Tech Topic | June 2020 Hearing Review

By Francis Kuk, PhD; Neal Ruperto, AuD; Christopher Slugocki, PhD, and Petri Korhonen, MSc

PureSound with ZeroDelay technology has been designed to shorten hearing aid processing delays to less than 0.5 ms across frequencies and therefore greatly enhance the sound quality, particularly for large-vent or open fittings—fittings that can reduce the effectiveness of directional microphones. This study shows the PureSound program on the MOMENT hearing aid is as effective as the premium products of two other manufacturers for managing speech understanding in noise at realistic SNRs.

The recently introduced Widex MOMENT™ hearing aid has a unique PureSound program that uses ZeroDelay technology to shorten the processing delay of the hearing aid to less than 0.5 ms across frequencies.1 This unique improvement in signal processing is designed to enhance the subjective quality of sounds processed by the hearing aid. 

Implementing the ultrashort delay necessitated a signal path design and reconsideration of the necessary processing features in the hearing aid. In order to ensure that such a reconfiguration would not compromise its effectiveness, we conducted a study to investigate if the PureSound program is as efficacious as other premium hearing aids when tested under conditions with realistic signal-to-noise ratios (SNRs).

The Inherent Problems of Hearing Aid Processing Delay

What is hearing aid processing delay? In a digital hearing aid, the input sound is split into many different frequency bands, and each band is manipulated by various algorithms before the bands are added together again and output from the hearing aid to the ear canal. This processing takes time and it delays the hearing aid sound with respect to the direct unprocessed sound reaching the ear canal. When digital signal processing was first introduced in hearing aids, some hearing aids exhibited processing delays as long as 30 ms, while today they are generally below 10 ms.

What are the consequences of delay? Hearing aid delay can significantly impact a wearer’s experience in different ways. Delays of 10 ms could lead to metallic or hollow sound quality due to the mixing of the direct natural sound and the processed delayed sound in the ear canal.2 This is a spectral distortion of the input sound called the comb-filter effect. Stone et al3 indicated that processing delays as short as 5-6 ms is needed for acceptable performance in an open-ear fitting. But, as Balling et al1pointed out, “acceptable” performance is not the same as natural performance, a longstanding design principle at Widex.

An Elegant Solution for Hearing Aid Delay

Despite the already low delay of about 3.2 ms in the Widex EVOKE™ hearing aid(5 ms below 500 Hz, 3.2 ms between 500 and 8000 Hz and 2 ms beyond 8000 Hz), we believe that an even shorter delay will preserve more of the temporal nuances in the input signal resulting in a more natural percept in an open-fit situation. 

For example, Figure 1 shows the KEMAR output for a broadband noise with various gap sizes (3 ms, 9 ms, 24 ms) when it was fit with 3 different commercial hearing aids using open-fit instant tips. These hearing aids differ in their processing delay. The MOMENT PureSound has a delay of <0.5 ms, while Manufacturers 1 and 2 have average delays of 8 ms and 6 ms across frequencies respectively. 

Figure 1. Summed RMS KEMAR output for a broadband noise with various gap sizes (3 ms, 9 ms, 24 ms) as processed by the PureSound feature in the MOMENT hearing aid (delay of 0.5 ms), as well as hearing aids from Manufacturer 1 (delay of 8 ms) and Manufacturer 2 (delay of 6 ms). All hearing aids were coupled to KEMAR with an open ear tip. The unaided condition (adjusted for amplitude) is also shown for comparison.

The unaided condition (first row) is shown as a reference. The second row shows the aided result using the PureSound program. One sees that the gap is well preserved at all gap sizes. The third row shows the output of the hearing aid from Manufacturer 1. Here, one sees that the gap­—which should be centered on 100 ms—is displaced to a later time and the gap is partially filled. The shape of the gap is also different from the original gap. Similar observations can be made in the fourth row which shows the output of Manufacturer 2’s hearing aid. 

These data clearly show that hearing aid delays doaffect the temporal characteristics of the input sound in an open-fit device. The impact of the delay is especially evident at a smaller gap size, where the delay “fills in” the temporal gaps. The shape of the gap is distorted more as the gap duration increases. 

