Real-Ear Measures
Challenges in verification of digital hearing instruments and particularly non-occluded or open fittings
Harvey Dillon: “We cannot know what a hearing aid does unless it’s performance is measured”
What are we trying to achieve when we verify a hearing aids performance?
That it matches a prescriptive target? That is makes speech audible? That it provides comfortable and safe amplification? That it does what the manufacturer says it does?
Open fittings are popular!
Minimise occlusion Light and inconspicuous Phase cancellation feedback suppression allows more high frequency gain.
So what about verification. Is it different for these hearing instruments than other digital hearing instruments?
Real-Ear Measures in Clinical Practice
In the U.S. studies and surveys routinely show only 30-40% of audiologists use Real-Ear measures. I recently surveyed 120 audiologists who fit hearing aids in New Zealand and received 82 responses (67% response rate).
What proportion of the surveyed audiologists use Real-Ear measures?
89% Always or Almost Always use Real Ear Measures.
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00% Never Occasionally Sometimes Almost alw ays Alw ays
Which Real-Ear Measure ?
70.00% 60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00% Real-Ear Insertion Gain Output SPL (e.g. Speech Mapping) Both Neither
Which prescription formula did the surveyed audiologists choose?
70.00% 60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00% NAL-NL1 DSL I/O DSL V 5 NAL-RP Custom Generated
Stimulus or signal used for Real-Ear Measures
35.00%
30.00%
25.00%
20.00%
15.00%
10.00%
5.00%
0.00% Sw ept tone Pink Noise Composite Noise ICRA Live speech Digitized Speech Signal Other (please specif y)
Real-Ear Measures and Open Fittings.
83% of the surveyed audiologists said they used Real-Ear Measures to verify open hearing instrument fittings.
Which Real Ear Measure do you use?
60.00%
50.00%
40.00%
30.00%
52% REIG 34% Output SPL 5% Both
20.00%
10.00%
0.00% Real-Ear Insertion Gain Output SPL (e.g. Speech Mapping) Both
Do you routinely fine-tune these fittings to meet the same prescriptive target as your other non-linear hearing aid fittings?
47% said yes 37% said no
Comments about verifying open fittings
“I usually find open fittings targets are not met yet the patient is very happy with the sound, and if I fine tune to meet targets the patient complains - usually that it is too loud.” “I generally expect a similar shape output to Nal-NL1, but less gain overall.”
Comments about verifying open fittings
“Some manufacturers just cannot be adjusted to match, and I have found that when you DO match the targets, it is over prescribed. so either its a manufacturer that I CAN match from first fit, and I match targets as best I can, or its another company that I cannot, so I go with the first fit and just smooth the response as best I can.”
Comments about verifying open fittings
“I don't tune to a target. I compare to NALNL1, but am really looking for the aids to be matched between ears and to be giving a 'sensible' amount of gain based on my own judgement. My open fittings tend to stay pretty near manufacturer's prescribed settings”
Real-Ear Measures and Open Fittings.
What reasons could there be for this experience?
Measurement errors Stimulus effects Features of the Prescriptive formula
REIG or Output SPL measures?
Real-Ear Basics
Real-ear loudspeakers
Probe Microphone Assemblies
Reference Microphone
Reference Microphone and Errors in RealEar measures.
Usually, output of the speaker is regulated concurrently with the measurement (Concurrent Equalisation). This allows for effects of patient movement and position. Deviation from the desired level will result in automatic adjustment of the speaker output. Larsby and Arlinger (1988) identified that there was a risk of errors with this method when verifying open fittings.
The alternative is to equalise prior to the measurement (Stored equalisation) . This means that the speaker level is set before the measurement rather than during the measurement. Patient must stay in the same position throughout.
Reference Microphone Associated Errors in Real-Ear Measures when using Concurrent Equalisation
Amplified sound from the ear canal can combine with the sound from the loud speaker (usually when higher levels of gain are being provided). This causes an automatic reduction in the output of the speaker at the frequencies where there is leakage. Consequently sound from the loudspeaker is less than the intended level at those frequencies. Therefore the REAR is also less than it would be without leakage at those frequencies. In this case, stored equalisation is a better option.
Measurement Errors
Results from a study by Lantz et al (2007) showing the mean difference in gain measured with stored and concurrent equalisation methods. MPSE= Stored Equalisation MPCE= Concurrent Equalisation
Interaction with digital feedback suppression
Lantz et al (2007) point out that a digital feedback system estimates the sound escaping from the ear canal and subtracts this from the hearing instrument microphone signal. Therefore only the sound from the speaker is amplified by the hearing instrument. As the speaker output is reduced at those frequencies where there is leakage this may also lead to a reduction in output SPL from the hearing instrument measured at the ear drum (REAR) .
Measurement Errors
The REIG measured with stored and concurrent equalization methods with four different gain settings in a single subject. The measuring error increases as a function of the digital feedback suppression benefit.
