Assessing FM transparency, FM/HA ratio with digital aids · 2020. 9. 11. · 4 Unaided Curve Speechmap/DSL - Single view 140 60 50-audioseem Mal Instrument FM ,.gj Mode Presentation

Assessing FM transparency,FM/HA ratio with digital aidsBy Jane Auriemmo, Denise Keenan, Denise Passerieux, and Francis Kuk

The advantages of frequency modulation (FM) systemsare well recognized.' In an ideal situation, an FM systemthat is connected to a hearing aid (HA) maintains the fre-quency-output characteristics of the hearing aid andenhances the signal-to-noise ratio (SNR) of the listeningenvironment.

That said, a clinician who works with FM and hearingaids must evaluate if the addition of the FM results in theoptimal use of the FM-HA combination. Specifically, theclinician must verify that: (1) the electroacoustics of FMand HA meet specifications (i.e., sufficiently broad band-width, no evidence of distortion, etc); (2) the coupling ofthe FM does not change the frequency response of thehearing aid in the various FM receiver positions (such asHA, FM, FM+HA) forthe same stimulus input(i.e., transparency); and (3)the ratio of the FM outputto the hearing aid output,when used in the FM+HAmode, is favorable.

This last verificationgoal is also known as theSNR advantage of the FMin the FM+HA mode. h isi mportant for cliniciansworking with these twodevices to know how todetermine the FM/HAratio and, when necessary, to adjust the sensitivity of thedevices to achieve the optimal ratio. The implicit expec-tation is that the FM+HA mode will increase wearer sat-isfaction.

Clinicians have sought to standardize the verificationof FM systems when used with hearing aids. However,with the advent of digital signal processing (DSP) hearingaids, the parameters and protocols previously used to ver-ify FMs and HAs may be inadequate. Digital processingcan interact with the type and duration of the stimulusused, resulting in equivocal results. Furthermore, the typeof FM system used, as well as the hearing aid analyzer usedto measure transparency and FM/HA ratio, may affect theoutcome of the verification.

Kuk and Keenan evaluated the impact of stimulus dura-tion and stimulus type on the reliability and validity ofverification. 2 Integrating their findings, we will present inthis article a three-step protocol for measuring both thetransparency and the FM/HA ratio in one measurementsession. Specifically, we will describe the protocol, provide

30 The Hearing Journal

a brief overview of Kuk and Keenan's findings pertinentto the measurements at hand, and demonstrate imple-mentation of the protocol with two commercial hearingaid analyzers and a new FM system.

WHAT NEEDS VERIFICATION?

Transparency

If a wearer's hearing aid has been adjusted for optimal fre-quency response, its use with an FM system should notchange that response. An FM system that does not alter thefrequency-response characteristics of the hearing aid at achosen input level is said to be transparent at that inputlevel.

A transparent FM sys-tem will ensure consistentaudibility, whether thehearing-impaired patientuses the HA alone, theHA with the FM con-nected but in the HA-alone mode, the HA withthe FM connected but inthe FM-alone mode, orthe HA with the FM con-nected in the FM+HAmode—provided thataudibility is assured in theHA-alone mode. Trans-

parency must be verified because real-ear verification ofthe FM/HA combination can be done with the couplerin the simulated mode only as long as real-ear verificationof the HA-alone condition proves sufficient. Althoughmany newer FM systems are transparent, some earlier oneswere not, which meant that the frequency output of thehearing aids coupled to the FMs was significantly differ-ent from their output in the stand-alone position.3

FM/HA ratio

An FM system, when used with a hearing aid in the FM-alone mode, yields an SNR advantage greater than 20 dB.But when it is used in the FM+HA mode, the SNR isgenerally poorer because both the FM transmitter micro-phone and the hearing aid microphone are active. A studyby Hawkins revealed no difference in children's word-recognition scores between the FM+HA mode and theHA-alone mode (i.e., the FM failed to improve the SNR).4The outcome might have been different if the output ofthe FM had been higher than that of the HA.

"... with the advent of digital`signal processing (DSP) hearing

aids, the parameters andprotocols previously used toverify FMs and HAs may be

inadequate..."

