There is little doubt that hearing aids with a directional microphone improve the signal-to-noise ratio (SNR) of the listening environments for their wearers at least some of the time (reviews on this topic are found in Ricketts1 and Valente2). To achieve SNR improvement, directional microphones are designed with higher sensitivity to sounds that originate from the front than those that originate from the sides or back. With this design rationale, SNR is improved when the speaker is in front and the background noise is to the sides and/or back of the hearing aid wearer.
Unfortunately, the same design that affords SNR improvement for the directional microphone is the same one that creates the loss of audibility for desirable sounds originating from the sides and back. Indeed, one would speculate that directional hearing aids with a higher directivity index (DI) are more likely to result in greater audibility loss than hearing aids with a lower DI.3 Considering that as much as 40% of an adults listening environments are devoid of background noise and as much as 20% of the time the signal of interest is from locations other than the front,4 a directional hearing aid should be designed with features to overcome the potential audibility loss in listening situations that may violate the assumptions of such a microphone.
The solutions to the limitations of a fixed directional hearing aid may include implementing the directional microphone on a wide dynamic range compression (WDRC) hearing aid that has a low compression threshold (CT). The use of WDRC (as opposed to linear) with a low CT (as opposed to a WDRC with a moderate or high CT) would provide extra audibility for soft sounds. Consequently, despite the lowered intensity level of the input sounds from the directional microphone, these sounds may still be audible to the wearer, whereas a linear hearing aid set to the same gain for conversational sounds may not.5,6 Of course, a directional WDRC hearing aid would still suffer audibility loss when compared to an omnidirectional WDRC hearing aid that used the same hearing aid settings.
A second solution may include both the omnidirectional and directional microphones on the hearing aid with a manual switch so the wearer may select the optimal microphone mode. This may be a good solution to preserve audibility and improve SNR on an as needed basis. However, its dependence on the dexterity of the wearers could limit or prevent its applications by young children and older adults. Furthermore, because many wearers frequently leave the hearing aids in the directional mode, they may not recognize the need to switch microphones if they are not made aware of the missing signals.
A third solution may include hearing aids with automatic directional microphones. In these hearing aids, the microphone automatically switches from an omnidirectional mode to a fixed directional mode depending on the input level. Typically it stays in an omnidirectional mode when the overall input level is below a conversational level. Above that level, it automatically switches into the directional mode. This may be beneficial in preserving the audibility loss from a fixed directional microphone. However, it may not be sufficient to overcome changes in the direction of the noise source. Thus, the SNR benefit may be compromised in some noisy situations.
An adaptive directional microphone changes its polar patterns in order to optimize the SNR when the noise source changes direction. However, it may not be adequate in ensuring audibility because none of the directional polar patterns is omnidirectional. A fully adaptive directional microphoneone that is both automatic (changes from an omnidirectional polar pattern to a directional polar pattern) and adaptive (changes amongst different polar patterns)may represent the best solution for minimizing the loss of audibility while enhancing the SNR of the listening environment. In this paper, we will report on some of our findings to support this assertion.
Fighting the Trade-Offs
A fully adaptive directional microphone known as the Locatoramong the first of its kindis a feature of the Senso Diva hearing instrument. Its polar pattern changes automatically from an omnidirectional pattern to any directional polar patterns (while keeping the same frequency response characteristics), depending on the intensity and azimuth of the stimulus.7 To ensure consistent directional performance, an Opti-Mic system continuously matches the sensitivity and phase of the dual microphones.
Typically, the microphone system stays in the omnidirectional mode for all sounds presented directly from the front. It stays in (or switches into) an omnidirectional mode when the sound level at the microphone is below 50-60 dB SPL, or when the noise source is wind. This mechanism preserves audibility of soft speech presented from the sides or back and minimizes the perception of circuit and wind noise. The polar pattern adapts its null (or minimum sensitivity) to the direction of single-noise sources. In a diffuse field where there may be multiple noise sources, a hypercardioid polar pattern results. Meanwhile, the response in the low-frequency is compensated such that the same frontal frequency response (as the omnidirectional microphone) is used in all directional polar patterns.
An important consideration in a fully adaptive directional microphone is the time it takes the omnidirectional microphone to switch to the directional mode. A short switching time means the hearing aid reduces its sensitivity to sounds from the sides and back quickly. A possible consequence is that one may not notice the presence of the sound.
