Trends in Amplirication, Vol. 1, No.

3,1996 i)1996 Woodland

Microcomputer Applications for Hearing Aid

Selection and Fitting
Robert de Jonge, Ph.D.
Professor of Audiology, Central Missouri State University,
Warrensburg, MO 64093

INTRODUCTION review or evaluation of program features. Indeed,

The purpose of this manuscript is to provide one program to be described was developed by the
the reader with a knowledge of which computer author who can hardly be considered impartial.
programs are available to the clinician for facili- A traditional literature search (i.e., Medline)
tating the selection and verification of hearing was not particularly useful in identifying software
aids. The objective is to provide a general concep- associated with hearing aid fitting. Even when
tual understanding of the theoretical basis behind fairly recent articles were found, often it was
each program, what features the program has to doubtful whether the program was still commer-
offer, and to share an appreciation for the "look cially available for the platform for which it was
and feel" of the programs. Since microcomputers designed (e.g., Apple Ile). Three or four years is a
can be used in so many areas related to hearing long time in the computer business. Identifying
aids, perhaps it would be useful to describe what what software is "out there" and currently avail-
will not be covered. Several topics are beyond the able, supported, and would run on a major plat-
scope of this text, such as: a review of software de- form (Macintosh or IBM PC compatibles, Mac or
signed by hearing instrument manufacturers to fit PC) was the goal, but it was also a hit or miss
their own programmable hearing aids, software proposition. So, apologies are offered in advance
that is related to instrumentation involving probe for the author's ignorance in inadvertantly omit-
microphone devices, or software for business man- ting someone's favorite program or creation. Also,
agement. The Hearing Instrument Manufacturers since this is not intended to be a treatise on the
Software Association's (HIMSA) product, NOAH, scientific aspects of hearing aid fitting, material
will not be reviewed, but interested readers can will be presented in a less formal tone.
find out more about NOAH and related applica- The main portion of the text will begin with a
tions by seeing Robertson (1996). NOAH is soft- general review which, hopefully, will point out the
ware designed to integrate hardware and software current need for including computer-based schemes
modules covering a wide range of applications into the selection and verification process, and
from a variety of manufactures. One prominent which issues need to be addressed in hearing aid
application involves programmable hearing aids selection. The underlying purpose of much of the
which use a microcomputer to control a hardware software is to take the theory from basic research
interface device for setting the parameters of the and to apply those concepts to the fitting of hear-
instrument. This is certainly an important use for ing aids. For the reader who would like more de-
microcomputers; however, most hearing aids sold tails than this general overview provides, please
are not programmable. Here, a more generic ap- see Studebaker and Hochberg (1980; 1993) and
proach will be taken. The emphasis is on selecting Valente (1994). A brief summary will be included
appropriate electroacoustic characteristics for hear- of what computer hardware would be needed to
ing aids whether they are linear, nonlinear, pro- support the software. It is assumed that the reader
grammable or not. Also, this paper is intended to is somewhat familiar with basic computer termi-
be non-judgmental and informative, not a critical nology and concepts. Developing a working un-
86
R. de Jonge Microcomputer Applications 87

derstanding of microcomputers is not easy. At- the features offered by the various programs (see
tempts have been made to keep descriptions Table 1).
simple for the beginner yet complete and accurate
for the reader with a more advanced understand- HARDWARE AND SOFTWARE
ing. A full explanation of all computer related REQUIREMENTS
terms, suitable for the novice, would be very
lengthy and beyond the scope of the paper. Fol- Most of the programs (also referred to as "ap-
lowing the hardware section will be a review of plications") run under MS-DOS (Microsoft Cor-

Table 1. Hearing aid selection and verification software. Minimum system requirements are included
Program Author(s) or Contact System

MSUv3. I Robyn Cox PC XT, 256

Department of Audiology and Speech Pathology kB RAM,
Memphis State University dual floppy
807 Jefferson Avenue
Memphis, TN 38105
Phone: 901-678-5831
E-mail: [Link]/harl
DSL3.1 Richard Seewald PC XT, 640
University of Western Ontario kB RAM. I
Communication Disorders, Elborn floppy, Mac
London, ON NBG I -HI version of
Canada DSL 3.1
Phone: 519-661 -3901

DSL 4.0 or DSL [i/o] E-mail: dsl@[Link] Windows.

386
processor

IHAFF IHAFF working group PC XT, 450

v 1.1 a Dennis Van Vliet kB RAM.
17021 Yorba Linda Blvd, Suite 130 0.5 MB
Yorba Linda, CA 92686 hard drive
Phone :714-579-0717:800-795-4288

Fig6 Mead Killion or Toni Gitles PC AT. 525

3.0 Rev L Etymotic Research kB RAM. I
61 Martin Lane floppy
Elk Grove Village, IL 60007
Phone: 847-228-0006
HAS 2.5 Robert de Jonge Macintosh,
SII Department of Speech Pathology and Audiology HyperCard
Central Missouri State University 2.2 or 2.3
Warrensburg, MO 64093
contact:
Support Syndicate for Audiology
108 S. 12th Street
Pittsburgh, PA 15203
Phone: 800-869-0758
SII ANSI Draft Standard (ANSI S3.79. proposed) PCXT, 128
Chaslav Pavlovic kB RAM. 1
Resound Corporation floppy
220 Saginaw Drive
Redwood City. CA 94063
Phone: 415-780-7867

HASP 2.07 Harvey Dillon PC AT. 384

National Acoustic Laboratories kB RAM.
126 Greville Street 0.7 MB
Chatswood 2067, NSW, Australia hard drive
Fax: 612-411-8273
88 Trends in Amplification

poration's Disk Operating System), the main op- class machines (including 486 and Pentium) to use
erating system for the PCs. An exception is the hard disk space as RAM. This "virtual memory"
most recent version of the Desired Sensation allows larger programs to operate. It also allows
Level (DSL 4.0), also called DSL [i/o], where "i/o" more programs to run concurrently. The main dis-
is an acronym for "input/output." DSL [i/o] is a advantage to using SoftWindows software emula-
Windows program. A Mac version of DSL 3.1 is tion is speed. Programs run slower. The DOS
also available. None of the other applications Compatibility Card, installed in the Mac 6100/60
were Windows based, but all ran without trouble PowerPC, is a 66 MHz 486 DX2 computer with 8
when they were invoked from the File Manager MB of RAM memory. Its performance was the
within Windows 3.1. The DOS programs are all same as a similarly configured stand-alone PC. The
modest in terms of the power of the microproces- DOS card is sold by Apple Computer, but other
sor required to run them, PC XT or AT with 640 third-party manufacturers offer similar products.
kB (kilobytes) of random access memory (RAM) All the figures of the DOS-based programs
or less. If an older computer is available, it could were taken from the Mac screen. The appearance
be dedicated to running these programs. In most of the programs was the same whether they were
cases the programs could run from a single floppy, viewed on the Mac or PC screen. Adobe Photo-
but all could be installed on a hard drive and used Shop 2.5 was used to convert the original displays
less than 1 MB (megabyte) of disk space. There (which were often in color) into black and white
was no difficulty installing them in their own di- images that were appropriate for viewing within
rectory or running them on a PC with an Intel 486 this text. This sacrifice lessens the esthetic appeal
processor. Only the DSL 3.1 software modified of the original programs, but at least gives the
the [Link] file, and this just added a line reader a feel for the features offered by the pro-
referencing a directory for storing patient files. grams.
So, removing the software, if necessary, is not For the Macintosh software, the Hearing Aid
complicated, and you needn't worry about the in- Selection (HAS) stack (program) required a Mac
stallation interfering with the operation of your and the HyperCard 2.2 application. HyperCard is
existing software. The [Link] file contains a a programming environment for the object-oriented
series of commands which are executed when the HyperTalk programming language. The Hyper-
computer is first starting up ("booting"). Chang- Card program can be used to create stacks, much
ing some of these commands, or the order in like a word processing program creates data files.
which the commands execute, can potentially in- The term "stack" is a metaphor for a stack of
terfere with the proper operation of programs. "cards." Each card can be thought of as a screen
Therefore, it's best that the [Link] file be image (Figures 19 to 33). HyperCard stacks take
left alone unless there is a good reason to change it. advantage of the Mac's intuitive user interface of
Even though these programs were designed to buttons, pull-down menus, etc. Stacks can be run
run on a PC, there are two ways that they can also from HyperCard or the HyperCard Player appli-
run on a Mac. The first method is by software em- cation that is distributed free of charge with each
ulation, and the second is by installing additional Mac. HyperCard Player allows existing stacks to
hardware in the Mac. This second approach es- be run, but new ones cannot be created. Stacks
sentially puts a PC on a card into one of the Mac can also be compiled into free-standing applica-
expansion slots. All of the PC programs, except tions. The HAS and Speech Intelligibility Index
DSL 4.0, ran on a Mac 6100/60 PowerPC com- (SLI) stacks will run on an older Mac Plus, but
puter running SoftWindows 1.02a. This is a pro- they perform significantly better if run on a Mac
gram that emulates the Intel 286 processor (PC with a faster processor (like a 16 to 25 MHz Mo-
AT class). It permits the Mac to run DOS 6.21, torola 68030 processor, or better). Currently all
Windows 3.1, DOS-based programs, and Win- Macs being sold perform at these levels or better.
dows programs in standard mode. Speed is com-
parable to entry level 486 machines. SoftWindows
3.0 emulates the 486 processor and runs Windows The Need For Computer-Based Applications
3.1 in enhanced mode and also runs Windows 95. In the early 1970s, when the author was first
No additional hardware is required. Since DSL exposed to hearing aid fitting, the selection and
4.0 requires, at minimum, a 386 processor, this verification process was much more of an art than
program needs to run in enhanced mode. En- a science. This was necessarily so, since so little
hanced mode mainly refers to the ability of 386 was known, or, at least, little was common knowl-
R. de Jonge-Microcomputer Applications 89

edge in the clinical, audiological community. The were being replaced by in-the-ear (ITE) hearing
entire concept of hearing aid prescription was aids. The 2-cc coupler could now be either an
viewed as questionable. It was doubted whether a HA-1 or HA-2 coupler depending upon whether
hearing aid could be prescribed with the same de- an ITE or BTE aid was being measured, respec-
gree of confidence as, for example, eyeglasses tively. The two couplers are slightly different
could be prescribed. If a prescriptive procedure physically and consequently measure different
was used, it was mainly the half-gain rule (Lybar- sound pressure levels (SPLs). In-the-canal (ITC)
ger, 1944). More commonly, hearing aids were se- aids grew in popularity as did completely-in-the-
lected at random, usually three or four aids, from canal (CIC) aids. Changing the location of the mi-
a pool of instruments available in the clinic. The crophone, as worn on the listener, would alter the
patient wore stock rather than custom earmolds. input to the hearing aid, and consequently the
Hearing aids were compared using a variety of output. These real-ear effects would not be mea-
speech perception tasks, or some variation of the sured on the 2-cc coupler. Hearing aids situated
Carhart procedure (Carhart, 1946). It was recom- deeply in the ear canal would produce an output
mended that the patient purchase (from a differ- SPL at the eardrum much greater than hearing
ent person, the hearing aid dealer) the hearing aid aids with a more shallow placement. Coupler gain
giving the best performance. When hearing aids would be different from real-ear gain for the aver-
were being compared, it was uncertain whether age ear. Individual differences in ear canal geom-
the listening environment, speech presented in a etry and middle ear impedance meant that the re-
sound-proof booth, was a valid simulation of the lationship between 2-cc coupler gain and real-ear
real-world listening experience. Objective mea- gain would be unique for every individual (de
surements of hearing aid performance were limited Jonge, 1996). Most certainly, hearing aids would
to 2-cc coupler measures. The differences between perform differently on children versus adults, nor-
decibels of coupler gain, audiogram thresholds ob- mal middle ears vs. pathological ears, and effects
tained in supra-aural earphones calibrated on a of normal variability within the population could
6-cc coupler, and real-ear gain were not appreci- be expected.
ated. The pronounced effects of earmold acous- Electronically, hearing aids were becoming
tics on the frequency response of the hearing aid more complex. Originally, most hearing aids had
were typically not assessed or controlled. No at- linear amplifiers. The frequency response of the
tempt was usually made to determine how the hearing aid was constant as the input level to the
hearing aid was shaping the spectrum (i.e., fre- hearing aid changed, at least up to the point
quency response) for the individual. The selection where saturation occurred. Two response curves
process was not logically connected to a prescrip- would describe the aid: one below saturation
tion. Verification of the selection was a self fulfill- where the hearing aid is in the linear range (the
ing prophecy since the Carhart procedure was used frequency response), and one at, and above, satu-
both to select and verify. It was doubtful whether ration (the saturation sound pressure level with a
speech-based measures could reliably choose the 90 dB input, OSPL90). Which frequency response
best hearing aid (Shore et al, 1960; Thornton and was appropriate for a patient was a question that
Raffin, 1978). And perhaps the best hearing aid had a single answer. This is not the case for hear-
was not chosen if it was not included in the origi- ing aids with frequency-dependent compression.
nal group of hearing aids randomly selected for Depending upon values selected for compression
evaluation. Follow-up evaluation was problematic ratio(s) and kneepoint(s), the frequency response
since the person fitting the hearing aid was not the of compression hearing aids can change dramati-
one making the recommendation, and there was cally with changes in the input level. Choosing a
no guarantee that the recommendation would be frequency response became a more complicated
followed or, indeed, that the audiologist's recom- question. What response is needed at which input
mendation was worthy of being followed. level? In addition the dispenser has been given
more control over the hearing aid. In the past
Increased Hearing Aid Complexity hearing aids may have had a low-frequency cut or
an output limiting control. Now, the audiologist
In subsequent years, hearing aids became more can select a hearing aid from dozens of possible
complex. The effect of this complexity on the per- matrices with a variety of trim pots. And, espe-
ception of sound became appreciated as a signifi- cially with programmable hearing aids, the audi-
cant issue. Behind-the-ear (BTE) hearing aids ologist has control over a bewildering array of pa-
90 Trends in Amplification