A protocol for using FM with DSP hearing aids March 2005 • Vol. 58 • No. 3

It is i mportant for clinicians to havean effective way to verify and adjust therelative gain settings between the FMand the HA if they wish to optimize boththe FM input (e.g., the teacher's voice)and the hearing aid input (e.g., fromclassmates, own voice, environmentalsounds).

The American Speech-Language-Hear-ingAssociation (ASHA) evaluated theissue of optimal FM/HA ratio and rec-ommended a +10-dB SNR advantage ofthe FM output (over the HA output) inan FM+HA mode. 5 For example, the out-put of the FM system from an 80-dB SPLinput signal should be 10 dB more thanthe output of the HA from a 65-dB SPLinput when the FM is in the FM+HAmode. A 65-dB SPL input level is chosento represent the level of conversationalspeech, while the 80-dB SPL signal reflectsthe level of the conversationalspeech of the speaker (i.e., theteacher) as measured at the FMmicrophone in the chest or lapelmic position.

While the +I0-dB advantagemay be a good compromise, itmay not always be achievablewhen the FM is in its defaultposition. Furthermore, theFM/HA wearer may require alower ratio to perceive environ-mental sounds and classmates'voices (from the HA mic) or ahigher ratio to further enhancethe teacher's voice (or attenuateclassmates' voices). Thus, knowing howto verify and adjust the FM/HA ratio willbe beneficial.

THE THREE-STEP PROTOCOLClinicians may follow this protocol toassess both the transparency and theFM/HA ratio of an FM-HA combina-tion in a single measurement session usingonly coupler measurements.v Step I: Obtain a reference. Measure

the output of the hearing aid in theFM+HA mode with a 65-dB SPLspeech or speechlike stimulus pre-sented to the hearing aid for 8 to 10seconds.Step II: Transparency. Measure theoutput of the hearing aid in theFM+HA mode with a 65-dB SPLspeech or speechlike stimulus pre-sented to the FM mic for 8 to 10 sec-

onds. The frequency-output curvesobtained in Steps I and II should besimilar if the FM is transparent.

+ Step HI: FM/HA ratio. Measure theoutput of the hearing aid in theFM+HA mode with an 80-dB SPLspeech or speechlike stimu l uspresented to the FM mic for 8 to 10seconds. The difference in outputbetween Steps III and I defines theFM/HA ratio (or relative output ofthe FM/HA combination).

Basis for the recommendations

Given the definitions of transparency andFM/HA ratio, it is intuitive why the dif-ference between Steps I and II shouldreflect transparency (or the lack thereof)and why the difference between the stepsshould reflect the FM/HA ratio. How-ever, the basis for the specific stimulus

seconds to stabilize. Furthermore, the useof a modulated signal (such as the digitalspeech signals used on the Frye) couldtake as long as 8 seconds to reach a sta-ble, compressed state. For a single proce-dure to be useful for most compressionsystems and most signal types, the selectedstimulus must be longer than 8 seconds.

The action of noise-reduction algo-rithms in many DSP hearing aids can con-found the results in two ways: introducingvariability in the output and reducing theoutput. These algorithms evaluate themodulation characteristics of the inputsignals to decide if additional gain reduc-tion should occur. It is necessary to useeither a modulated signal (speech orspeechlike signals) or a non-modulatedsignal shorter than the activation time ofthe noise-reduction algorithm to preventthe hearing aid from initiating additional

gain reduction. For Widex hear-ing aids with such an algorithm,the duration of a continuous sig-nal must be no longer than 15seconds to avoid the action of thenoise-reduction algorithm.

The action of an active feed-back-cancellation mechanismmust be considered as well. Thesealgorithms examine the input sig-nals over time and identify repet-itive signals or sinusoidal signalsas feedback. The use of speech orspeechlike stimuli or signalsshorter than 10 seconds willbypass the action of these algo-

rithms and result in a reliable output.The action of an adaptive directional

microphone may also affect the outcomeof the evaluation. The azimuth of stimu-lus presentation, the nature of the stim-ulus (continuous versus modulated), andthe duration of the stimulus can all affectthe measured output. This is especiallycritical when measuring the output of thehearing aid alone in its normal mode ofoperation. In such a case, a modulatedsignal of 8-10 seconds presented at thecorrect azimuth yields the most reliableresponse. However, since some hearingaids, including the Widex Senso Diva,change to an omnidirectional mic in theFM+HA mode during the evaluation, nospecial considerations are necessary tobypass this feature during FM evaluation.

