

Characteristics (IEC 118-0). Geneva, Switzerland: International Electrotechnical Commission.
International Electrotechnical Commission (1983b). Hearing Aids. Part 2: Hearing Aids with Automatic
Gain Control Circuits (IEC 118-2). Geneva, Switzerland: International Electrotechnical Commission.
Jenstad, L. M., Van Tasell, D. J., & Ewert, C. (2003). Hearing aid troubleshooting based on patients’
descriptions. Journal of the American Academy of Audiology, 14(7): 347-360.
Kamm, C., Dirks, D., & Mickey, M. (1978). Effect of sensorineural hearing loss on loudness discomfort level
and most comfortable loudness judgments. Journal of Speech and Hearing Disorders, 21, 668-681.
Killion, M. C., & Fikret-Pasa, S. (1993). The 3 types of sensorineural hearing loss:
Loudness and intelligibility considerations. Hearing Journal, 46(11), 31-36.
Kochkin, S. (1991). Hearing professionals’ views on market expansion. Hearing Instruments, 42(12), 6-8.
Montgomery, A. A., & Edge, R. A. (1988). Evaluation of two speech enhancement techniques to improve
intelligibility for hearing-impaired adults. Journal of Speech and Hearing Research, 31, 386-393.
Mueller, H. G. (2002). Hearing Aids. Featured presentation at the 9th Annual Convention of the Illinois
Academy of Audiology, Chicago, IL.
Nabelek, A. K., & Pickett, J. M. (1974). Monaural and binaural speech perception through hearing aids
under noise and reverberation. Journal of Speech and Hearing Research, 17, 724-739.
Neuman, A. C., Bakke, M. H., Mackersie, C., Hellman, S., & Levitt, H. (1998).
The effect of compression ratio and release time on the categorical rating of sound quality.
Journal of the Acoustical Society of America, 103(5 Pt. 1), 2273-2281.
56
References (continued)
Pascoe, D. (1988). Clinical measurements of the auditory dynamic range and their relation to formulas for
hearing aid gain. In J. Jensen (Ed.), Hearing Aid Fitting: Theoretical and Practical Views, Proceedings of
the 13th Danavox Symposium (pp. 129-152), Copenhagen: Danavox.
Plyler, P., Hill, A., & Trine, T. D. (2005a). The effects of expansion on the objective and subjective
performance of hearing instrument users. Journal of the American Academy of Audiology, 16, 101-113.
Plyler, P., Trine, T., & Hill, A. (2005b). The effects of expansion time constants on objective performance.
Poster presented at the 17th annual convention of the American Academy of Audiology, Washington, DC.
Scharf, B., & Fishken, D. (1970). Binaural summation of loudness reconsidered.

Journal of Experimental Psychology, 86, 374-379.
Silman, S., Gelfand, S. A., & Silverman, C. A. (1984). Late-onset auditory deprivation: Effects of monaural
versus binaural hearing aids. Journal of the Acoustical Society of America, 76(5), 1357-1362.
Skinner, M. W. (1976). Speech intelligibility in noise-induced hearing loss: Effects of high-frequency

compensation. Journal of the Acoustical Society of America, 67, 306-317.
Trine, T. D., & Van Tasell, D. J. (2002). Digital hearing aid design: Fact vs. Fantasy.
Hearing Journal, 55(2), 36-38, 40-42.
Valente, M., & Van Vliet, D. (1997). The Independent Hearing Aid Fitting Forum (IHAFF) protocol.
Trends in Amplification, 2(1), 6-35.
Walker, G. (1997). Conductive hearing impairment: The relationship between hearing loss,
MCLs and LDLs. Australian Journal of Audiology, 19(2), 71-80.
Woods, W. S., Frush, R. A., & Van Tasell, D. J. (1999). “Preference space” for parameters of two-channel
speech compression. Poster presented at the IEEE Workshop on Applications of Signal Processing to
Audio and Acoustics, New Paltz, NY.
World Health Organization (2005). Facts about deafness. Retrieved on October 17, 2005,
from http://www.who.int/pbd/deafness/facts/en/index.html.
Yund, E. W., & Buckles, K. M. (1995). Multichannel compression hearing aids:

Effect of number of channels on speech discrimination in noise.
Journal of the Acoustical Society of America, 97, 1206-1223.
57
Notes
58
Notes
59
A global hearing technology
company headquartered in
Eden Prairie, Minnesota, U.S.A.
Starkey Laboratories, Inc.

6700 Washington Avenue South
Eden Prairie, MN 55344-3476
800.328.8602
StarkeyPro.com
© 2011 Starkey Laboratories, Inc. All Rights Reserved.

60 BKLT1668-04-EE-ST 10/11

Compression Handbook PDF

Uploaded by

Copyright:

Available Formats

Compression Handbook PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Compression Handbook PDF

Uploaded by

Copyright:

Available Formats

What are the main topics covered in the document?

What are the main topics covered in the document?

What are some of the applications of compression discussed?

What are some of the applications of compression discussed?

The compression

Shilpi Banerjee, Ph.D.

Chapter 2: Basics of Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Chapter 3: Applications of Compression. . . . . . . . . . . . . . . . . . . . . . 21

Chapter 4: Fitting Compression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Chapter 6: Troubleshooting & Fine-tuning . . . . . . . . . . . . . . . . . . . . 42

Intense Too Loud Too Loud

Weak Too Soft Too Soft

Range of Environmental Sounds Dynamic Range of Hearing

125 250 500 1K 2K 4K 8K a loudspeaker, the receiver converts the

Basics components of a hearing aid: microphone, amplifier, receiver, and battery

the ordinate (Y‑axis) as a function of the input 110

in an output of 80 dBSPL, while an input of 50

(page 6) shows sample I/O functions for

Gain = Output – Input = 80 – 50 = 30dB

Figure 1-5) has a gain of 15dB. 30

30 that is distorted, often described as sounding

Harmonic distortion is said to occur when

Maximum Output of Device

Too Loud Too Loud

Too Soft Too Soft

Too Loud Too Loud

Too Soft Too Soft

Range of Environmental Sounds Dynamic Range of Hearing

it is tempting to alleviate the problem simply 50

90 reduction in gain for input levels greater than

threshold and compression ratio, indicate the

the constant hub-bub of a noisy restaurant). 70 TK

On the other hand, dynamic features, such

than it would be if no compression had occurred 90 TK

like the knee of a bent leg, the point at which 70

Depending on the purpose, a compression Figure 2-5

output. Because the reference condition is

Note that, in Figure 2-6, 30dB of gain is

when the CR is greater than 1:1. 50

Depending on the purpose, a compression Input (dBSPL)

a hearing aid so that it does not exceed the Figure 2.7

[Gain is reduced when

B Gain is reduced when

of any compression circuit. The position of this 80

As shown in Figure 2-12 (page 18), rotating the

Microphone Pre-Amplifier Volume Control Output Amplifier Receiver

70 Kneepoints types of circuits, the TK changes only in

50 Channels and Bands

Microphone Pre-Amplifier Volume Control Output Amplifier Receiver

90 acoustic events in a discrete frequency region

all frequencies. For example, Figure 2-15, shows

the response of a 1-channel and 4-channel

hearing aid in the presence of a 90 dBSPL tone

at 500 Hz. It is clear that the high-frequency

50 response is considerably reduced for the

Although on the surface it appears that a large

30 the ability to further fine-tune the frequency

6. Maintain listening comfort, 60

7. Maximize speech recognition ability, and 50

• ATs and RTs may be fast or slow, but should

• AGC-I compression is used because amount 110