rameters, for multiple channels and memories, actually performing speech audiometry it is possi-
that can create a multitude of different frequency ble to predict the effects of hearing loss (i.e., dif-
responses. The audiologist faces the dilemma of ferent audiometric configurations) on speech in-
being able to deliver almost any frequency re- telligbility. It is possible to anticipate how modifying
sponse, but being uncertain of which will maxi- the speech spectrum through changing the fre-
mally benefit the patient. Consequently, knowing quency response of the hearing aid will alter speech
what to deliver (the hearing aid prescription) be- recognition. It is also possible to predict how am-
comes a very important consideration. plified environmental noise might interfere with
speech perception (ANSI S3.79, proposed).
Advances in Basic and Applied Research
So, information is available that would make it
possible for the audiologist to take a much more
In the last twenty years, much progress has structured approach to the hearing aid selection
been made in a number of areas. Hearing aid pre- and verification process. For example, after the
scription techniques have been developed. Proba- audiogram is obtained, a set of calculations speci-
bly the most well known, widely used, and experi- fied by a prescriptive procedure would generate a
mentally verified is that developed at the National target REIR. The effect of earmold modifications
Acoustic Laboratories in Australia (NAL-R, Byrne and venting could be addressed (Dillon, 1985; Dil-
and Dillon, 1986; Byrne et al, 1990). Important lon et al, 1992). The effect of the REIR on speech
acoustical effects have been defined: the role of intelligibility could be calculated. If a satisfactory
the head, neck, torso, pinna, concha, ear canal and result is obtained, then correction factors could be
middle ear in transforming sound from the free- applied to each frequency to convert the REIR to
or diffuse field to either the entrance to a hearing a target 2-cc coupler frequency response. The cor-
aid microphone or the eardrum (Shaw, 1974; Shaw, rection factors would adjust for the effect of hear-
1980; Bentler and Pavlovic, 1989; Table 2). Differ- ing aid microphone placement, the average REUR
ences between hearing aid output in the real ear and RECD, and type of coupler (HA-1 vs HA-2).
versus the 2-cc coupler (real-ear coupler differ- Experimental data could be used to generate a
ence, RECD) have been measured, and it is possi- target RESR based upon average loudness dis-
ble to convert one to the other (Killion and Revit, comfort levels (LDLs) for individuals with vary-
1993; Revit, 1994). The factors producing individ- ing amounts of sensorineural hearing loss (Kamm
ual differences in real-ear gain and output are bet- et al, 1978; Pascoe, 1988). The target RESR, again
ter understood. And, most importantly, equip- based upon the average ear, could be converted
ment for measuring the in-situ performance of a to a OSPL90 curve for a 2-cc coupler. Once the
hearing aid is common and readily available com- frequency response and SSPL have been identi-
mercially. Real-ear probe-microphone systems fied, an appropriate BTE aid could be ordered, or
can sample the SPL for a wide range of frequen- a matrix could be selected for ordering a custom
cies at the eardrum of individuals whether they ITE/ITC/CIC hearing aid.
are unaided (REUR, the real-ear unaided re- For a more individualized fit, measurements of
sponse) or wearing the hearing aid (REAR, the the patient's own loudness perceptions (i.e., LDL
real-ear aided response). The real-ear insertion or most comfortable level, MCL) could be made.
response (REIR) can be measured as the differ- The patient's own REUR and RECD could be
ence between aided and unaided listening (REAR used to develop customized corrections. When
- REUR). The real-ear saturation response (RESR) the hearing aid arrives, any adjustments could be
can be compared to the OSPL90 of the hearing made (based upon 2-cc coupler measurements) to
aid in the 2-cc coupler. fine-tune the instrument in the absence of the pa-
The characteristics of the average conversa- tient. When the patient is wearing the aid, probe
tional speech spectrum have been measured for microphone measures would verify (or refute)
different vocal efforts (Pearsons et al, 1976; Pav- that the real-ear fit is as expected based upon the
lovic, 1993). The relationship between the speech initial prescription, and more fine-tuning could be
spectrum and normal audibility is known so that performed. The effect of any residual error, of the
the dB SPL of speech in the sound field can be measured REIR not matching the target, upon
compared to an audiogram measured in hearing speech intelligibility could be calculated and a de-
level (dB HL). Techniques like the articulation cision made whether to proceed with the fitting.
index (Al) and speech intelligibility index (SII) A self report questionnaire could be used before
can be used to make useful predictions. Without and after a period of hearing aid use to evaluate
 Table 2. A summary of acoustical correctilions often made by hearing aid selection software
and some psyclhoacoustical measures.
Variable Name Comment

Microphone location micLoc Increased SPL at mic location due to baffle of effects of head and
external ear; varies for BTE, ITE, ITC, CIC aids

Real-ear SPL RESPL SPL developed in a real ear canal, close to the eardrum

Real-ear aided response REAR The RESPL created during aided listening
Real-ear saturation response RESR The maximum RESPL, measured with the aid in saturation, 90 dB SPL
RESR= OSPL90 + RECD - 3mm bore
2-cc coupler SPI 2ccSPLI SPL developed in the HA- 1, 2-cc coupler,
2ccSPL2 SPL developed in the HA-2, 2-cc coupler
2ccSPL Refers to either 2ccSPLI or 2ccSPL2
18 mm long, 3 mm diameter 3mm bore Produces greater high frequency output in HA-2 coupler versus HA- 1:
earmold bore 3mm bore = 0 dB for HA- I coupler measures
2ccSPL2 = 2ccSPL I + 3mm bore

6-cc coupler SPL 6ccSPL SPL developed in the NBS-9A coupler used to calibrate circumaural
earphones
2-cc coupler frequency 2ccCR Gain as a function of frequency for an HA- 1 or HA-2 coupler
response

Output saturation sound OSPL90 The maximum 2ccSPL, measured with the aid in saturation, 90 dB SPL
pressure level OSPL90 = RESR + RECD -3 mm bore
Real-ear coupler difference RECD RECD = RESPL -2ccSPL
for an HA- I coupler Load impedance of ear canal and middle ear increases output, especially
for higher frequencies in the real ear
Real-ear dial difference REDD REDD = DialReading -RESPL
The difference between the audiometer dial reading (DialReading, in dB
HL) and the RESPL

Real-ear unaided response REUR The difference in SPL developed by the reference mic and the RESPL
due to external ear resonances

Real-ear insertion response REIR Real-ear insertion gain frequency response

REIR = REAR -REUR

NBS-9A. 6-cc coupler 6cc Converts 6-cc coupler SPL to RESPL

RESPL = 6ccSPL + 6cc
Audiometric zero, circumaural ANSI Converts dB HL to dB SPL in the 6-cc coupler, according to ANSI
earphones S3.6- 1989
6ccSPL = dB HL + ANSI
ER-3A insert earphone ER3A Converts dB HL to dB SPL in the HA-1, 2-cc coupler, according to
ANSI S3.6-1989
2ccSPL = dBHL + ANSI
Coupler response for flat CORFIG Converts REIR to 2ccCR
insertion gain CORFIG = REUR -RECD + 3 mm bore - micLoc
2ccCR = REIR + CORFIG
The opposite of CORFIG GIFROC Converts 2ccCR to REIR
GIFROC =-CORFOG
REIR = 2ccCR + GIFROC

Audiogram threshold dBHL Conventional audiometric threshold measured in dB HL

Most comfortable level MCL MCL is usually viewed as the target output for amplified average
conversational level speech, measured in dB HL, RESPL, etc.; MCL
is often calculated from dBHL, or could be measured directly.

Loudness discomfort level LDL LDL is usually set to slightly less than RESR or OSPL90; LDL is often
calculated from dBHL, or could be measured directly.
92 Trends in Amplification

subjective benefit from the hearing aid (Cox and with 256 kB of memory, a hard drive or dual
Alexander, 1995; Fabry and Schum, 1994). floppy, color or monochrome monitor, and a printer
All of the above may seem complicated, too for text output. The documentation file can be
complicated to be worth the effort even if a better printed to serve as a user manual. This manual is
fit and, presumably, a more satisfied hearing aid restricted to an explanation of how to use the pro-
user was the result. Or, at the very least, such at- gram; there is no information about the (consider-
tention to detail would be tediously time consum- able) theory nor empirical research behind the
ing. It would take away from time spent with the MSU procedure. However, references are given
patient, and it would be impractical in the real to several articles which thoroughly explain the
world of hearing aid fitting. And this would be rationale and verification of the approach. To ef-
true if it were not for a remarkable invention that fectively use and understand the software, I found
emerged in the late 1970s and has now become the articles to be very helpful and essential (Cox,
commonplace, the microcomputer. A microcom- 1983; Cox, 1984; Cox, 1985; Cox and Alexander,
puter, running the appropriate software, can take 1990).
the results of basic research and apply these re- The program is text-based, no graphics. Fea-
sults to the practical task of hearing aid fitting. tures are easily selected via the keyboard using
The computer can reduce the complexity of add- pull-down menus. The user can create, save, and
ing correction factors, gain conversion, calculating load patient files containing basic information
speech intelligibility, etc. to a simple matter of en- such as name, date, date of birth, clinician, ear
tering some information about the patient and tested, comments, etc. All information relating to
choosing the appropriate menu item. While the the hearing aid prescription and calculations made
microcomputer cannot have empathy nor replace by the program can be saved. The information en-
clinical judgment, perhaps it can be used as an as- tered, a table containing the prescription, along
sisting tool, allowing more science to be added to with a worksheet for comparing three hearing
the process, freeing the clinician to pursue the art aids, can be printed.
of hearing aid fitting. The principle used for selecting insertion gain
is based on the long term listening range (LTLR).
The LTLR range is defined by the patient's
SOFTWARE PACKAGES threshold at the low end and the HCL (highest
The following is a brief description of the soft- comfortable listening level, formerly referred to
ware packages available for assisting the clinician as ULCL, the upper limit to the comfortable lis-
in the selection and verification of the electroa- tening range) at the high end. Cox and Bisset
coustic characteristics of hearing aids. The de- (1982) found that the preferred listening level is
scription includes hardware requirements, theo- close to the midpoint of the LTLR. If the LTLR is
retical basis of the procedure, and details of how 30 dB or greater, conversational speech, produced
the clinician interacts with the software. Gener- with a raised vocal effort (70 dB SPL overall
ally, earlier procedures that were primarily de- level), is amplified so that 1/3 octave band SPLs
signed to fit linear hearing aids are described first. are positioned at the midpoint of the LTLR. For
More recent procedures tend to follow. Often, the patients with a reduced LTLR, speech is ampli-
more recent programs address issues involving fied to a level 15 dB below the HCL. Insertion
frequency-dependent compression (including pro- gain can be converted to coupler gain by applying
grammable) aids, issues such as target input-out- the appropriate corrections for either a BTE or
put functions and level dependent frequency re- ITE aid.
sponses. For this reason, version 4.0 of the DSL The OSPL90 curve is selected with consider-
procedure is considered separately from version 3.1. ation given to both the loudness and clarity of
speech:
Memphis State University Hearing Aid OSPL90 - (HCL -12 dB) + speechPeaks
Prescription (MSUv3.1) where the difference between the dB level ex-
Version 3.1 of the University of Memphis (for- ceeded by 1% of the speech peaks and the rms
merly called Memphis State University) software level is indicated by speechPeaks (Cox, 1988).
comes on a single floppy disk which includes both The formula is designed to calculate an output
the program and documentation. Hardware re- level that would not exceed the user's comfort
quirements are modest: PC XT class machine range, yet only allow 1% of the speech peaks to
 ds
R. de Jonge-Microcomputer Applications 93

be distorted by clipping (harmonic distortion), if Data in dB Hearing Leuel

that is the method used for output limiting by the
hearing aid. OSPL90 can be displayed for either Freq hz THE USFT HCL
a
an HA-1 or HA-2 coupler.
Thresholds can be entered for supra-aural or 250 40 41 * 83 *
500 45 48 * 86 *
insert earphones (such as the Etymotic ER-3A). 800 50 51 * 90 *
FoI the supra-aural earphone, thresholds can be 1000 55 57 * 90 *
expressed in either dB HL or the equivalent real- 1600 55 56 * 89 *
2000 60 62 * 88 *
ear SPL (RESPL), called the sound pressure 2500 65 64 * 91 *
hearing level (SPHL). Insert earphone thresholds 3000 70 70 * 93 *
require SPHL. The HCL can be measured di- 4000 75 75 * 95 *
6000 80 87 * 94 *
rectly or it can be estimated. The method for mea-
suring HCL uses a descending procedure and
stimuli commonly available on standard audiome- * denotes calculated values
ters. The patient is presented with either warble
tones or 1/3 octave bands of pulsed noise. Loud- Figure 1. Subject data for an individual with moderate
ness of these stimuli is comparable to the loudness hearing loss given by the MSUv3.1 program.
associated with bands of speech. Initially, the pre-
sentation level is relatively high, close to LDL. surements, you can change HCL at 500 and 4000
The stimuli are reduced in level until the patient Hz, based upon your judgment, to modify output
responds, indicating that the comfort range has levels for an individual you suspect to be able to
been reached. A series of descents are performed, tolerate more, or less, loudness. Figures 1 and 2
and the highest level producing a comfort re- show screen displays created by the MSUv3.1
sponse is HCL. The day-to-day reliability of this program. For this example, HCLs and unaided
procedure is such that standard deviations (SDs) sound field thresholds (USFTs) were predicted
of test-retest differences are 5 to 6 dB. Long term from threshold (THR). In addition to coupler and
(more than 17 months) SDs of test-retest differ- real-ear measures, the target aided sound field
ences for this procedure are greater, about 8 dB threshold (ASFT) is given.
(Cox, 1989).
Since loudness testing can be time consuming Desired Sensation Level (DSL 3.1)
and difficult, if not impossible, for some patients
to perform, HCL can be estimated using one of The DSL software comes on a single disk. Like
two predictive techniques. The first uses regres- the MSU program, hardware requirements are
sion equations to predict HCL from SPHL and modest: IBM XT, 640 kB RAM, a single double-
take the following form: density floppy drive, DOS 3.3, and at least Her-
cules compatible, CGA video. This was a very
HCL = m (SPHL) +b common video standard which pre-dated EGA
where m ranges from 0.25 to 0.45 and b from 68 to
85 dB, depending upon frequency (Cox, 1988). PRESCR IPT ION
Using this procedure produces errors in predict-
ing the true HCL. Standard errors of estimate are Freq hz Zcc Gain OSSPL 90 ASFT Goal IG
I I I~~~~~
about 9 to 10 dB. The second procedure was de- Z50 14 102 38 19
veloped to reduce this error and requires that 500 12 104 42 16
HCLs be measured at only two frequencies, 500 800 18 107 37 21
1000 ZZ 106 38 25
and 4000 Hz. The difference between predicted 1600 21 104 35 24
and measured HCL (called the residual) is calcu- 2000 26 108 36 29
lated by the program and used to adjust HCLs at 2500 29 111 38 Z8
3000 31 113 39 32
other frequencies. Standard errors of estimate us- 4000 29 115 41 36
ing this technique were about 4 to 6 dB, which is 6000 34 113 47 44
comparable to the day-to-day test-retest reliabil-
ity (Cox, 1989). So, if the patient can perform the Based on the use of an ITE hearing aid.
Zcc Gain and OSSPL 90 measured on HA-1 Coupler
HCL measurement, and if the dispenser wishes to
take the time, a more accurate prescription can be Figure 2. Prescription for the hearing loss shown in Fig-
made. Also, without making any loudness mea- ure 1.
94 Trends in Amplification