These considerations prompted us tochoose a speech or speechlike signal to

"...it is important for clinicians tohave an effective way to verify

and adjust the relative gainsettings between the FM and theHA if they wish to optimize boththe FM input ...and the hearing

aid input..."characteristics (i.e., speech or speechlikestimuli for 8-10 seconds) requires furtherexplanation.

A key difference between today's DSPhearing aids and traditional analog linearhearing aids is the use of non-linear sig-nal processing, which also includes manyadaptive features. Rather than respond.ing instantaneously to the input signalsas in linear processing, non-linear signalprocessing requires time to effect change.

For example, the attack time on a com-pression hearing aid decides how long thehearing aid takes to reach a stable, reducedgain state. The longer the attack time, thelonger it takes the non-linear hearing aidto reach the compressed state. In a fast-acting compression system, it may takeless than a second for the output to sta-bilize, whereas in a slow-acting compres-sion system, it may take as long as 2

32 The Hearing Journal A protocol for using FM with DSP hearing aids March 2005 • Vol. 58 • No. 3

4

UnaidedCurve

Speechmap/DSL - Single view140

60

50-

audioseemMal Instrument FM ,.gj

ModePresentation Single view :pjFormat Graph .C11Scale (dB) SPL :pj

AudiometriAge AdultTransducer insert-WO-W--UCL AverageRECD Average

130 -

120 -

110-

100-

90 -

eo

-70 -

REAR Stimulus Level SII

2j OlSpeech-7emale±Avgj...1_Q_J$Peeeli-Temate fAvg rry.,

(65) 623101Speech-femaleIM Chest 70

40 -

30-

20 -

10

0-

-10-230 5O0 1000 2000 4000 • 8000

Figure 1. (A) Placement of hearing aid in theAudioscan test chamber. The female speechsignal was presented from the left loudspeaker.(B) Output of the Senso Diva with a 65-dBSPL female speech signal averaged over thecourse of 10 seconds with the signal presentedto the hearing aid (in FM+HA).

Connect coupler and instrument to coupler microphone. Select one of REAR I through REAR 4

bypass the noise-reduction and feedbackalgorithms, and a signal duration of 8-10seconds to ensure a stable output.

Implementation of the protocol

The FM system used in this illustrationis the Scola, an FM system recently intro-duced by Widex that is compatible withmost behind-the-ear hearing aids. Likemany other FM systems today, it uses aminiature FM receiver that can be cou-pled to the base of a BTE. However, atransparent audio shoe is used with a slide-on plug instead of the Euro-plug for afirmer attachment to hearing aids. Fur-thermore, the sensitivity of the FM inputcan be adjusted via a screw potentiome-ter on the receiver.

The FM transmitter is a hand-helddevice with several directional-microphoneoptions. During evaluation, the FM trans-mitter's microphone must he set to theomnidirectional mode and it must directlyface the sound outlet of the hearing aidanalyzer for a reliable output. The proto-col described should be applicable to otherFM systems as well.

The Diva BTE hearing aid is used inthe demonstration. This is an advancedDSP hearing aid with 15 channels of inde-pendent enhanced dynamic range com-pression, each with an adaptive attack andrelease time (SoundStabilizerTm). A noise-reduction algorithm, a fully adaptive direc-tional microphone, and an activefeedback-cancellation algorithm areamong the non-linear adaptive featuresthat must be controlled during the HAand FM measurements.

To avoid output variability, one canset the hearing aid to a special test mode(Test Mode 2) and deactivate some of thefeatures during real-ear/coupler measure-ments. However, doing so requires thehearing aid to be programmed to that con-dition. This may he difficult for clinicianswho lack the software for the program-ming of a special test mode. In acknowl-edgment of clinical realities, the currentprotocol was written so that the hearingaid can he in either its normal mode orthe special test mode. Because this rec-ommendation is based on the Diva hear-ing aids, minor deviations may occur withother hearing aids. However, we believethis recommendation is inclusive enoughto be suitable for most of today's hearingaids, both digital and analog.