The significance of these measurements is that hearing aid delays affect the temporal characteristics of the input sounds also, with a smaller delay better preserving the temporal characteristics of the input signals. A benefit of an ultrashort delay is a more natural perception of the input sounds from preservation of the temporal nuances and minimization of the spectral distortion from comb-filtering.1Moderately loud impulse sounds will be more natural and tolerable. Also improved may be the perception of certain speech sounds which is dependent on the preservation of temporal cues (such as stops and fricative consonants).  

ZeroDelay™ Technology. Widex has been persistent in preserving the naturalness of input sounds. This requires not just sufficient amplification over an extended range of frequencies, but also preserving as many of the temporal characteristics of the input signal as possible. One testament to this is our implementation of the slow-acting compression of the first digital in-the-ear (ITE) hearing aid, the SENSO™.4 The introduction of variable-speed compression in our UNIQUE hearing aid is another testament to our commitment to preserving the natural temporal characteristics of sounds.5 Both slow-acting and variable-speed compression were designed with the aim to preserve the temporal envelope of the input waveform while minimizing potential artifacts.6 

ZeroDelay technology in the MOMENT hearing aid follows the same design philosophy of preserving the naturalness of the input signal. Rather than targeting the compression speed of the hearing aid, ZeroDelay targets the processing delay of the hearing aid to minimize comb-filtering and temporal smearing. 

To understand how ZeroDelay technology works in the MOMENT, it is important to understand where processing delays occur. In most digital hearing aids, the signal processing path starts with the input signal going through a filter bank, which splits the input into different channels for subsequent processing. Features that include multichannel processing—such as compression, directional microphones, noise reduction, feedback cancellation, and frequency lowering—are typically integrated into the filter bank design to function as one unit. Hence, the design and efficiency of integrated processing in the multichannel filter bank contributes most to the processing delay of the hearing aid. In the case of our previous product, the EVOKE, this delay averaged about 3.2 ms.

Achieving an even shorter delay requires a completely new paradigm in signal processing. This is because the delay in a filter bank cannot arbitrarily be lowered without degrading the frequency resolution—and thus the number of channels—in the filter bank. The challenge in the new PureSound signal processing algorithm was to maintain the number of channels while reducing the delay to below 0.5ms. 

The PureSound program operates in parallel to the current filter bank path (known as the Universal or master program). In the MOMENT hearing aid, the wearers have the option to switch between the Universal program with all the classic Widex features, and the PureSound program, which is adapted to a new set of tightly integrated features to the filter bank. The Dual-Mic Input Control allows the user to seamlessly switch between the two signal pathways. 

What is the Dual-Mic Input Control?

A key consideration in feature specification is to understand the intended candidates for the program. The advantage of ultrashort delay is most noticeable by listeners who use a large vent or open-ear fittings. This is because the large vent allows most of the direct natural sound to enter the ear canal and mix with the delayed processed sounds. This situation results in the most comb-filtering with the most discernible effect. Thus, the target group of hearing aid wearers for ZeroDelay technology are people who are candidates for open fitting. People with a mild-to-moderate hearing loss and first-time hearing aid wearers are ideal candidates. 

If the intended use of the PureSound program is in open-fitting, it is reasonable to ask if a directional microphone is necessary. It is commonly accepted that a directional microphone improves the SNR for speech by as much as 6 dB in an occluded earmold, as reported by Bentler.7 Indeed, Kuk et al8 reported that the directional microphone on the EVOKE hearing aid improved SNR by 6.5 dB. On the other hand, as the vent opening (or leakage) of the ear mold/tip increases, so does the amount of direct sound entering the ear canal via the vent opening.9 This leakage “dilutes” the relative amount of processed sound in the ear canal and reduces the effectiveness of certain signal processing algorithms.10 

Kuk et al11 showed that the magnitude of directional benefit measured in the laboratory varies proportionately with the degree of hearing loss. This is because greater hearing loss necessitates a smaller vent diameter, and directional benefits likely vary inversely with vent diameter. Hence, one would expect the least directional benefit in an open-ear fitting. Indeed, many studies12-14 showed that the benefit of a directional microphone in an open fitting ranged between 1 and 2 dB.  