and the newer VGA and SVGA standards. The to permit adjustment of key parameters (i.e., REUR
software should work with almost any monitor. and RECD) to make the hearing aid recommen-
The 137 page DSL manual, "A Computer-Assisted dation applicable to older children and adults.
Implementation of the Desired Sensation Level The theory behind the DSL 3.1 procedure is
Method for Electroacoustic Selection and Fitting similar to that of MSUv3.1. An approximate bi-
in Children" is clear, well-written, quite compre- secting of the dynamic range is the target level
hensive and covers both the use of the program (i.e., desired sensation level re: threshold) for the
and much of the theory behind DSL's procedures. long-term average conversational speech spec-
The interested reader can find additional infor- trum (LTASS). Speech should remain audible,
mation about DSL in other publications (Seewald above threshold, but not exceed the upper limit of
and Ross, 1988; Seewald, 1992; Seewald et al, the comfort range which then becomes the target
1993a; Seewald, 1994; Moodie et al, 1994). Phone RESR. The target RESR is calculated by the pro-
numbers are given for technical support, and dif- gram, but it is also listed in a series of tables
ferent contact people are listed depending upon printed in the manual. RESR increases systemati-
the type of request, i.e., theoretical development, cally with increases in hearing loss. For a normal
software/hardware issues, or clinical application. threshold, the RESR varies from 94 to 102 dB
The tutorial information and references should SPL, depending upon frequency. For a 40 dB HL
make the manual appealing to faculty teaching in- hearing loss the RESR increases by about 5 dB
dividuals first being exposed to hearing aid fitting, over the normal level. Beyond 40 dB HL the
such as graduate audiology students. Data is en- RESR increases by about 1 dB for each additional
tered by keyboard and manipulated by pull-down 2 dB of hearing loss. The DSL procedure empha-
menus. Navigation through the program is easy. sizes the value of measuring, and representing, all
What is expected of the user is straightforward. variables with a common reference: in-situ ear ca-
The screen displays are easily understood. Help nal SPLs. However, all predictions can be made
screens are available where the program requests from audiogram thresholds obtained with conven-
information that might not be intuitively obvious. tional earphones, so this is the minimum informa-
For operating the program, it was seldom neces- tion that needs to be entered into the program.
sary to consult the manual. In addition to the au- Thresholds can also be entered for measurements
diometric and prescriptive results, client demo- calibrated for sound field listening (dB HL), ER-
graphic information can be entered. All relevant 3A insert earphones (dB HL), or RESPLs.
information can be stored, retrieved, or printed. Once thresholds are entered the option is avail-
A one page report, a hearing aid recommendation able for supplying additional information about
form, can also be printed. LDLs, REUR, and RECD. The manual describes
As the name of the manual implies, the pro- how the RECD can be obtained. As shown in Fig-
gram is focused upon the amplification needs of ure 3, I have defined custom values for each of
children. For example, the speech spectrum used these measures. While the program modifies its
as the input to the hearing aid is a compromise be- prescription for changes in REUR and RECD,
tween adult values for males and females and also LDLs are not used to calculate the target RESR.
considers the spectrum of the child's voice when
self-monitoring speech (Cornelisse et al, 1991).
However, the speech spectrum is not too different LEFT Ear
from one described by Cox and Moore (1988) that THRESHOLD MEASURES
would be applicable to adults. Concerning the size
of the external ear, young children tend to have Measure: Thresholds (TDH Head Phone - HL) Press F3 for choices

ear canals which are shorter and narrower than

dB 30 30 35 40 45 50 55 60 65

adults. A shorter canal will tend to have a REUR ADDITIOINAL ASSESSMEPIT MEASURES

peaking at a higher frequency. A smaller sized ca- easur LDL TDH Head Phon HEL) c ! :s
Medsure: LDL ( TDH Hedd Phone HL )
nal will have a larger RECD. The program calcu- dB 80 90 95
-
1N0 102 105 108 110 liz
lates prescriptions for young children based upon Measure: Real Ear Unaided Response (REUR)
dB 2 2 2 2 6 12 16 14 1Z
age specific values for REUR and RECD. This Measure: Real Ear Coupler Difference (RECD)_ HA-ZSTD
data is given at yearly intervals from birth to 5 dB 3 3 3 5 6 8 10 12 13
[End]=Calc. A Disp.. [Esc]=Abort, FP1]=SPECIFIC HELP, [F5]=XEYS HELP
years. While the program is specifically geared to
these unique aspects of pediatric fittings, the au- Figure 3. Thresholds, LDLs, REUR, and RECD values
thor's opinion is that the software is flexible enough entered into the DSLv3.1 program.
R. de Jonge Microcomputer Applications 95

DSL Selection Results

Ear: LEFT LEFT Ear
DE JONGE, JAMES Measure: Thresholds 1TDH Head Phone - HL)
GAIN MEASURES
Z50 500 750 1000 1500 Z000 3000 4000 6000
THRESH. ENTERED
Measure: Insertion Gain Press F3 for choices
MAP' (SPL) Meas 10 15 16 Z 24 31 30 Z2 27
Targ 6 7 11 19 23 Z8 34 37 44
DSL (dB) 1 !i!6 sL | ff " Diff 4 8 5 1 1 3 -4 -9 -17

AMP. LTASS (SPL) MAXIMUM OUTPUT MEASURES

IM SITU GAIM (dB)
LDL (SPL) Is GM Measure: RESR 90
Meas 101 106 110 112 108 109 111 112 113
|~lx 93W_LI i [|. L LM
DSL RESR (SPL) Targ 97 105 106 107 109 114 116 115 115
Diff 4 1 4 5 -1 -5 -5 -3 -2
IMSERTIOM GAIM
[End] =Cac. B Disp., [Esc1=Abort, [Fl]=SPECIFIC HELP, [F5]=KEYS HELP
AIDED SF (HL) 11] 1I
iPress AMY key to coont inue ..-
Figure 6. Measured versus target gain (REIG) and out-
Figure 4. Selection results calculated by DSLv3.1 for put (RESR) computed by DSLv3.1.
the patient in Figure 3.

However, a subsequent display allows measured fied by any one of five procedures: aided sound
LDLs to be compared to predicted RESRs so that field thresholds, REIR, REAR, amplified level of
adjustments can be made to the hearing aid, if de- LTASS, or 2-cc coupler. The amplified LTASS
sired. Figure 4 shows the calculated selection re- procedure developed by Hawkins et al (1989)
sults. Audiogram thresholds are converted to a measured the amplified level of frequency modu-
minimum audible pressure (MAP) which is the lated (FM) tones adjusted to have a level equiva-
RESPL corresponding to threshold. The target lent to 1/3 octave bands of speech. The two op-
DSL and amplified LTASS are given along with tions for output verification are a RESR to a
the in-situ gain (REAR), insertion gain defining pure-tone sweep at 90 dB, or coupler measures of
the REIR, and aided sound field thresholds OSPL90. It is recommended that a RECD be ob-
needed to verify the insertion gain. At this point tained if the coupler method is used. Figure 6
comments can be entered as to the type of hearing shows the difference between the target and mea-
aid and earmold. These comments are stored with sured gain and output for the two methods that were
the patient's file and are printed with the hearing selected.
aid recommendation form, but they are not used Figures 7 and 8 are graphs displaying the un-
in any calculations. For example, a Libby horn aided and aided results associated with the selec-
could be selected, but the REIR would not reflect tion and verification procedure (SPLograms ). All
the expected increase in high frequency gain. Fig- variables are represented in RESPLs. For each
ure 5 shows the target 2-cc coupler use gain and graph, variables can be added to the display se-
OSPL90 needed to supply the desired REIR and quentially. This should simplify the counseling
RESR. process if, for example, this graph were to be used
The next step in the process is verification of
gain and output. The manual describes how the UN§AIDED - DE JONGE, JAMtES - LEFT - 23-NOU-1995

measures can be obtained for commonly available 140 LEGEND:

probe-microphone equipment. Gain can be veri- 130

--b.-
6
AUG. NORMAL
THRESHOLD
LDL
120
.. LTASS
110 LTASS-12
-LTASS-18
DSL Zcc Coupler Information 100 . AMP. LTASS (T)
1 Ear: LEFT 90
. RESR-90 (T)
DE JOMGE, JAMES Hearing Aid: BTE m
so
z50 see 750 ieee 200 0 300 1000 6000 ..............

70
HA_2 Coupler
J
60

USER GAIM a 50

OSPL r 40

30

I 20
AMY key to continue.... [FI]4HELP
10

-10 n
250 500 1000 2000 4000 8000 = ADD
FREQUENCY (H.) - = REROUE
P = PRINIT
If LDLs hane been measured, please consult HELP message [FP] S PACE = EXI T

Figure 5. Target 2-cc coupler gain and OSPL90 recom- Figure 7. Unaided results for the subject data described
mended by DSLv3.1 for the patient in Figure 3. in Figures 3 to 6.
96 Trends in Amplification

AIDED - DE JONGE, JAMES - LEFT - 23-NOU-1995

versity holds the copyright. The program disk is
140 LEGEND: distributed free of charge and contains a compre-
130

120
--
a
THRESHMD
LDL hensive user manual that can be printed. An over-
110
.. AMP. LTASS (M>
AIW. LTASS (M)-12
AMP. LTASS (M-18
view of the IHAFF protocol can be found in the
manual or from Van Vliet (1995). The version de-
_-
A XA
1W0 . AMP. LTASS (T>
. RESR-90 (T)
t41 90 ------ FtESR-90 (M)
scribed in this paper was [Link] and the software is
W 70 ,
still considered to be under development; i.e., new
J
Jr
60

50
features will be added and modifications will be
40
made. The existing software runs quite well, has
a 30 an attractive screen layout, and is easy to use.
20

10
Help screens are available and commands are
0
chosen via pull-down menus which can be ac-
-10
2 50 5 00FREQUENICY
1000 2 000
(Hz.>)
4000 8oo
00 = ADD
cessed either by keyboard or a mouse. Data files
- = RE11OVE
SPACE
P =
=
PRINT
EXI T can be created, stored, loaded and most of the es-
sential information can be printed. Some graphics
Figure 8. Aided results for the subject data described in
Figures 3 to 6. screens cannot be printed without a third party
program (such as that which comes with the Fig6
software).
to describe the fitting to the parents. First, normal Perhaps the most unique feature of the IHAFF
thresholds, the thresholds associated with the protocol is that it addresses the issue of targets for
hearing loss, and LDL are displayed. This defines hearing aids that have frequency responses that
the LTLR. By adding the LTASS, it can be seen vary as a function of input level (i.e., frequency-
what part of the speech spectrum is audible and dependent compression hearing aids). Prior to
what is inaudible. Adding the amplified LTASS saturation, the performance of linear hearing aids
and RESR targets can then portray the effect the can be described by one frequency response.
hearing aid should have on the speech spectrum. Compression aids, especially those with circuitry
Figure 8 then shows how the measured speech that is activated at low thresholds (kneepoints)
spectrum actually fits into the LTLR, and how the must be described by a family of frequency re-
measured RESR compares to the target. Overall, sponse curves. The target frequency response for
this seems like a convenient tool for taking a any hearing aid wearer may be different for soft,
structured approach to expressing to the parents moderate, or high SPL listening levels. Another
the purpose, and success (or failure), of the hear- way of viewing the issue is that, for any given fre-
ing aid fitting. quency, what is needed is a target input-output
function that matches the needs of the listener.
The Independent Hearing Aid Fitting Forum
Generally, IHAFF is based upon the guideline
that, as with most procedures, speech should be
(IHAFF) Protocol audible, but not reach LDL. Also, the aided loud-
Dissatisfied by the lack of procedures and guide- ness experience should be comparable to loud-
lines for fitting more technologically complex hear- ness experienced by normal hearing individuals.
ing aids (i.e., two channel compression, program- Speech that normally sounds soft should, when
mable), the IHAFF group met in March of 1993 amplified, also sound soft. Intense speech that is
to begin the process of developing a comprehen- normally experienced as loud, should evoke a
sive set of protocols for selecting and fitting hear- similar sensation in the hearing aid wearer. The
ing aids. The IHAFF group is composed of twelve Contour Test and VIOLA (Visual Input-Output
notables in the hearing aid field: Lu Beck, Gail Locator Algorithm) are responsible for convert-
Gudmundsen, Gus Mueller, Ruth Bentler, David ing this theoretical concept into the reality of a 2-
Hawkins, Larry Revit, Robyn Cox, Mead Killion, cc coupler input-output function for a given hear-
Michael Valente, David Fabry, Michael Marion, ing aid.
and Dennis Van Vliet. The result of their effforts The first step in the process is illustrated in Fig-
is a set of software tools that run on an entry level ure 9 where threshold information is entered.
PC such as those described for the DSL and MSU Here values have only been entered for the left
procedures. Robyn Cox and associates at the Uni- ear. Thresholds must be entered for insert ear-
versity of Memphis seem primarily responsible phones (like the ER-3A) which have been cali-
for guiding software development, and this uni- brated on an HA-1 coupler. To allow the program
R. de Jonge-Microcomputer Applications 97

COITRR Stimulus Test Print Graph Main Menu DIOLA Confiumo Grnsh Print Main Menu

THR Audiometer Correction (CORR) Left Ear US SS C CSL L UCL

Left RhtLeft Rih [1I250 Hz -
250 Hz 250 Hz EI_ [X] 500 Hz *- Enter
590
100
Hz
Hz 100 Hz
50 Hz_Resp. TineC(s) I I
Hz
I I 1000 Hz -a names in dB HL
20
300
Hz
Hz
0Z0 Hz
300 Hz
EX]
300
I[14000
Hz
Hz m m
400 Hz 100 Hz
Right Ear UsIS S C CSL L LiCL
NOTE: CORR = SPL(HA-1) - DIAL [ ] 250 Hz -
[XI 500 Hz -
THR units enterHed in: I I 1ONHz
()dB HL (ER3fA) I Z11 Hz
()dB SPL (HA4-1 coupler) [X] 30 Hz
I I 40 Hz-