A long switching time is designed to provide wearers enough time to detect the presence of the sound source. The wearers can make a decision if they want to turn towards the sound source. If the source is undesirable and they choose to ignore it, the fully adaptive microphone will switch to the correct polar pattern by forming a null at the direction of the sound source. If the sound source is desirable and the wearers wish (and are able) to turn their heads, the sound source that once originates from the sides or back now originates from the front. This keeps the fully adaptive microphone in an omnidirectional mode and minimizes the loss of audibility.
It is estimated that 5-10 seconds is a good compromise between giving wearers enough time to respond (thus preserving audibility) and sufficient responsiveness to the changing listening situation. For these reasons, the Locator uses a switching time that varies between 3-10 seconds.
This paper reports on studies that provide additional evidence relative to SNR and audibility efficacy, and discusses clinical implications of directional microphone designs for both adults and children.
Demonstration of SNR Improvement with Different Noise Azimuths
Valente and Mispagel8 reported on the SNR efficacy of the Diva Locator in-the-canal (ITC) hearing aid on 20 adult hearing-impaired subjects who had a mild to moderately severe sensorineural hearing loss and were experienced hearing aid users. Subjects wore binaural Senso Diva for a month prior to data collection. The test conditions included measuring the subjects SNR on the HINT test at a 68 dB SPL-A noise level with the uncorrelated HINT noise presented continuously from the back, the sides, and the sides plus the back (see Figure 1). Subjects repeated the HINT sentences, and the averaged signal level for 50% identification was measured to determine the absolute SNR for the specific test condition. Subjects were tested without any hearing aids, with their own aids, and with the Diva in the omnidirectional and Locator modes. Subjects also completed an APHAB questionnaire at the beginning of the trial and after 1 month experience with the Diva hearing aids in the fully adaptive directional mode. Figure 1 shows the averaged absolute SNR for the different test conditions.
FIGURE 1. Averaged SNR for 50% performance on the HINT with the noise presented from behind (180°), to the sides (90° and 270°), and sides and back (90°, 180°, 270°). Hearing aid conditions included unaided, own aids, Diva-omnidirectional, and Diva-Locator modes.
Figure 1 demonstrates that an SNR improvement of over 7 dB was noted between the adaptive directional microphone and the omnidirectional microphone when the noise source was at 180°. An SNR improvement of almost 5 dB was noted when the noise sources were to the sides (90° and 270°) and sides and back (90°, 180°, and 270°) of the wearers. Such a magnitude of benefit will probably decrease if the Locator were in a fixed directional mode.
Figure 2 summarizes subjects benefit scores on the APHAB questionnaire. Included were items from Ricketts et al9 that were specific to the use of a directional microphone. Items on the SoundFront category were items that measured the advantages of a directional microphone, while those on the SoundBack/localize category measured the disadvantages of a fixed directional microphone.
FIGURE 2. Comparison on APHAB benefit scores (including the SoundFront and SoundBack subscales from Ricketts et al9) between the subjects own hearing aids and aids in adaptive directional mode.
Figure 2 shows that the benefits provided by the subjects own hearing aids were about 20% across most subscales (except Aversiveness of Sounds). The benefit scores reported on the Diva were 10-20% higher than the subjects own aids; in the Background Noise and Reverberation subscales, the benefit scores exceeded 40%.
Despite the Locator being a fully adaptive directional microphone, its benefits on the Ease of Communication and Sound Back subscales were higher than those reported of the subjects own aids. This would suggest that the fully adaptive directional microphone system may not result in (at least significant) audibility limitations that could affect the real-world benefits experienced by the subjects.
Demonstration of Audibility Preservation
In order to evaluate if the Locator design could ensure the audibility of sounds presented from the sides and back, we tested the speech recognition scores of 17 hearing-impaired subjects with the Computer Assisted Speech Perception Assessment (CASPA)10 test. There were a total of 10 monosyllabic word lists, and they were presented at a distance of 1 meter from the subject at 50 dB SPL, 65 dB SPL, and 75 dB SPL in quiet. A carrier phrase, Please say the word —– preceded each word. Testing at multiple levels was done to evaluate the intensity effect of the audibility limitation (of a fixed directional microphone). Testing was conducted only on the right ear with the speech stimuli presented at 0°, 45°, 90°, 135°, and 180°. The left ear was occluded with a foam plug. A Senso Diva SD-9 hearing aid was used in the fixed directional mode, the omnidirectional mode, as well as the fully adaptive directional mode. The hearing aid was coupled to a foam earmold and programmed to the subjects hearing loss with the standard sensogram and feedback tests. An occluding foam earmold was used in order to prevent sounds from directly entering the earcanal and biasing the measurement. All test conditions and word lists were counterbalanced (for a more detailed description of the procedure, see Kuk et al11).