DEMONSTRATION WITH THEAUDIOSCAN SYSTEMThe Audioscan Verifit is a hearing aid ana-lyzer that allows real-ear, simulated real-ear, and coupler measurement of hearingaids and FM systems. The output is dis-played with a 1 /3-octave bandwidth. Thissystem allows the use of real-life speechstimuli (child, female, male) and ICRAdigitally distorted speech stimulus, as wellas a continuous pink noise as stimuli forevaluation. Each speech stimulus is about10 seconds in duration, and the outputof the hearing aid can be "captured" oraveraged over the course of the stimulus.

For a reliable output, it is importantto use averaging by pressing the "continue"function. The "capture" function capturesthe instantaneous output spectrum of thehearing aid at only one point in time. Thiscan be extremely variable over time.Because the spectra of various real-lifestimuli differ significantly, we choose oneswhose spectra may be similar to thoseavailable on other hearing aid analyzers.For this reason, we recommend the"female speech" and the "ICRA" signalson this system for verification purpose.

Pre-evaluationIn the following example, the Audioscanis used in the "simulated real-ear" modefor ease of implementation. This is a rec-ommended mode to use after the outputof the hearing aid alone has been verifiedto meet one's criteria of acceptability (e.g.,meeting a particular prescriptive gain tar-get, ensuring that soft sounds are abovethresholds) using real-ear or other mea-surement tools. If the FM is transparentto the HA, the real-ear output of theFM+HA should be similar to the real-earoutput of the hearing aid alone.

When considering a criterion ofacceptability, it is important to realize thatmany prescriptive gain/output targets havenot considered the issue of channel sum-mation in their formulation. Thus, thetarget values may be higher than theintended values. 6 Rather than matchingthe average output to the gain targets forconversational level sounds, one mayinstead estimate how much of the softersignal is above the wearer's threshold as acriterion estimate of audibility. Any adjust-ment of the hearing aid to achieve thisaudibility should be made at this stage.

March 2005 • Vol. 58 • No. 3 A protocol for using FM with DSP hearing aids the Hearing Journal 33

Figure 2. (A) Recommended position ofthe Scola Fiti1 mic/transmitter

in theAudioscan test chamber. The FM micshould he in the onMtZrectional mode,(8) Output of the Diva with a 65-dB

female speech signal aver aged over.the coum of 10 seconds with the signalpresented to the FM mic (in FM+HA).

Evaluation

Step II of the three-step protocol is, as isexplained above, to obtain a reference. Fig-ure lA shows the Position of the hearingaid (connected to the FM receiver) in thetest chamber. A 65-dB SPL female speechsignal is directed to the hearing aid micro-phone with the FM receiver in theFM+HA position. The female speech Sig-nal is presented in its entirety (about 10seconds) and the "continue" option isselected to average the output over time.Figure 1B shows the output of the hear-ing aid. The FM transmitter is in the omni-directional mode and positioned outsidethe test chamber. Since the FM micro-phone is active, it is important to performthis testing in a quiet environment.

Step II of the protocol is to determinetransparency. Replace the hearing aid (andFM receiver) with the FM mic/transmit-ter and Position it in the test chamber asillustrated in Figure 2A. The position ofdie FM transmitter is critical in avoidingpotential büffle or shadow effects. In addi-tion to the illustrated position, the FMmic may be placed in parallel to the Sig-nal source. Also, the FM mic should beset in die omnidirectional mic mode. Onceproperly placed, the 65-dB SPL femalespeech signal may be presented in itsenrirety (using the "continue" option) anddie output from the HA measured. Mean-while, the hearing aid is connected to theHA coupler and left outside the test cham-ber. The output is shown in Figure 2B.

34 The HearingJournal

If one examines Figurcs 1B and 2B,they appear to be very similar. Indeed, ifone superimposes the two curves, they arewithin 1-2 dB of Bach other. When diecurves determined in Steps I and II areidentical, transparency of the FM system(with die HA) is confirmed.

Step III is to measure the FM/HA ratio.With the hearing aid outside the test boxand the FM transmitter still inside the testchamber, select "FM" and choose theappropriatc stimulus to present to the FMtransmitter. Since the FM mic used in thisexample is designed to be worn on dieehest, a "ehest mic" position should bcselected to correct for the increase in soundpressurc level as the microphone is movedcloser to the talken Another alternative isthe "boom mic," which incorporates dif-ferent corrections.