Fl Help Fl Help

Figure 9. Threshold and audiometer calibration infor- Figure 10. Contour data entered for 500 and 3000 Hz
mation is entered for the IHAFF procedure. for the left ear.

to accurately convert dB HL to coupler SPL, cali- categories 1-7. This information can be gathered
bration corrections can also be entered. It is be- manually, but IHAFF contains software to drive a
lieved that the use of insert earphone is preferra- set of common, commercially available audiome-
ble to conventional earphones since the former ters equipped with a serial interface. The com-
can be calibrated directly on a 2-cc coupler, the puter controls the presentation of stimuli by the
same device used to measure hearing aids. The audiometer, the operator enters the category
next step is to perform the Contour Test which number of the subject's response, and the soft-
utilizes a categorical scaling procedure to measure ware computes the median SPLs. Figure 10 shows
the individual subject's loudness perceptions. It is the median SPL data, entered manually, for each
assumed that individual variability is too great to loudness category. Figure 11 portrays this infor-
use group data on loudness perceptions to predict mation graphically and is as close as IHAFF
the performance of the individual, at least with an comes to giving a target frequency response. The
acceptably low amount of error. It is also recog- next step in the procedure involves relating this
nized that performing loudness measurements loudness data to the perception of speech and cre-
can be very time consuming, especially if a great ating a target to which a hearing aid's input-out-
number of frequencies are included. As a com- put function can be matched.
promise (about 1.5 hours estimated for the entire Three speech spectra, associated with differing
evaluation) loudness measurements are recom- levels of vocal effort, are used to create targets for
mended for two frequencies, i.e., a low frequency so-called soft, average, and loud speech. The
and a high frequency such as 500 and 3000 Hz. overall level of the speech is 50, 65 and 85 dB
These two frequencies would typically be a good
choice for a two-channel programmable aid with a
crossover frequency of about 1500 Hz. Left Ear
Each frequency is tested by presenting four as- 120

cending series of pulsed warble tones (narrow 110

I

band noise can also be used if the audiogram is 100

not very steeply sloping) encompassing the lis- 90

tener's dynamic range. Step size can be 5, 2.5, or 2 a 80

dB. Smaller step sizes are preferred for smaller U
70
7 - U.- f.

dynamic ranges. After giving the listener standard X 0

40
6
5
-

-
L..d b. t OX
C-f/L..d
instructions, the test begins. At each presentation 60
4 - C-f.,tbl.

level the patient chooses which loudness category 20

3

2
-

-
C..f/[Link]
[Link]
best describes the stimulus: 0) inaudible, 1) very I - V."j [Link]

soft, 2) soft, 3) comfortable, but slightly soft, 4) 10

.0 - Th,..[Link]

P .... EPgD.3
comfortable, 5) comfortable, but slightly loud, 6) 250 500 1000 2000 3000 4000
f" E.,
Right

loud, but OK, and 7) uncomfortably loud (Hawk- FREQUENCY (H.)

ins et al, 1987). The data is evaluated to determine Figure 11. Threshold and equal loudness contours for
the median SPLs corresponding to each of the the data shown in Figure 10.
98 Trends in Amplification

SPL, respectively. SPLs have been measured for

1/3 octave bands of each speech spectrum. Nor-
mal hearing subjects have rated the loudness of
these bands, and these judgments have been
equated to the loudness categories associated
with warble tones used in the contour test (Cox et 120 ,X
L O UD
al, 1994a). Categories 1-3 were collapsed to define J 110

the soft range, 3-5 define the comfortable range,

and categories 5-7 represent the loud range. Con-
sequently, it is known where in the normal dy-
.~~~~~~~~~OF
namic range speech falls. For example, a 2000 Hz X : o: :
o
X
:to
t~~~~~~~~~~~~~~~~~~~F9 3
:50
warble tone judged equally loud as a 2000 Hz 1/3 40 -o
1 : //
1 z .o
y re~~~~~~~~~~d...
r9a~~~~~~~~~~g,phs

octave band of average speech lies at 82% of the 30 ,. ., .