Figure 3 shows the absolute speech (phoneme) recognition scores in each microphone mode at each intensity level as the stimuli were presented from different azimuths. At a 50 dB SPL input level (soft speech), the omnidirectional microphone and the fully adaptive microphone both yielded around 80% performance regardless of the azimuths of presentation (Figure 3A). The fixed directional microphone showed a distinctly poorer performance as the azimuth of presentation deviated from 0°. Indeed, as the azimuth of presentation moved away from 0°, phoneme recognition decreased steadily and reached the lowest at an azimuth of 135° where the averaged absolute performance was around 35%. This was less than 50% of the performance of the omnidirectional or fully adaptive directional microphones. Performance improved beyond the 135° where the score was almost 60% at an azimuth of 180°. The poorer scores at a 135° azimuth coincided with the null of the hypercardioid polar pattern of the fixed directional microphone. This demonstrates that a fixed directional microphone can degrade the recognition of soft speech by as much as 50% when it is presented from the null of the polar pattern.
FIGURE 3. Absolute speech (phoneme) recognition scores for each microphone mode (fixed=blue; adaptive=red; omni=yellow) at each intensity level (50 dB, 65 dB, and 75 dB SPL) as the stimuli were presented from different azimuths. The error bars indicate the magnitude of 1 standard deviation.
The fact that the performance of the fully adaptive directional microphone was similar to the omnidirectional microphone is related to the activation threshold (around 50-60 dB SPL) of the directional microphone. Consequently, the directional microphone mode was never activated when the stimulus level was at 50 dB SPL. Audibility for soft speech was preserved because the microphone system behaved like an omnidirectional microphone.
On the other hand, the limitation of the fixed directional microphone was not as evident at the 65 dB SPL and 75 dB SPL presentation levels (Figures 3B-C). Compared to the omnidirectional microphone, performance with the fixed directional microphone was about 15% poorer at 65 dB SPL and 5% poorer at 75 dB SPL when the stimuli were presented at 135°. This suggests that the limitations of a fixed directional microphone will be lessened at higher stimulus levels.
It is interesting to note that the performance of the fully adaptive directional microphone was similar to that of the omnidirectional microphone at the 65 dB SPL and 75 dB SPL conditions. This may seem puzzling, because it was indicated earlier that the activation threshold of the directional mode was about 50-60 dB SPL. The 65 dB SPL and 75 dB SPL presentation levels should have been above the activation threshold and switched the system into a directional mode. Furthermore, in the Valente and Mispagel8 report, it was demonstrated that the directional microphone was effective at a 68 dB SPL input level.
The activation time (eg, the time it takes for the microphone to switch from an omnidirectional mode to a specific directional mode) may be responsible for the observations at the 65 dB and 75 dB SPL stimulus levels. Recall that it takes the directional system between 3-10 seconds to switch from an omnidirectional mode to a specific polar pattern in the directional mode. Additionally, the speech test materials were monosyllabic words with a carrier phrase, Please say the word —–. The duration of the carrier phrase and target word was no more than 3 seconds. This means that the stimulus was never long enough to have activated the directional mode. The hearing aid remained in the omnidirectional mode the whole time during speech testing in quiet. This explains why it performed like an omnidirectional microphone in quiet.
There are several conclusions to this study. First, the use of a fixed directional microphone limits the amount of audible information (and consequently intelligibility) to the hearing aid wearer. Secondly, the audibility limitation was especially noticeable at the lowest input level and when the stimulus was presented at the null of the polar pattern. At a higher stimulus level and when the stimuli were presented away from the null, the audibility limitation decreased. This explains why few complained about the limitations of a fixed directional microphone for typical everyday use because speech understanding is only mildly affected. Although a fixed directional microphone can severely limit the audibility of soft sounds, few actually complained because many did not hear the soft sounds. In other words, how can one complain about something that one did not hear?
Although the study was conducted on adult wearers, there is no reason to expect that children would fare better than adults with a fixed directional microphone. Indeed, one would expect children to be at even more of a disadvantage by this limitation because they are still developing their speech and language skills and because they may require a higher sensation level than adults for speech discrimination and identification. The results of this study add to the literature that supports a cautious approach to the choice of a fixed directional microphone for children.