Onc may also wonder if the same stim-ulus (i.e., female speech) should be pre-sented as the Signal of interest. Forexample, one may argue that if a child'steacher is a man, the clinician should selectmale speech as the stimulus to representthe real-life situation more accurately.

However, there are two other consid-erations. First, female speech is selectedhecause it is closer in spectral character-istics to the charactcristics recommendedby ANSI 3.42. Cluusing die signal wouldresult in a different audibility estimate.Secondly, becausc of the spectral differ-ence between stimuli, the use of differentstimuli for FM and HA would confound

the estimated FM/HA ratio. In this case,it would be lower than if the same stim-ulus was used in both measures. For therereasons, the same female speech in thechest position is recommended.

Figure 3 shows the output of the hear-ing aid when the input is presented to dieFM mic in the chest position and theFM mic/transmitter is positioned as inFigure 2A. When the curve obtained inStep III is superimposed on the curvcobtained in Step I, one obtains an esti-mate of the FM/HA ratio as the differ-ence between the two curves. In this case,the estimated ratlos vary from 0 to 12-dBcross frequencies, with the most typicaldifference around 9 dB.

The Audioscan also offers anorher wayof estimating the FM/HA ratio throughdie use ofa table display (Figure 4), whichlists the numeric output at several fre-quencies. In this example, REAR 1 is thenumeric output at the specific frequen-cies when the signal is presented to diehearing aid; REAR 2 is the output whenthe signal is presented to the FM; andREAR 3 is the output of die HA whenthe female speech is measured at the FM-ehest position. The difference betweenREAR 1 and REAR 3 defines the FM/HAratio at a particular frequency.

Adjustments in direct audioinput sensitivityThe previous steps reveal the FM/HATatios at the default setting of the FM.

A protocol for using FM with DSP hearing oids March 2005 • Vol. 58 • No. 3

However, a +10-dB FM SNR may notalways be obtainable or dcsirable. Form-nately, today's FM receivers frequentlyhave gain controls/potentiometers thatallow adjustments to the sensitivity of theFM microphone or output of the FM.For example, the Scola receiver incorpo-rates a gain parameter that can be adjustedby +/-7 dB. The above measurement (StepIII) can be repeated alter an adjustmentto the FM receiver sensitivity to achievethe desired FM/HA ratio. In some hear-ing aids, such as die Senso Vita BTE, onemay also adjust the direct audio input(DAI) sensitivity directly through die hear-ing instrument using the fitting software.

DEMONSTRATION WITH

THE FRYE SYSTEM

The Frye Fonix 6500 system typicallyallows coupler and real-ear measurementsusing primarily composite (or syntheticconiplex) signals that are speechlike. Thesesignals include the "speech-shaped com-posite noise" (or composite), which is acon rinuous signal with spectral charac-teristics that comply with ANSI 53.42-1992 specifications; the digital speech-ANSI signals, which are modulated sig-nals (randomly, about Tour rimes a sec-ond, to approxirnate die temporal parternsof speech) that have the spectra of diespeech signal specified by ANSI; and dig-ital speech-ICRA signals, which have the

are speechlike signals, we recommend theANSI-wcighted speech signal because itis similar in spectrum to die female speediused in the Audioscan system.

Pre-evaluation

Again, it is important to ensure that diesettings on die hearing aid alonc meet one'scriteria of acceptability when using real-ear measurement. As explained before, ifthe real-ear results are acceptable, a trans-parent FM system will also result in anacceptable output, even though couplermeasurements are conducted with theFM-HA comhinations. Real-ear outputof the hearing aid alone should bc mea-sured and displayed on die SPL-0-gramscreen to esti mate acceptability before pro-ceeding to FM measurements.

Figure 5 is a real-ear display of the hear-ing aid output (SPL-O-Gram) from dieFrye 6500 system. The output of the hear-ing aid with 50-dB and 65-dB SPL, ANSI-weighted digital speech signals is displayedin relation to the patients thresholds andUCLs. The signal was presented for 8 sec-onds at 0° azimuth. One can see that theoutput for die softer stimulus was abovethe thresholds of the wearer through 4000Hz, ensuring the criterion of audibility.Adjustments may be made on the hearingaid at this stage to reach one's criterion ofacceptability. Aftenvards, all subsequentevaluation may be done in a coupler.

spectra of the speech signal specified bythe International Collegium of Rehabil-itative Audiologists. Output from the hear-ing aid is displayed using a 100-Hz fixedbandwidth.