o: : :
o t , .
40 50 60 70 80 90 100110120o
X
o
S
..
40 50 60 708 010102
~~~~~~~~~~~P....
[~~~~~~~~~~~PgDn
fo
ight
]

distance bounding the soft range. The goal of the IN*PUT dB SPL e HA MIC Ear

hearing aid fitting would be to amplify this same Figure 12. VIOLA graph showing possible choices for
speech signal to the 82% level of the soft range compression parameters defining the input-output func-
measured for the hearing-impaired listener. For- tion needed to hit the targets based on the loudness
tunately, algorithms have been developed for ex- judgment shown in Figures 10 and 11.
pediting this procedure and are incorporated into
the VIOLA section of the IHAFF suite (Cox et
al, 1994b). The final section of the IHAFF suite is a self-
Figure 12 shows the VIOLA screen for the assessment scale, the Abbreviated Profile of Hear-
contour data plotted in Figure 11. Two graphs are ing Aid Benefit (APHAB) as described by Cox
shown for the frequencies at which loudness data and Alexander (1995). The APHAB consists of
were generated, 500 and 3000 Hz. Output SPL in 24 items drawn from the 66-item profile of hear-
a 2-cc coupler is plotted as a function of the SPL ing aid benefit (PHAB) developed by Cox et al
input to the hearing aid microphone. Corrections (1991) as a research tool. This shortened version
are made for type of hearing aid (BTE, ITE, ITC, takes approximately 10 minutes to administer and
or CIC). Targets are shown as asterisks represent- the results can be used to determine how well an
ing levels needed to normalize the loudness of individual perceives that he or she is doing with
soft, average and loud speech. The shaded areas and without the hearing aid, thereby deriving an
represent the soft (1-3), comfortable (3-5), and estimate of benefit. There are six questions re-
loud (5-7) range of the Contour Test. At the top lated to each of four areas: ease of communica-
of Figure 12 are a series of boxes where the user tion (EC), reverberation (RV), background noise
can enter parameters which describe the input- (BN), and aversiveness of sounds (AV). The first
output function of the hearing aid: gain at 40 dB three scales evaluate speech communication in
SPL, two kneepoints, two compression ratios, and quiet, reverberant, and noisy situations. The last
maximum output. The user adjusts these parame- scale provides input on perception of intense
ters until there is an acceptable match between sounds. Responses to the APHAB have been evalu-
the input-output function and the targets. Also, ated on a group of 90 men and 38 women with mild-
the aid should saturate at a level below LDL (the to-moderate, gradually falling hearing loss who
top of the loud range). After the ideal parameters wore linear hearing aids. The mean age was 68
are selected, the program's work is finished. It is years, the range from 30 to 87 years. A similar
the dispenser's job to find a hearing aid with spec- group of 55 successful (defined as hearing aid use
ifications that match. The hearing aid should be of at least 4 hours per day for at least one year)
adjusted in a test box until the input-output func- wearers of linear amplification were also adminis-
tion matches the target. If the aid were to be tered the APHAB questions (Cox and Rivera,
placed on an average ear, there should be a close 1992). Individual performance on the APHAB
approximation to what is needed to normalize can be compared to this reference group. Gener-
loudness for the three speech signals. However, ally, the APHAB has been shown to be a reliable
since IHAFF currently has no provision for enter- instrument.
ing custom REUR or RECD data, there may The client can take the APHAB by entering
need to be some final adjustments when the aid is responses to the questions via the keyboard (Fig-
placed on the individual's ear. ure 13). A pencil and paper version can be used
R. de Jonge-Microcomputer Applications 99

Subscale Scores (AIDED) Respono ise [Link]

APHAB Questionnaire Options Main Mlenu tloo _ a°
o90A
El] Ln in0 0| t-
070 F
60 C
ABBREVIATED PROFILE OF HEARHIG AID BENEFIT -- FORN A ~50 [Link] 2
Please select the answer that comes closest to your cueryday 40 XWs- E
experience. If you haue not experienced a particular situation,
imagine how you would respond In a similar situation. - _ 0 111.44
EC RV BN au l.b.-
1. When I am in a crowded grocery store, talking with the Soubscalles
cashier, I can follow the conuersation. -20th Pctl 0 - 80th Pctl S -JAMIE
Responses us ing Forn A
G

Without My Hearing Aid With Ny Hearing Aid

A Always (99z) A Always (99Y)
w1
E x XX x1- 1-x" X x
B Almost Always (87z) B Almost Always (87z) Q1
D
\;// XC/v \,/ \;

C Generally (75z) => C Generally (75z)

D Half-the-time (5&/a) D Half-the-time (50z)
=> i l_.11RE E Occasionally (25z) N

F Seldom (12,.) F Seldom (1Z<) 10 ZO

I tens
G Neuer (1z) G Neuer (117) Press ..q keg to see next graph or press <Es.> to ..it

F1 Help Figure 15. Display show aided results of the APHAB.

Figure 13. Screen showing first question of the
APHAB when the client enters data.
tion, it seemed appropriate to discuss it within the
context of IHAFF. RAB is similar in concept to
for individuals who cannot use the computer, and VIOLA in that an attempt is made to restore nor-
the clinician can then enter the responses. Once mal loudness perception (Ricketts, 1996). Re-
all responses are entered, three graphs are gener- search conducted in the development of VIOLA
ated. The graph in the upper left corner of Figure (Cox et al, 1994a) suggests that corrections need
14 summarizes responses for the four subscales to be applied for differences in loudness percep-
during unaided listening. The client's responses tion associated with differences in the bandwidth
are compared to the 20% and 80% percentiles for of speech and warble tones. Ricketts and Bentler
the successful hearing aid wearers. A similar (in press), however, have found that these differ-
graph is generated for aided listening (Figure 15), ences do not exist provided that the overall SPL
and the resulting benefit is shown in Figure 16. of the stimuli are the same and that the bandwidth
From a statistical standpoint, the 90% critical dif- of the stimuli are less than the critical band, as is
ferences for benefit are 26% (EC), 28% (RV), the case with 1/3 octave bands of speech and war-
27% (BN), and 31% (AV) as reported by Cox ble tones. RAB estimates hearing aid parameters
and Alexander (1995). Consequently, the results in two frequency regions, i.e., such as those associ-
shown in Figure 16 can be interpreted to indicate ated with a two channel hearing aid. Gain in the
statistically significant benefit on the EC, RV, and linear portion of the input-output function is:
BN scales. When used in this way, the APHAB
can be a useful tool for documenting hearing aid
(Si - Sn) - (WBf - NBf)
Linear Gain =

benefit. where Si and Sn are the differences, respectively,

Ricketts and Bentler (RAB) have developed in signal level needed to elicit a "soft" loudness
software which builds upon the IHAFF approach category rating for the impaired ear and the nor-
and extends its usefulness; RAB provides level- mal ear. WBf and NBf are SPLs needed to elicit a
dependent frequency response targets. Although soft rating for wide band and narrow band stimuli
the RAB software was not available for evalua- at the frequency tested. The quantity WBf NBf -

S ubsca le S cores <(UNA IDED ) Response Distribut io0 S ubsca le S co re.

,/////-8A I
2

BtRi- C 7

H - F 4
G 4
_
omum toN

-20th PctlIU 80th Pctl1 S - JAMIE

Responses us ing F.,. A
G X X X
EC RV am auXX
ki E * - IcDED SubscalesU - -060
UNIDED

| EC HU BN AwU -8 0
00
* 41 .7 7 7 .0 87 .0 4. 7 B - BENEF I T
* 14.2 41.7 43.7 20.7 -10 B
EC RV AU

B 27.5 35.3 43.3 -16.0 Subscales

Press ..y kox to noot graph or Press <Esc> to ..it

Figure 16. Display comparing the unaided and unaided

APHAB. results (benefit) of the APHAB.
 .2s_X1onrtwe~3:ugpl.
100 Trends in Amplification

represents a reduction in gain needed to compen- cell population and virtually complete destruction
sate for loudness summation. Gain at other fre- of the outer hair cells. This type of hearing loss re-
quencies is estimated using a procedure similar to quires no gain for more intense signals (80 dB HL
the revised National Acoustic Laboratories' NAL-R and above) because recruitment is complete and
(Byrne and Dillon, 1986), adjusted by the mea- loudness perception is normal. There is a loss in
sured gain. This is an improvement over the loudness for faint signals and gain is needed to
IHAFF protocol, which provides no target fre- compensate for their lack of audibility. A type II
quency response other than the two measured fre- loss involves destruction of all outer hair cells and
quencies. Compression thresholds are fixed at 48 a proportion of the inner hair cells, too. Conse-
dB SPL in the low frequency band and 42 dB SPL quently, with a loss of this magnitude, say 60 dB
in the high frequency band. Compression ratios HL, recruitment will not be complete, even for in-
are calculated by the slope of the regression line tense sound exceeding 90 dB HL. However, less
fitting the targets for soft, comfortable, and loud gain will be needed at higher levels, than at lower
as would be displayed on the input-output func- input SPLs, to compensate for loss in loudness.
tion for VIOLA. Type III hearing losses are greater than 70 dB HL
and require the patient to listen at levels close to
Figure 6 (Fig6) Program the LDL in order to understand speech. The Fig6
manual cautions the user to consider a low distor-
Fig6 is a spreadsheet approach, delivered as an tion aid with compression limiting (variable re-
executable program. It is designed to calculate in- lease time) for hearing losses exceeding 70 dB
sertion and coupler gain frequency response tar- HL. Figure 6 (from which Fig6 gets its name) of
gets for three input levels (40, 65, and 95 dB SPL). the Killion and Fikret-Pasa (1993) article is a plot
These targets are calculated from the audiogram of required gain for normal loudness perception
thresholds alone, and the calculations are based as a function of hearing loss. The graph contains
on empirical research involving loudness percep- three curves; the parameter of the curves is the in-
tion in groups with varying degrees of sensorineu- put level to the hearing aid: 40,65, and 95 dB SPL.
ral hearing loss (Hellman and Meiselman, 1993; From the graph it is possible to read target inser-
Lyregaard, 1988; Pascoe, 1988; Lippman et al, tion gain needed to normalize loudness for soft,
1981). The program comes with an 18-page user moderate and intense inputs, for differing magni-
manual and a number of reprints of articles. Col- tudes of hearing loss. The Fig6 program automati-
lectively, this material thoroughly describes the cally calculates these targets and uses the appro-
theory behind Fig6, formulae for the calculations, priate CORFIGs (Table 2 for a definition of
and how to install and use the software. The pro- CORFIG) to convert insertion gain to coupler
gram and related files come on a single 3.5" gain for BTE, ITE, ITC, or CIC instruments. Fig-
floppy disk, and they can be automatically in- ure 17 shows the data entry screen, and a moder-
stalled on a hard drive by executing a batch file ate hearing loss has been entered for the right ear.
called "Setup." Fig6 also contains an online help After each threshold has been entered, Fig6 auto-
facility. Patient files, which consist of mostly the matically calculates the target insertion gain for
patient's name and audiogram, can be saved. All that frequency at each of the input levels. Fig6
screens can be printed. A second disk, unrelated
to Fig6, contains the ELCVIL8 program. This
software is based on Villchur's (1973) work and is F I G 6 Rev L

designed to facilitate selecting parameters for fit- 4

Fix-st Na
60
_
40 4 M
_
__
p&lnu
- p *nu -
-
1eLT+F1
EALT+G3
ting the ReSound programmable hearing aid. The :" iSPC14
t
ELCVIL8 program has similar hardware require- FT te
Fil Na.
H.-
fB
:G o
on
n-t
.^HIF
Qui t - 1LTQ:
- E F13

ments as Fig6. rPUDIOGRAM ESTIMATED GAIN

Killion and Fikret-Pasa (1993) describe three | Freq. RICHT ~LIEFT Freq. I4tU PL NU a SgsL,

types of sensorineural hearing loss, differing

mostly in degree of loss. The Fig6 procedure pro-~
. ~
RIGHTgBBd
~ ~
12oo4@@ HZN
4~ ~ ~ ~ ~
Note ~ ~ ~
9.Cmrsinrtofr4-5u-
41
~ 4
:
L-2
:

vides input level dependent frequency response

targets intended to normalize loudness for type I
and type II hearing loss (Killion, 1996). The type I
-
hearing loss has a magnitude of up to about 40 dB m R.". Wgo 0 EKU

HL and is characterized by a normal inner hair Figure 17. Data entry screen for the Fig6 program.
R. de Jonge-Microcomputer Applications 101

RlG_6 ITE Z.. TARGiE7T,;

HFE413= 25
-FIG6L.,,
G40hf= 37
-
how intelligible amplified speech will be for the
hearing-impaired listener. The SI1 draws upon a
HFE95= a G95hf= 5
considerable body of research validating the tech-
50
C65h,r= 22 nique (Dugal et al, 1980; Pavlovic, 1993; Stude-
baker and Sherbecoe, 1993), and is based upon an
30 older standard describing procedures for calculat-
20
ing the articulation index (ANSI S3.5-1969). Briefly,
both SIl and Al calculations result in a number
ranging from 0 to 1 (or, if expressed as a percent-
age, 0 to 100%). This number is a value estimating
-10 the proportion of the long term speech spectrum
which is audible. Higher numbers are associated
25tR 5fGU logo
Frequency in M.
zoosG 4000s 8001D 16000
with better intelligibility since more of the rele-
- 441.65. 95-dB SPL
vant speech cues are audible. Calculation of the
Figure 18. Target coupler gain for 40, 65, and 95 dB in- Al takes into account a number of factors: the
puts calculated by Fig6. sensation level of the speech relative to an audio-
gram threshold or masked threshold, the band-
also calculates compression ratios for soft-moder- width of the speech, the relative importance of
ate (40-65 dB SPL) and moderate-intense levels different portions of the spectrum to intelligibil-
(65-95 dB SPL) for a low and high frequency band. ity, and upward spread of masking. SI! calcula-
These compression ratios could be used for pro- tions additionally include: self-masking of the
gramming two-channel instruments. Figure 17 also speech spectrum by adjacent speech bands and a
shows the "Graphs" menu from which the audio- correction for distortion created by high speech
gram and various target curves can be displayed. presentation levels, an important issue in amplifi-
Figure 18 illustrates a typical graph showing cation. Since calculation of the SI! is very compu-
coupler gain targets (input levels of 40, 65, and 95 tationally intensive, it is convenient that a com-
dB SPL) for an ITE hearing aid. Graphs for inser- puter program is supplied with the draft standard.
tion gain, or for other instruments (e.g., BTE or The program requires only a PC running DOS 2.1
CIC), are similar in appearance. Figure 18 also and 128 kB of RAM. To calculate the SII, the user
contains a table which provides information which first creates a text file containing three lines. The
should facilitate selecting a matrix. The label first line of the text file contains numerical data
G40hf, which is 37 dB, gives the average high fre- about the speech spectrum level, the second line
quency (hf) gain (G) needed for a 40 dB SPL in- is the equivalent noise spectrum level, and the
put. The frequencies used in the average are 3 and third line is the equivalent threshold level. In the
4 kHz. HFE40 is a slope parameter, indicating following examples, a text file is created specify-
that there should be a difference of 25 dB be- ing average conversational level speech, quiet lis-
tween the hf average gain and the gain at 500 Hz tening (no noise), and a flat 35, 40, and 50 dB HL
for a 40 dB input. Similar information is given for hearing loss. The first text file is named
gain for moderate level signals of 65 dB (G65hf) "[Link]," the name of the program is
and for intense input levels of 95 dB SPL (G95hf). [Link], and both the program and data files are in
An additional slope parameter (HFE95) helps to a directory named "ai." Typing
define the saturation response of the aid. The Fig6
procedure produces results very quickly since au- c:\ai> sii [Link]
0.54
diogram thresholds are the only information en-
tered into the program. The targets represent val- c:\ai> sii [Link]
ues that should be appropriate for the average
0.382
listener, because the formulae were developed c:\ai> sii [Link]
from group trends. Since individuals vary, some 0.081
fine tuning of the hearing aid would be expected. yields the above results, values of 0.54, 0.382, and
0.081 for the SII. The speech intelligibility corre-
Speech Intelligibility Index (Sl1) sponding to these SlIs depends upon specifics
such as characteristics of the speech material (i.e.,
The speech intelligibility index is a proposed word or sentence) and talker (male or female),