Real-Ear Measurement Equipment Options
AURICAL Plus’ has an open-REM mode Fonix equipment allows the reference microphone to be disabled for open fittings. Audioscan uses a stored equalisation method in it’s Open fit option within Speech Mapping. Or locate reference microphone away from the ear canal opening if it is not possible to switch off the reference microphone during the measurement.
Stimulus characteristics effect the output of digital hearing instruments.
These signals differ in: Bandwidth
Narrow (tone) vs. broad band (speech) Short/brief (tone) vs. longer (e.g., speech) Pink noise (little fluctuation) vs. Speech (greater fluctuation)
Pink noise (flat) vs. Speech (roll off in highs)
Duration
Fluctuation over time
Spectrum
Stimulus Spectra Examples
Non-Linear instrument
Showing less gain for a swept tone than for a speech stimulus
Prescriptive Formulae Considerations
Speech-based verification signals are recommended for use with both NAL-NL1 and DSL V 5. Both formulae also allow for the effects of multi-channel compression.
Manufacturers are able to make allowance for this in their pre-calculation within programming software. Most Real-Ear measurement equipment does not allow for this.
Prescriptive Formulae Considerations
A speech-based verification signal interacts with multi-channel compressors in a more realistic way than tones. Targets at moderate input levels show less effect from “channelization” than targets for very high or low inputs. (Moodie et al) Both formulae allow for the effects of binaural summation. A gain reduction of up to 5-6 dB may occur in NAL- NL1 when binaural is selected (dependent on input level and symmetry of hearing loss) In DSL Vsn 5 a 3 dB gain reduction is implemented when the binaural option is chosen.
REIG or Output SPL Measures?
A Real-Ear Insertion Gain measure is the difference between the aided and the unaided measure: REIG = REAG – REUG The assumption behind this measure is that natural amplification provided by the resonances of the concha and ear canal are lost through insertion of the hearing instrument. Before the hearing instrument can provide additional signal it must provide at least as much gain as the unobstructed ear. (Dillon, Hearing Aids 2001)
REIG
Many audiologists have used this routinely for many years and may have equipment that only allows this measure. However meeting a prescriptive target in REIG does not ensure audibility Should we be combining a hearing threshold measure obtained with insert earphones with a real-ear unaided gain measure obtained in free-field?
Output SPL Measures
SPLogram format e.g. Speech Mapping in Audioscan aims to fit the hearing aid response within the residual dynamic range of the hearing instrument wearer.
Audibility Comfortable loudness of important speech cues Recommended measure for DSL prescription formulae
Real-Ear Aided Gain (REAG) measures look at the output in SPL at the ear drum with prescription formulae targets available for REAG.
Understanding an SPLogram The Unaided SPLogram
Sounds get louder as you go UP the scale
dB SPL Eardrum reference
Maximum output targets Loud speech Avg. speech Soft speech Threshold (dB SPL TM)
Normal hearing
An Open-Fit Verification Protocol suggested by Audioscan
Measure REAR with instrument on ear but turned off Measure REAR with instrument on ear and turned on Adjust gain to maximize SII without invoking feedback
Example of Open-Fit REAR Result
Aid’s contribution
Pink banana = REAR with aid OFF Green banana = REAR with aid ON
Speech Intelligibility Index
The Speech Intelligibility Index, or SII is calculated from acoustical measurements of speech and noise. The SII is highly correlated with the intelligibility of speech as evaluated by speech perception tests given to a group of talkers and listeners. American National Standard ANSI S3.5-1997 (“Methods for Calculation of the Speech Intelligibility Index”).
The SII Score
Recommendations for Verifying Open Fittings
Do use Real-Ear measures! Use a speech-based signal Select a binaural prescription calculation. Consider using an Output SPLogram If possible use a stored equalisation rather than concurrent equalisation method to set you speaker level
Recommended Reading:
The Hearing Journal November 2006 “Open-canal fittings: Ten takehome tips” H. Gustav Mueller and Todd A. Ricketts
The Hearing Journal is available online free.
References
Dillon, H. 2001“Hearing Aids”, Thieme Larsby and Arlinger,1988 “Comparison of different types of equipment for Insertion gain measurement” in: J.H Jensen Hearing aid fitting: Theoretical and practical views. Copenhagen 13th Danavox Symposium Lantz et al 2007“Real-ear measurement verfification for open,nonoccluding hearing instruments” IJA, 46:1,11-16 Lindley,G. 2007 “Real-Ear Measures and Advanced Technology” eSeminar on www.audiologyonline.com Moodie et al “The DSL Method for paediatric and Adult Hearing Instrument Fitting: Version 5”, Phonak Focus 37 www.phonak.com Mueller and Ricketts 2006 “Open canal fittings: 10 take home tips” 2006,The Hearing Journal, 59:11, 24-39 www.audioscan.com