Once the output of the hearing aid isstabilized, the clinician presses a stop But-ton on the control panel to freeze die fre-quency-output response. Although bothANSI- and ICRA-weighted digital speech

36 The Hearing Journal A protocol for using FM with DSP hearing aids March 2005 • Vol. 58 • No. 3

Subject DMK; SD-9 with 50 dB flat loss4120 -u- u– 4—

tt

100

20

0 100 1000 10000

40 — 50 dB ANSI weighted digital speech

— 135 dB ANSI weighted digital speech

— 90 dB puretone sweep

T = Threshold= UCL

Frequency (Hz)

Figure 5. Real-ear output of the Diva evaluated with the Fiye 6500.

EvaluationStep I: Obtain a reference. Figure 6Ashows the position of the hearing aid (con-nected to the FM receiver) in the testchamber. A 65-dB SPL ANSI-weighteddigital speech signal is directed to the hear-ing aid microphone with the FM receiverin the FM+HA position. The modulatedANSI-weighted digital speech signal ispresented for 8 seconds to allow the out-put of the hearing aid to stabilize. Thestop button is pressed to freeze the screenand capture the response.

Figure 6B shows the output of the hear-ing aid. The FM transmitter is in the omni-directional mode and positioned outsidethe test chamber. Since the FM micro-phone is active, it is important to performthis testing in a quiet environment.

Step II: Determine transparency.Replace the hearing aid (and FM receiver)with the FM mic/transmitter and posi-tion it in the test chamber as illustratedin Figure 7A. Because the signals are pre-sented from a loudspeaker underneath thedevice being evaluated, the position of theFM transmitter is less critical as long asthe microphone is within the indicatedcircle in the test box. In addition, the FM

40 The Hearing Journal

mic should be set in the omnidirectionalmic mode. Once properly placed, the 65-dB SPL ANSI-weighted speech signal maybe presented for 8 seconds and its outputfrom the HA measured. Meanwhile, thehearing aid is connected to the HA cou-pler and left outside the test chamber. The

OPL......dE70

60

E3

Lio

1

A protocol for using FM with DSP hearing aids

output is shown in Figure 713.If one compares Figures GB and 7B,

they appear to be very similar. Indeed, ifone superimposes the two curves, they arewithin 1-2 dB of each other. When thecurves determined in Steps I and II areidentical, transparency of the FM system(with the HA) is verified. An advantageof the Frye system is that it also reportsthe overall output level on the right-handside. In this case, the output in Step I is79.6 dB and 79.9 dB in Step II.

Step III: Measure the FM/HA ratio.With the hearing aid outside the testchamber and the FM transmitter stillinside the test chamber, increase the levelof the ANSI-weighted digital speech sig-nal to 80 dB SPL and present it to theFM transmitter for 8 seconds. Freeze thescreen when a stable response is reached.This level of stimulus intensity is selectedto correct for the increase in sound pres-sure level as the microphone is movedcloser to the speaker.

Figure 8 shows the output of the hear-ing aid when the input is presented to theFM mic at 80 dB SPL and the FM/trans-mitter is positioned as in Figure 7A. Whenthe curve obtained in Step III is super-imposed to the curve obtained in Step I,an estimate of the FM/HA ratio isobtained as the difference between thetwo curves. Furthermore, one may sub-tract the overall levels reported on theright-hand side berween Steps I and III.In this case, the estimated ratio was 8 dB

Figure 6. (A) Placement of hearing aid inthe Frye test chamber. The 65-dB SPLANSI-weighted digital speech signal waspresented from the loudspeakerunderneath the hearing aid. (B) Outputof the Diva with a 65-dB SPL ANSI-weighted digital speech signal capturedafter it has been presented for 8 secondsto the hearing aid (in .17M+HA).