ANSI standard which can be useful for predicting and proficiency of the individual. For W-22 mono-
102 Trends in Amplification

syllables, intelligibility would be roughly 95%, Band

Center
Freg dB HL
Coupler
Gain
Noise
REIR Spectrum REUR RECD
Aid Type
160 29 21 25.1 -50 0.3 4.0
80%, and 5% correct for the SIls of 0.54, 0.382, 2 200 29 21 25.0 -50 0.4 4.0 O BTE
i ITE
and 0.081, respectively (Studebaker and Sherbe- 3
4
250
315
29
29
2
22
25.0
25.8
-50
-50
0.5
1.0
4.0
4. 0 ITC
coe, 1993). There are limitations to how well ab- 5
6
400
500
29
29
22
22
25.7
25.6
-50
-50
1.3
1.7
4.0
4.2
REUR
O Free Field
solute intelligibility can be predicted for a particu- 7 630 31 24 27.4
28.2
-50 2.2 4.3 i DiffUSe Field
Speech Spectrum
8 800 32 25 -50 2.9 4.5
lar individual (as opposed to group data). However, 9 1000 34 28 31.3 -50 3.6 5.2 ONormal
10 1250 38 30 33.0 -50 5.3 6.1 i Raised
it is assumed that higher Al or SlI scores predict 1 1600 41 32 33.8 -50 7.2 6.6
0 LoUd
0 Shouted
better intelligibility for the individual. From a 12
13
2000
2500
45
49
34
34
34.6
31.8
-50
-50
10.2
14.9
8.0
9.3 Calculate REIR
EI
practical standpoint, it is quite convenient to use 14
15
3150
4000
52
56
36
37
36.2
41.5
-50
-50
14.4
12.9
10.5
12.2
GoTo-Rudiogram Card *

the Al or SII to quickly predict intelligibility for 16 5000 61 34 42.3 -50 10.8 13.6 ( Colculate Sll )
17 6300 65 32 43.1 -50 6.7 1 4.7
multiple changes in parameters affecting the fre- 186 000 70 30 41.6 -50 8.5 15.0 0-782
slb= (tUSpee)h
quency response. Measuring intelligibility using Example for a mild, sloping hearing loss. Sl= 06.36 t Threshold)
SIId= 0.521 (Both)
traditional techniques is extremely time consum-
ing, especially considering the large number of Figure 20. An example of aided SII calculations.
words that must be given to ensure that results are
statistically significant (Thornton and Raffin, 1978).
To make calculation of the SII more conve- adjustment for high level speech), Sllb includes
nient and more applicable to hearing aid evalua- the high level speech adjustment (and is the calcu-
tion, the author developed a HyperCard stack lation recommended by the standard), SlIc includes
called "SII" to implement the 1/3 octave band a correction proposed by Pavlovic, Studebaker,
method for computing SI. Figure 19 shows the and Sherbecoe (1986) for additional distortion
card used to enter audiogram information. There created by hearing loss, and SlId combines both
are similar cards for entering 2-cc coupler gain, corrections for speech levels and hearing loss.
noise spectrum levels, RECD, and REUR. Data
are entered using the mouse to drag thresholds to Hearing Aid Selection (HAS 2.5)
their desired level. The program will calculate in-
termediate frequencies by interpolating from the The current version of the Hearing Aid Selec-
octave frequencies. Figure 20 shows the results of tion stack is based upon older programs originally
the calculations for the audiogram shown in Fig- created for the Apple II platform (de Jonge, 1985;
ure 19. Noise levels were set to a minimum value de Jonge, 1987). The objective has been to create
(quiet) as recommended by ANSI S3.79. HA-1 a structured, scientifically defensible approach to
coupler gain was modified by corrections for an selecting hearing aid characteristics. The program
ITE microphone location, REUR was set to nor- should incorporate relevant findings from the lit-
mal values for a diffuse field, RECD was normal, erature (i.e., RECD, microphone placement ef-
and speech was produced with increased vocal ef- fects, prescriptive protocols, etc.) and allow the
fort (68 dB SPL). Four SII values are calculated: user to apply them in an easy, intuitive manner.
SSIa is more similar to the original Al (there is no The goal was not to espouse any one philosophy
but to facilitate the application of different ap-
Hz dB
proaches. A primary aim was to introduce gradu-
125 250 500 1000 2000 4000 8000 160 29 ate audiology students to hearing aid selection, to
-10: - 200 29
0 .- 250 29 give them a place to begin. HAS also contains tu-
10 - 315
400
29
29
torial material. The HyperCard stack uses the
20 l-
30 --- -0-013-
-

500 29 Macintosh interface, pull-down menus, buttons,

40 -
50 n 0
630
800
31
32
and so on. On-line help and explanations were in-
60 ___ ___
1000 34 cluded to make the "user manual" internal to the
70 1250 38
00 - 1600 41 program itself, hopefully to improve ease of use.
2000 45 The software and a user manual is available from
100 __ _ ___ 2500
3150
49
52 the Support Syndicate for Audiology (Table 1).
|10 n-rpoaton _
-bot
120 -- 4000
5000
56
61
HAS can be used to create, save, and print patient
6300 65 files. The files contain all demographic, audio-
fbut I nterpo lati~o:n]v 6000 70
logic, and prescriptive information including a
narrative not to exceed 30,000 characters. Target
Figure 19. Audiogram data entry card for the calcu- frequency responses, SSPL curves, etc, can also
lations. be printed on a single sheet of paper as mini-
R. de Jonge-Microcomputer Applications 103

graphs. As with any HyperCard stack, any screen other prescriptive procedures are also available:
can be printed. POGO (McCandless and Lyregaard, 1983); bi-
All information in the program is entered for secting the dynamic range; a 1/3, 1/2, 2/3 gain rule
the fifteen 1/3 octave bands contained in the oc- adopted from Libby (1986); Berger's procedure
taves from 250 to 4000 Hz. The highest frequency (Berger et al, 1979; Berger et al, 1989); DSL 3.1
is about 6000 Hz, corresponding to the high fre- (Seewald et al, 1993b); or the loudness density
quency cut-off of the last 1/3 octave band. The au- normalization (LDN) procedure (Leijon, 1991).
diogram shown in Figure 21 was created by drag- Using LDN in conjuction with different input lev-
ging (with the mouse) the thresholds at octave els, it is possible to generate level dependent tar-
intervals to the desired dB HL calibrated for su- gets, such as those given by DSL [i/o]. The program
pra-aural earphones. The program contains de- also allows for three custom selection procedures
fault real-ear dial differences (REDDs) for Ety- to be used providing the format is that of:
motic ER3-A earphones as well as the TDH-39 REIG = m(T) + b
headphones. Since the individual REDD can be
entered, thresholds are generalizable to any trans- where T is the audiogram threshold, m is a multi-
ducer. To speed data entry, the intermediate fre- plier (e.g., m = 0.5 for a half gain rule) and b is a
quencies were estimated by clicking the "Interpo- frequency dependent correction. Figure 22 illus-
late" button. In a similar manner, MCLs and trates the effect upon the ACL speech spectrum
LDLs can be entered manually, or they can be of the insertion response specified by NAL-R.
predicted. LDLs (labeled "U" for UCL) were cal- The NAL-R REIR is shown by the curve connect-
culated using the formula suggested by Pascoe ing the symbols in Figure 23. Other curves shown
(1988), but LDLs recommended by Seewald et al in Figure 23 were generated using the LDN pro-
(1993b) can also be used. In this example, it is as- cedure to calculate targets for different speech in-
sumed that 65 dB SPL, average conversational put levels (50, 65, and 90 dB SPL, overall level)
level (ACL) speech (as defined by ANSI S3.5- for an ear with complete recruitment. Targets can
1969), will be the input to the listener. The speech also be calculated for partial recruitment or for
level can be raised or lowered covering a range from recruitment based upon LDLs calculated using
45 to 95 dB SPL, or the input level can be adjusted Pascoe's (1988) data. Figure 22 also shows the ef-
manually, for each 1/3 octave band, to whatever fect of the REIR on speech intelligibility using Al
level desired. The effect of the hearing loss on the calculations for the unaided (19%) and aided
speech spectrum can be visualized by drawing the (58%) conditions (ANSI S3.5-1969). Figure 24
audible (lightly shaded) and inaudible (darkly shows the program's display of the relationship
shaded) portion via a menu selection, and the cor- between Al values and speech intelligibility. One
responding Al value is calculated and displayed. curve is for monosyllables and the other is for sen-
The MCL was estimated using the NAL-R tence material. The added contextual cues pro-
(Bvrne and Dillon, 1986) procedure. A variety of vided by sentences make them more intelligible

Gain... SSPL... ather... T 9-3 HAS version2.5

125 250 500 1000 2000 4000 B0oo Set audiogram for:
(Normal Hearing
[ep. 10
IHelp...
10 _ _ ( Interpolate
20
_ t-{Set
D Input Leuel...
40 Set MCL To...

0_|Prescript Methods

1 10 _ Hid UL
1 20 _ = = = = _ s0Hide Input Leuels
Articulation Index values are for speech at: Set UCL By Method:
MCL or Input Level Pascoe _ I
So/ie. speh_ F-9--
Figure 21. A sample screen from the HAS 2.5 program Figure 22. The same audiogram in Figure 21 after the
illustrating audiogram thresholds (0), loudness discom- speech spectrum has been amplified by the REIR speci-
fort levels (U), and the audibility of the speech spec- fied by the NAL-R procedure.
trum.
104 Trends in Amplification

in... I SSPL...I Other...I Curues... ISet Inse specifications (e.g., a matrix). First, the REIR is
r4
converted to an estimated 2-cc use-gain curve
Real-ear Insertion Response
@2

ITE vs. STE

@

RER...
(not shown), and then the use-gain curve is con-
verted to a full-on gain (FOG) curve by specifying
the desired amount of reserve gain. The FOG
curve shown in Figure 25 corresponds to 11 dB of
reserve gain. So, for an ITE aid, this FOG curve
5- will produce a use-gain creating the target REIR
resulting in the previously estimated intelligibility.
At this point it is possible to work the procedure
'!50 500 1000 2000 4000 in reverse: drag the FOG thresholds to, say, corre-
lBTEIE ITC Ie
0 CIC a Uentp cSpecial Earmol spond to a particular 2-cc frequency response
curve. The FOG will then change the use-gain
Figure 23. Target REIR needed to produce the ampli- curve, target REIR curve, and the effect upon the
fied speech spectrum shown in Figure 22. Three addi- speech spectrum and Al values will be redrawn
tonal level-dependent targets are displayed. These tar- on the screen shown in Figure 22. Thus, it is possi-
get were created by the LDN procedure.
ble to visualize the effects of error in selecting a
frequency response. Or, the appropriateness of a
than monosyllables. The data for each curve is patient's hearing aid can be estimated.
stored so that the user can get numerical esti- The LDL curve (Figures 21 and 22) is then
mates of the speech intelligibility corresponding converted to a RESR (not shown, but available in
to the Al values by selecting a menu item from the program). Corrections are applied to adjust
the cards shown in Figures 21 and 22. Figure 24 the real-ear SPL to an HA-1 SSPL90 curve and
also displays critical differences needed to assure the result is given in Figure 26. The largest 2-cc
that two speech recognition scores are not likely coupler ouput corresponding to LDL is estimated
to be due merely to chance. For example, the 58% to be 116 dB SPL at 3000 Hz. It is then at the user's
Al values for amplified speech should create a discretion to decide how close to this curve the out-
monosyllable word recognition score of 82% cor- put should get. Perhaps a matrix of 110/40/20
rect. For another word recogniton score to be sta- would be appropriate given the calculated gain and
tistically different it must be greater than 94% or output. These target responses can be customized
less than 66%, at least for a 50-word list. The pro- for the individual by making adjustments to the
gram also gives estimates for 10, 25, and 100-item default (normal, or average) REUR or RECD
lists (Thornton and Raffin, 1978). curves. Figure 27 gives an example of an adjusted
If one is satisfied with the predicted REIR (the RECD, and changes can be made to the REUR in
REAR curve can be displayed also) and esti- a similar way. The solid curve shows the RECD
mated intelligibility, then the next step is to con- used for an average individual (Sachs and Burkhard,
vert this information to 2-cc coupler terms so that 1972). The symbols were dragged to a higher posi-
a hearing aid can be found from a manufacturer's tion, reflecting an individual with a larger than nor-

and Intelligibility 66%]

_go __0_ X _ t HjC >i~~~~~~~~~~~999Crtical

S e | Di~~~94%
| Critical fference

S@ 540=; [Link] wrd s }t I

60--
720 Co- _c_t_ -

10 20 30 40 50 60 70 80 90 100
Articulation Index Values (Al) to re

Return I ,. ...

Figure 24. The relationship between articulation index Figure [Link] HA-1 2-cc full-on gain needed to give
values and speech intelligibility and critical differences. the REIR shown in Figure 23.
R. de Jonge-Microcomputer Applications 105

Gain SSPL Other I Curues

TEvs BTE 2 cm3 SSPL

11
Hz ORC

31o .......1.;1
15:1.
w . .. . ..............
..... .. t.5... ..........
.... .g7........ihS.;^
794. ......... 1.:1

!..........
I~...6...0...,,,...,.-.2-.....
771 1 ZIWn
A.21
,4W0........22:1.;
2:1
2:1
. .....,2.l...

.........a23.1
L- 1- r--r-T-
500
0 ITC 0 CIC

Figure 26. Target HA-1 2-cc SSPL90 corresponding to Figure 28. Dynamic range compression ratios needed
the LDLs shown in Figures 21 and 22. to map the normal dynamic range into the dynamic
range shown in Figures 21 and 22.
mal RECD, as would be expected for someone
with a smaller ear canal and higher eardrum im- trated in Figure 30. The input-output function
pedance. The program will automatically alter 2- (heavy line) is displayed with the input defined by
cc coupler curves to account for these changes. the location of a reference microphone used dur-
Figure 28 shows target dynamic range compres- ing real-ear measures. The output is referenced to
sion curves and compression ratios needed to map the eardrum. Microphone location corrections are
the dynamic range of the normal ear into the re- being applied for an ITE aid in this example. Mi-
sidual dynamic range of the impaired ear for each crophone corrections are assumed to be level de-
of the 15 bands. The concept behind this display is pendent. The full correction is used when the aid
similar to that described, in the next section, for is operating in the linear range. When the com-
DSL [i/o]. By clicking on the "More Info .. . " but- pressor is activated, the correction is reduced by
ton, the user is taken to the card that provides a the compression ratio (e.g., half as great a correc-
brief tutorial on the topic of compression and com- tion for a 2:1 ratio). No microphone correction is
pression ratios (Figure 29). This card is animated applied when the aid is in saturation. Compres-
to the extent that the user can vary the input level sion threshold (kneepoints), ratio, and type (input
by dragging the rectangle in the normal range, or output) can be defined for two compressors for
and the rectangle in the reduced range will move the currently selected frequency, 1000 Hz in this
according to a 2:1 compression ratio. Figure 29 is example. A target input-output function is given
shown as just one example of many such tutorials. based upon Leijon's loudness density normaliza-
The VIOLA procedure, which is part of the tion procedure (the circles connected by the thin
IHAFF suite, and Killion's Fig6 procedure pro- line). Targets (the heavy plus symbols) are also
vided the motivation for creating the card illus- shown for output levels corresponding to 40, 65,

Gain ..I SSPL. I Other... I Curues ...

Real-ear Coupler Difference 3
(RECD) About the RECD..
0
1.

2
e

E
0 1

1- 1-
250 500 1000 2000 4000
® ITE 0 ITC RIdl CUent 5 Special 3W up~ w thin the dynamic range would bIeuu
Return ~

Figure 27. Card where custom values for the RECD Figure 29. Tutorial information related to dynamic
can be entered. range compression.
106 Trends in Amplification

and 95 dB SPL inputs. These latter targets (shown ..III

.. SSPL
reni
.-I. I ...
.. III ther
n*I.-
. ... .. ...
... I urues...
I .....

as 60, 84, and 104 dB RESPL) are not calculated l I audio4f

Spectrum Level of Noise 4noise
by the program but are assumed to relate to loud- Al
-, Function..
In Ear Canal
-1 . Info
More .. .. ....
Change..
Masked
ness measurements obtained by another means, Audiegram
Thresbold
such as one similar to the IHAFF contour proce- 11 -
23,--
..-ii5
dure. Figure 30 also shows a picture of the mea- P
....G.
_25

surement paradigm, to remind the user of the ref-

erences for input and output. By dragging the ...........

control, indicated by the large triangle under _31

34
29

"(dBin)," the input level increases or decreases. 40

33

_ 42
-1 j --A
Output and headroom numerically change based 250
--- 500
--- 000
1---- 2000
---- 4000
.---
---- .---
43_

upon the gain and SSPL selected. A large dot CO ITE C ) ITIC
f--%
f lTr
%-j f--l lTr
..N_/ . MAM
... IIant M C-i-l
Rid_ M 11--4
I--J~
moves along the input-output function to show