More importantly, however, the results of this study showed that with careful design considerationsspecifically in the appropriate choice of activation threshold (above 50 dB SPL) and switching time (longer than 3 to 10 seconds)the limitation of a fixed directional microphone may be minimized with a fully adaptive directional microphone. In this study, it was shown that the fully adaptive directional microphone behaved liked an omnidirectional microphone when the listening condition was quiet, and a fixed directional microphone when the listening condition was filled with a continuous noise. The favorable benefit scores reported on the APHAB questionnaire attest to the real-world acceptance of such a design.8
This study was conducted using Senso Diva hearing aids only. Although it is possible that other manufacturers may have used the same criteria (level of activation and switching time) in the design of their adaptive directional microphone, it is important that the clinicians understand the particular design criteria of the directional microphone. Specifically, they should realize how the design may improve SNR while minimizing audibility loss and choose the optimal system accordingly. If the optimal system cannot be chosen (eg, because of cost concerns, etc), the wearer should be counseled properly so the right expectations may be set and the appropriate communication strategy developed.
Knowledge of the design criteria relative to the switching of the directional system is important for another reason. The accurate and reliable verification of a fully adaptive directional microphone is dependent on the duration and level of the chosen stimulus. Because the activation threshold for this device is around 50 dB SPL, any stimulus used must be higher than 50 dB SPL in order to activate the directional system. Otherwise, only an omnidirectional microphone will be evaluated. In addition, the stimulus duration must be longer than 10 seconds in order to fully switch the system into the appropriate directional mode. These considerations were discussed in an earlier article.12
The findings from these studies prompted us to reconsider childrens candidacy for a directional microphone. As has been repeatedly stressed, hearing-impaired children need a more favorable SNR than normal-hearing children to perform satisfactorily. On the other hand, the potential audibility limitations of a fixed directional microphone prevented many clinicians from recommending a directional microphone to children, despite the knowledge that directional microphones can improve the SNR of a childs listening environment.13 Some limited the use of a directional microphone only to older children, and only for formal learning environments. The current study showed that it is possible to maintain audibility and still achieve SNR improvement in adults. If this finding can be extrapolated to children, then a fully adaptive directional microphone is not only a desirable feature in a hearing aid for children, but a necessary and standard feature as well. Further validation studies are planned with this technology in children.
|This article was submitted to HR by Francis Kuk, PhD, director, and Denise Keenan, MA, and Chi Lau, PhD, research audiologists, at the Widex Office of Research in Clinical Amplification (ORCA), Lisle, Ill. Correspondence can be addressed to Francis Kuk, Widex Office of Research in Clinical Amplification, 2300 Cabot Dr, Ste 415, Lisle, IL 60532; email: [email protected].
1. Ricketts T. Directional Hearing Aids. Trends Amplif. 2001;5(4):139-176.
2. Valente M. Use of microphone technology to improve user performance in noise. Trends Amplif. 1999;4(3):112-135.
3. Gravel J, Fausel N, Liskow C, Chobot J. Children’s speech recognition in noise using omni-directional and dual- microphone hearing aid technology. Ear Hear. 1999;20(1):1-11.
4. Walden B, Surr R, Cord M, Dyrlund O. Predicting hearing aid microphone preference in everyday listening. J Am Acad Audiol. 2004;15:365-396.
5. Lee L, Lau C, Sullivan D. The advantage of a low compression threshold in directional microphones. The Hearing Review. 1998;5(8):30-32.
6. Kuk F. Optimizing compression: The advantages of a low compression threshold. In: Kochkin S, Strom KE, eds. High Performance Hearing Solutions, Vol 3: Marketing and Technology. The Hearing Review. 1999;[Suppl]6(1):44-47.
7. Kuk F, Baekgaard L, Ludvigsen C. Using digital signal processing to enhance the performance of dual microphones. Hear Jour. 2002;55(1):35-43.
8. Valente M, Mispagel K. Performance of an automatic adaptive dual-microphone ITC digital hearing aid. The Hearing Review. 2004;11(2):42-46, 71.
9. Ricketts T, Henry P, Gnewikow D. Full time directional versus user selectable microphone modes in hearing aids. Ear Hear. 2003;24(5):424-439.
10. Mackersie C, Boothroyd A, Minniear D. Evaluation of the Computer-Assisted Speech Perception Assessment Test (CASPA). J Am Acad Audiol. 2001;12(8):390-396.
11. Kuk F, Keenan D, Lau C, Ludvigsen C. Performance of a fully adaptive directional microphone to signals presented from various azimuths. J Am Acad Audiol. 2005;16(6):335-349.
12. Kuk F, Peeters H, Keenan D, Baekgaard L. Timing is (almost) everything: Verifying the performance of fully adaptive directional microphones. The Hearing Review. 2005;12(8):24-29.
13. Kuk F, Kollofski C, Brown S, Melum A, Rosenthal A. Use of a digital hearing aid with directional microphones in school-age children. J Amer Acad Audiol. 1999;10(10):535-548.