DIG. SPEECH IN NOISE

ANSI SIGNAL - WT PWR

Prr OUT SPLRMS SOURCE SPL:

SPEECH SIG LUL 6SdB

BIAS SIG LUL OFF

BIAS SIG FREQ OFF

NOISE REDUCTION 4 X

4 AHr LEVELED

March 2005 • Vol. 58 • No. 3

(87.6-79.6). Note that this level isreported as an overall level change (i.e.,across all frequencies and not I /3-octavefrequencies, as reported in the Audioscan).One may adjust the FM receiver to achievea more desirable FM/HA ratio.

(Note: Frye Electronics offers softwarecalled "Press and Go" that allows auto-mated testing of the FM/HA ratio simi-lar to die test described in this paper whenthe Frye system is connected to a personalcomputer via an RS232 cable.)

CONCLUSIONSDespite the differente in stimuli used inthe two hearing aid analyzers (one fcmalespeech and one ANSI-weightcd speech-shaped noise), the estimated transparencyand FM/HA ratios measured in theFM-HA combination are remarkably sim-ilar. For both the Audioscan and the Fonix,

the output spcctrum obtained in Step IIwas similar to that in Step I, suggestingtransparency. In addition, the FM/HAratio was 8 to 9 dB for both test systems,although it is important to note that dieAudioscan reports in 113-octave levels andFrye reports an overall level.

The result may differ if different stim-uli are used during die evaluation. Indeed,during our evaluation, we realized thatthe use of different stimuli from the samehearing aid analyzer or use of the "same"stimulus on different hearing aid analyz-ers often yielded a different estimate ofthe FM/HA ratio.

This raises an important question forclinicians to consider. If the noted FM/HAratio varies depending on test stimuli andtest systems, which stimulus (and test sys-tem) should be used to obtain results thatbest represent the real-life perfortnance of

the FM-HA combination? While moreresearch is neccssary to reach a consensus,clinicians should use the same equipment,test stimulus, and test method each timethey make the measurements in order toavoid matching one FM/1-1A ratio targetin the first test session and anotherFM/HA ratio target in the nexr testsession.

Prior to making hearing aid programchanges to accompl ish a desired FMadvantage, the clinician should, wheneverpossible, supplement results with perfor-mance-hased measures.

Although it is impossiblc to eliminateall sources of error when making thesemeasurements, using a consistent testmethod will at least limit the variahilirythat would otherwise occur. The three-step procedure descrihed in rhis article isa sequential approach that can achieveboth the goals of assessing transparencyand of determ ining a beneficial FM/HAratio. Although results displayed wereobtained with the Widex hearing aids andFM system and only two commercialhearing aid analyzers, its principles shouldbe applicable to other hearing instrumentsand hearing aid analyzers.

Acknowledgment

We want to thank Bill Cole of Audio-saunCo. and George Frye of Frye Electronicsfor dick helpful comments on an earlierversion of this manuscript.

Jane Auriemmo, huO, is Pediotric Partnership Program Manager,Widex USA; Denise Keenan, MA, is Research Audiatag'st, WidexORCA; Denise Passerieux, MA, is en audiologist eith Widex USA;and Francis Kuk, PhD, is Director of Audiology, Widex. Correspon-dence to Dr. Aunemmo ot jounemmoCeidexmail.corn.

REFERENCESI. Crandcll C, Sinaidino J: I ntproving classroom acoustics:

Utilizing, hcaring assistive technology and communi-ration stratcgics in the cducational scnings. Volta Rev200 I ;101(5):47-62.

2. Kuk Keenan D: lssues in FM verification. Presenta-tion at the 2005 Illinois Arademy of Audiology Con-vcntion, Chicago.

3.Thibocleatt L, Sauccdo K: Consistcncy of clectroarousdrcharacteristics across components of FM systems. JSp Hear Res 1991;34:628-635.

4. Hawkins Comparisons of speech recognition in noiscby mildly-to-moderarely hearing-impaired childrenusing hcaring aids and FM systems. J Sp Hear Dis1984;49:409-418.

5. American Speech-Language-Hearing Association: Guide-lines for lining and monitoring FM systems. ASHALeader 2000;5(20).

6. Kuk E Ludvigscn C: Variables JFerring the use of gen-eral-purpose prcscriptive formulac to fit modern non-linear hearing aids. JAAA 1999; I 0(8):458-465.

42 The Hearinglaumtu' A protocol for using FM with DSP hearing aids March 2005 • Vol. 58 • No. 3