the user the corresponding point on this curve. Figure 31. An example of calculations showing ex-
Gain and SSPL90 can also be changed by drag- pected masking effect of noise.
ging the appropriate triangle. In this example, the
targets are not being matched. The gain should be cles and squares show the Al function for ampli-
increased somewhat. fied speech. "Quiet" means that the Al is calculated
A major concern is how well hearing aids per- for the audiogram thresholds reflecting the hearing
form in noisy situations. HAS provides two tools loss. In this example they are not visible since they
for evaluating the effects of noise on speech intel- are overlaid by the solid squares. "Quiet+Noise"
ligibility. Figure 31 shows a spectrum level of means that the Al is calculated for whichever is
noise measured in the ear canal with a probe mi- larger: the audiogram threshold in quiet or the
crophone system (the Fonix 6500). Since this sys- masked threshold due to noise. "Noise Only" (open
tem displays the overall level in a 100 Hz band- squares) means that the Al is calculated based on
width (BW), subtracting 20 dB from the measured masked thresholds. For this example, this curve
noise converts the measurements to spectrum lev- indicates that noise is not a factor in limiting per-
els (spectrum level = 10 log(BW)). The masked formance. The magnitude of the hearing loss is
audiogram thresholds, calculated from these mea- controlling intelligibility.
surements using critical band theory, are shown in Earmold (or shell) acoustics can have an ap-
Figure 31. The effect of this noise level upon esti- preciable effect upon the frequency response of
mated speech intelligibility is illustrated in Figure the hearing aid. Adjustments can be made to the
32 using Al calculations. The curve indicated by frequency response by selecting from a variety of
the open circles shows how the Al varies as a special earmolds, when a BTE aid is selected.
function of presention level relative to the input Venting can be applied to either BTE, ITE, or
spectrum (which is 65 dB SPL average conversa- ITC styles. Figure 33 shows a typical response for
tional level speech in this example). The filled cir- a 1.0 cm long, 0.3 cm diameter parallel vent. The

Compression Type 1 Input | Gaoin... SSPL...I Other...

Threshold 1 Ratio 1
145 dB 1.13 L-1
Draw 10 HidelShow dB Values < Gain... SSPL... other... 100.
Compression Type 21 Input I Aid ITE * Hz 1000 | Data|
Threshold 2 Ratio 2
185 dB 1-12 1I
Input) Reference Mie OutputQ oCoupler (w1) More Info 90
O Eardrum Unaided Real-Ear Aided

130 Reference
dB re Quiet + Hoise
120- Input srij *O oic Input Unaided Quiet Noise Only e 70J
-40 0 0 0 23
10 40 dB -30 0 1 1 54 60
-20 1 19 19 186
Wdin) 9 49 49 1971
90 BTE 0 20 e1 99'
40,60
4-10 46 97 97 0
80 3 165.841
ITE
+20 70 100 100 100
70I +30 92 100 100 100 30 /
60 Output SPL = 62 dB (dBout) BRuto Calculate 20/
50- dBout - dBin= 14 dB =ain o0
Plot
40-

300
SSPL9O =11 26 d 0 0 Unaided 0
O40 -20 +20
f0*
E *
A 3 Quiet Preseontation Level. re: Input Spectrum
30 40 50 60 70 90 100 110 120 130
Input, dB SSPL90 - dBout = [64 dB -Headroom _Quiet + Noise

P 0lot -0 Noise Only ..j

Figure 30. Sample input-output functions for a set of
compression parameters and the target needed to nor- Figure 32. Articulation index functions predicting the
malize loudness. effect of noise upon speech intelligibility.
R. de Jonge-Microcomputer Applications 107

IGain... SSPL... Other ... I Curues... user-friendly Windows program. A detailed fifty
Length: 0.5 cm Diameter: J0.m eight page spiral bound paper manual accompa-
10.2 am
|. |eVent
11.5
12.0 cm
cm
10.4 am
0.4cm
nies the software, and a very comprehensive on-
line help facility makes the program easy to use
and understand. There is an extensive references
section that integrates background theory with
the more practical aspects of the fitting. This
should appeal to instructors who would like to en-
courage students to study the rationale behind the
fitting procedure. DSL 4.0 has most of the fea-
250 500
---
1000
. ---
2000
- - .--- tures of version 3.1. Since those features have
EITCRid0UentE]S been described in a previous section of this manu-
script, that discussion won't be duplicated here.
Figure 33. Options for parallel vents. The focus will be on improvements or enhance-
ments. Figure 34 shows the "Specification" win-
program would modify the REIR and RESR to dow, and illustrates what a typical DSL 4.0 screen
account for sound bleeding out of the vent and image looks like. This window displays desired 2-
also sound entering the vent. cc coupler SSPL90, gain (10 dB reserve gain),
speech levels, and compression ratios for a mod-
Desired Sensation Level (DSL 4.0 or DSL[i/o]) erate hearing loss. All commands are selected
from the menus shown toward the top of the
The DSL 4.0 software, also called DSL [i/o], screen, just below the title bar. Commonly used
has all the features of version 3.1 but in a more menu items are represented as icons in the button

G DSL 4.0 - File: c:%[Link] 1--l1

File Edit Assess HA Fit Verify Display Window Help
I~~~~~~~~~~~~~~~~~~~~~~l
Transducer : TDH I Circuit T ype: WDRC I Max. Out: Predicted I RE to 2cc: Predicted I Speech: Cox&Moor
HL to SPL: Predicted I HA Style: ITE I Compr. Type: Linear Input DR Extended Compr. Thresh: 1Mi
[DSL 4.0 Icon Help Line - Press the RIGHT MOUSE BUTTON on an icon
Specification
Frequency (Hz)
250 500 750 1000 1500 2000 3000 4000 6000
HAl Coupler
OSPL 94 102 101 102 102 106 106 103 101
R es. Gain
10 Full-On Gain 18 22 22 23 26 32 33 32 31

User Gain (65 dB Tone) 8 12 13 16 22 22

Speech Output 70 75 72 72 74 79 80 79 80

Speech Input 657 55 53 49 4 46 45

Compression Ratio 1.6 16 1.6 1.7 1.8 2.0 2.3 2.5 3.1

Client : James deJonge; Birth Date: 02-jun-1 979; Test Date: 06-May-1 996
Figure 34. The "Specification" window from DSL 4.0 showing target OSPL90, gain, and compression ratios.
108 Trends in Amplification

bar, located just below the menus. Clicking on measured thresholds and estimated RESRs and
these icons has the same effect as making a selec- 2) providing target compression ratios and input-
tion from a menu. The status bar, located just be- output functions. Figure 35 shows the "SPLogram
low the button bar, contains a series of boxes giv- Verification" window for tonal signals measured
ing the setting of each of the parameters used in in the real ear for an ITE aid. Normal thresholds
the DSL algorithms. For example, the box con- are shown at the bottom of the graph as inverted
taining the text, "Speech: Cox & Moore," indi- triangles. The patient's thresholds are given as
cates that a speech spectrum appropriate to adult squares, and the target RESR is the curve at the
listeners has been used for predicting target aided top of the graph. Between threshold and RESR
output levels (Cox and Moore, 1988). The exten- are a series of curves representing target fre-
sion of DSL 4.0 to adult hearing aid fittings is one quency responses for each of the input levels, 45
of the major enhancements to the software. DSL to 90 dB. This display can also be shown for target
4.0 will also allow custom REDDs to be entered, speech levels rather than tones.
and the software supports a variety of transduc- The compression ratio at each frequency is cal-
ers: TDH-style headphones, ER-3A earphones, culated so that the normal dynamic range is com-
and sound field loudspeakers. pressed into the hearing-impaired patient's dy-
In addition to linear hearing aids, DSL 4.0 is par- namic range. The SPL of sound which is barely
ticularly concerned with fitting wide dynamic range audible for the normal hearing listener (normal
compression (WDRC) instruments. The usefulness threshold, Tn) is barely audible for the hearing
of version 3.1 is extended in two important ways: impaired (impaired threshold, Ti), and sounds
1) by calculating a family of input-level-depen- which are at the upper limit of comfort for the
dent target frequency response curves based upon normal (ULn) are amplified to the upper limit of

DSL 4.0 - File: c:%[Link] 1I-W I

File Edit Assess HA Fit Verify Display Window Help
[212e1R~~~~~~~% L46.1V
21ll9EEl b
[ ransd cer: TDH Circuit Type: WDRC ax. editd to 2cc Predicted -Spe_ech : Cox&M ore
HL to SPL: Predicted I HA Style: ITE Compr. Type: Linear Input DR: Extended I Compr. Thresh: Il
rDSL 4.0 Icon Help Line - Press the RIGHT MOUSE BUTTON on an icon
SPLogram - Verification
Verification SPLogram
140-- Thresholds Upper Limits
z Normal 1 iP;edicted.
120- M easured E [Link]
100- r Targets Measured
Real Ear ,.45 dB El-455dB
SPL (dB) 80__-
,El
5d
|e SPLtdB] 80; ~~ - t_60 dB [ 60 dB
l 75 dB E 75dB
60- 90 dB E90 dB
40-

20

l~~ 126 250

_I1
600 1000 2000 4000 8000
Frequency (Hz]

|Client: James deJonge; Birth Date: 02-jun-1 979; Test Date: 06-M ay-1 996
Figure 35. The "SPLogram - Verification" window. Target level-dependent frequency response curves are shown.
R. de Jonge-Microcomputer Applications 109

comfort (ULi) for the hearing-impaired listener the Audioscan hearing aid test system (Etymonic
(Cornelisse et al, 1994). The compression ratio is Design, Inc.) implements a subset of the DSL 4.0
then: program. Audiometric data can be entered and
Compression Ratio = (ULn-Tn)/(ULi-Ti) target values for RESPL are calculated for ACL
speech and also for the RESR. While the Audio-
In this formula, compression thresholds (knee- scan does not calculate target level-dependent
points) are set very low, corresponding to Tn, but frequency response curves, the screen does show
DSL [i/o] permits other compression thresholds amplified ouput levels for soft, average, and loud
to be entered, from these minimum values up to speech. This display is not shown, but it is similar
65 dB. Figure 36 shows a sample input-output to the SPLograms that in Figures 7 and 8. The
function for a 2000 Hz tone. An example of a typ- user can easily visualize whether soft sounds are
ical "Help" window is also shown partially overly- above threshold, if ACL speech is hitting the tar-
ing the "Input/Output" window. get frequency response, and whether loud speech
One complication to using hearing aid selec- is below the RESR. The Audioscan does not give
tion/verification software is the difficulty integrat- a target input-output function, but a display is
ing the software's recommendations with actual shown of how well the hearing aid compresses the
hearing aid measurements, either real ear or cou- speech signal into the residual dynamic range.
pler. For example, if a target frequency response These displays can represent the actual real-ear
curve (or input-output function) is calculated, it output for the individual, or the output can be
would be very useful for that target to be dis- simulated from 2-cc coupler measures. The hear-
played on the equipment used to measure the ing aid can be adjusted on the 2-cc coupler, and
hearing aid. The Audioscan helps to overcome then the fitting can be verified on the individual.
some of these difficulties. Software version 2.6 of This is a very convenient approach with patients

II &
. I
Q
I )DSL 4.0 - File: c:%dsl40win1 _n -1 _ .
File Edit Assess HA Fit Verify Display Wind DSL v4.0 for Windows Help
[le Edit Bookmark Help
2131<1 ^|\X 61 AAI 21<l 1 nIl ontentsl S eatc (lckc Hsi
Transducer: TDH I Circuit Type: WDRC I Max. Out: Predicte
HL to SPL: Predicted I HA Style: ITE Compr. Type: Line
jDSL 4.0 Icon Help Line - Press the RIGHT MOUSE BUTTON on DSL[i/o] Algorithm
Input/Output
Input/Output
The gain ,cra e c,omr,,,,j,,,
(and hence output)
of a
wijd,e,-
i.d..y...dyna,,.r?.
niq [Link]
140 ,,[Link])
the level ofhearing aid is dependent upon
the input signal. The DSL[i/o]
algorithm has extended the original DSL
20-
_ algorithm so that the DSL Method can
be applied to the fitting of
100-_ wide-dynamic-range compression
Real Ear I- vi I I I-I hearing aids. The original DSL algorithm
SPL (dB) 80 calculated amplified target levels for
ii;;; l speech - the DSL[i/o] algorithm can
60- calculate the target output level for any
input.
40-
More details
20-
Caution
0 20 Some of the DSL[i/o] parameter settings
0 26 40 60 80 100 1 can have a large effect on the calculated
Input SPL (dB) targets. If you are not certain use the
DSL default settings.
IClient: J ames deJonge; B irth D ate: 02-jun-1 979; T est D ate: 06-M ay-1 996
Figure 36. A sample DSL[i/o] input-output function along with the online help manual and tutorial.
110 Trends in Amplification

who don't readily cooperate, such as children. The computer "understands" the effect of tone
The simulation can be made more accurate by controls and output limiters for the hearing aids
taking into account the patient's RECD and the within its database, and the program will search
REDD. for aids which achieve an acceptable match.
HASP calculates the root-mean-square deviation,
adjusted by Al weights, between the target and
Hearing Aid Selection Program/lProcedure the hearing aid's response. Acceptable matches
(HASP 2.07)
can be ranked for "closeness of fit."
The emphasis of the hearing aid selection pro- Figure 37 shows the results of an automatic
grams previously discussed has been upon identi- search when an ITE hearing aid was requested for
fying appropriate prescriptions. These prescrip- a moderate hearing loss. The required REIG
tions usually account for differences in REIG and (RREG) and required aided sound field thresh-
coupler gain, but often they do not account for olds (RAT) are displayed along with the audio-
venting or earmold acoustics. The HAS program gram (HTL). HASP identified eleven possible
does make predictions concerning venting and hearing aids and displayed a variety of useful in-
earmold effects, but it is still left to the dispenser formation. Vdev estimates the dB deviation of the
to find an actual hearing aid that will meet the volume control from a mid rotation, COF (close-
prescription. Here is one area where HASP de- ness of fit) is the rms deviation in dB of the REIR
parts significantly from the previous approaches from the target response, and FBM (feedback
(Dillon et al, 1992). The program comes with a margin) indicates the number of dB of additional
small core database of hearing aids and, addition- gain available before feedback will occur. HASP
ally, two large databases of hearing aids from can also be used in a manual mode so the dis-
manufacturers (Miracle Ear and Starkey). Once a penser can visualize the effects of modifying the
target prescription is identified, the computer will hearing aid. A particular hearing aid is selected
automatically search the databases to find hearing from the database, its parameters (such as vent-
aids which match. These databases were created ing, dampers, tone control) are defined, and
by the manufacturers so the dispenser is spared HASP displays the expected REIG (EREG) and
the tedious chore of entering this information the amount of deviation from the target (Figure
manually. However, if desired, the dispenser does 38). HASP also has a menu option for "coupler
have the option of including hearing aids from gain mode" so that a prescription can be calcu-
other manufacturers. It is recommended that lated. Here the required coupler gain (RCG)
HASP run on a PC AT compatible computer using needed to match the REIR target is displayed for a
a color monitor and a minimum of 384 kB of RAM particular style of hearing aid, venting, and tubing
memory. The HASP suite of programs and data- options (Figure 39). The RCG is useful for adjust-
base files comes on a single disk and should be in- ing the hearing aid in the test box prior to fitting
stalled on a hard drive. This requires 0.7 MB of the client and performing the real-ear verification.
space. The fifty one page manual is attractively Since HASP has no mechanism for entering cus-
bound, has an index, references section, descrip- tom REUR or RECD values, real-ear verification
tion of the theory behind the approach, along
with detailed instructions for using the program. HASP - Hearing Aid Selection Procedure
HASP calculates target insertion gain curves Frequency (Hz) 250 500 1000 1500 2000 3000 4000 6000

using the NAL-R procedure (Byrne and Dillon, HTL

BC
.
.
35 40 45 50 55 60 65 70

1986; Byrne et al, 1990). The minimum SSPL90 is AB .........

BREG 1.......
0 0
11
0
22
0
23
0
23
0
2:4
0
25 27
0

selected so that 70 dB SPL speech will not satu- RAT ........ 50 38 Z9 31 34 331 33 51
Selected Aids ---- Audiogram Number:
rate the hearing aid. The maximum SSPL90 MPO Reduction
AidName Tone Steps dB Vent Tube Damperr Oden COF FB11
should not exceed LDL (which is estimated from MEK31E TFC 3 9 Occluded 1 1.9 10
ISTA600/0735 15 Tight 4 1.5
threshold). The target SSPL90 is the midpoint of MEX86E TFC
ISTA500/0235 10
Z 6 Occluded
Occluded
_2
2
1.8
1.9
0
9
these two curves. HASP allows the user to select ISTA8OO00Z35 10 N
ISTA600/0735 15
Occluded
Occluded
2
4
1.9
1.6
9
7
vent size, earmold configurations and dampers for MENZ7E TFC
ISTA800/0740 20 N
1 3 Occluded
Tight
4 1.6
