With the advent of programmable and digital signal processing (DSP) hearing instruments, many dispensing professionals who have embraced the routine utilization of probe microphone (probe-mic) measurements over the past 15 years are nowand have been for awhilefaced with some new problems (or challenges if one is inclined to take the glass is half full approach). The first challenge relates to using a computer with fitting software to fit programmable/DSP hearing instruments while using an older stand-alone probe-mic system (i.e., problems of compatibility and system integration). The second challenge relates to staff training and the time it can take to gain a working understanding of these systems.
The goal of this article, which uses the Siemens Unity PC Probe System1 as an example of a real ear measurement system, is to address these two challenges and provide information on the usefulness of utilizing probe-mic measurements in a dispensing practice. It is hoped that, upon reading this article, the reader will also understand how ICRA signals and an integrated probe-mic system provide new and easier procedures for evaluating the gain and output of advanced hearing instruments.
Probe Microphones and Fitting Programs
The solution to the first challenge of using fitting software with an older probe- mic system lies in answering the following question: Is your probe-mic equipment NOAH and computer compatible? If not, in the authors opinions, it is time to replace the systemperiod! A more professional approach that makes the task of programming, fitting and verifying todays DSP instruments easier and faster requires many of the capabilities only available with modern probe-mic test instruments. These new test instruments are powerful, computer-based fitting tools that are NOAH compatible, integrate manufacturers fitting software, provide new testing procedures and are fast and easy to use.
The second challenge relates to the question: What is the best way to test and verify DSP circuits with probe mics? Two new test features available in the Siemens Unity PC Probe and a few other probe-mic systems (e.g., Madsen Auricle and MedRx OtoWizard2) are the basis for this article and will provide readers with some answers to both of these two challenges. First, the new set of real ear measurement (probe mic) test signals called ICRA signals will be explained. And then, we will do the same for a new probe-mic test called the Universal Probe Mic-Link test procedure.
Probe-Mic Testing: Believers vs. Skeptics
For over 15 years, Hawkins, Mueller and Northern3,4 have been some of the many hearing care professionals and industry experts who have recommended the routine utilization and practical value of performing probe mic measurements.
Like many things in life, a user curve exists for probe mic testing. There is a group of hearing health care professionals who have become true believers and routinely use this objective hearing instrument fitting technique. There is also a middle group (perhaps the majority), who use probe mic testing about 50% of the time, reserving it only for tough cases or when they have the time. Finally, the ardent skeptics group have reasons for not using real ear measurements that are as varied as the colors of the rainbow.
The two new probe microphone fitting tools that are going to make the task of fitting digital and programmable hearing instruments easier and faster (and maybe even better) are:
- A new set of test signals to use for testing digital hearing instruments called ICRA signals, and
- The Universal REM-Link test procedure, a simple new procedure that links all hearing aid fitting software programs with the probe-mic test system.
ICRA Signals and Why Theyre Important to You
Here is a brief overview of the what, who, why and when of the new ICRA test signals.
What? ICRA signals are speech-modulated sounds used in the testing and fitting of non-linear hearing instruments and for fitting hearing aids in background noise situations.5
Who? ICRA is an abbreviation for International Collegium of Rehabilitative Audiology, a European audiology organization. ICRA signals were developed by HACTES (Hearing Aid Clinical Test Environment Standardization), a work group within ICRA.6
Why? Unlike broadband signals (e.g., composite noise, speech or white noise), the speech-modulated ICRA signals will not activate the environmental noise reduction type features in digital signal processors.
When? These signals should be used to perform probe-mic verification measures with non-linear analog and DSP hearing instruments.
Modulated vs. Unmodulated Test Signals
Since one of the primary objectives of probe mic hearing instrument verification is to ensure that speech is appropriately amplified, it is necessary for the digital hearing instrument to recognize the test signal as if it were actual speech.7 When fitting DSP or non-linear analog circuits, the use of a speech modulated type of signal (i.e., ICRA Groups 2 and 3) will provide a better, improved, more real-world evaluation of this type of hearing instrument performance during probe mic verification. These signals are unlike unmodulated broadband signals (speech, composite and white noise) and narrow-band signals (pure tone, warble and NB), which will cause a reduction in signal amplification or compression.
Fig. 1. Three REIG gain curves which demonstrate the varied amounts of reduction in gain of a DSP circuit when using a 65 dB unmodulated speech noise signal. Curves: Blue = 1 second, red = 5 seconds and brown = 10 seconds.
Fig. 1 demonstrates the importance of using a modulated signal as compared to an unmodulated signal. In this example, an unmodulated signal was presented in real-time to a DSP circuit, and the dB gain was captured at 1-, 5- and 10-second intervals. The constant unmodulated signal significantly reduced the gain of the DSP hearing instrument by varying amounts during the 10 seconds. Varied gain activity like this would make it very difficult for the dispenser to properly evaluate and fit this instrument to a real ear insertion gain (REIG) target.