Moreover, there are reasons to suspect that the 1-2 dB directional benefit reported with open fitting is an artifact of the test condition and may not be observable in typical listening environments. In all reported studies, directional benefit was measured as the difference in SNR required to correctly identify 50% of the speech materials. Often, the reception thresholds for speech at this performance criterion are at SNRs that are at or below 0 dB. This means that the purported benefit offered by a directional microphone will only reveal itself if the SNR that the hearing aid wearer encounters in real-life situations is at or below 0 dB. However, recent research from Smeds et al15and Wu et al16 on the topic of realistic SNRs suggests that the SNRs typically encountered by listeners with a mild-to-moderate hearing loss are between 5 dB and 15 dB. Thus, in an open fitting, the potential benefits of a directional microphone may not be realized in typical real-life listening situations.17,18 Together, these lines of research suggest that a directional microphone may not be a crucial feature for the PureSound program.

Widex developed the Dual-Mic Input Control system in the MOMENT hearing aid to allow for the use of the PureSound program under typical listening conditions, while still ensuring the use of directional microphone in very challenging situations. When the wearer selects the PureSound program for typical daily use (under realistic SNRs), the Dual-Mic Input Control automatically switches to an omnidirectional microphone, and the input signal will pass through the low delay path. In a difficult listening situation, the wearer may switch to the Universal program which will direct the Dual-Mic Input Control system to switch to the HD Locator adaptive directional microphone (or whatever microphone mode selected by the dispenser/wearer) so the wearer can benefit from the use of the appropriate microphone mode.    

The current study examined if the PureSound program of the MOMENT hearing aid, when tested in an open-ear fitting mode under realistic SNRs conditions, yields similar speech-in-noise performance to other premium digital hearing aids that include an adaptive directional microphone.

Methods 

Subjects. A total of 21 hearing-impaired listeners participated in the study. Subjects ranged in age from 58 years to 82 years, with a mean age of 71.8 years. A total of 13 listeners were current hearing aid wearers with experience ranging from 1-15 years. Two listeners had participated in hearing aid studies previously but were not wearers at the time of this study, and 6 had never worn hearing aids. Subjects had a bilateral symmetrical (within 10 dB) mild-to-moderate sensorineural hearing loss (see Figure 2). All subjects passed cognitive screening using the Montreal Cognitive Assessment tool (MoCA average score = 26.4). 

Figure 2. Individual (thin lines) and average (thick lines) audiograms of the test subjects.

Hearing aids. The PureSound program in the Widex MOMENT hearing aid was compared to the premium products from two other manufacturers (Manufacturers 1 and 2) using their default gain fit to the NAL-NL2 target and feature settings with open ear-tips. All three hearing aids have top-of-the-line processing features that would be expected on premium products (eg, adaptive directional microphones, noise reduction, and feedback management). An exception is that the PureSound program uses an omnidirectional microphone. The PureSound program has a delay of 0.5 ms, while the delay of Manufacturers 1 and 2 averaged 8 ms and 6 ms across frequencies respectively. 

Procedure. The Repeat portion of the Quick Repeat-Recall Test (Quick RRT)19 was used to compare subjects’ speech understanding ability in noise under realistic SNRs. The Quick RRT is a shortened version of the RRT.20 Briefly, the RRT evaluates speech-in-noise understanding for high-context (syntactical and meaningful, eg, “Keep the ice cream in the freezer”) and low-context (syntactical but meaningless, eg, “Keep the ice foods in the lemon”) sentences at SNRs of 0, 5, 10, 15 dB, and quiet. Listeners are also required to recall the sentences after all are repeated, rate their listening effort, and estimate how long they are willing to listen in the specific noise condition.  

The Quick RRT is limited to low context sentences presented at SNRs of 5, 10, and 15 dB. These SNRs are most representative of the typical listening environments that listeners with mild-to-moderate hearing loss encounter.15 Low-context sentences are more difficult and will likely not plateau even at 15 dB SNR. This makes them more sensitive (than high-context sentences) to differences in hearing aid signal processing while testing at a meaningful, realistic SNR. Because our interest was in the speech understanding ability of the test subjects, we focused our attention on the Repeat portion of the Quick RRT.