2.2
0

BTE instruments. The computer adjusts the fre- ISTA001/1040

ME49AHSE
3
1 2 4
Occluded
Occluded
-2
4
2.1
1.8
0
6
4 2.0
quency response and SSPL90 for sound entering MEX86E TFC 2 6 Tight
Do You wish to saue AudiograM (Y/N)
and exiting the vent (Dillon, 1991), considers the
occlusion effect, and also can predict the maxi- Figure 37. A sample HASP screen showing the results
mum gain available before feedback will occur. of an automatic search for an ITE aid.
R. de Jonge-Microcomputer Applications ill

HASP - Hearing Aid Selection Procedure HASP - Hearing Aid Selection Procedure
Frequency (Hz) 250 500 1000 1500 2000 3000 4000 6000 Frequency (Hz) 250 500 1000 1500 2000 3000 4000 6000
HTL ........ 35 40 45 50 55 60 65 70 HTL ........ 35 40 45 50 55 60 65 70
BC BC .
0B ......... 0 0 0 0 0 0 0 0 ADB ........ 0 0 0 0 0 0 0 0
RREG ... 1 11 22 23 23 Z4 25 27 ROEG ....... 1 11 2Z 23 23 Z4 Z5 27
HAT ........ 50 38 29 31 34 31 33 51 RAT ........ 50 38 Z2 31 34 31 33 51
Press Enter Key to evaluate -- Space to Change Up, Down to noue Press Enter Xey to eualuate -- Space to Change -- Up, Down to none
Uent type Vent type 2 mm
Tube type Mot Used.. Tube type Not used...
Damper No Micr.
Tone ctrl TFC
Manual Selection 1
Audiogran Humber: Required Coupler Gain Selection ---- Audiogran Humber: 1
AidHade Type Tone Vent Damper (Idea
Tube COF FBM AidType Vent Tube
M1EX31E ITE TFC 1 mm No H 1 1.7 8 ITE 2nm ----------
MPO should be 9 dB below maximum
Frequency (Hz) 250 500 1000 1500 2000 3000 4000 6000 Frequency (Hz) 250 500 1000 1500 2000 3000 4000 6000
EREG 1 15 Z1 23 24 24 22 Z1 RC G 19 Z4 31 32 32 35 3Z 25
Deviiation 0 4 -1 -1 1 0 -3 -6
This aid's Power reduction will affect the gain and VDea
Press (G)raph. (P)rintScreen, Space to Continue -- Escape to Quit Press (P)rintScreen, Space to Continue -- Escape to Quit

Figure 38. HASP display illustrating information avail- Figure 39. Coupler gain required to match the target
able when used in the manual mode. REIR.

is an important aspect of the fitting process. RCG of the test box running the hearing aid's input-
is also useful if real-ear measures cannot be per- output function. The same software used to select
formed. ideal characteristics of the hearing aid would also
The database stores relevant information about be used to adjust the aid, as in the case of a pro-
the electroacoustic characteristics of the hearing grammable. Or, for conventional hearing aids, a
aid. The computer, however, has no knowledge direct link to the manufacturer's database would
about issues that are important to the fitting, but be useful for matching a matrix. It would also be
which are difficult to quantify: manipulation skills useful, when trying to match a target coupler gain
of the client, years of hearing aid use, and other curve, to simulate the effect of making an adjust-
special needs that the client may have. The dis- ment to a trim pot. Then, the hearing aid could be
penser can then combine this factual information ordered online, and billing would be facilitated by
with clinical judgment and experience to provide the software. Perhaps the NOAH software plat-
the best instrument. In this way the human is left form, together with cooperation from hearing aid
to do what he or she does best, while the com- manufactures and software developers, will help
puter can handle the chore of calculating correc- make such high level integration a practical reality.
tions and remembering all the details of the hear- While improvements will surely come, a vari-
ing instrument inventory. ety of hearing aid selection programs are cur-
rently available. Each program has its own set of
features, philosophy, and is geared toward differ-
SUMMARY ent aspects of the fitting process. Some of the pro-
An issue, and perhaps a limitation, which is cedures are more abstract and theoretical, others
common to all the software packages reviewed is: are practical and concrete. Certain procedures
How well does the program fit into the dis- emphasize the prediction of hearing aid charac-
penser's practice? Ideally the programs would be teristics from threshold data, while others require
integrated so that, for example, the patient's measurement of loudness perceptions. Hearing
name and address would not have to entered re- aid dispensers also have preferences and a per-
peatedly into different software modules. An au- spective shaped by their personality and experi-
diogram entered once would not have to be dupli- ences. This dictates what they value. Some per-
cated. An audiogram stored in a computer file ceive the hearing aid fitting to be a conceptual
would not also have to be stored on paper. Tym- experience to be researched or taught to students,
panogram information or other diagnostic tests others see it as a social experience shared with the
(e.g., auditory brainstem response, word recogni- patient for whom they care. Others may visualize
tion scores) would be stored within the same pa- it as an integral part of the business they have la-
tient record. Ideally, the software would be part bored to create. Regardless of the perspective,
of the hearing aid test systems so that real ear and each person can find something to appreciate in
coupler measures could be easily compared to the the software that has been reviewed. Powerful, in-
targets. The dots (representing loudness calcula- expensive microcomputers are readily available.
tions) from VIOLA would appear on the screen A minimal amount of time (less than a few min-
112 Trends in Amplification

utes) is often needed to enter data and receive Cornelisse LE, Gagn6 JP, Seewald RC. (1991). Ear
prescriptions that incorporate decades of re- level recordings of the long-term average spectrum
search. Patients have more confidence in the of speech. Ear Hear 12:47-64.
opinions of their dispensers when they can see the Cornelisse LE, Seewald RC, Jamieson DG. (1994).
reasoning that went into the recommendations. Wide-dynamic-range compression hearing aids: The
Allowing patients to watch the computer screen, DSL[i/o] approach. Hear J 47(10):23-29.
Cox RM. (1983). Using ULCL measures to find fre-
even though they cannot appreciate most of the quency/gain and SSPL90. Hear Instr 34(7):17-21,39.
details, is an engaging process and allows them to Cox RM. (1984). Relationship between aided preferred
feel they are an active participant. Hopefully, listening level and long-term listening range. Ear
readers will be encouraged to try some of these Hear 5:72-76.
software tools to add more science and objectivity Cox RM. (1985). A structured approach to hearing aid
to the art of hearing aid fitting. selection. Ear Hear 6:226-239.
Cox RM. (1988). The MSU hearing instrument proce-
dure. Hear Instr 39(1):6-10.
Cox RM. (1989). Comfortable loudness level: Stimulus
ACKNOWLEDGEMENTS effects, long-term reliability, and predictability. J
I would like to thank Robyn Cox, Richard See- Speech Hear Res 32:816-828.
wald, Mead Killion, Toni Gitles, and Harvey Dil- Cox RM, Alexander GC. (1990). Evaluation of an in-
lon for sending software to assist with the review. situ output probe-microphone method for hearing
Ruth Bentler provided details concerning the re- aid fitting verification. Ear Hear 11:31-39.
Cox RM, Alexander GC. (1995). The abbreviated pro-
search she and Todd Ricketts undertook in creat- file of hearing aid benefit. Ear Hear 16:176-183.
ing the RAB procedure. Dennis Van Vliet clari- Cox RM, Alexander GC, Gilmore C. (1991). Objective
fied some details relating to IHAFF. Larry and self-report measures of hearing aid benefit. In:
Humes and Joe Smaldino helped to make me Studebaker GA, Bess FH, Beck L, eds. The Vander-
aware of the availability and changes in software. bilt Hearing Aid Report II. Parkton, MD: York
Press, 201-213.
Cox RM, Bisset JD. (1982). Prediction of preferred lis-
REFERENCES tening levels for hearing aid gain prescription. Ear
Hear 3:66-71.
American National Standards Institute. (1969). Ameri- Cox RM, Goff CM, Martin SE, McCloud LL. (1994a).
can National Standard Methods for Calculation of The contour test: Normative data. Presented at the
the Articulation Index. (ANSI S3.5-1969). New York: American Academy of Audiology Meeting, Rich-
ANSI. mond, VA, 1994.
American National Standards Institute. (Draft). Ameri- Cox RM, Moore JN. (1988). Composite speech spec-
can National Standard Methods for The Calculation trum for hearing aid gain prescriptions. J Speech
of the Speech Intelligibility Index. (ANSI S3.79- Hear Res 31:102-107.
Draft). New York: ANSI. Cox RM, Rivera IM. (1992). Predictability and reliabil-
Bentler RA, Pavlovic CV. (1989). Transfer functions ity of hearing aid benefit measured using the PHAB.
and correction factors used in hearing aid evaluation J Am Acad Audiol 3:242-254.
and research. Ear Hear 10:58-63. Cox RM, Taylor IM, Gray GA, Brainerd LE. (1994b).
Bentler RA, Pavlovic CV. (1992). Addendum to "Trans- The contour test: Applications to hearing aid selection
fer functions and correction factors used in hearing and fitting. Presented at the American Academy of
aid evaluation and research." Ear Hear 13:284-286. Audiology Meeting, Richmond, VA, 1994.
Berger KW, Hagberg EN, Rane RL. (1979). Determin- de Jonge RR. (1985). A microcomputer-based hearing
ing hearing aid gain. Hear Instr 30:26-28,44. aid selection strategy. J Comp Users Speech Hear
Berger KW, Hagberg EN, Rane RL. (1989). Prescrip- 1:117-138.
tion of Hearing Aids: Rationale, Procedures and Re- de Jonge RR. (1987). Hearing aid selection using a mi-
sults, ed 5. Kent, OH: Herald. crocomputer. Hear Instr 38(5):126-130.
Byrne D, Dillon H. (1986). The national acoustic labo- de Jonge RR. (1996). Real-ear measures: Individual
ratories new procedure for selecting the gain and variation and measurement error. In: Valente M, ed.
frequency response of a hearing aid. Ear Hear 7:257- Hearing Aids: Standards, Options, and Limitations.
265. New York: Thieme Medical Publishers, 72-125.
Byrne D, Parkinson D, Newall P. (1990). Hearing aid gain Dillon H. (1985). Earmolds and high frequency re-
and frequency response requirements of the severely/ sponse modification. Hear Instr 36(12):8,11-12.
profoundly hearing-impaired. Ear Hear 11:40-49. Dillon H. (1991). Allowing for real ear venting effects
Carhart R. (1946). Tests for the selection of hearing when selecting the coupler gain of hearing aids. Ear
aids. Laryngoscope 56:780-794. Hear 12:406-416.
R. de Jonge-Microcomputer Applications 113

Dillon H, Byrne D, Battaglia J. (1992). Hearing instru- Pascoe DP. (1988). Clinical measurements of the audi-
ment selection: By dispenser or by computer?. Hear tory dynamic range and their relation to formulas for
Instr 43(8):18-21. hearing aid gain. In: Jensen JH, ed. Hearing Aid Fit-
Dugal RL, Braida LD, Durlach NI. (1980). Implications ting: Proceedings of the 13th Danavox Symposium.
of previous research for the selection of frequency- Copenhagen: Danavox, 129-152.
gain characteristics. In: Studebaker GA, Hochberg I, Pavlovic CV. (1993). Problems in the prediction of
eds. Acoustical Factors Affecting Hearing Aid Per- speech recognition performance of normal-hearing
formance. Baltimore: University Park Press, 379-403. and hearing-impaired individuals. In: Studebaker
Fabry DA, Schum DJ. (1994). The role of subjective GA, Hochberg I, eds. Acoustical Factors Affecting
measurement in hearing aid fittings. In: Valente M, ed. Hearing Aid Performance, 2nd ed. Boston: Allyn
Strategies for Selecting and Verifying Hearing Aid Fit- and Bacon, 221-234.
tings. New York: Thieme Medical Publishers, 136-155. Pavlovic CV, Studebaker GA, Sherbecoe RL. (1986).
Hawkins DB, Morrison TM, Halligan PLW, Cooper An articulation index based procedure for predicting
WA. (1989). Use of probe tube microphone mea- the speech recognition performance of hearing-im-
surements in hearing aid selection for children: Some paired individuals. J Acoust Soc Am 80:50-57.
initial clinical experiences. Ear Hear 10:281-287. Pearsons KS, Bennett RL, Fidell S. (1976). Speech lev-
Hawkins DM, Walden BE, Montgomery AA, Prosek els in various environments. BBN Rep. No. 3281
RA. (1987). Description and validation of an LDL (Bolt, Beranek, and Newman).
procedure designed to select SSPL90. Ear Hear Revit L. (1994). Using coupler tests in the fitting of
8:162-169. hearing aids. In: ed. Strategies for Selecting and Veri-
Hellman RP, Meiselman CH. (1993). Rate of loudness fying Hearing Aid Fittings. New York: Thieme Med-
growth for pure tones in normal and impaired hear- ical Publishers, 64-107.
ing. JAcoust Soc Am 93:966-975. Ricketts TA, Bentler RA. (In Press). The effect of test
Kamm C, Dirks D, Mickey MR. (1978). Effect of senso- signal type and bandwidth on the categorical scaling
rineural hearing loss on loudness discomfort level (CS) of loudness. J Acoust Soc Am.
and most comfortable loudness judgments. J Speech Ricketts TA. (1996). Fitting hearing aids to individual
Hear Res 21:668-681. loudness-perception measures. Ear Hear 17:124-131.
Killion MC. (1996). Talking hair cells: What they have Robertson PR. (1996). A guide to NAOAH-compatible
to say about hearing aids. In: Berlin CI, ed. Hair programmable fitting software. Hear Rev 3(2):12.14,
Cells and Hearing Aids. San Diego: Singular Press, 16,19,28 30.
125-172. Sachs RM, Burkhard MD. (1972). Zwislocki couipler
Killion MC, Fikret-Pasa S. (1993). The 3 types of senso- evaluation with insert earphones. Report No. 20022-
rineural hearing loss: Loudness and intelligibility 1. Franklin Park, IL: Knowles Electronics.
considerations. Hear J 46(11):31-34. Seewald RC. (1992). The desired sensation level method
Killion MC, Revit LJ. (1993). CORFIG and GIFROC: for fitting children: Version 3.0. Hear J 45(4):36-41.
Real ear to coupler and back again. In: Studebaker Seewald RC, Moodie KS, Zelisko DLC. (1993a). Cri-
GA. Hochberg I, eds. Acoustical Factors Affecting tique of current approaches to the selection and fit-
Hearing Aid Performance, 2nd ed. Boston: Allyn ting of hearing aids. J Speech Lang Path Auidiol
and Bacon, 65-85. Monogr. Suppl. 1, 29-37.
Leijon A. (1991). Hearing aid gain for loudness-density Seewald RC, Ramji KV, Sinclair ST, Moodie KS,
normalization in cochlear hearing losses with im- Jamieson DG (1993b). A compluter-assisted imple-
paired frequency resolution. Ear Hear 12:242-250. mentation of the Desired Sensation Level method for
Libby ER. (1986). The 1/3 2/3 insertion gain hearing
- electroacoustic selection and fitting in children: DSL
aid selection guide. Hear Instr 37:27-28. 3.1 luser's manual. The Hearing Health Care Re-
Lippman PR, Braida LD, Durlach NI. (1981). Study of search Unit Technical Report 02. The University of
multichannel amplitude compression and linear am- Western Ontario. London, Ontario.
plification for persons with sensorineural hearing Seewald RC, Ross M. (1988). Amplification for young
loss. J Acoutst Soc Anm 69:524-534. hearing-impaired children. In: Pollack MC. ed. Anm-
Lybarger SF. (1944). US Patent application SN 543.278. plification for the Hearing Impaired. New York:
Lyregaard PE. (1988). POGO and the theory behind. Grune and Stratton, 213-267.
In: Jensen JH, ed. Hearing Aid Fitting: Proceedings Seewald RC. (1994). Fitting children with the DSL
of the 13th Danavox Symposium. Copenhagen: Dan- method. Hear J 47(9):10, 49-51.
avox,81-94. Shaw EAG. (1980). The acoustics of the external ear.
McCandless GA, Lyregaard PE. (1983). Prescription of In: Studebaker GA, Hochberg I, eds. Acouistical Fac-
gain/output (POGO) for hearing aids. Hear Instr tors Affecting Hearing Aid Performance. Baltimore:
35(l): 16--21. University Park Press, 109-125.
Moodie KS, Seewald RC, Sinclair ST. (1994). Proce- Shaw EAG. (1974). The external ear. In: Keidel WD.
dure for predicting real-ear hearing aid performance Neff WD, eds. Handbook of Sensory Physiology V/
in young children. Am J Auidiol 3:23-31. 1. Berlin: Springer-Verlag, 455-490.
114 Trends in Amplification

Shore I, Bilger RC, Hirsch I. (1960). Hearing aid evalu- Allyn and Bacon, 185-204.
ation: Reliability of repeated measurement. J Speech Thornton AR, Raffin MJM. (1978). Speech-discrimina-
Hear Dis 25:152-167. tion scores modeled as a binomial variable. J Speech
Studebaker GA, Hochberg I. (1980). Acoustical Factors Hear Res 21:507-518.
Affecting Hearing Aid Performance. Baltimore: Uni- Valente M. (1994). Strategies for Selecting and Verifying
versity Park Press.
Hearing Aid Fittings. New York: Thieme Medical
Studebaker GA, Hochberg I. (1993). Acoustical Factors
Publishers.
Affecting Hearing Aid Performance, 2nd ed. Boston:
Allyn and Bacon. Van Vliet D. (1995). A comprehensive hearing aid fit-
Studebaker GA, Sherbecoe RL. (1993). Frequency- ting protocol. Audiol Today 7:11-13.
importance functions for speech recognition. In: Villchur E. (1973). Signal processing to improve speech
Studebaker GA, Hochberg I, eds. Acoustical Factors intelligibility in perceptive deafness. J Acoust Soc
Affecting Hearing Aid Performance, 2nd ed. Boston: Am 53:1646-1657.