Fig. 2. REAG Curves: Red = modulated signal, blue = unmodulated signal after a 10- second test. Note the significant differences in measured gain.
So what happens to the gain and output of DSP circuits when a modulated test signal is used? Fig. 2 illustrates the real ear aided gain (REAG) differences between an unmodulated signal and a modulated signal.
Use of ICRA Signals
First, it is important to mention that, because the ICRA signals are relatively new as a test signal, there is no standard for how to use them. Therefore, this article will provide suggestions for use of these signals during probe-mic testing.
|Table 1: ICRA Signal Names and Descriptions
There are nine different ICRA signals. At first glance, a persons eyes can start to glaze over at the acronyms involveduntil one realizes that the names of the signals are actually good descriptions of them. For the purpose of this article, the signals have been placed into three different groups. Group 1 consists of three unmodulated signals, while Groups 2 and 3 consist of modulated signals. Signals from Groups 2 and 3 will be the primary focus in this article.
Siemens Hearing Instruments8 recommends using the ICRA signal 2PB-1F1M-N from Group 2 (see Table 1) as the first signal of choice during probe-mic testing when fitting a digital aid for the primary listening situation with their probe mic system. This signal reflects one male and one female speaking simultaneously, and it should most closely resemble the modulation of conversational speech. Because this signal is not specifically male or female, it should be the most appropriate signal to reflect a typical conversational setting.
If attempting to simulate a listening situation that has mostly female speakers or mostly male speakers, then the most appropriate signal would be 3BSMN-F-N or 3BSMN-M-N, respectively, from Group 2. If attempting to fit to a listening situation in which there are many speakers, both male and female, then it would be suitable to use the 6PB-N signal from Group 3. The signals from Group 3 are also good signals to use if you are making subjective measures as to how well the end user understands speech in noisy settings. Allow the signals from Group 3 to run as you speak to end users to assess their performance with different levels of multi-talker babble.
When fitting digital hearing instruments, do not use the signals in Group 1. Signals from Group 1 are not speech-modulated, so they will be recognized as noise by digital hearing instruments.
Determining the REIG with the REUG & REAG Tests
A typical real ear insertion gain (REIG) test is composed of three steps:
- Step 1: The real ear unaided gain (REUG) measurement, then
- Step 2: The real ear aided gain (REAG) measurement.
- Step 3: The REAG minus the REUG equals REIG (REIG = REAG REUG).9 In the following examples (Figs. 3-5), three or four different test signals are used to demonstrate the differences between modulated signals (ICRA) and unmodulated signals (WT, SN).
|Fig. 3. Three REUG comparison curves with a 65 dB input level: Purple = warble tone difference, Blue = speech noise and Red = ICRA/2PB-1F1M-N. Note that there are no significant differences.|
|Fig. 4. Four REIG curves at 65 dB input level with an analog non-linear BTE. Red = WT, Brown = SN, Green = ICRA (2PB-1F1M-N) and Pink = ICRA (6PB-N). Note the significant differences from 500 Hz and above.|
|Fig. 5. Three REIG curves at 65 dB input level for a digital BTE, Blue = SN, Red = ICRA (2PB-1F1M-N) and Brown = ICRA (6PB-N). Again, note the significant differences from 500 Hz and above.|
For the REUG test (Fig. 3), three different input signals are comparedwarble tone, speech noise and the ICRA Group 2 2PB-1F1M-Nto see if any differences between the signals could be demonstrated. As suspected, because the ear canal resonance operates in a linear function, the type of input signal used does not seem to make any significant difference (Fig. 3). (Authors Note: An input level of 60-70 dB is recommended to avoid any test result artifacts in the REUR curve. Using the same test for both the REUG and the REAG has been recommended.4)
For the REAG measurement, a non-linear analog circuit (Fig. 4) and a digital circuit (Fig. 5) are used to illustrate the significant differences between unmodulated and modulated test signals. (Note: No attempt was made to smooth or modify the gain curves of either hearing instrument used in the figures.) The modulated ICRA signals, 2PB-1F1M-N from Group 2 and 6PB-N from Group 3, and the unmodulated warble tone (WT) and speech noise (SN) test signals were used to demonstrate different listening situations (Fig. 5).
As an example of the Universal Probe Mic Link system, both the ICRA test signals and the Universal REM-Link procedure were used. The patient, BH, was a 65-year-old male who was a previous hearing instrument user. BH had no measurable thresholds on the right ear, and a moderate-to-severe, steeply sloping, high-frequency sensorineural configuration on his left ear. He had been fit previously at a different facility with a Siemens MusicD Digital CIC on the left ear. His major complaints were problems hearing in large groups and a lack of directional hearing.