All testing was conducted with the speech presented from the front (at 0° azimuth) at 75 dB SPL and noise from the back (at 180°) at levels to produce the desired SNRs. The noise was a continuous speech-shaped noise with the same average spectrum as the speech stimulus. Continuous noise in a noise-back configuration was chosen to optimize any potential benefits of the directional microphones used in the competitors’ aids. 

At each of the 3 SNRs tested, participants listened to each of 6 sentences and repeated them back to the experimenter as they were heard. For each set of 6 sentences, 20 target words were scored for correct repetition. The order in which SNRs were tested was randomized. The order of the test lists and the test hearing aids were counterbalanced across subjects. Subjects were blinded to the identity of the hearing aids. 

Results

Figure 3 shows the average repeat performance (along with the 95% confidence interval of the average) for all three hearing aids at SNRs of 5, 10, and 15 dB. The results are shown superimposed on a normative performance template of unaided normal-hearing listeners measured under the same test conditions. The legend on the right indicates how the different shaded regions on the plot can be used to compare the performance of the aided hearing-impaired listeners to the expected performance of normal-hearing listeners. For example, if the score falls in the dark blue area, it suggests performance similar to the top 25% of normal-hearing listeners. 

Figure 3. Repeat score on the Quick RRT for low context sentences tested at SNRs of 5, 10, and 15 dB with a speech-front (fixed at 75 dB SPL) noise-back configuration. Error bars show 95% confidence intervals of the means.

Figure 3 shows an average score of around 60% at 5 dB SNR, 80% at 10 dB SNR, and 85% at 15 dB SNR. Importantly, the average scores of all three hearing aids were within 5% of each other. This suggests that despite the differences in features among hearing aids (and the use of an omnidirectional microphone in the PureSound), the speech-in-noise performance of all three hearing aids under realistic SNRs is similar. Most importantly, the average aided repeat performance of all three hearing aids put the listeners’ performance within the average range of normal-hearing listeners (the medium blue area, between 25-75 percentile).  

Discussion

This study shows the PureSound program on the MOMENT hearing aid is as effective as the premium products of two other manufacturers for managing speech understanding in noise at realistic SNRs. More importantly, listeners’ performance with the PureSound program is similar to that of normal-hearing listeners under the same realistic SNRs.

The performance similarity between the PureSound program and the competitors’ aids confirms our design rationale for the action of the Dual-Mic Input Control. The leakage from the use of an open ear tip reduces the potential benefits offered by signal processing algorithms (including directional microphone). Testing at realistic SNRs circumvented a potential artifact of adaptive testing at unrealistic SNRs. The current study design is likely more reflective of daily hearing success in typical real-life situations. 

Despite the lack of a measurable benefit at realistic SNRs, it must be stressed that directional microphones—even in an open fitting—are beneficial in challenging SNRs. Indeed, our literature review suggests a 1-2 dB SNR benefit with open fitting in the laboratory. 

In a previous study, we demonstrated that the benefit of the directional microphone in the EVOKE (same used in the MOMENT) decreased with SNR.8 However, the improvement in repeat scores attributable to the directional microphone at 0 dB SNR was over 50%. This suggests that even though we observed similar performance among the three hearing aids under typical SNRs, there may be more challenging situations in real-life when a directional microphone would be helpful. The MOMENT acknowledges this fact and includes the Dual-Mic Input Control system that automatically switches to the adaptive HD Locator mode when the wearer switches from the PureSound program to the Universal program. This gives wearers of the MOMENT hearing aid full control of the features best suited to different listening environments. 

The current results should be interpreted with findings reported by Balling et al1who showed an overwhelming preference for the PureSound program over the Universal program. Indeed, 85% of hearing-impaired listeners and 100% of normal-hearing listeners preferred the PureSound program in a majority of listening environments ranging from noisy to quiet/soft sound situations. Listeners reported the PureSound program to be “more natural,” “clearer,” and “more detailed.” 

