Two-in-One Custom Hearing Aids Made Possible By Software-Based DSP Platform

In conventional, hardware-based DSP hearing instruments, the algorithms that determine how sound signals are interpreted and processed are hard-wired into the chip design. In contrast, software-based DSP is designed to transform a hearing aid into a PC— the algorithms are not hardwired. This makes it possible to completely change the signal processing in a hearing aid by downloading new software.

The Canta 7/Open is the first commercially available hearing instrument to take advantage of this software-based DSP. After receiving the instrument from the manufacturer, the dispenser can download traditional Canta processing (an FFT processor) or Canta 7/Open processing that is best suited to fit individuals with high frequency hearing loss.

In addition to being “open” in terms of the DSP platform, the new custom hearing instruments are equipped with stepped venting with larger vent sizes than a traditional Canta instrument. This article provides a general overview of this custom hearing instrument system and sound processing approach.

Design Considerations to Suit Patient Needs
Patients want cosmetically appealing hearing instruments that fit comfortably and sound “natural,” without annoying feedback. For many patients, the model of choice is a small, custom instrument that is inconspicuous in the ear. Additionally, a successful fitting must provide ample gain for hearing soft, high frequency sounds. This is the prescription for millions with presbycusic high frequency hearing loss. For many patients and clinicians, however, this has posed a dilemma—how to provide adequate gain and larger venting to ensure a “natural” sound quality without acoustic feedback?

A time-honored method of reducing occlusion-related complaints (such as “my voice sounds hollow,” “my voice echoes,” or “I sound like I’m talking in a barrel”) is to increase the venting. More venting allows low frequency energy to escape, and thereby imparts a more natural sound to the patient’s own voice. The problem arises when high frequency gain is necessary to meet the patient’s audiometric needs. Gain in the high frequencies can cause annoying feedback, and clinicians are then forced to compromise between providing sufficient gain on the one hand, and providing acceptable voice quality on the other. The Canta 7/Open custom products offer greater venting along with feedback control to help solve this dilemma.

The custom instrument is a digital multi-band compression hearing aid. The WarpOpen compression system utilized in the system uses fast-acting WDRC with processing delays of <3.5 msec, and low-level expansion for microphone noise suppression. Additionally, a signal processing technique called “frequency warping” enables variable crossover frequencies in the compressor filter bank. The frequency warping technique is designed to offer better processing efficiency and virtually no crossover region distortion. (For a complete description of the WarpOpen compression system, see Groth & Pedersen.¹)

Figure 1. 15/70/15 stepped vent configuration in Canta 7/Open custom instruments.

Feedback control is achieved through adaptive Digital Feedback Suppression (Stabilizer DFS™) to extend the amount of usable gain. Standard venting for all of these models is one size larger than the current GN ReSound Canta vent template recommends. Additionally, vents are stepped (tapered) in a 15/70/15 fashion (Figure 1): 15% of the length of the vent is tapered from the faceplate inward; an additional 15% of the length of the vent is tapered from the end of the canal inward. The remaining 70% of the venting cavity is within the body of the shell. The resulting 15/70/15 configuration shortens the vent; combined with the added vent diameter, this results in less low frequency energy build-up within the ear canal, and fewer occlusion problems for the patient.²

Clinical Trials
Clinical trials were conducted in the US to assess the performance of the Canta 7/Open custom CIC [710] and ITC [730]. Trials were conducted in two phases: a research phase (“alpha”) and a field-trial phase (“beta”). The alpha trial examined hardware and software performance using patients seen in GN ReSound’s Chicago research facility. The beta trial involved patients fit by ReSound dispensers, and tracked the responses of both patients and clinicians to the new product.

The research trial examined: 1) maximum gain before feedback in one-size larger than current factory-select vent size, with the additional modification of stepping (ie, tapering) the vent; and 2) occlusion relief with the modified one-size larger than current factory-select vent size. In addition, patients were surveyed on issues regarding sound quality and overall performance and satisfaction with the hearing instruments. The field trial concentrated on verifying the functionality of the fitting software and obtaining patients’ subjective responses to sound quality, performance, and overall listening satisfaction.

Figure 2. Average thresholds for 18 ears wearing CICs and 18 ears wearing ITCs.

I. Research Trial (Alpha Trial)
Nine subjects (18 ears) were fit with CIC instruments, and 9 subjects (18 ears) were fit with ITC instruments in the research trial. All subjects were previous hearing aid users, were fit binaurally, and had sensorineural hearing losses ranging from mild to moderately severe (Figure 2).

Methods
Patients were seen for three visits over a 4-week period. At the initial visit patients filled out the APHAB survey,³ providing responses about their current instruments. Real-ear measurements of use gain for their current instruments were obtained. Patients were then fit with the Canta 7/Open test instruments programmed to prescribed target gain. Settings were fine-tuned as necessary. Patients were given a take-home questionnaire to bring back at the second visit (after 2 weeks), which surveyed sound quality, occlusion, and feedback issues.