NAL-R with speech noise was used for the initial probe-mic verification. Later, NAL-R and ICRA (2PB-1F1M-N and 6PB-N) were used for this case study, and appropriate changes were made in the fitting. The patient, after using the new settings for one month, noted that he is hearing more clearly and coping more successfully in social settings.
In the before and after REIG tests, a combination of modulated and unmodulated curves were performed at a 50 dB input level. In Fig. 7a (the before probe-mic verification REIG test), the two curves of most importance are the blue curve, which is the ICRA-2PB signal, and the green curve, which is speech noise (SN). The other curves displayed are various modulated and unmodulated REIG curves. While all the REIG curves generally follow the NAL-R target, the REIG at 250 Hz is over-amplified, and under-amplified at 500Hz. Then, between 1000-3000 Hz, it is under-amplified again.
Fig. 7b illustrates the after REIG tests following the appropriate modifications and the resulting ICRA 2PB REIG curves that match the target better. It can be seen in this case study that REIG differences between a modulated ICRA signal and an unmodulated speech noise signal are significant.
Analog Non-linear Hearing Instrument Test
Fig. 4 illustrates the ICRA 2PB-1F1M-N and 6PB-N signals along with warble tone and speech noise (SN) signals at 65 dB input level when performing a REIG test with a Siemens Music BTE analog non-linear hearing instrument. Note that for the REIG differences in 500-6000 Hz range, the gain differences between these signals types are from 2-10 dB. As the real-time modulated ICRA signal was presented, the gain curve did modulate (i.e., fluctuate up and down +/-5 dB) across the frequency response which demonstrated the non-linear function of the circuit. The unmodulated WT and SN signals did not demonstrate this same type of temporal activity.
In Fig. 5, a Siemens Signia digital BTE was used with unmodulated SN and two modulated ICRA (2PB-1F1M-N) and ICRA (6PB-N) signals at a 65dB input level are used. Again, as in Fig. 2, note the significant differences demonstrated between modulated and unmodulated input signals.
REM and Instrument Adjustments
The Universal Probe Mic-Link test procedure is designed to integrate NOAH and all of the manufacturers hearing instrument fitting software modules with the Siemens Unity PC Probe System. A three-step procedure takes dispensers from the active mode to any manufacturers fitting mode. Hearing professionals can now have an active real ear measurement display and manufacturers fitting module open and working at the same time in NOAH. With the REM module window super-imposed on the fitting software screen, the dispensing professional can perform real ear probe measurements and adjust hearing instruments on the same screen at the same time using available target gain formulae (e.g., NAL-NL1, DSL i/o, FIG6, etc.), any type of test signal (i.e., ICRA, speech noise, warble, etc.) at various input levels ranging from 40-90 dB.
This article was submitted by Bud Majest, BGS, a hearing industry consultant who has many years of experience in special testing instruments, and Cynthia L. Ellison, MA, a private practice audiologist and president-elect of the Academy of Dispensing Audiologists (ADA). Both authors maintain their respective offices in Bella Vista, AK. Correspondence can be addressed to HR or Bud Majest, 13 Fenchurch Lane, Bella Vista, AR 72714; email: [email protected] or [email protected].
1. Equipment used: Siemens UNITY Probe Microphone (v. 2.3) system, MicroSound Artificial Ear, Comply disposable earmold and Siemens Music and Signia BTE hearing aids, both of which were programmed with a flat 50 dB threshold using Connexx First Fit procedure. Probe mic tests were conducted at a 45° azimuth and at 24-in from the probe mic speaker.
2. Madsen Auricle, Madsen Electronics, Inc., Minnetonka, MN, and Oto-Wizard, MedRx Inc, Seminole, FL.
3. Mueller HG, Hawkins DB, Northern JL: Probe Microphone Measurements, Hearing Aid Selection and Assessment. San Diego: Singular Publishing Group, Inc, 1992.
4. Mueller HG: Probe-mic assessment of digital hearing aids? Yes, you can! Hear Jour 2001; 54 (1): 10-17.
5. Dreschler W (chairman): International Collegium of Rehabilitative Audiology (ICRA). Amsterdam: ICRA, 2001.
6. Westermann S (chairman): Hearing Aid Clinical Test Environment Standardization (HACTES). Copenhagen: Widex, 2001.
7. Siemens Hearing Instruments, Inc. UNITY PC Probe software Version 2.3. Piscataway, NJ: Siemens, 2001.
8. Siemens Hearing Instruments. UNITY PC Probe Version 2.3, Help Menu. Piscataway, NJ: Siemens, 2001.
9. American National Standard Institute (ANSI). ANSI S3.46-1996, Methods of Measurement of Real-Ear Performance Characteristics of Hearing Aids. New York: ANSI, 1996.