Their findings, coupled with the current findings, suggest that the ZeroDelay technology in the Widex MOMENT hearing aid optimizes sound quality in daily situations while preserving speech understanding in noise.  

Correspondence can be addressed to Dr Kuk at: [email protected].

Citation for this article: Kuk F, Ruperto N, Slugocki C, Korhonen P. Efficacy of directional microphones in open fittings under realistic signal-to-noise ratios using widex moment hearing aids. Hearing Review. 2020;27(6):20-23.

References

1. Balling LW, Townend O, Stiefenhofer G, Switalski W. Reduced hearing aid delay for optimal sound quality: A new paradigm in processing. Hearing Review. 2020;27(4):20-26. 

2. Bramslow L. Preferred signal path delay and high-pass cut-off in open fittings. Int J Audiol.2010;49(9):634-644. 

3. Stone MA, Moore BCJ, Meisenbacher K, Derleth RP. Tolerable hearing aid delays. V. Estimation of limits for open canal fittings. Ear Hear.2008;29(4):601-617.

4. Kuk F. Rationale and requirements for a slow acting compression hearing aid. Hear J.1998;51(6):45-53, 79.

5. Kuk F, Hau O. Compression speed and cognition: A variable speed compressor for all. Hearing Review.2017;24(3):40-48.

6. Kuk F, Korhonen P, Slugocki C. Preserving the temporal envelope in hearing aid processed sounds. Hearing Review.2018;25(10):40-44.

7. Bentler RA. Effectiveness of directional microphones and noise reduction schemes in hearing aids: A systematic review of the evidence. J Am Acad Audiol. 2005;16:473-484.

8. Kuk F, Slugocki C, Korhonen P. An integrative evaluation of the efficacy of a directional microphone and noise-reduction algorithm under realistic signal-to-noise ratios. J Am Acad Audiol.2020;31(04):262-270.

9. Kuk F, Keenan D. How do vents affect hearing aid performance? Hearing Review.2006;13(2):34-42.

10. Bentler R, Wu Y-H, Jeon J. Effectiveness of directional technology in open-canal hearing instruments. Hear J. 2006;59(11):40-47.

11. Kuk F, Keenan D, Ludvigsen C. Is real-world directional benefit predictable? Hearing Review. 2004;11(11):18-25.

12. Klemp EJ, Dhar S. Speech perception in noise using directional microphone in open-canal hearing aids. J Am Acad Audiol. 2008;19:571-578.

13. Magnusson L, Claesson A, Persson M, Tengstrand T. Speech recognition in noise using bilateral open-fit hearing aids: The limited benefit of directional microphones and noise reduction. Int J Audiol.2013;52(1):29-36.

14. Valente M, Mispagel KM. Unaided and aided performance with a directional open-fit hearing aidInt J Audiol.2008;47(6):329-336.

15. Smeds K, Wolters F, Rung M. Estimation of signal-to-noise ratios in realistic sound scenarios. J Am Acad Audiol.2015;26(2):183-196. 

16. Wu Y-H, Stangl E, Chipara O, Hasan SS, Welhaven A, Oleson J. Characteristics of real-world signal to noise ratios and speech listening situations of older adults with mild to moderate hearing loss. Ear Hear. 2018;39(2):293-304.  

17. Cord MT, Surr RK, Walden BE, Dyrlund O. Relationship between laboratory measures of directional advantage and everyday success with directional microphone hearing aids. J Am Acad Audiol. 2004;15:353-364.

18. Palmer C, Bentler R, Mueller GH. Evaluation of a second-order directional microphone hearing aid: II. Self-report outcomes. J Am Acad Audiol.2006;17(3):190-201.

19. Kuk F, Korhonen P, Slugocki C, Ruperto N. How to use the Repeat-Recall Test (RRT) to increase hearing aid satisfaction. Hearing Review.2019;26(9):32-40.

20. Slugocki C, Kuk F, Korhonen P. Development and clinical applications of the ORCA Repeat and Recall Test (RRT)Hearing Review.2018;25(12):22-26.