At the second visit, any fine-tuning necessary to address occlusion, feedback, and sound quality was performed. Real-ear measures of maximum gain before feedback (REAR at maximum gain setting) were obtained to calculate useable headroom. At the third and final visit, the APHAB was completed for the open-canal instruments. Real-ear measurement of use gain was obtained. Patients also completed a final questionnaire about occlusion, feedback, and sound quality.

Figure 3. Real-ear use gain—Real Ear Insertion Response (REIR) for 50 dB inputs comparing subjects’ own instruments to research instruments.

Results
Real-ear measurements of use gain: Figure 3 shows real-ear gain for 50 dB SPL ANSI digital-speech-in-noise input. Subjects fitted with both CIC and ITC models were able to utilize more gain than with their own instruments. The average overall difference was 4.7 dB for CICs and 3.0 dB for ITCs; from 2000-4000 Hz, the average difference was 5.5 dB for CICs and 4.0 dB for ITCs. For 80 dB input levels, use gain was lower for the test aids. The overall difference in gain for loud sounds was 1.6 dB less for CIC subjects and 4.4 dB less for ITC subjects, compared to their own instruments. In addition to achieving more gain, the average increase in vent size over their previous instruments was 0.4 mm for CIC fittings, and 0.9 mm for ITC fittings.

Figure 4. Real-ear use gain—Real Ear Insertion Response (REIR) for 50dB inputs comparing subjects’ own instruments to research instruments.

Real-ear measurements of gain just before feedback (GBF): Figure 4 shows gain before feedback when compared to use gain. The amount of reserve gain for each subject was estimated by taking the real-ear frequency response curves (50 dB SPL, digital speech-in-noise input) obtained at the point of maximum “gain before feedback” (GBF) and compared with subjects’ preferred use gain. Figure 4 shows an average of 8.8 dB of “reserve gain” from 2000 Hz to 4000 Hz for CIC subjects, and 9.0 dB of “reserve gain” from 2000 Hz to 4000 Hz for ITC subjects.

Figure 5. Gain Before Feedback (GBF): Gain values shown in software (“gain handles”) at point of maximum gain before feedback.

Gain before feedback—DFS: The Digital Feedback Suppression system in the test hearing aid utilizes Stabilizer DFS™ to control feedback, allowing a wide range of gain levels, even in the high frequencies. To measure the effect of the Stabilizer DFS, maximum gain before feedback was measured with and without DFS activated. Gain before feedback (GBF) was determined by setting the hearing aid to have a flat 50 dB response curve from 2000-6000 Hz, then simultaneously raising the gain for those frequencies and noting when feedback occurred. Gain was then reduced until feedback stopped. The gain values for 2000 Hz, 4000 Hz, and 6000 Hz at the point just before feedback occurred were considered “maximum GBF.” These values were determined with DFS on and DFS off, and the difference was referred to as “DFS headroom.” For patients wearing CICs, average DFS headroom was 11.8 dB at 2000 Hz and 4000 Hz, and 10.0 dB at 6000 Hz. For patients wearing ITCs, average DFS headroom was 14.6 dB at 2000 Hz and 4000 Hz, and 11.8 at 6000 Hz. With DFS on, 15-of-18 ears (83%) wearing CICs reached maximum instrument gain without experiencing feedback. For ITC wearers, 6-of-18 ears (33%) reached maximum instrument gain without experiencing feedback.

APHAB: The APHAB (Abbreviated Profile of Hearing Aid Benefit)³ is a 24-item self-assessment inventory in which patients report the amount of trouble they are having with communication or noises in various everyday situations. Benefit is calculated by comparing the patient’s reported unaided difficulty with the amount of difficulty they experience when using amplification relative to Ease of Communication (EC), Reverberation (RV), Background Noise (BN), and Aversiveness (AV).

Figure 6. APHAB results comparing benefit ratings for subjects’ own aids to open-canal CIC (alpha research trial).

As shown in Figure 6, CIC patients showed a (non-significant) tendency toward greater benefit with the open-canal instruments vs previous instruments. As shown in Figure 7, there was no difference for the ITC patients in reported benefit with the open-canal vs their own instruments. It should be noted that most subjects in both groups were already wearing high-end technology (primarily DSP) instruments, with generally good satisfaction levels. Therefore, a dramatic increase in overall benefit was not necessarily expected.

Figure 7. APHAB results comparing benefit ratings for subjects’ own aids to open canal ITC (alpha/research trial).

Take-home questionnaire: The take-home questionnaire was given to patients at the end of the first visit. Their instructions were to complete the questionnaire before the second visit (ie, after wearing the new instruments for 2 weeks). The questionnaire elicited information about occlusion and feedback in two ways: 1) rating the occurrence of occlusion and feedback (“always,” “sometimes,” or “never”); and 2) judging instrument performance in five areas (voice quality, feedback, distortion, sound quality, and fit). Subjects reported that their own voices were almost always natural-sounding, with little distortion or “muffled” perception. Additionally, feedback in quiet was reported by only 4-of-18 subjects (22%).

Figure 8. Final rating questionnaire. All subjects’ performance ratings of their own aids compared to open-canal custom instruments.

Final rating questionnaires of own vs open-canal instruments: At the end of the 4-week trial period, patients were asked to compare the new instruments to their previous instruments in six areas. Subjects used a 5-point scale, where 1 equalled “own much better,” 2 equalled “own somewhat better,” 3 equalled “no difference,” 4 equalled “new somewhat better,” and 5 equalled “new much better.” Subjects compared the performance of the open-canal instruments to their own instruments (Figure 8). The average score for all subjects was 3.5, indicating a slight preference for the new instruments. The greatest preference was shown for “overall sound quality performance” (average score of 3.7) and “naturalness of your own voice” (average score of 3.6).

II. Field Trial (Beta Trial)
The field trial was completed at 9 dispenser sites throughout the United States. Both new and previous users were included in the study. Twenty-five patients (12 new users) were fit with CICs, and 22 patients (13 new users) were fit with ITCs. This portion of the investigation concentrated on verifying the functionality of the fitting software and obtaining subjective responses to sound quality, performance, and patient satisfaction.

Methods
At the first visit, the APHAB was completed by the previous users relative to their current hearing instruments. Patients were then fit with their new instruments, and the test aids were programmed to prescribed target gain. Settings were fine-tuned as necessary. All patients were given a take-home questionnaire to bring back at the second visit, which surveyed sound quality, occlusion, and feedback issues. At the second visit (after 2 weeks), any fine-tuning necessary to address occlusion, feedback, and sound quality was performed.

At the third and final visit (after 4 weeks use), the APHAB was completed by all patients, referring to their new instruments. Patients also completed a final questionnaire surveying occlusion, feedback, and sound quality.

Results
Occlusion: In general, survey responses of the experienced patients in the field trial concurred with those participating in the research trial relative to overall sound quality and performance. In comparing the responses of the new and previous users, however, we did note some marked differences in subjects’ ratings of perceived occlusion.

Figure 9. Percentage of subjects’ responses to “my own voice sounds natural.”

As shown in Figures 9-10, 42% of experienced users responded “always” to the question: “my own voice sounds natural,” compared to 28% of new users; 12% of experienced users responded “never,” compared to 42% of new users. A similar distribution of responses was seen to the question: “my head is in a barrel.” Overall, new users had more negative reactions to questions about the naturalness of their own voice and hollowness.

Figure 10. Percentage of subjects’ responses to “my head is in a barrel.”

There are several possible theories that would explain these findings. New users may have higher expectations of hearing aid performance regarding the sound quality of their own voice (ie, they assume no change in voice quality while wearing hearing aids). Previous users, on the other hand, know from their experience to expect a change in voice quality with amplification. Additionally, experienced users may be comparing the voice quality with their own aids to the aids worn in the study, and therefore may actually be rating the improvement in voice quality with the open fittings. A controlled study with separate groups of new and experienced users is warranted to shed more light on this finding.

Figure 11. APHAB results comparing benefit ratings of open-canal instruments to subjects’ own aids.

APHAB: For previous users, APHAB results show a preference for the new instruments over their previous instruments (Figure 11). Previous users rated their new instruments the same as their previous instruments, with the exception of one question (“how much difficulty do you have in a situation where you most wanted to hear better”) where the new instruments were rated slightly better than their previous instruments.

Discussion
Canta 7/Open gives a dispensing professional the opportunity to order one hearing aid and, in effect, try two on the patient. This is particularly beneficial for the case of the patient with borderline normal low-frequency hearing loss. This patient might do well with the traditional Canta and its features, or might do better with Canta 7/Open and its features. Stepped-vent fittings combined with advanced DSP can be beneficial to patients desiring a small, custom hearing aid that offers adequate feedback control and relief from occlusion.

In addition, if a patient’s hearing loss becomes worse over time, software-based DSP allows the option of changing the sound processing without buying a new aid. The processing is available in a hearing instrument with a software-based DSP, meaning that a Canta 7/Open can be configured to a traditional Canta by the dispenser if the patient experiences a deterioration in hearing or the traditional sound processing is preferred.

References
1. Groth J, Pedersen BD. A high-definition compression system for open fittings. Hearing Review. 2003; 10(4): 68-81.
2. Stuart A et al. The effect of venting on in-the-ear, in-the-canal, and completely–in-the-canal hearing aid shell frequency responses: real-ear measures. J Speech Hear Lang Res. 1999: 42: 804-813.
3. Cox RM, Alexander, GC. The Abbreviated Profile of Hearing Aid Benefit (APHAB). Ear Hear. 1995; 16:176-186.